The relationship between total food intake (g) - CEBC - CNRS

Université de POITIERS
UFR des Sciences Fondamentales et Appliquées
Thèse présentée par Olivier Lourdais
Pour obtenir le grade de docteur de l’Université de POITIERS
Coûts de la reproduction, gestion des ressources et
fréquence des épisodes reproducteurs
chez la vipère aspic (Vipera aspis)
Soutenue le 19 novembre 2002 devant la commission d’examen :
Miaud C Maître de conférence, CNRS Rapporteur
Gasc JP Professeur de rang 1, CNRS Rapporteur
Baehr JC Professeur à l’Université de Poitiers Examinateur
Mazin JM Directeur de recherche au CNRS Examinateur
Zuffi MAL Conservateur, Muséum de Pise Examinateur
Bonnet X Chargé de recherches CNRS Directeur de thèse
Invités:
Shine R Co-directeur de thèse, Professeur à l’Université de Sydney
Naulleau G Chargé de recherches CNRS
Weimerskirch H Directeur de recherches CNRS
1

Sommaire
Remerciements 3
Remarque 4
Publications 5
Résumé 8
Introduction générale
A. Sélection, évolution et traits d’histoire de vie 11
B. Optimisation de l’investissement et coût de la reproduction 13
C. Dimension physiologique des compromis et des systèmes de gestion
de la ressource 15
D. Connexion entre effort reproducteur et coût de la reproduction 18
E. Intérêt d’une perspective ectothermique 21
I. Présentation de l’espèce et des méthodes d’étude
A. Résumé du Chapitre 26
B. Position systématique 27
C. Répartition de l’espèce et des populations l’études 28
D. Biologie de la reproduction 29
E. Méthodes d’étude 37
II. Le système d’allocation de l’énergie
A. Résumé du chapitre 47
B. Article 1 : short-term versus long-term effects of food intake on
reproductive output in a viviparous snake (Vipera aspis) 50
C. Article 2 : when does a reproducing female viper (Vipera aspis) “decide"
on her litter size 81
D. Article 3 : capital breeding and reproductive effort in a variable
environnment: a longitudinal study in the aspic viper (Vipera aspis) 98
2

III. Les coûts de la reproduction: amplitude et degré de dépendance
avec la fécondité
A. Résumé du chapitre 124
B. Article 4 : reproduction in a typical capital breeder, costs, currencies
and complication in the aspic viper 128
C. Article 5 : costs of anorexia during pregnancy in a viviparous snake
(Vipera aspis) 156
D. Article 6 : thermoregulation and metabolism in a viviparous snake,
Vipera aspis: evidence for fecundity-independent costs 170
E. Article 7 : what is the appropriate time scale for measuring costs of
reproduction in a capital breeder such as the aspic viper ? 198
IV. Les déterminants de la tendance semélipare femelle: description et
implications démographiques
A. Résumé du chapitre 217
B. Article 8 : comparaisons des tactiques demographiques de la vipère aspic
(Vipera aspis): y’a t’il un avantage a etre semélipares ? 220
C. Article 9 : do sex divergences in ecophysiologie translate into sexdimorphic
demographic patterns? 240
V Discussion-conclusion
Reproduction sur réserves, coûts de la reproduction et évolution vers la
seméliparité 264
Bibliographie 274
Annexe
Article : Natural thermal conditions influence embryonic development in a
viviparous snake (Vipera aspis) 311
3

Remerciements :
Plutôt que d’écrire de très longs remerciements, un peu pénibles à lire, j’ai préféré
opter pour une formulation plus concise. Les raisons d’un tel choix sont
multifactorielles et j’offre ci-dessous des versions alternatives plus ou moins
crédibles :
Je tiens à remercier l’ensemble de l’humanité pour m’avoir fournit une aide et un
réconfort quotidien dont l’ampleur est telle qu’une page ne suffirait évidemment
pas. Je vous remercie donc tous : je vous aime.
Il existe un véritable risque à l’écriture de remerciements trop longs. En effet, si je
dépasse la fin de cette page 3, l’ensemble de ma pagination (rentrée à la main)
va être modifiée. Les contraintes énergétiques associées à la remise à jour du
sommaire sont telles que de fortes pressions de sélection favorisent des
remerciements courts et non répétitifs (à tendance semélipare) : Merci.
Enfin, d’un point de vue historique, au moment où je me suis engagé dans ce
travail avec mon ami Xavier Bonnet, l’herpétologie n’était pas vraiment une
discipline académiquement reconnue et le département d’herpétologie de Chizé
était au bord de la fermeture. La bonne conduite de ce projet est donc liée à une
poignée d’humains ayant joué un rôle fondamental. J’ai déjà pris soins de
remercier individuellement ces personnes et en rajouter deviendrait de la
flagornerie.
De façon plus proximale, je tiens à remercier mon co-équipier François
Brischoux pour m’avoir fournit une aide indispensable à la finition du manuscrit
et au respect des délais.
Merci à mes deux rapporteurs Claude Miaud et Jean Pierre Gasc, pour avoir
accepté de lire et corriger un tel pavé.
Merci à mon ami Hassan pour la phrase qu’il m’a dite un soir, l’air profondément
dubitatif en autopsiant le cadavre d’une ancienne version de mon schéma de la
page 267 : « Olivier, tu sais, la vie c’est comme ton schéma, c’est compliqué »
…à mes parents et à Pierre pour avoir rendu possible cette véritable odyssée.
4

Remarque :
Le format de cette thèse s’inscrit dans le cadre actuel des efforts de production
scientifique (publications) associés à un doctorat en écologie. Ce travail repose donc
sur la présentation d’une série d’articles (7 publiés, 1 soumis et 1 en cours
d’élaboration). Les articles publiés ou soumis sont présentés sous leur structure
classique et rédigés en anglais. L’introduction, la présentation de l’espèce, les
résultats récents ainsi que la discussion-conclusion ont été spécifiquement rédigés
en français pour ce mémoire. Les différents travaux ont été regroupés en trois
grands chapitres intimement connectés, qui retracent l’approche mise en oeuvre pour
examiner la stratégie reproductrice de l’espèce et élaborer un scénario évolutif. Afin
de faciliter le cheminement, chaque chapitre est précédé d’un résumé des principaux
résultats et éléments de réflexion. Afin de limiter l’hétérogénéité associée à un tel
regroupement d’article, tous les travaux ont été mis au même format et une
pagination continue a été choisie. Enfin, une bibliographie globale est présentée à la
fin du mémoire. J’espère que la présentation choisie sera à la hauteur d’un défi fort
difficile, celui de donner à une thèse sur articles une organisation formelle cohérente
et fonctionnelle.
5

Publications
Mes premières implications dans l’étude des serpents débutent très tôt dans mon
cursus Universitaire (1996, stage de DEUG). Dès mon entrée en DEA à Chizé (1998)
j’ai pu commencer à me consacrer au travail de publication. Ma position d’auteur
associé dans ces travaux publiés trouve ainsi son origine dans une participation
active de ma part à la fois dans la récolte des données, les analyses statistiques et la
rédaction. Par la suite, je me suis impliqué de façon plus personnalisée avec la
publication de quatre articles dont je suis l’auteur principal.
Liste des publications sur la thèmatique de thèse :
1. Lourdais O, Bonnet X, DeNardo D, Naulleau G. (2002) Does sex differences in
reproductive eco-physiology translate in different demographic patterns ?
Population Ecology, in press
2. Lourdais O, Bonnet X, Shine R & Taylor E. (2002) When does a reproducing
female viper (Vipera aspis) "decide" on her litter size? Journal of Zoology, London
in press
3. Lourdais O, Bonnet X, Shine R , DeNardo D, Naulleau G & Guillon M. (2002).
Capital-breeding and reproductive effort in a variable environment: a longitudinal
study of a viviparous snake. Journal of Animal Ecology 71 : 470-479.
4. Lourdais O, Bonnet X, Doughty P. (2002). Costs of anorexia during pregnancy in
a viviparous snake (Vipera aspis). Journal of Experimental Zoology 292 : 487-
493.
6

5. Bonnet X, Lourdais O, Shine R. & Guy Naulleau. (2002). Reproduction in a
typical capital breeder : costs, currencies and complications in the aspic viper
Ecology 83 : 2124-2135.
6. Bonnet X., Naulleau G. & Lourdais O. (2002). The benefits of complementary
techniques: using capture-recapture and physiological approaches to understand
costs of reproduction in the asp viper. Biology of the Vipers, in press
7. Bonnet X., Shine R, Lourdais O & Naulleau G (2002) Measures of reproductive
allometry are sensitive to sampling Bias. Functionnal Ecology, in press
8. Aubret F, Bonnet X, Shine R & Lourdais O. (2002) Fat is sexy for females but
not males: the influence of body reserves on reproduction in snakes (Vipera
aspis). Hormones and Behaviors 42 : 135-147.
9. Bonnet X., Naulleau G., Shine R. & Lourdais O. (2001). Short-term versus long-
term effects of food intake on reproductive output in a viviparous snake (Vipera
aspis). Oikos 92 : 297-308.
10. Bonnet X., Naulleau G., Shine R. & Lourdais O. (2000). What is the appropriate
time scale for measuring costs of reproduction in a capital breeder?
Evolutionary Ecology 13 : 485-497.
11. Bonnet X., Naulleau G., Shine R. & Lourdais O. (2000). Reproductive versus
ecological advantages to larger body size in female Vipera aspis. Oikos 89 : 509-
518.
12. Naulleau G., Bonnet X., Vacher-Vallas M, Shine R. & Lourdais O. (1999). Does
less-than-annual production of offspring by female vipers (Vipera aspis) mean
less-than-annual mating? Journal of Herpetology 33 : 688-691.
7

13. Bonnet X., Naulleau G., Lourdais O. & Vacher-Vallas M. (1999). Growth in the
asp viper (Vipera aspis L.): insights from long term field study. Current Studies in
Herpetology. C. Miaud et R. Guyetant eds. pp. 63-69.
Autres publications
14. Bonnet X, Shine R, Lourdais O. (2002). Taxonomic Chauvinism. Trends in
Ecology and Evolution 17 : 1-3.
15. Bonnet X, Pearson D, Ladyman M, Lourdais O, & Bradshaw D. (2002). Heaven
for serpents? A mark-recapture study of Tiger Snakes (Notechis scutatus) on
Carnac Island, Western Australia. Austral Ecology 27 : 442-450.
16. Pearson D., Shine R., Bonnet X., A. Williams, B. Jennings & O. Lourdais. (2000).
Ecological notes on crowned snakes, Elapognathus coronatus, from the
Archipelago of the Recherche in southwestern Australia. Australian Zoologist 31 :
610-617.
Enfin, signalons deux travaux actuellement soumis:
17. Lourdais O, Shine R, Bonnet X, Guillon G, & Guy Naulleau. Natural thermal
conditions influence embryonic development in a viviparous snake (Vipera aspis).
Functionnal Ecology
18. Ladyman M , Bonnet X, Lourdais O, Bradshaw D & Naulleau G. Gestation,
thermoregulation and metabolism in a viviparous snake, Vipera aspis: evidence
for fecundity-independent costs. Physiological and Biochemical Zoology
8

Résumé :
Le nombre d'épisodes reproducteurs au cours de l'existence d'un organisme
constitue un trait d'histoire de vie majeur. On distingue ainsi des espèces
semélipares (une seule reproduction et la mort de l’organisme), et d'autre itéropares
(reproductions répétées). La vipère aspic occupe une position intermédiaire avec une
faible fréquence de reproduction (tous les 2-4 ans) et une tendance marquée vers la
seméliparité. D'un point de vue évolutif, il est légitime de s'interroger sur les
avantages d’une telle stratégie où les dépenses reproductrices sont accrues et peu
fréquentes.
Nos travaux sur la vipère aspic suggèrent une relation directe entre la
fréquence reproductrice et la nature des contraintes énergétiques et écologiques de
la reproduction. Dès l’engagement dans la folliculogénèse, la vipère aspic va être
confrontée à des activités très coûteuses (exposition aux prédateurs, coûts
métaboliques) qui reflètent des changement profonds de la physiologie et du
comportement. De façon surprenante, si ces changements sont directement liés au
statut reproducteur, ils ne sont pas dépendants de l’effort reproducteur et du nombre
de jeunes produits. En outre, ces coûts particuliers, s’expriment sur un pas de temps
complexe impliquant des composantes directes (l’année de la reproduction), et des
composantes délayées (post-reproduction). Ces résultats viennent donc confirmer
l’hypothèse de Bull et Shine (1979) selon laquelle les systèmes à faible fréquence
reproductrice émergent lorqu’il existe des contraintes reproductrices (coûts) dont
l’amplitude est élevée et indépendante de la fécondité.
Notre idée originale repose sur une connexion du modèle de Bull et Shine
avec les stratégies d'acquisition et d'allocation de l'énergie. En effet, si le nombre de
reproduction est réduit, l’organisme aura un intérêt évident à investir massivement
9

son énergie pour garantir le succès de ses quelques opportunités de reproductions.
Une possibilité de répondre à une telle demande passe par la sélection de système
de gestion de la ressource particuliers, impliquant notamment le stockage de
réserves corporelles. Nos résultats supportent largement l’existence d’une telle
relation évolutive entre les coûts indépendants de la fécondité, les systèmes à faible
fréquences de reproductions et les stratégies de capitalisation de l’énergie (“Capital-
breeding”). Cette étude apporte donc des éléments de réponses pertinents sur les
conditions d’émergence des systèmes de reproduction “extrêmes” et sur la transition
évolutive vers la seméliparité.
10

Introduction générale
Un saumon rouge (Oncorhynchus nerka) mourant d’épuisement à la suite
d’un épisode reproducteur unique (seméliparité)
“...expenditures on reproductive processes must be in functional harmony with each
other and worth costs, in relation to the long range reproductive interest; and the use
of resources for somatic processes is favored to the extent that somatic survival, and
perhaps growth, are important for future reproduction.”
Willians 1966b: 687
11

A. Sélection, évolution et traits d’histoire de vie
Le monde vivant est caractérisé par une étonnante diversité qui se manifeste sur une
suite de niveaux hiérarchisés, depuis l’échelon moléculaire jusqu’au fonctionnement
des écosystèmes. Une telle diversité a depuis longtemps attiré l’attention des
philosophes et scientifiques. C’est avec Darwin (1859) qu’une explication puissante
et unificatrice - la théorie de l’évolution - a été formulée afin de comprendre et
interpréter cette variabilité du vivant. L’évolution des organismes fait intervenir
l’action clé de la sélection naturelle qui opère par multiplication différentielle des êtres
vivants selon leurs aptitudes plus ou moins grandes à transmettre leurs gènes à la
génération suivante. Les processus évolutifs s’opèrent par modifications et
diversifications continuelles des organismes, face à des pressions sélectives variées.
La théorie de l’évolution est la seule capable de fournir un sens commun à tous les
domaines de la biologie (Mayr 1963). En outre, elle offre le fondement conceptuel de
la biologie évolutive actuelle, un champs d’investigation très étendu dont la
puissance explicative tire profit de l’intégration de disciplines complémentaires
(génétique, écologie, physiologie).
L’étude des traits d’histoire de vie (Lessells 1991; Roff 1992; Stearns 1992)
est un domaine central en biologie évolutive qui connait un véritable essor depuis
une trentaine d’année. Par “traits d’histoire de vie” , on désigne un ensemble
complexe de caractères directement “impliqués“ dans la reproduction et la survie des
organismes et donc la contribution en terme de descendance. Si tous les êtres
vivants doivent survivre et se reproduire, il existe cependant des variations majeures
dans certains traits comme la fécondité, la taille des jeunes, l’investissement
reproducteur ou la durée de la vie. Ces variations sont particulièrement évidentes si
l’on compare des espèces présentant des “stratégies d’histoire de vie” très
12

contrastées comme un oiseau longévif et un insecte ayant une durée de vie
inférieure à quelques jours (Stearns 1992).
La théorie des traits d’histoire de vie cherche donc à fournir une explication
évolutive pour interpréter la diversité et la complexité des cycles de vie entre les
espèces, à élucider le mécanisme commun d’une relation générale liant l’âge, la taille
et la mortalité avec les performances reproductrices. L’évolution et la partition de
l’effort de reproduction constitue dans ce cadre un centre d’intérêt majeur. Pourquoi
certaines espèces investissent une quantité énorme d’énergie dans la production
d’un nombre très élevé de jeunes alors que d’autres espèces plus prudentes vont
allouer une quantité plus réduite d’énergie dans des reproductions peu fréquentes
et/ou des tailles de portées réduites? Le schéma d’investissement âge-spécifique
dans la reproduction et notamment la fréquence des épisodes reproducteurs est
particulièrement variable selon les espèces. Classiquement, on réalise une
distinction entre les organismes dits itéropares et d’autres semélipares (Cole 1954).
Alors que chez les premiers la vie, reproductrice est constituée d’une succession
d’épisodes reproducteurs, il n’existe chez les seconds qu’une seule opportunité de
reproduction associée à la mort systématique de l’organisme. La seméliparité est
considérée comme un état dérivé de l’itéroparité. Les avantages de ces stratégies
contrastées ont été largement discutés (Cole 1954, Stearns 1992) et de nombreux
modèles mathématiques ont été formulés pour comprendre les facteurs favorisants
l’une ou l’autre (Cole 1954 ; Gadgil & Bossert 1970 ; Bryant 1971 ; Charnov &
Schaffer 1973 ; Ranta et al. 2002a,b). L’élucidation des pressions sélectives
favorisant la transition entre itéroparité et seméliparité reste néanmoins une question
complexe, notamment du fait de l’apparition de systèmes semélipares dans des
groupes phylogénétiques très éloignés comme les insectes, les annélides, les
mollusques ou encore les poissons ( Boyle 1983, 1987 ; Corkum et al. 1997 ; Andries
13

2001, Crespi & Teo, 2002). La reconstitution de scénarios évolutifs satisfaisant pour
comprendre la transition vers un mode de reproduction induisant la mort de
l’organisme est un problème situé au cœur même de l’étude des traits d’histoire de
vie (Crespi & Teo 2002). L’objectif principal du présent travail est d’apporter des
éléments de réponses et proposer un scénario évolutif possible.
B. optimisation de l’investissement
reproducteur et coût de la reproduction
On constate généralement de fortes variations dans les performances reproductrices
des individus (Darwin 1859). De telles variations suggèrent l’existence de contraintes
sur les traits d’histoire de vie (Williams 1966b ; Lessells 1991; Stearns 1992). Pour
comprendre ces variations, il est important d’adopter une approche intégrative de
l’organisme et de son environnement. En effet, le milieu de vie est caractérisé par
d’importantes fluctuations ou limitations à la fois dans des facteurs biotiques
(nourriture) et abiotiques (paramètres physiques: température, hygrométrie). Dans
cette situation, les organismes ne vont disposer que d’une quantité limitante d’énergie
qui devra être allouée dans des fonctions très différentes comme la croissance, la
reproduction et la maintenance. La ressource investie dans la reproduction va ainsi
entrer en conflit avec les autres fonctions. En effet, selon le principe d’allocation de
Williams (1966a,b) et Levins (1968), tout investissement supplémentaire dans un
aspect quelconque de la vie d’un organisme ne pourra se faire qu’au dépend d’un
autre aspect. Le fait que les ressources soient généralement limitantes et que les
organismes doivent investir l’énergie dans des voies concurrentielles est à la base de
la notion de compromis ou “trade-off” entre les traits d’histoire de vie. Ainsi, la valeur
reproductrice totale d'un organisme peut être considérée comme la somme entre le
14

succès d’une reproduction donnée et celui attendu dans les reproductions futures.
Selon la théorie de l’effort reproducteur (Williams 1966a,b), toute augmentation de
l'investissement dans la reproduction courante se fera au détriment de la valeur
reproductrice résiduelle. Dans un tel contexte, la maximisation du succès reproducteur
à vie passera par une optimisation de l'investissement dans chaque épisode
reproducteur. L’effort reproducteur optimal sera donc déterminé par un équilibre entre
les bénéfices attendus d’une reproduction et les coûts pour les reproductions futures.
L’existence d’un conflit entre le succès d’une reproduction donnée et la valeur
reproductive résiduelle d’un organisme est une proposition clé, à la base du concept
de “coût de la reproduction” formulé par Williams (1966b). Cette notion théorique
simple et très attractive s’est très vite révélée difficile à tester dans la pratique et il en
a résulté une intense controverse sur l’existence même des coûts et sur les
méthodes adaptées pour les mesurer (Reznick 1985). En dépit de ces difficultés,
cette notion occupe désormais une position clé dans la théorie des traits d'histoire de
vie : depuis une quinzaine d'années, une littérature abondante est venue confirmer
l'existence des coûts de la reproduction (Bell 1980 ; Bells et Koufopanou 1986 ;
Stearns 1992 ; Lessells 1991) et leurs très fortes implications évolutives (Clutton-
Brock 1998). Les coûts de la reproduction sont désormais considérés comme des
contraintes majeures qui vont déterminer la réalisation de compromis adaptatifs entre
les traits d’histoire de vie et favoriser l'émergence de stratégies reproductrices
optimales (Williams 1966a,b ; Reznick 1992 ; Niewiarowski & Dunham 1994 ; Shine
& Schwarzkopf 1992 ; Clutton Brock 1998). Le point majeur des réflexions actuelles
ne repose donc plus sur l’existence des coûts de la reproduction mais plutôt sur
l’importance et la valeur de ces coûts, leur origine proximale et les mécanismes
physiologiques impliqués. On réalise ainsi une distinction fondamentale entre deux
grands types de coûts (Calow 1979) : on parle de coûts en survie lorsque la
15

eproduction affecte les probabilités de survie de l'organisme et d'autre part de coût
en fécondité si l'événement reproducteur influence les capacités reproductrices
futures de l'individu. Ces derniers peuvent être directs, via l’épuisement des réserves
mais ils peuvent aussi s’exprimer de façon plus détournée en affectant par exemple
le taux de croissance (Williams 1966b ; Calow 1979 ; Shine 1980). Cette
classification est en fait très artificielle car les deux composantes des coûts sont
rarement indépendantes. Ainsi, une dépense énergétique importante peut, par
exemple, affecter la survie d'un organisme si l'augmentation des prospections
alimentaires (nécessaires à la reconstitution des stocks de réserves corporelles)
l’expose d’avantage aux prédateurs (Bauwens & Thoen 1981 ; Brodie 1989). Les
coûts de la reproduction sont donc de natures multiples et un organisme réalisant un
effort reproducteur donné pourra être affecté de façon complexe par des coûts
d’origines variées
C. Dimension physiologique des compromis et
des systèmes de gestion de la ressource
Ces différents éléments indiquent que la façon de répartir la ressource disponibles a
d’importantes conséquences sur la valeur sélective (ou fitness) d’un organisme. Il est
important de bien comprendre que la réalisation de “compromis d’allocation” va
exister à l’échelle individuelle (Höglund & Sheldon 1998) et notamment au niveau de
son fonctionnement physiologique (Sinervo & Licht 1991). En effet, au sein d’une
espèce à reproduction sexuée, en raison d’une importante variabilité génétique, les
individus vont différer par de nombreux points de leur biologie et notamment au
niveau des “réglages physiologiques” fins qui soutendent l’allocation de l’énergie
dans des voies concurrentielles (comme reproduction donnée et future). L’existence
16

de compromis entre les traits d’histoire de vie souligne la présence de mécanismes
endocrines contraignant la covariation entre ces traits (Sinervo & Litch 1991 ;
Sinervo & DeNardo 1996). Les contraintes entre traits d’histoire de vie sont
directement liées à des actions hormonales dépendantes de nombreux gènes de
régulation ayant des effets multiples, opposés et pléïotropiques sur l’expression de
un ou plusieurs des traits (Rose & Bradley 1998). L’investissement reproducteur va
donc impliquer le rôle clé d’une intégration physiologique, s’exerçant au niveau de
l’organisme. Les systèmes neuroendocrines vont offrir une certaine plasticité de
fonctionnement. A l’échelle de l’organisme, ils vont permettrent l’intégration
d’informations complexes (environnementales et/ou endogènes) et vont régir en
conséquence les aspects majeurs de la reproduction comme l’engagement/
désengagement, le degré d’investissement parental. Sinervo & Svensson (1998)
soulignent que dans de nombreux cas, la plasticité observée dans les traits de vie
est véhiculée par les mêmes mécanismes endocrines (gonadotropines, stéroïdes
sexuels, glucocorticoïdes) qui soutendent les compromis adaptatif existant entre ces
traits. A l’échelle de l’espèce ou de la population, l’évolution ultime de compromis
adaptatifs impliquera l’action de la sélection sur ces facteurs de régulation. Ainsi, il
existe des interactions évidentes entre les causes proximales et les causes ultimes
des compromis (Sinervo & Svensson 1998). Les pressions de sélection vont donc
agir sur les mécanismes proximaux qui seront alors “modelés” au cours des temps
évolutifs. Cette approche “physiologique” et mécanistique des compromis est très
pertinente et constitue un complément nécessaire à l’approche génétique et
populationnelle classique (Rose & Bradley 1998).
De façon conjointe, les pressions de sélection vont également agir sur le
niveau global d’énergie disponible pour la reproduction et ce, en façonnant les
systèmes d’acquisition et de gestion de l’énergie. En effet, pour compenser les
17

esoins énergétiques particulièrement élevés de la reproduction, la plupart des
organismes vont tenter d’acquérir de plus grande quantité de ressource (Jönsson
1997). Deux tactiques de compensation sont généralement identifiées : les
systèmes de reproduction basés sur l’alimentation courante (“income breeders”) et
d’autres sur la réalisation d’un capital de réserves (“capital breeders”) (Drent & Daan
1980, Jönsson 1997). Dans le premier cas, la reproduction est accompagnée d’une
augmentation de l’acquisition des ressources alimentaires qui vont être directement
allouées dans la reproduction. Dans le second, les ressources sont stockées sous
forme de réserves qui serviront ultérieurement de support énergétique principal pour
la reproduction. La formulation de Drent & Daan est en fait quelque peu artificielle car
il existe un continuum entre ces stratégies (Doughty & Shine 1997). Cependant, la
réalisation d’une telle dichotomie est très utile (Price et al. 1988, Martin 1995) et peut
être appliquée à de nombreux organismes. En outre, elle permet de souligner un
aspect important des stratégies reproductrices basée sur la dimension temporelle de
l’acquisition et de l’allocation de l’énergie (Fisher 1930). En effet, les reproductions
sur revenus ou sur réserves vont impliquer des voies biochimiques identiques (dans
les deux cas il y a mise en réserve) mais vont différer dans la durée du stockage des
molécules avant utilisation (Bonnet et al. 1998). Une telle variation repose
directement sur la sélection de soubassement physiologiques qui vont orienter les
compromis d’allocations entre reproduction, stockage de réserves et besoins de
maintenance (Dought & Shine 1998). L’identification des stratégies d’utilisation de la
ressource est évolutivement très importante, notamment parce que les avantages et
inconvénients de la reproduction sur revenus ou sur réserves vont varier selon les
taxons (Jönnson 1997 ; Bonnet et al. 1998). En outre, elle est indispensable lorsque
l’on cherche à évaluer les coûts de reproduction, qui vont alors se manifester de
18

façons différentes selon que la reproduction impose ou non une longue période de
constitution de réserves.
D. Connexion entre effort reproducteur et coût
de la reproduction
La compréhension de l’évolution des traits d’histoires de vie, tel l’investissement
reproducteur (facteur ultime), implique donc l’intégration des facteurs proximaux
sous-jacents (physiologie de l’organisme). Dans ce cadre, il convient d’examiner
finement la manière dont les modifications éco-physiologiques de la reproduction
vont pouvoir se manifester en terme de coûts démographiques. Les coûts de la
reproduction peuvent avoir des origines multiples (prédation, dépenses énergétiques,
effets pathologiques de taux hormonaux élevés, les effets de sénescence...). Dans
les populations naturelles, les coûts peuvent être d’origine écologique, physiologique
ou de façon plus rationnelle une combinaison de ces deux composantes. Pour
étudier l’influence des différentes composantes des coûts de la reproduction sur la
partition de l’énergie et de l’effort reproducteur, il est donc fondamental d’identifier la
nature des coûts impliqués et les mécanismes par lesquels ils se manifestent
(Sinervo & Svensson 1998). Dans un tel contexte, il devient crucial de bien
distinguer les deux notions, a priori très proches, d’effort reproducteur et de coût de
la reproduction (Tuomi et al. 1983 ; Niewiarowski & Dunham 1994). L'effort de
reproduction représente l’allocation “brute” réalisée par l'organisme en temps,
énergie ou matière dans la transmission de ses gènes (Clutton Brock 1991; 1998).
Par extension, cet effort a souvent été assimilé à des coûts car l'animal va "allouer"
de l'énergie et du temps dans la production de jeunes. Cependant, pour bien
percevoir la nuance entre ces deux notions, il est nécessaire d’avoir une perspective
19

à long terme intégrant la vie reproductrice de l’organisme. En effet, à la différence de
l’effort reproducteur qui est simultané à la reproduction, les coûts peuvent s’exprimer
avec un certain décalage temporel, via les effets démographiques sur la survie ou
sur les capacités reproductrices ultérieures (Williams 1966b ; Clutton-brock 1998).
Ces composantes (coût et effort) sont donc liées mais leur degré de connexion peut
être variable. Ainsi, on peut envisager des situations où coût et effort sont
directement connectés, c’est-à-dire où l’augmentation de la fécondité engendrera
une augmentation proportionnelle de l’effort et des coûts. Ce type de situation est
particulièrement fréquent dans des systèmes avec soins parentaux où le nombre de
jeune va directement affecter le degré d’investissement parental (Clutton-brock
1991). Dans d'autres, cas la relation peut être de type “tout ou rien”, notamment
quand la reproduction impose des activités qui sont d’emblée coûteuses (comme des
migrations) et ce indépendamment de l’effort reproducteur courant et du nombre de
jeunes produits. La nature de la relation entre coût et effort n'est donc pas forcément
linéaire ou constante (Niewiarowski & Dunham 1994). En outre, elle dépend
largement de facteurs environnementaux et un effort reproducteur important ne se
traduit pas nécessairement par des coûts élevés si le contexte est favorable
(Congdon et al. 1982 ; Tuomi et al. 1983).
La réalisation de cette distinction est importante car la relation existante entre
investissement reproducteur et les coûts associés peut profondément déterminer
l’évolution d’une stratégie d’acquisition et d’allocation optimale de la ressource (Bull
& Shine 1979 ; Shine & Schwarzkopf 1992; Niewiarowski & Dunham 1994,
1998 ; Sinervo & Svensson 1998). L’intégration des différentes composantes des
coûts (dépendantes ou indépendantes de la fécondité) peut être un facteur clé dans
la compréhension de l’évolution de la fréquence de reproduction dans les systèmes
itéropares. En effet, si la majorité des organismes se reproduisent de façon
20

annuelle, il existe néanmoins des espèces où les différents individus de la population
vont “sauter” des opportunités de reproduction et se reproduire de façon asynchrone
tous les deux ans ou plus (Bull & Shine 1979). Ce type de système à faible
fréquence de reproduction (LFR = Low Frequency of Reproduction) est souvent
observé chez des espèces dont le cycle impose des activités particulières,
caractérisées par des dépenses en énergie, temps et/ou survie qui s’expriment de
façon fixe c’est à dire indépendamment de la fécondité. Parmi ces activités, Bull &
Shine (1979) citent les migrations de reproduction, les soins/défense des oeufs (à la
ponte dans son ensemble), la rétention des oeufs dans les voies génitales
(viviparité). Les migrations de reproduction imposent par leur nature, des contraintes
indépendantes de la fécondité et du succès reproducteur de l’organisme. Une telle
déconnexion existe aussi dans le cas de soins prodigués à la progéniture dans son
ensemble (défense des pontes, gestation) au cours desquelles les femelles cessent
souvent de manger et réduisent leur déplacements (Bull & Shine 1979). Si ces
composantes des coûts sont très élevées, l’organisme aura avantage à se reproduire
de façon alternée, éviter ainsi ces coûts ”fixes” et profiter d’une année de repos pour
investir plus d’énergie et d’assurer un meilleur succès dans l’évènement reproducteur
suivant. Ce système est particulièrement avantageux dans un habitat pauvre en
ressource, où la dépense énergétique dans les activités de reproduction est très
élevée par rapport au niveau de ressource disponible. L’évolution vers un
système “LFR” permettrait alors une augmentation de la fécondité moyenne. Si
l’hypothèse de Bull et Shine (1979) n’a pas été testée empiriquement, elle suggère
néanmoins que l’identification des différentes composantes des coûts de la
reproduction et de la relation entre ces composantes et l’effort reproducteur est très
importante à réaliser. Une telle approche permet de déterminer la signification
adaptative de la répartition de l’effort reproducteur dans la vie de l’organisme et doit
21

donc être intégrée dans l’analyse de l’évolution des stratégies de reproduction
extrêmes (seméliparité).
E. Intérêt d’une perspective ectothermique
Le test d’hypothèses évolutives sur les stratégies reproductrices et notamment la
fréquence des reproductions va impliquer le recours à une démarche à la fois
méticuleuse et intégrée. En effet, considérant la complexité des systèmes vivants et la
diversité des pressions de sélection auxquelles sont soumis les organismes, l’examen
de l’influence évolutive des coûts de reproduction sur l’effort reproducteur sera
optimisé si l’on sélectionne des situations biologiques simples où le nombre de
variables confondantes sera limité. Idéalement, pour élucider la nature de la relation
entre effort de reproduction et coût, il va être nécessaire de caractériser clairement les
sources d’investissements (Fisher 1930) et d’identifier leur chronologie exacte (Bonnet
et al. 2000a). Malheureusement, dans la plupart des systèmes biologiques, la
reproduction est un phénomène très complexe et multifactoriel. Chez les
endothermes par exemple (oiseaux et mammifères), les grandes étapes d’allocation
directe de l’énergie dans la progéniture (formation des oeufs, incubation, gestation)
sont souvent réalisées en synchronie avec des activités complexes et moins
étroitement liées à l’investissement gamétique comme par exemple des interactions
sociales avec les congénères (hiérarchie de dominance, défense du territoire) ou des
comportements sexuels très élaborés (construction d’un nid). Il va donc résulter une
“superposition” entre ces sources de dépenses énergétiques et les aspects plus
proximaux de l’investissement reproducteur. Une telle complication va rendre difficile,
voir impossible, l’identification des contributions respectives des différentes activités
reproductrices en terme de demande énergétique et de coûts associés. En outre, les
endothermes manifestent généralement des soins parentaux très élaborés pouvant
22

impliquer une collaboration complexe entre les partenaires (Lequette & Weimerskirch
1990 ; Clutton-Brock 1991). La mesure des coûts de la reproduction va faire intervenir
l’évaluation de ces soins et de leurs variations avec la fécondité. Ceci va
sérieusement compliquer l’étude des déterminants du succès reproducteur individuel.
Enfin, les endothermes sont caractérisés par de très forts besoins métaboliques et
l’organisme est contraint de subvenir, en parallèle à toutes ces activités, à des
besoins de maintenance très élevés (Else & Hulbert 1981; Hulbert & Else 1989). Les
vertébrés ectothermes diffèrent fondamentalement des endothermes par de faible
besoins énergétique pour la maintenance et des niveaux métaboliques qui ne
dépassent pas 20% de ceux d’un mammifère ou d’un oiseau de taille comparable
(Pough 1980). La régulation de la température corporelle est principalement réalisée
de façon comportementale et n’implique pas de dépenses accrues dans la production
de chaleur endogène. En conséquence, l’énergie assimilée va pouvoir être facilement
convertie en biomasse (Bradshaw 1997). Cette grande efficience de conversion va
faciliter d’autant l’étude du budget énergétique (plus grande lisibilité des prises
énergétiques). De plus, il existe une grande diversité dans les soins parentaux chez
ces organismes avec généralement des soins post-nataux très réduits, voir inexistants
(Clutton-Brock 1991). Un tel contexte offre de nombreuses simplifications avec
notamment une concentration de l’effort reproducteur avant la mise-bas ou la ponte
Ce cadre s’avère donc très favorable pour l’identification précise des activités
reproductrices et de leurs implications énergétiques et écologiques.
Dans ce travail nous nous intéressons aux stratégies reproductrices
“extrêmes” (effort reproducteur élevé associé à un petit nombre de reproduction de
l’organisme). Notre objectif principal est d’identifier des éléments de réponse
originaux sur les facteurs favorisant l’évolution vers des systèmes semélipares. Dans
ce cadre théorique, il faut noter que les possibilités d’investigations sont très
23

variables selon que l’on s’adresse à des modèles endothermes (oiseaux et
mammifères) ou ectothermes. En effet, la grande majorité des systèmes semélipares
au sens strict (mort après un seul épisode reproducteur chez un ou les deux sexes) a
été observée chez des organismes ectothermes avec de nombreux exemples chez
des invertébrés (Boyle 1983, 1987 ; Andries 2001) et certains vertébrés (Crespi &
Teo 2002) . Les rares systèmes semélipares décrits chez les endothermes sont
limités à quelques espèces de marsupiaux avec une restriction de la seméliparité aux
mâles (Oakwood et al. 2001). Les vertébrés ectothermes sont des modèles
particulièrement intéressants qui offrent une grande variabilité dans le “degré”
d’itéroparité avec l’existence d’un continuum entre seméliparité et l’itéroparité
(Schaffer & Elson 1975 ; Crespi & Teo 2002). Il existe ainsi des espèces de lézards
avec des populations itéropares et d’autres semélipares (Bradshaw 1997) : un cas de
figure exceptionnel pour une analyse comparative des forces sélectives et des
mécanismes physiologiques impliqués dans cette transition. Le degré d’itéroparité
peut aussi varier au sein d’une même population, notamment dans des systèmes où
la fréquence de reproduction est faible ( “LFR”) et varie selon les individus. A de très
rares occasions, on pourra ainsi observer la coexistence intra-populationnelle,
d’individus à trajectoire itéropares et d’autres semélipares. C’est le cas notamment
de la vipère aspic (Vipera aspis) au Nord de son aire de distribution. Une telle
situation fournit des opportunités uniques pour tester l’hypothèse de Bull et Shine
(1979) sur la nature des coûts de la reproduction et l’évolution de systèmes a faible
fréquence des épisodes reproducteurs. Du fait des simplifications énergétiques de
l’ectothermie, il devrait être possible de décrire finement le système d’allocation de
l’énergie et le mettre en relation avec ces aspects démographiques. Dans cette
perspective, nous avons examiné la stratégie reproductrice déployée par la vipère
24

aspic femelle dans l’0uest de la France et ce travail se structure en cinq grandes
parties:
1. Présentation de l’espèce, écophysiologie et cycle reproducteur
2. Le système d’allocation de l’énergie dans la reproduction : description et
avantages dans un environnement contraignant
3. Caractérisation des coûts associés à la reproduction, amplitude et relation avec la
fécondité
4. Examen des facteurs déterminant la seméliparité femelle et les conséquences
démographiques
5. Discussion-conclusion: reproduction sur réserves, coûts de la reproduction et
évolution vers la seméliparité
25

I. Présentation de l’espèce
et des méthodes d’études
Vipère aspic mon amie, qu’as tu fais de ta vie ?
Noël Guillon, Poète-philosophe de Haute Saintonge.
26

A. Résumé du chapitre:
Si la biologie de la reproduction des serpents est encore mal connue, la famille des
viperidés est probablement la plus intensément étudiée. Ce groupe très homogène est
caractérisé par une morphologie trapue, un appareil venimeux très perfectionné et un
mode de reproduction généralement vivipare. Le mode de vie est basé sur la chasse
à l’affût et il existe une prédispostion au stockage de réserves corporelles pour la
reproduction (“capital-breeding”). La vipère aspic (Vipera aspis) est un petit serpent
européen qui présente, dans l’ouest de la France, une stratégie de reproduction très
particulière. Dans cette région, la maturité des femelles est tardive, l’espérance de vie
courte et la fréquence des reproductions faible (tous les 2-4 ans). Les réserves
semblent jouer un rôle très important et il existe un seuil minimum de condition
corporelle pour permettre l’engagement reproducteur. La reproduction est
caractérisée par de fortes contraintes énergetiques et écologiques pendant la
vitellogénèse et la gestation. L’investissement maternel est donc généralement très
élevé et la masse des portées produites dépasse souvent celle de la mère après la
mise bas. Le nombre des reproductions est réduit et la vie reproductrice des femelles
présente une véritable tendance semélipare. Cette situation contraste beaucoup avec
celle des mâles chez qui l’investissement reproducteur est graduel et repété au cours
de la vie (itéroparité). En occupant une position intermédaire entre itéroparité et
seméliparité, la stratégie des vipères aspic femelles offre une excellente opportunité
d’examiner les facteurs favorisant un investissement reproducteur élevé et la
transition vers les systèmes semélipares.
27

B. Position systématique et phylogénie
L’origine évolutive des serpents est discutée depuis plus de 130 ans et leur position
au sein de l’orde des squamates (qui regroupe lézard et serpents) est encore l’objet
de débats (Coates & Ruta 2000). Il existe actuellement plus de 2700 espèces de
serpents qui se répartissent en trois groupes trés inégaux: les scolécophidiens, les
anilioïdes et les macrostomates. Au sein des macrostomates sont regroupées les
formes de serpents les plus “dérivées”, caracterisées par une très grande mobilité
des mâchoires supérieures et inférieures permettant l’ingestion de proies de grande
tailles. Ce groupe phylétique comprend la plus grande diversité des espèces
actuelles avec notamment des taxons ayant développés des appareils venimeux
élaborés comme la famille des viperidés, à laquelle appartient la vipère aspic (Vipera
aspis, Linné 1758). Cette famille s’individualise par de nombreux critères
anatomiques, ostéologiques, histologiques et son monophyllétisme n’a jamais été
remis en question par les données moléculaires (Ineich 1995). Cette famille est
caracterisée par une morphologie assez trapue, un mode de chasse basé sur l’affût
et une prédisposition au stockage de réserves lipidiques pour la reproduction
(“capital breeding”, Madsen et Shine 1992a, Brown 1993; Martin 1993). La répartiton
actuelle des vipéridés suggère une origine asiatique avec une différentiation pendant
l’ère Tertiaire à la fois dans la partie orientale (crotalinae) et occidentale (viperinae)
du continent Eurasien (Ineich 1995). En europe occidentale, la vipère aspic (Vipera
aspis), la vipère ammodytes (Vipera ammodytes, Linné 1758) et la vipère de lataste
(Vipera latastei, Boscà 1879) constituent un groupe d’espèces très proches qui se
sont probablement différenciées durant le Pleistocène dans chacune des grandes
péninsules méditerranéennes (Saint Girons 1997). Parmis ces espèces, la vipère
aspic est celle qui occupe la partie la plus nordique de l’aire ouest-paléarctique.
28

C. Répartition de l’espèce et des populations
d’études
La vipère aspic est un petit serpent venimeux (en moyenne 55 cm de longueur
totale) avec une morphologie trapue et une tête relativement bien distincte du corps.
Cette espèce localement abondante, est caracterisée par un haut polymorphisme qui
s’exprime dès le niveau intra-populationnel. Cinq sous-espèces sont classiquement
reconnues : Vipera aspis aspis (nord et centre de la France), Vipera aspis
francisiredi, (Nord et centre de l’Italie), Vipera aspis atra (Centre ouest de la Suisse,
Nord-ouest de l’Italie), Vipera aspis hugyi (Sud de l’Italie et Sicile), Vipera aspis
zinnekeri (Pyrénées Françaises et Espagnoles). Le statut de ces sous-espèces reste
cependant incertain et fait l’objet de discussions actives (Zuffi 2002). Ce travail porte
sur la sous espèce nominale (Vipera aspis aspis) dans des populations de l’Ouest de
la France, au Nord de l’aire de répartition de l’espèce (carte ci-dessous).
Figue 1. Répartition de la vipère aspic (d’après Naulleau 1997)
29

D. Biologie de la reproduction
1. Généralités chez les squamates
Les squamates sont des vertébrés amniotes qui produisent des oeufs de grandes
taille et chargés en vitellus (type mégalecithe). La reproduction impose de fortes
contraintes à l’organisme femelle qui doit investir d’importantes quantités d’énergie
dans des pontes de grande taille. Ce groupe est caratérisé par une importante
diversité du mode de reproduction avec un véritable continuum entre oviparité et
viviparité (Shine 1985). La rétention des oeufs dans les voies génitales et la
production de jeunes (viviparité) génère des contraintes très spécifiques qui viennent
augmenter la hauteur de l’investissement maternel. Pendant la gestation, il existe
notamment un déplacement des préférences thermiques vers des températures
élevées (Shine 1980). Les femelles vont alors présenter un changement dans leur
schéma d’activité avec de longues périodes d’exposition. En outre, le
développement embryonnaire va souvent affecter la mobilité des femelles qui sont
ainsi plus exposées à la prédation (Shine 1980, Bauwens et Thoen 1981). Les soins
parentaux post-nataux sont la plupart du temps très réduits et les jeunes sont
autonomes dès la naissance. L’investissement paternel dans la reproduction est
beaucoup moins complexe et il se limite souvent à la recherche et la fertilisation de
partenaires sexuels (avec le cas échéant des combats avec des compétiteurs). Le
système d’appariement est en général simple (Duval et al. 1992, 1993) et les
investisssements énergétiques spécifiques des deux sexes ne sont pas masqués
par des interactions comportementales complexes entre partenaires.
L’éco-physiologie de la reproduction chez les squamates est un domaine
encore très jeune et en pleine émergence (Crews & Gans 1992). Si de nombreuses
30

études ont été conduites sur les lézards (physiologie comportementale notamment),
la biologie de la reproduction des serpents est caracterisée par d’énormes lacunes.
Ainsi, en dépit de l’étonnante diversité des modes de reproduction des serpents,
l’état des connaissances actuelles ne permet pas une approche physiologique
comparative des soubassements endocrines. Toute tentative de synthèse serait
d’ailleurs très risquée du fait de la concentration des travaux sur des espèces de
zones tempérées alors que l’immense majorité des serpents occupe des zones
tropicales et équatoriales (Seigel et Ford 1987).
Les vipéridés des zones tempérées sont probablement les serpents les plus
intensément étudiés (Saint Girons 1949, 1952, 1957a,b, 1975 ; Fitch 1960 ; Klauber
1972 ; Brown 1991 ; Madsen & Shine 1993 ; Martin 1993) et il existe de bonnes
informations sur les cycles reproducteurs en général très uniformes et fortement
saisonniers. La vipère aspic a fait l’objet de nombreux travaux et constitue un des
serpents les mieux connus au monde (Saint Girons 1949, 1957a,b; Saint Girons &
Duguy 1992; Bonnet 1996; Naulleau et al. 1999; Bonnet et al. 1999b, 2000a,b,
2001b, 2002b) notamment en ce qui concerne les soubassements endocrines de la
reproduction (Bonnet et al. 1994; Bonnet 1996; Bonnet et al. 2001a).
2. Eco-physiologie de la vipère aspic
La femelle vipère aspic peut être considérée comme un reproducteur sur réserves
“typique”. L’entrée dans la reproduction (recrutement des follicules) s’effectue dès la
sortie d’hivernage au début du printemps et la condition corporelle (masse ajustée
par la taille) calculée à ce moment offre une bonne estimation du niveau des
réserves lipidiques (corps gras). Il existe un seuil minimal de réserve pour permettre
l’engagement reproducteur. Ainsi, seules les femelles ayant accumulé des réserves
31

vitellogéniques supérieures à un niveau minimum vont se “lancer” dans la
folliculogénèse. Il en résultera des reproductions asynchrones avec, chaque année,
la coexistence d’une fraction de femelles reproductrices (40% en moyenne) avec des
femelles non-reproductrices (Bonnet & Naulleau 1994,1995).
L’accouplement a lieu dès la sortie d’hivernage uniquement chez les femelles
avec une condition corporelle suffisante (Bonnet & Naulleau 1994). La mobilisation
du capital de réserves lipidiques va fournir les substrats nécessaires pour la synthèse
de vitellogénine par le foie et la déposition du jaune dans les follicules en croissance
(Bonnet 1996). Il existe notamment une relation positive entre la masse des corps
gras abdominaux et le nombre de follicules en croissance (Saint Girons & Naulleau
1981). La vitellogénèse (synthèse du vitellus) constitue une étape clé de l’effort
reproducteur et s’accompagne de profonds bouleversements physiologiques. Les
travaux empiriques et expérimentaux indiquent le rôle clé de l’oestradiol dans la
mobilisation des réserves et le déclenchement de la croissance folliculaire (Bonnet et
al. 1994, Bonnet 1996). De très nombreux follicules sont en général recrutés et
après cette phase initiale (début du printemps), la mort d’une fraction des follicules
(atrésie, Méndez de la Cruz 1993) va déterminer de façon proximale le nombre
d’oeufs ovulés. Lorsque la nourriture est disponible, les femelles reproductrices
continuent a s’alimenter pendant cette période (Saint Girons & Naulleau 1981).
L’ovulation et la fécondation ont lieu pendant la première quinzaine de juin. Les oeufs
ovulés vont alors s’hydrater de façon importante et le développement embryonnaire
dans les voies génitales va s’étendre ensuite jusqu’au début de l’automne.
Pendant la gestation, la vie des femelles reproductrices va être affectée par de
profonds changements. Elles vont presenter des optimum thermiques élevés et
passer significativement plus de temps que les non reproductrices en
thermorégulation (Bonnet & Naulleau 1996). En combinaison avec la
32

thermorégulation les femelles reproductrices deviennent nettement plus sédentaires
et leur domaine vital passe ainsi de 3000 m 2 (en début de gestation) à 300 m 2 à
partir de la mi-juillet (Naulleau et al. 1996). Enfin on constate durant la gestation, une
réduction, voir un arrêt des prises alimentaires (Saint-Girons 1952, 1979). La
gestation entraîne donc des modifications profondes (écologie, physiologie) et non
graduelles chez les individus qui ont pris la "décision" de s’engager dans la
reproduction (Bull et Shine 1979). Un travail récent (Bonnet et al. 2001a) suggère
l’importance de la progestérone dans le maintien de la gestation chez cette espèce.
La mise bas ne dure que quelques minutes et les femelles produisent de 2 à 22
vipéreaux de 6 g en moyenne. La masse de la portée représente de 30 à 120% de la
masse post-parturiente femelle (Bonnet 1996) ce qui constitue un effort reproducteur
énorme pour un vertébré amniote. Après la mise-bas, les femelles sont très
amaigries et la condition corporelle post-partum constitue un bon indicateur de leur
degré d’émaciation. Une à plusieurs années seront alors nécessaire pour accumuler
un stock de réserves suffisant pour se reproduire à nouveau.
La mise en place d’un suivi populationnel aux Moutiers en Retz (Loire
Atlantique) depuis 1992 a permi d’apporter des informations précises sur le nombre
d’opportunités de reproduction pour les femelles dans cette région. Le temps
nécessaire à la constitution des réserves corporelles induit une maturité tardive (2.5
à 3.5 ans avec une durée de vie de 5 ans en moyenne). Les données de suivi
individuel suggèrent un faible nombre d’opportunités reproductrices avec une
majorité d’individus à trajectoire semélipare (une seule reproduction) et d’une fraction
d’individus itéropares. Ces derniers ne présentent en général que deux à trois
reproductions dans leur vie. Pour une minorité d’individus, quatre reproductions ont
été enregistrées. Dans tous les cas de figure, la fréquence des reproductions est
faible avec des cycles compris entre deux et quatres ans. Nous sommes donc en
33

présence d’un système à très faible fréquence de reproduction (“LFR”, Bull et Shine
1979) avec une tendance semélipare.
Femelles reproductrices
Accoupl
Alimentation
Hibernation Vitellogénèse O Gestation MB
Hibernation
Janvier Février Mars Avril Mai Juin Juillet Août Septembre Octobre Novembre Décembre
Femelles non-reproductrices
Mâles
Hibernation
Alimentation
Hibernation
Janvier Février Mars Avril Mai Juin Juillet Août Septembre Octobre Novembre Décembre
Hibernation
Accoupl
Alimentation
Spermiogénèse continue?
Hibernation
Janvier Février Mars Avril Mai Juin Juillet Août Septembre Octobre Novembre Décembre
Figure 2. Phénologie des cycles reproducteurs mâle et femelle dans l’ouest de la France. Les
grandes étapes ont été représentées par des rectangles dont la position temporelle est variable selon
les années et les individus. Seule l’ovulation semble être un phénomène assez fixe (première
quinzaine de Juin). Chez les mâles, le cycle reproducteur est annuel et la production de
spermatozoïdes semble continue. Dans la zone d’étude le cycle reproducteur des femelles est
toujours supérieur à un an. Chaque année, la population est donc composée d’un fraction de femelles
reproductrices et d’une autre de non reproductrices. (Accoupl: accouplement; O: ovulation, MB: mise
bas).
34

La situation des femelles contraste beaucoup avec celle des mâles. En effet,
chez ces derniers, la spermiogénèse (maturation des spermatozoïdes) semble plus
ou moins continue dans l’année avec deux vagues : la première en Avril-Mai alors
que les accouplements ont déja commencé, et la seconde en Août -Octobre
(Naulleau 1997). Les spermatozoïdes produits pendant l’automne sont stockés dans
les canaux déférents jusqu’aux accouplements de printemps. Le système
d’appariement est basé sur la polygynie simultanée et il n’existe pour ce sexe aucune
forme de soin parental. La période sexuelle se limite à un ou deux mois d’activités de
recherche et de fertilisation de partenaires avec parfois des combats ritualisés entre
rivaux (Bonnet et al. 2002b). Pendant cette phase, les mâles deviennent
anorexiques et les réserves lipidiques vont fournir le support énergétique nécessaire.
Les mâles sont donc également des reproducteurs sur réserves , cependant,
l’engagement dans la reproduction ne va pas être contrôlé par un seuil élevé de
condition corporelle. Chez ce sexe, le système l’allocation de l’énergie est beaucoup
plus graduel (Aubret et al. 2002, Bonnet et al. 2002b). Ainsi, un individu avec peu de
réserves pourra s’accoupler efficacement avec une partenaire. En outre, si le niveau
d’amaigrissement pendant la phase sexuelle devient trop élévé, un mâle pourra
aisément se désengager de la reproduction et entreprendre des prospection
alimentaires (Aubret et al. 2002). Les coûts de la gamétogénèse étant très réduits,
les mâles sont donc beaucoup moins contraints que les femelles en terme
d’investissement énergetique et temporel dans la reproduction. Le cycle
reproducteur est généralement annuel (Vacher-Vallas 1997).
35

Effort:
Résumé des grandes étapes de l’effort
reproducteur et des contraintes associées:
I. La folliculogénèse (3 mois):
Figure 3. folliculogénèse observée par résonance magnétique nucléaire (RMN)
Mobilisation des réserves lipidiques, protéiques pour la synthèse de vitellus
Recrutement et croissance folliculaire: allocation de temps, d’énergie et de
Contraintes:
matière dans la formation des oeufs
Besoins thermiques élevés, thermophilie,
Exposition aux prédateurs.
36

Effort:
II. La gestation (2 à 3 mois):
Figure 4. Vipère aspic gestante en thermorégulation
Déplacement vers des optimums thermiques élevés pour l’embryogénèse
Epuisement des réserves et catabolisme protéique
Contraintes:
Figure 5. Comparaison de la partie postérieure d’une vipère aspic avant
(en haut) et après (en bas) la mise-bas
Thermophilie accentuée, exposition très fréquente.
Diminution des déplacements et des prises alimentaires.
Etat physiologique critique après la parturition
37

E. Méthodes d’étude
Le présent travail est basé sur la combinaison de deux approches complémentaires :
le suivi longitunidal d’une population et la mise en place d’expérimentation en
captivité.
1. Approche longitudinale:
Un suivi par capture-marquage-recapture d’une population de vipère aspic a été
lancé par Guy Naulleau en 1992 aux Moutiers en Retz (47 o 03N'; 02 o 00W'). La
zone d’étude de 33 hectares est une mosaîques de prairies, de plantations et de
friches en régénération. Le climat est de type océanique-tempéré. Chaque année, la
zone est régulièrment patrouillée par une à quatre personnes, essentiellement au
printemps et en fin d’été (capture de femelles reproductrices). Ce suivi constitue la
plus grande base de données existante sur la vipère aspic avec 1032 individus
marqués et plus de 10000 données de captures-recaptures. L’effort de recherche
dépasse 700 jours et 550 heures de terrain. Mon implication active dans cette
étude (terrain et analyse de données) a débuté en 1996.
Les animaux sont capturés à vue, sexés par éversion des hémipénis. Tout
individu capturé est pesé au gramme près, mesuré au demi cm près (en longueur
totale: “LT” et distance museau-cloaque: “SVL”) et marqué individuellement. Le
marquage des adultes s’effectue à l’aide de puces electroniques (PIT-TAG : passive
integrated transponder, TX1400L, Rhône Mérieux, 69002 LYON France, product of
Destron/IDI Inc). Les jeunes individus sont marqués par ablation codifiée d’écailles
ventrales. Chaque serpent est ensuite relaché au point exact de capture. La
population d’étude est relativement bien isolée des populations avoisinantes
(Vacher-Vallas 1997, Bonnet et al. 2000a). Cette espèce est très philopatrique
38

(Naulleau et al. 1996) et tout individu non capturé pendant une longue période (> 2
ans) est considéré comme mort (Bonnet et al. 2002a)
Le statut reproducteur des femelles est déterminé selon différentes méthodes
qui dépendent de la période de capture. Au début du printemps, les femelles
présentant une condition corporelle supérieure au seuil sont considérées comme
reproductrices (voir Bonnet & Naulleau 1994 pour des détails sur la méthode). Plus
tard dans l’année, la palpation de l’abdomen permet de détecter et compter des
follicules en croissance (vitellogénèse) ou des embryons (gestation) (Fitch 1987,
Naulleau & Bonnet 1996). A la fin de l’été, les femelles reproductrices sont capturées
et transportées au laboratoire. Elles sont alors maintenues en captivité dans des
boîtes individuelles et pesées tous les deux jours jusqu’à la parturition. Cette
méthode nous a permis de récolter des données détaillées sur plus de 190 mises
bas. Les différents éléments de la portée sont caractérisés (nouveau-né, mort-né,
embryon, oeuf non-développé) et pesés au dixième de gramme près. Les nouveaux-
nés et les morts-nés sont mesurés au demi cm près (en longueur total: ‘LT” et
distance museau cloaque:” SVL”). La masse tolale de la portée (“litter mass”) est
calculée en sommant l’ensemble des différents élements produits. Pour calculer la
masse de la portée viable (“fit littermass”) seuls les nouveaux-nés ont été considérés
(Gregory et al. 1992). Une fois la mise-bas achevée, la femelle est palpée pour
déterminer la présence éventuelle d’éléments non expulsés.
De 1993 à 1996 certains individus ont été équipés avec des émetteurs placés
par ingestion forcée dans l’estomac et suivis par télémétrie. Les animaux équipés
(principalement des femelles reproductrices et quelques non-reproductrices) ont été
suivies sur des périodes variables (inférieure à deux mois) jusqu’à régurgitation de
l’émetteur ou capture en vue d’obtenir la mise bas. Pendant toute la période de suivi
télémétrique, 1 à 4 relevés journaliers de positions et de températures ont été
39

éalisés. Les données sur les déplacements n’ont pas été exploitées dans le cadre
de ce travail. En revanche, les données de température nous ont permis d’étudier
l’influence du statut reproducteur sur les préferences thermiques.
Facteurs environnementaux examinés
Pendant la période d’étude, nous avons obtenu des données précises sur des
variables environnementales particulièrement importantes: l’abondance des proies et
les fluctuations climatiques.
Abondance des proies : la vipère aspic se nourrit essentiellement de campagnols
(Microtus arvalis Pallas) dont les populations fluctuent avec un cycle typique de trois
à quatre ans (Delattre et al. 1992). Ces variations d’abondance affectent directement
la consommation des serpents et donc la proportion d’animaux capturés avec des
indices de repas (palpation de proies, crottes avec poils). La proportion annuelle de
serpents capturés avec des restes de repas est disponible pendant toute la durée
d’étude à l’exception de 1992 et nous avons utilisé ce paramètre comme un indice
d’abondance des proies. Cet estimateur donne des resultats très consistants avec
les données de piégages de rongeur dans cette région (Bonnet et al. 2001b;
Salamolard et al. 2000). De plus amples précisions et une discussion de cette
méthode sont disponible dans la section matériel et méthode de l’article 3.
Températures : la zone d’étude est située au nord de l’aire de répartition de l’espèce
et les conditions climatiques sont contraignantes en comparaison avec les
populations méridionnales. Ainsi, en Italie, les mises bas sont observées à la mi-
Juillet, soit deux mois plus tôt que dans l’ouest de la France (Zuffi et al., 1999).
Pendant toute la période d’étude, les maximum thermiques journaliers sous abris
(°C) ont été mesurés (données Météo France, station d’enregistrement de Pornic).
40

Deux périodes biologiques ont été identifiées: la saison active (de Mars à Octobre) et
la gestation (de mi-Juin à fin Août).
2. Approche expérimentale en captivité
En parallèle avec les analyses des données des Moutiers, nous avons realisé
plusieurs expérimentations dans les enclos d’élevages de la station biologique de
Chizé (Forêt de Chizé, Deux-Sèvres, 46°07’ N, 00°25’ W). Dans tous les cas, les
vipères ont été capturées dans des populations peu éloignées de la zone d’étude
des Moutiers et principalement en Vendée. Après la capture, certaines femelles sont
ramenées au laboratoire, pesées, mesurées et marquées individuellement par
ablation d’écailles ventrales.
Conditions de captivité
Les femelles sont ensuite placées dans des enclos extérieurs (5 x 3m) recréant le
milieu naturel de la vipère et exposés aux conditions climatiques de la station
biologique de Chizé. Chaque enclos est équipé avec le même nombre de plaques
en Fibrociment (50x50cm) qui servent de refuges. L’eau est fournie ad-libitum et la
végétation (Poacées) est maintenue haute pour fournir abris et fraicheur. Dans le cas
des captures au sortir de l’hivernage (i.e. au début des activités sexuelles), les
accouplements ont été obtenus en captivité. Pour chaque femelle, un séjour de dix
jours en contact avec de nombreux mâles a été organisé dans un enclos intérieur
(2.5 x 1.5m) avec une source de chaleur (radiants 1000w) et de l’eau ad-libitum.
Cette méthode permet d’obtenir des copulations très facilement avec des
accouplements observés quelques heures après la capture.
41

Alimentation
Un fois placés dans les enclos extérieurs les animaux sont nourris avec des souris
d’élevage fraichement tuées et posées, quand les conditions météorologiques sont
jugées favorables, à proximité des abris. A chaque nourrissage les proies offertes
sont standardisées au sein de chaque enclos (± 1g). La consommation des proies
est enregistrée par observation directe ou par des moyens plus indirectes quand
l’observation n’est pas possible (palpation de souris dans l’estomac ou changement
brutal de masse). Les souris non ingérées sont retirées très rapidement des enclos
(6 à 12 heures plus tard). En combinant la masse des proies offertes avec les indices
de consommation, il est possible de connaître la quantité totale de proies (g)
ingérées pour chaque serpent dans chaque enclos pendant l’experience (voir détails
sur la méthode, article 2. La quantité et la masse des proies offertes par enclos
varient selon les dispositifs expérimentaux et la problématique examinée. Dans tous
les cas, les femelles sont réparties de façon aléatoire dans chaque enclos et lots
experimentaux.
Mesures
Les pesées ont lieu regulièrement et varient selon les expériences. Le nombre de
follicules en développement (vitellogénèse) ou d’embryons (gestation) sont
déterminés par palpation. Peu de temps avant la mise bas, les femelles sont toutes
recapturées et installées au laboratoire dans des boites individuelles jusqu’à la
parturition. Le protocole appliqué aux mises bas est identique en tout point à celui
décrit pour les femelles reproductrices des Moutiers en Retz.
Afin d’étudier le métabolisme aérobie des vipères pendant la gestation, j’ai adapté la
chambre calorimétrique du CEBC pour la mesure de petites consommations en
42

oxygène (avec l’aide technique de Guy Merlet et Laurence Pastout). Ce travail de
mise au point a fournit un support de base pour l’étude de la relation entre fécondité
et niveau métabolique, menée par Mitchell Ladyman (Honours Degree 2000). Des
informations détaillées sur le protocole et les caractéristiques techniques sont
données dans la section matériels et méthodes de l’article 6.
3. Les paramètres de la gestion de la resource : condition
corporelle et changements de masse
Comme nous l’avons vu, de nombreux travaux indiquent que la vipère aspic est un
reproducteur sur réserves qui accumule d’importants stocks de lipides (Bonnet 1996 ;
Bonnet & Naulleau 1994 ; Bonnet et al. 1994). Il existe donc un décallage temporel
entre les phases d’aquisition (accumulation de l’énergie) et d’allocation dans la
reproduction. Dans le cadre de notre problématique, la compréhension du système
de gestion de la resource (alimentation versus réserves) est donc d’un interêt crucial
pour déterminer les influences spécifiques des réserves et de la nourriture sur
l’expréssion des traits d’histoire de vie de l’espèce. Dans un pareil cadre, il est
nécessaire de pouvoir estimer précisement l’état et la dynamique des réserves au
cours du cycle reproducteur. Ces informations vont être fournies par différentes
mesures complémentaires : la condition corporelle et les variations de masses.
Condition corporelle, mesure et signification
D’une facon générale, la mesure de l’état physiologique ou “condition corporelle” des
organismes est d’un interêt majeur en éco-physiologie en apportant des informations
sur le succès en chasse et l’état des réserves (Jakob et al. 1996). La signification du
terme condition corporelle est évidemment très variable selon les organismes et
notamment selon le système d’allocation de l’énergie (reproduction sur revenu ou sur
43

éserves). Il existe de nombreuses méthodes de mesures invasives ou non qui ont
toutes pour objectif de s’affranchir de la taille corporelle lorsque l’on désire comparer
la masse ou l’état de nutrition des individus.
Chez la vipère aspic femelle, la condition corporelle peut être calculée sous la
forme d’un ratio entre masse et taille ou en appliquant la méthode des résidus
(Bonnet & Naulleau 1994 ; Bonnet 1996). Dans ce cas, on réalise une régréssion
entre masse et taille corporelle après tranformation logarithmique. La distance
résiduelle entre un point et la ligne de régression fournit un estimateur de condition.
Cette méthode offre de nombreux avantages en permettant notamment une
distinction nette entre l’effet de la taille corporelle et l’effet de la condition. C’est la
méthode idéale pour tester des hypothèses sur la condition corporelle d’individus
provenant d’une même population. L’estimation de la condition à partir des résidus
permet souvent une interprétation biologique très précise (Jakob et al. 1996).
Garcia-Berthou (2001) a récemment formulé une critique sur l’utilisation des résidus
pour la comparaison de differents groupes d’individus. Cet auteur préfère alors
l’utilisation d’analyses de covariance (avec la masse en variable dépendante et la
taille corporelle en co-facteur) plus adaptées pour la comparaison de groupes.
Dans le présent travail, l’utilisation des deux méthodes de calculs (résidus et
covariances) a conduit à des résultats similaires n’affectant pas les conclusions.
Chez la vipère aspic, la condition corporelle est un paramètre central qui fournit des
informations très précises et très différentes selon la période biologique examinée.
Nous avons ainsi distingué trois grandes périodes de mesures :
Condition corporelle initiale : au début du printemps (de Mars à Avril avant les
premières prises alimentaires) la condition corporelle renseigne sur l’état des
réserves et notamment les stocks de corps gras pré-vitellogéniques (avant toute
croissance des folliculles). Il s’agit donc de l’état “initial” des réserves réalisées les
44

années précédentes. La qualité de l’estimation est largement confimée par des
méthodes de quantification absolue non invasive (RMN).
Condition corporelle pre-partum : la condition corporelle pre-partum est calculée juste
avant la mise bas. Cet estimateur complexe intègre à la fois la masse de la portée et
la masse de la “carcasse” de la femelle. La comparaison entre les masses ajustées
initiales et pre-partum fournit un bonne indice de l’alimentation pendant cette période.
Condition corporelle post-partum : la condition corporelle post-partum est basée sur
la masse corporelle juste après la mise bas. A cette période, les stocks de réserves
sont très réduits, voir absents et les femelles présentent un amaigrissement très net
au niveau des muscles squelletiques. Le calcul de la condition corporelle post-partum
constitue alors un bon indice du degré d’émaciation maternelle après la reproduction
(Bonnet et al. 2000a).
Les variations pondérales
Nous bénéficions donc d’un estimateur précis de l’état physiologique des vipères au
cours du cycle reproducteur. Les variations pondérales vont nous fournir des
informations complémentaires sur la balance énergétique et notamment la cinétique
des réserves. La vipère aspic est un vertébré ectotherme qui ne produit pas de
chaleur de façon endogène. En dehors de toute contraintes énergétiques
(reproduction, croissance) les prises alimentaires vont donc être associées à une
augmentation proportionnelle de la masse corporelle et des stocks de lipides. Les
variations pondérales observées entre le printemps et l’automne chez les individus
non-reproducteurs vont donc nous renseigner très finement sur les prises
alimentaires et la dynamique des réserves lipidiques (voir explications détaillées
article 1). Chez les individus reproducteurs, la situation est moins simple car il existe
45

des demandes énergétiques accentuées pendant la vitellogénèse et la gestation.
Néanmoins, la situation énergétique d’un ectotherme est très contrastée par rapport
à un organisme endotherme (Bradschaw 1997). Ainsi et en dépit des contraintes
énergétiques de la reproduction, les variations pondérales des femelles vont aussi
nous renseigner sur l’alimentation (Lourdais et al. 2002a,c). Il va alors être possible
d’estimer la quantité de nourriture consommée par les femelles reproductrices dans
l’année (différence entre la masse initiale et la masse pre-partum Lourdais et al.
2002b) ou bien de façon spécifique à la vitellogénèse (différence entre la masse
initiale et la masse à l’ovulation Bonnet et al. 2001b, Lourdais et al. 2002c) ou à la
gestation (différence entre la masse à l’ovulation et la masse pre-partum ; Lourdais et
al. 2002a).
Figure 6. Mise-bas de Vipère aspic en septembre
46

II. Le système d’allocation
de l’énergie
47

A. Résumé du Chapitre:
Chez la vipère aspic, les réserves corporelles sont déterminantes pour la “décision”
de reproduction (Naulleau & Saint Girons 1981 ; Bonnet & Naulleau 1994,1995 ;
Bonnet 1996). Par ailleurs, cette espèce se nourrit essentiellement de campagnols
dont la démographie est très fluctuante. Ce premier chapitre présente des résultats
qui précisent les effets de l’acquisition de nourriture sur la reproduction à différentes
échelles temporelles. Ce chapitre est composé de trois articles : le premier est basé
sur des données de terrain, le second présente des résultats expérimentaux qui
procurent un lien fonctionnel avec le troisième lui aussi basé sur le travail de terrain.
Dans un premier temps (article 1), nous avons cherché à clarifier, en
situation naturelle (aux Moutiers-en-Retz), l’influence de l’alimentation sur la
fréquence de reproduction et le succès reproducteur des femelles de vipère aspic.
Ce travail suggère l’existence d’un système d’allocation complexe combinant à la fois
des aspects de reproduction sur réserves (”capital breeding”) et de reproduction sur
revenu facultative (“facultative income breeding”). Ainsi, si la fréquence des
reproductions est influencée par les réserves corporelles, la nourriture obtenue
pendant la vitellogénèse va directement influencer la masse des jeunes produits. La
taille des portées semble quand à elle déterminée par les influences conjuguées des
réserves corporelles et la consommation de proies au cours de la reproduction. Ces
résultats indiquent que si la vipère aspic est un reproducteur sur réserve “typique”
pour certains aspects, d’autres aspects (dynamique folliculaire et la masse des
jeunes) pourront être influencés par la nourriture obtenue pendant la vitellogénèse.
Nous avons ensuite cherché à examiner expérimentalement les effets des
réserves et des prise alimentaires sur le recrutement folliculaire et le nombre de
follicules qui parviennent à maturité (article 2). Nous avons soumis un groupe de
48

femelles reproductrices capturées en début de saison de reproduction à différentes
conditions d’alimentation. Les résultats obtenus confirment l’importance de la
nourriture obtenue pendant la vitellogénèse en tant que modulateur des processus
de déclenchement de la reproduction et de recrutement des follicules qui sont
contrôlés par les réserves corporelles initiales. En limitant le phénomène de
régression folliculaire, cette source d’énergie va agir sur le nombre final de follicules
qui arrivent à l’ovulation. La vipère aspic n’est non pas un reproducteur sur réserve
“rigide” mais va pouvoir ajuster son investissement en fonction des ressources
disponibles pendant la phase d’allocation de l’énergie. Une telle flexibilité peut être
très avantageuse, particulièrement dans un système ou la nourriture fluctue
beaucoup et de façon imprévisible d’une année à l’autre. Ainsi, elle permettrait aux
femelles d’améliorer leur succès reproducteur en profitant d’opportunités
alimentaires. Le capital de réserve minimum assure, quand à lui, une reproduction
efficace, y compris lorsque la nourriture est peu abondante au cours de la
reproduction et/ou les années précédentes : les réserves corporelles pouvant être
stockées sur des périodes prolongées (années).
Dans la troisième partie (article 3), nous avons examiné les bénéfices de
cette flexibilité dans le système d’allocation en prenant appui sur dix ans de suivi
longitudinal de la population des Moutiers-en-Retz ; notamment en examinant les
conséquences des fortes variations trophiques et climatiques. Dans un
environnement fluctuant, les avantages du système de gestion des ressources
observé chez les femelles vipères apparaissent nettement : 1) La mise en place de
réserves corporelles longtemps avant la reproduction va assurer décalage temporel
entre acquisition et allocation en permettant aux vipères de se reproduire avec
succès même les années ou la nourriture est rare. L’efficience de reproduction est
donc toujours élevée et partiellement indépendante des conditions d’alimentation
49

courante. 2) Lorsque la vipère aspic se reproduit une année où les rongeurs sont
abondants, l’énergie acquise ne va pas être uniquement allouée dans la portée. Une
fraction de cette énergie va ainsi influencer les caractéristiques maternelles en
limitant le niveau d’amaigrissement après la mise bas. Enfin, les contraintes
thermiques associées à la gestation sont aussi clairement révélées. Durant les étés
chauds, les opportunités de thermorégulation sont accentuées et la durée moyenne
de gestation est plus courte. Parallèlement, les coûts métaboliques associés au
maintien de régimes thermiques élevés augmentent, ce qui accélère
l’amaigrissement des femelles après la mise bas.
Chez les femelles non-reproductrices, les fluctuations d’abondance en proies
affectent directement la dynamique des stocks lipidiques et la croissance corporelle.
La température joue aussi un rôle positif sur les gains de masse en influençant
probablement les conditions de digestion. Les plus forts gains en masse et en taille
sont observés l’année de plus forte abondance en campagnols (1996). De façon
logique, les femelles qui s’engagent dans la reproduction en 1997 présentent des
réserves corporelles très importantes. Pour cette année particulière, l’investissement
reproducteur n’est donc pas limité par l’état des réserves corporelles. Les contraintes
spatiales (volume abdominal) sur l’effort reproducteur sont alors révélées et pour
cette année seulement on détecte une relation significative entre la taille maternelle
et la taille de la portée. Cela signifie aussi que la plupart des années dans notre
étude, l’état des réserves corporelles accumulées avant la reproduction combinée à
la disponibilité alimentaire au cours de la reproduction constituent les principales
limites de la fécondité.
50

B. Article 1
Short-term versus long-term effects of food
intake on reproductive output in a viviparous
snake (Vipera aspis)
X. Bonnet 1 2 , G. Naulleau 1 , R. Shine 4 & O. Lourdais
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 School of Biological Sciences A08, University of Sydney, NSW 2006, Australia
Published in OIKOS 92: 297-308
(2001)
51
1 2 3

Abstract
Feeding rates influence reproductive output in many kinds of animals, but we need to
understand the timescale of this influence before we can compare reproductive
energy allocation to energy intake. A central issue is the extent to which reproduction
is fuelled by long-term energy stores ("capital" breeding) versus recently-acquired
resources ("income" breeding). Our data on free-living aspic vipers show that there
is no simple answer to this question: reproductive frequency is determined by long-
term energy stores, offspring size is influenced by maternal food intake immediately
prior to ovulation, and litter size is influenced by both long-term stores and short term
energy acquisition. Thus, offspring size in free-living vipers reflects the mother’s
energy balance over the preceding year (via a trade-off between litter size and
offspring size) as well as her energy balance in the current breeding season. Hence,
different components of a given reproductive output (litter) are not only functionally
linked, but also respond to different temporal scales of prey availability. A female's
body size has little effect on her reproductive output. Attempts to quantify
reproductive energy allocation must take into account the fact that different
reproductive traits (such as offspring size versus number) may respond to energy
availability over different timespans. Thus, although the aspic viper is a typical
"capital breeder" in terms of its reliance on stored reserves for maternal "decisions"
concerning reproductive frequency, it is to some degree a facultative "income
breeder" with respect to the determination of offspring size and litter size.
Key words: annual variations, body condition, energy allocation, litter size, offspring
size, prey availability, reproduction, trade-off, snakes, vitellogenesis.
52

Introduction
The relative allocation of resources between maintenance, growth, reproduction and
storage is a central theme of studies on life-history evolution (e.g., Stearns 1989,
1992; Roff 1992), but we still cannot provide a convincing answer to R. A. Fisher’s
classic (1930) challenge: "It would be instructive to know not only by what
physiological mechanism a just apportionment is made between the nutriment
devoted to the gonads and that devoted to the rest of the parental organism, but also
what circumstances in the life history and environment would render profitable the
diversion of a greater or lesser share of the available resources towards
reproduction". Modelling these kinds of allocation decisions is a relatively
straightforward task, and much has been accomplished in this respect (e.g., Charnov
1982). Measuring energy allocation among these pathways, in a form that is directly
relevant to those life-history models, has proved to be a more challenging
proposition. This is especially true for long-lived organisms living in places where
food availability fluctuates on a seasonal or annual basis. Quantifying rates of food
intake and various expenditures is logistically feasible, but simply measuring the
magnitude of these pathways (although of interest in its own right) does not enable
us to answer Fisher’s question about the mechanisms of allocation. In order to
answer the question, we need to understand how fluctuations in food supply
influence allocation decisions by the organism.
One central problem in this respect involves the timescale at which those
fluctuations occur. Attempts to quantify allocation pathways necessarily invoke a
timescale for measurement: for example, we must compare energy gain to energy
expenditure over some specified period of time. Clearly, the relevant timescale will
differ among taxa: for example, it will be longer for an animal that stores energy for a
53

long period prior to expending it on reproduction. The timescale will also differ
among habitats: for example, it will be longer in a habitat with low food availability,
where the animal must forage for longer to gain enough energy to initiate
reproduction. Such difficulties may be overcome by choosing the appropriate
timescale for the system in question. However, another complication may also arise,
that is more difficult to address. For a given reproductive bout by a single animal,
some aspects of reproductive output may be governed by energy availability over a
long timescale whereas other aspects are determined by short-term energy intake.
In this paper we document such a case: a female aspic viper’s "decision" as to
whether or not to reproduce, and how large a litter to produce, are largely governed
by long-term energy stores; but her "decision" as to offspring size is driven by short-
term rates of food intake. Unless we know about such relationships, we cannot
meaningfully compare resource availability to reproductive output, or apply concepts
such as "capital" versus "income" breeding to biological systems.
The division between "capital" and "income" breeders refers to the source of
nutrients used to support reproductive expenditure (Drent & Daan 1980; Jönsson
1997). Income breeders derive these resources directly from food consumed during
the reproductive season, whereas capital breeders derive these resources from
reserves that are developed prior to the reproductive season. This concept has
proved to be useful in studies on birds (e.g. Chastel et al. 1995), but has been rarely
explored in other vertebrates (Jönsson 1997; Doughty & Shine 1998). Several
features of ectothermy preadapt reptiles to reliance upon capital breeding (Bonnet et
al. 1998). Among reptiles, snakes are excellent models to study such strategies
because most species provide no parental care (Shine 1988a) and embryonic
development is primarily lecithotrophic, with minimal placental transfer of energy
(Stewart 1992). Thus, the resources allocated to offspring are fully committed prior to
54

ovulation, rather than being provided over a long period during gestation or after
hatching (or birth). Because the number and size of offspring in snakes are
determined during vitellogenesis, it is easier to quantify maternal investment than
would be the case in many other kinds of animals. Furthermore, the wide range in
body-sizes of adult females within a single snake population (Andrews 1982)
provides an opportunity to examine the influence of maternal size on reproductive
output.
Previous studies suggest that two main factors are likely to play an important
role in determining reproductive output in these animals: (i) maternal body size,
because space to hold the clutch depends upon female size (Vitt & Congdon 1978;
Shine 1988b; Seigel & Ford 1987; Ford & Seigel 1989a); and (ii) energy availability
(Ford & Seigel 1989b; Seigel & Ford 1991). In turn, the latter factor can be separated
into two components: food intake during follicular growth ("income") versus maternal
reserves at the onset of vitellogenesis ("capital"). Long-term energy stores are likely
to influence reproductive decisions in some snake species (Diller & Wallace 1984;
Blem & Blem 1990; Brown 1991; Bonnet et al. 1994; Naulleau & Bonnet 1996), but
not in others (Plummer 1983; Ford & Seigel 1989b; Whittier & Crews 1990; Naulleau
& Bonnet 1995). For example, preliminary studies report a positive correlation
between maternal reserves and offspring number in Vipera aspis but no such
relationship in Elaphe longissima (Bonnet & Naulleau 1994; Naulleau & Bonnet
1995).
Published data suggest that the aspic viper, Vipera aspis, offers a classic
example of a typical "capital breeder". In snakes, body-condition (mass scaled by
size) reflects their long-term foraging success during the preceding year(s) (Forsman
1996 ; Shine & Madsen 1997). Adult female aspic vipers need to accumulate very
large body reserves to reach the body-condition threshold necessary for the induction
55

of vitellogenesis, and hence postpone reproduction for a long period (up to 4 years)
while they are accumulating those reserves (Naulleau & Bonnet 1996).
Vitellogenesis in this species involves an intensive mobilisation of maternal reserves
(Bonnet et al. 1994) and maternal energy reserves dictate whether or not a female
will reproduce in a given year (Naulleau & Bonnet 1996). Captive female vipers can
reproduce without feeding during the whole reproductive period (i.e., vitellogenesis
plus gestation), demonstrating that reproduction can be entirely supported by the
energetic "capital" stored prior to reproduction.
The aim of our study is to examine the influences of maternal size, energy
stores and current energy intake, on the number and size of offspring in a natural
population of aspic vipers. How is the very large capital of stored energy invested
(packaged) during vitellogenesis? Because female aspic vipers continue to feed
during vitellogenesis, we can also examine whether the energy thus obtained has
any influence on reproductive output. Does vitellogenesis rely only on maternal
reserves, or does this snake adopt a more flexible strategy using supplementary
"income" to improve the quality of the litter? In other words, is the aspic viper a
"strict" capital breeder with respect to offspring size and litter size as well as
reproductive frequency?
Materials and methods
Study site and animals
Over a period of seven years (1992-1998), we studied a large (more than 1000 adults
individually marked) population of aspic vipers, Vipera aspis, at Les Moutiers en
Retz in western central France near the Atlantic Ocean (47°03'N; 02°00'W). This
population is near the northern limit of the geographic range of the species. The
56

study area of approximately 33 ha consists of fields and paths bordered by hedges.
In fields not used for farming, brambles, brushwood and small trees are common.
The study population is separated from adjacent populations by several roads and a
village (see Bonnet & Naulleau 1996 for a more detailed description).
The aspic viper is a medium-sized (typically to 55 cm body length, 100 g), slow-
moving, terrestrial, venomous snake species (Naulleau et al. 1996). Autopsies and
NMR imaging confirm that in this species, good body condition (high mass relative to
body length) is related to large body reserves (abdominal fat bodies: Bonnet and
Naulleau 1994, 1995; Naulleau & Bonnet 1996; Villevieille 1997). Females more
than 41 cm snout-vent length (= 47 cm total body length, the minimal body size
where parturition has been observed) were considered sexually mature (henceforth
referred to as adult).
Procedures and measurements
Any study to investigate the influence of maternal reserves on reproduction in free-
living animals must meet the following criteria:
(i) maternal condition must be recorded at the onset of vitellogenesis; this requires
precise information about the timing and kinetics of follicular growth, ideally in relation
to mobilisation of maternal reserves;
(ii) relationships between maternal size, body condition and energy reserves must be
quantified, so that we can estimate the amount of reserves a female can devote to
her clutch;
(iii) quantitative data about food intake during follicular growth would be useful;
failing this, we need a index of food consumption over this period,
57

(iv) we need to recapture the monitored females before laying or parturition (which
typically occurs 2 to 6 months after the start of vitellogenesis in snakes: Seigel & Ford
1987), in order to obtain data on reproductive output.
In the present study, snakes were caught by hand and individually marked by
scale clipping or (later in the study:1993) with electronic tags (11±1mm length [mean
± 1S.D. as in all subsequent results]; 2.1±0.1mm diameter; 0.25±0.1g total
mass;125KH, sterile transponder TX 1400L, Rhône Mérieux, Destron/IDI INC). The
study area was intensively searched almost every day throughout the active season
(early February to October) by one to three people; the total searching effort
represents more than 3500 h in the field. More than 500 adult female vipers have
been marked since 1992, and their snout-vent lengths, total body lengths (± 0.5cm),
and masses (± 1g with an electronic scale) recorded in the field. Body condition was
calculated as the residual scores from the regression of the natural logarithm of body
mass against that of body length (Jayne & Bennett 1990; Madsen & Shine 1992a;
Naulleau & Bonnet 1996). The snakes were released at the exact point of capture
within 15 min, and more rapidly after subsequent recaptures (except before
parturition, see below). Measurements were made at three stages during the
reproductive season: (i) at the onset of vitellogenesis, (ii) at the end of vitellogenesis
(late May, i.e. close to ovulation), and (iii) before parturition (Figure 1). In the aspic
viper in western central France, vitellogenesis begins in March (Bonnet et al. 1994),
but a major increase in follicular size does not occur until late April (Saint Girons
1957b; Saint Girons & Duguy 1992; unpublished data from nuclear magnetic
resonance; Figure 1). Vipers are easier to catch at the beginning of the reproductive
season, and many females were caught in March or April. Feeding activity is
reduced during this first period of vitellogenesis, and there is generally a slight
decrease in body mass. Thus, data gathered in March and April were pooled. Body
58

masses obtained during the periods of rapid follicular growth (May and June) and
pregnancy (July-August), or from individuals with a prey item in the stomach, were
not used to calculate body condition.
Figure 1. In the aspic viper (Vipera aspis), vitellogenesis begins in early March, with ovulation
occurring three months later in early June. Follicular growth (yolk deposition) accelerates in late April
and is complete in early June. Embryonic development and egg hydration (dashed zones) occur later,
from June to August during gestation until parturition three months later. In the present study, 44
reproductive females were captured at three different physiological states: (i) at the onset of
vitellogenesis (arrow 1), (ii) at the end of vitellogenesis (arrow 2), and (iii) before parturition (arrow 3).
Maternal body condition index (an estimate of previtellogenic body reserves) was calculated at the
onset of vitellogenesis.
Food intake and mass gain during vitellogenesis
We do not have data on actual prey masses from female in the field, because we
were unwilling to stress these animals by forced regurgitation. Thus, we used the
increase in female body mass during vitellogenesis as an index of food intake. The
validity of this method was tested using two data sets:
59

(i) food consumption versus change in body mass for 30 captive female vipers
(collected from a variety of locations in France). The captive females were kept in
individual cages (40x40x40cm, an electric bulb [60W] provided a thermal gradient,
water was at libitum). The snakes were offered laboratory mice (weighed to the
nearest 0.1g) each week, and were weighed and palpated regularly. Ovulation date
was determined from palpation, radiography (Naulleau & Bidault 1981) and
parturition dates. The changes in body mass recorded during vitellogenesis were of
the same order of magnitude in captive (9.1 ± 25.0; range -33.0 to 56g) versus free-
ranging female vipers (see results).
(ii) meals records versus change in body mass in free-living snakes. Recent meals
in snakes can be detected by palpation (e.g., Fitch 1987). Although we could not
quantify precisely prey consumption in the field, we can compare mass change in
snakes recorded to feed during vitellogenesis compared to those that were not
recorded to feed over this period.
Reproductive output
One hundred and forty six gravid females were caught one to 21 days before
parturition, in late August-early September (Figure 1), and placed in individual cages
in the laboratory until they gave birth. To avoid a selection towards obviously gravid
females during capture (distended body, and well-developed embryos easily
detected by palpation), all females with identifiable items in the abdomen (detected
by palpation, excluding the stomach region) were also collected. Palpation enabled
us to detect objects as small as 2g (corroborated by dissection, unpublished data).
Captive females were monitored every day until parturition, and weighed (± 0.1 g)
every two days and immediately after parturition.
60

We recorded the number, mass (± 0.1 g), and length (± 0.5 cm) of healthy
offspring, stillborn and relatively undeveloped embryos, and the number and mass of
unfertilised eggs (± 0.1 g). To analyse reproductive output, we made a distinction
between the “classical” measures of total litter size (henceforth LS) or total litter mass
by including healthy neonates, stillborn offspring and undeveloped eggs as well (Farr
& Gregory 1991; Gregory et al. 1992); whilst we included only viable neonates in our
measures of “effective” reproductive output. Thus, fourteen females that did not
produce any healthy offspring were excluded from several analyses. The other 132
females gave birth to 699 healthy offspring. The mean body mass of healthy
offspring per litter was used to test the predicted negative relationship between
offspring number and size. To control for the effect of female body length, we used
partial correlation analyses (Ford & Seigel 1989a).
Statistics
Data gathered in 1992 were excluded from several analyses on reproductive output
because we selected five obviously gravid, and large, females this year. This bias
should be important for inter-annual comparisons, but probably did not influence
other analyses. Estimates of population size were obtained using the program
CAPTURE (see Bonnet & Naulleau 1996 for further details on methods). In every
case, the first models suggested by the goodness-of-fit tests were Mth (population
estimate under individual heterogeneity in capture probabilities) or Mh (population
estimate under time variation and individual heterogeneity in capture probabilities;
Chao et al. 1992), and they were systematically adopted. The differences between
the two estimates were small, and we conserved the first selected model.
Snakes are very secretive animals (Seigel 1993), and the breeding frequency of
female aspic vipers is low in western central France (Bonnet & Naulleau 1996).
61

Thus, despite a large initial sample size (> 300) and relatively high recapture rates,
we obtained complete data for only 44 females. In this complete data set, each of
these animals was captured at least three times during the 6 months reproductive
period: early in vitellogenesis (before the 15th of April for any given year), close to the
time of ovulation (from 15 May to 15 June), and close to parturition (3 weeks to one
day before; figure 1). The low probability of recapturing a snake at three precise
occasions separated by long time intervals combined with the exclusion of individuals
caught with prey in the stomach were excluded from analyses, explains why only 44
females were included in the "complete" data set. These females did not differ from
other females in mean snout vent length or reproductive characteristics (Table 1).
However, this sub-sample of females tended to over-represent particular years
particularly those with high proportions of reproductive animals, and a low feeding
rate because we did not use data on body masses of animals containing recently-
ingested prey.
Table 1. Characteristics of 146 reproductive female aspic vipers (Vipera aspis) and their litters. SVL =
snout-vent length, Fit litter size = number of healthy neonates, Offspring mass = mean mass of healthy
neonates (calculated as the mean of the means gathered on 132 litters). * df = 1,130 for mean
offspring body mass. "Complete data set" = females monitored throughout the entire reproductive
year, from the onset of vitellogenesis until parturition. "Others" = females caught after the onset of
vitellogenesis. There was no difference between the two subsets of data ("complete" versus "others")
in maternal size or in the reproductive characteristics that we measured.
SVL
(cm)
Total litter
size
Fit litter
size
Total litter
mass (g)
Offspring
mass (g)
Complete n = 44 49.1 ± 3.2 5.8 ± 1.6 4.7 ± 2.3 32.0 ± 12.9 6.3 ± 0.9
Others n = 102 48.7 ± 3.3 6.3 ± 2.2 4.9 ± 2.9 35.6 ± 16.2 6.3 ± 1.1
F(1,144)* 0.45 2.0 0.10 1.64 0.15
p 0.50 0.16 0.75 0.20 0.69
62

To ensure that these biases did not substantially affect our conclusions, we also
analysed the larger data set (n = 146) in which many more animals were
incorporated, but with incomplete data for some individuals. We used these data to
examine annual variation in reproductive output. Fourteen of these 146 reproductive
females were represented twice (at intervals of 2 to 4 years, due to the low breeding
frequency), raising the possibility of pseudoreplication. However, none of our results
were modified when we randomly excluded duplicate records from these animals. In
the complete data set (n = 44), every female was represented only once. The
following results are derived from analyses on these two data sets. All the tests were
performed using STATISTICA 5.1.
Results
Relationship between food intake, growth and weight gain
Females did not increase in body length during reproductive years, except for a slight
increase (1 - 2 cm) in a few individuals in one year when food availability was
exceptionally high (1996, see below). However, females gained appreciably in mass
during vitellogenesis (mean mass change was +11.7 ± 15.3g, +12 ± 16% of initial
mass, n = 44), with considerable variation among individuals (range -12 to +51g, -12
to +61% of initial mass). Such body mass variations were not related to the female's
SVL (r = 0.08, P = 0.58, n = 44), nor to her pre-vitellogenic body mass (r = -0.13, p =
0.41, n = 44) or initial body condition (r = -0.14, p = 0.54, n = 44).
A causal link between maternal mass gain and feeding was suggested by the
data set on captive snakes: female viper's food intake during vitellogenesis was
highly correlated with her change in body mass over this period (r = 0.83, n = 30, p <
0.0001; Figure 2). This result was also supported by the data gathered on the 44
63

free-ranging snakes. Using palpation, we recorded prey in 13 reproductive females
in late April – May. All of these animals increased in body mass during vitellogenesis
(mean mass change was +20.7 ± 15.0 g, +24 ± 17% of initial mass; range +2 g to
+51 g, +2 to +60% of initial mass), whereas the other 31 females in which we found
no evidence of feeding showed a lower gain in body mass (mean mass change +8.0
± 14.0g, +8 ±14% of initial mass; range -12g to +51 g, -11 to + 54% of initial mass;
comparing the two groups, F(1.42) = 7.28, p = 0.01). Although many of these 31
females probably fed at some time during vitellogenesis, the difference in mass
change is consistent with the hypothesis that mass change reflects feeding rate.
Thus we conclude that the increase in body mass constitutes a simple and reliable
index of prey consumption over the period of vitellogenesis.
Figure 2. Relationship between changes in maternal body mass and food intake during vitellogenesis
in 30 captive female asp vipers. Food intake was calculated as the sum of the mice consumed during
the whole vitellogenic period (3 months). Change in body mass was the difference between the
female's mass at ovulation minus her initial mass as recorded at the onset of vitellogenesis. Although
change in body mass reflects the influences of complex phenomena (egg hydration + maternal
metabolic expenditure + individual physiological differences [e.g. litter size]), food intake strongly
influences maternal somatic weight changes (see text for statistics).
64

Influence of maternal body length on reproductive output
The first likely correlate (determinant?) of litter size and offspring size is maternal
body size; numerous researchers have documented strong allometry in these traits in
a wide variety of reptile species (Dunham et al. 1988; Wilbur & Morin 1988). If
maternal body size strongly affects reproductive output, then we need to know this at
the outset so that we can factor out allometric effects in all subsequent analyses. We
used the entire data set (all females and their litters) for this analysis.
Maternal snout-vent length was weakly correlated with total litter size (r= 0.17, p=
0.036, n= 146), but not with the number of healthy neonates (r= 0.13, p= 0.12, n=
n=146, Figure 3a). Hence, maternal body size has little effect on litter size in this
population. Larger females tended to produce heavier offspring (r= 0.22, n= 0.01, p=
132, Figure 3b). The effect was stronger in a partial correlation analysis where we
held constant the effect of litter size on offspring size (partial correlation: r= 0.24, p =
0.005, n = 132). Thus, our data suggest that a female's body size influences the
mass of her young, but has less effect on the total number of neonates that she
produces.
The second plausible influence on offspring size and number is the relationship
between these two variables. Given finite resources, an increase in litter size will
reduce mean offspring size. We looked for this effect in the entire data set, using
mean values for offspring size for each litter.
Trade-off between developing offspring for maternal resources
Offspring snout-vent length and offspring mass were strongly correlated (r= 0.75, n=
699, p< 0.0001); we use body mass to characterise neonatal size, rather than length,
because it offers a more direct measure of maternal investment. Stillborn offspring,
relatively undeveloped embryos, and unfertilised eggs weigh much less than healthy
65

neonates (mean masses of healthy neonates, stillborn and undeveloped eggs were
6.3 ± 1.1 g [n = 699], 4.4 ± 2.0 g [n = 80], and 2.0 ± 1.1 g [n = 103] respectively), and
so were not included in our test of the influence of litter size on offspring size.
Figure 3. The relationship between maternal size and reproductive output in Vipera aspis. Larger
females do not produce more viable offspring (a), but tend to have slightly larger neonates (b).
66

The predicted negative relationship between offspring number and the mean mass
of healthy neonates was not evident from our raw data (Figure 4a). Using total litter
size rather than the number of healthy neonates did not change the significance of
this result (r= -0.15, p= 0.09, n = 132). However, when female body length was held
constant, mean offspring mass was negatively correlated with offspring number
(Figure 4b). Thus, our data demonstrate a weak trade-off between offspring size and
number in aspic vipers.
We now turn to the relationship between energy balance and viper reproduction.
First, we use the data on the 44 females measured at each stage of their
reproductive cycles (the "complete data" females) to assess relationships between
energy stores, food intake, and reproductive output.
Influence of pre-vitellogenic body condition on reproductive output
Female vipers that were in good body condition early in vitellogenesis, tended to
produce large litters (r = 0.47, n = 44, p = 0.001; Figure 5a). In contrast, early body
condition had no influence on offspring size (r = 0.09, n = 42, p = 0.56, note that
sample size was reduced because 2 females did not produce any living offspring;
Figure 5b) even when female snout-vent length was held constant (r= -0.09, n= 42,
p= 0.57).
67

Figure 4. The relationship between litter size and offspring size in Vipera aspis. a. The number of
viable offspring produced by a female viper is not significantly correlated with the mean mass of her
offspring, unless the analysis factors out the effects of maternal body size on reproductive output (see
Figure 3). If the effect of maternal size is removed (by using a partial correlation analysis), the
underlying tradeoff between offspring size and litter size is revealed (b).
Influence of food intake on reproductive output
The change in maternal body mass during vitellogenesis (food intake) was correlated
with litter size (r= 0.42, n= 44, p= 0.005, Figure 5c), even when female snout-vent
length was held constant (r= 0.44, n= 44, p= 0.002). Similarly, mass gain during
follicular growth strongly influenced offspring size (r= 0.39, n= 42, p= 0.01, Figure
5d). Controlling for the influence of maternal snout-vent length on offspring mass did
not change this result (r= 0.40, n= 42, p= 0.01). When the number of neonates was
68

included as a correcting factor, the influence of maternal mass change on offspring
size was strengthened (r= 0.47, n= 42, p= 0.002; and r= 0.50, n= 42, p= 0.001
including maternal snout-vent length as an additional correcting factor).
Figure 5. The relationship between previtellogenic maternal reserves (initial body condition) (a, b),
maternal somatic weight change during vitellogenesis (an indication of food intake; see text) (c, d), and
reproductive output in Vipera aspis. Females with high initial body condition produce larger litters (a)
but do not have larger offspring (b). Females that increase substantially in body mass during the
vitellogenic period produce larger litters (c) of larger neonates (d).
These latter results presumably reflect the fact that litter size influences the
amount of energy from a given prey item that can be allocated to each offspring.
The rank correlation between number and size of offspring in the "complete data" set
(r= -0.20, n= 42, using number of healthy offspring) was similar to that from the larger
69

data set (r= -0.14, n= 132). Because offspring size was positively correlated with
changes in maternal mass during vitellogenesis, we tested for the presence of this
trade-off including change in maternal body mass as a correcting factor. These
analyses confirmed the underlying negative relationship between offspring mass and
number of healthy neonates (r= -0.35, n= 42, p= 0.025; and r= -0.40, n= 42, p= 0.01
with female snout-vent length held constant).
The body condition of a post-parturient female was affected by her change in
body mass during vitellogenesis (r= 0.43, n= 44, p= 0.004). Thus, energy and
materials obtained from the prey during vitellogenesis were not exclusively allocated
to growing follicles, but were also directed to maternal reserves.
Combined effects of maternal length, maternal reserves, food intake and offspring
competition for energy on reproductive output
Our data allow us to quantify the four main factors that seem likely to influence
reproductive output (number and size of offspring) in female vipers. These factors
are independent of each other; for example, two of them (maternal length and
maternal body condition) are independent by definition. The third variable, change in
maternal body mass during the vitellogenic period (food intake), was not influenced
by either of these variables (see above). The fourth independent variable is the
trade-off between number versus size of growing follicles, given the finite amount of
energy allocated for vitellogenesis.
Stepwise multiple regression analyses were performed on the "complete data" set
(n= 44) with these four independent variables, and the number or size of offspring as
the dependent variables. The highest proportion of the variation in litter size was
explained when all four predictor variables were included in the analyses (r 2 = 0.61,
n= 42, p < 0.00001). However, variation in mean offspring mass was largely
70

explained by maternal weight change and litter size in the regression model (r 2 =
0.25, n = 42, p = 0.003), with no further significant increase in explained variance by
including the other variables (r 2 = 0.30, n = 42, p = 0.008 with the four variables).
Hence, we conclude that the reproductive output of a female aspic viper is
determined by the combined effects of at least four factors: her body length, her pre-
existing energy reserves, her food intake during vitellogenesis (as estimated by her
weight change over this period) and the trade-off between offspring size versus
number (i.e., competition among the offspring for energy).
Annual variation in energy availability and viper reproduction
In order to evaluate the generality of our results on 44 females, we can also examine
patterns of annual variation in the traits of interest in the all females. If our results
from the "complete data" sample are robust, they should enable us to understand the
ways in which annual variation in prey availability influences the reproductive output
of vipers.
Mean maternal body size remained consistent among years, but significant
annual variation occured in some of the reproductive traits (Table 2). The
consistency in maternal body size, combined with the lack of strong body-size effects
on reproduction (Table 2, Figure 3), substantially simplifies the analysis of correlates
of annual variation in reproductive biology. It also supports the notion that factors
other than maternal size influence reproductive output in our study population.
71

Table 2. Annual variation in the body sizes and reproductive output of female aspic vipers (Vipera
aspis). SVL = snout-vent length, Fit Litter size = number of healthy neonates, Litter mass = total litter
mass, Offspring mass = mean mass of healthy neonates. Data for the first four variables (columns)
were analysed by one way ANOVAs with year of the study (1992-1998) as the factor. For the final trait
(mean offspring mass), we used a two-factor ANOVA with litter number nested within year (so df = 6,
557 for offspring mass). These ANOVAs reveal no significant variation in maternal body size (SVL) of
146 reproductive female aspic vipers, but significant variation in some of the characteristics of their
litters. Significant values are indicated in bold faces. Note that excluding 1992 from analyses (see
text) did not change any results.
SVL
(cm)
Total litter
size
Fit litter size Litter mass
F(6,139)* 0.44 3.42 1.58 2.36 7.14
(g)
Offspring
mass (g)
p 0.85 0.01 0.16 0.03 0.0001
Prey availability varied substantially over the course of our study (see Bonnet
et al. 2001b). Vipers feed mainly on voles, whose populations typically fluctuate in
this region in a 3 to 4 year cycle (Krebs and Myers 1974; Delattre et al. 1992). Such
fluctuations directly influence feeding rates of the snakes (as revealed by the
proportion of snakes with a prey in the stomach, Bonnet et al. 2001b). Restricting the
analyses to the vitellogenesis period and to adult females, we can divide the years of
our study into three categories in this respect. Two years (1994 and 1998) were poor
(9% [n = 228] and 10% [n = 81] respectively of the snakes with a prey in the
stomach), three were "medium" (1993, 1995 and 1997 with 14% to 21% [99

i) Reproductive frequency showed substantial variation, but is difficult to quantify
precisely because population sizes also varied (Figure 6; Chi-square = 99.9, df = 5, p
< 0.0001). The number of reproductive females strongly decreased after a low food
year (1994), and increased after a good year (1996). The proportion of reproductive
females in comparison to the total number of adult females also varied significantly
among years (Chi-square = 46.3, df = 5, p < 0.0001), and was apparently related to
prey availability in the preceding year (Figure 6).
ii) Litter sizes varied among years (Table 2), with an increase of mean litter size
during the "high-food" year (1996) and the next ones as well (Figure 7). This is what
we would predict from the notion that litter sizes are mainly affected by existing
energy stores, and by current feeding rates also (see above). Similarly, during a year
of low prey availability (1994) we observed a reduction of mean litter size, and in the
following year as well (Figure 7).
iii) Offspring mass was higher in the "good" year (1996) than at any other time in the
study, and a nested ANOVA (with litter number nested within years) detected
significant variation in offspring body mass among years (Figure 8, Table 2).
Interestingly, during the 1994 "poor" year we did not observe any reduction of
offspring sizes, but such effect may have occurred in 1998 (Figure 8). Thus we
speculate that females allocate additional nutrients to offspring only in exceptionally
"good" years, and hence that modest annual variation in food supply will not be
automatically reflected in detectable changes to mean offspring mass. Instead,
changes to offspring mass will be apparent only in occasional very good years
(Figure 8).
73

Figure 6. Annual variation in the total number of adult females in a closed population of asp vipers
(black dots ± S.D., black line). The observed number of reproductive females (a subset of the total
number of adult females) is represented with hatched bars (± S.D). The expected number of
reproductive females (dotted circles), simply calculated as a constant proportion (33%, see Bonnet et
Naulleau 1996 ) of the total number of adult females, is indicated to better visualise the annual
fluctuations in the relative number of reproductive females in comparison to the total number of adult
females. The arrows with the sign "-" indicates a low food availability year, the arrow with the sign "+"
indicates a exceptionally high food availability year. Population estimates (± S.D.) were performed
using the program CAPTURE (see text for statistics).
Figure 7. Annual variation in litter size (annual mean ± S.E.) in a closed population of asp Vipers.
Number above each symbol indicate sample size (number of females). The arrows with the sign "-"
indicates a low food availability year, the arrow with the sign "+" indicates a exceptionally high food
availability year. See text for statistics.
74

Figure 8. Annual variation in mean offspring mass (annual mean ± S.E.) in a closed population of asp
Vipers. Number above each symbol indicate sample size (number of neonates). The arrows with the
sign "-" indicates a low food availability year, the arrow with the sign "+" indicates a exceptionally high
food availability year. See text for statistics.
Discussion
Our data suggest that the reproductive outputs of a female aspic viper (i.e., the size
and number of offspring in her litter) are affected in complex ways by her energy
acquisition over the months and years preceding the actual birth of the offspring. Her
foraging success over an interval of one to three non-reproductive years determines
the magnitude of her energy stores, and hence the "decision" as to whether or not
she will reproduce in a given year (because reproduction is initiated only after
females attain a precise body-condition threshold: Naulleau and Bonnet 1996). Our
data show that these energy stores (measured by female body condition prior to the
onset of vitellogenesis) determine litter size: females with larger energy stores initiate
vitellogenesis of a greater number of follicles. To our knowledge (see below), this
study demonstrates for the first time (using strict methodological criteria to calculate
75

early body condition) that in snakes, initial maternal reserves gathered over years
before reproduction positively influence reproductive output. Food intake after this
time, but prior to ovulation, also seems to affect the female’s reproductive output not
only by changing her litter size, but by modifying the mean mass of her offspring.
Recruitment of follicles starts at the initiation of vitellogenesis in early March (Bonnet
et al. 1994), feeding activity is rare before April, and vitellogenesis culminates in May-
June (see Figure1). Thus, we might expect that food intake during vitellogenesis
occurs too late to modify the number of growing follicles. However, we found a
positive relationship between food intake and litter size. This counterintuitive result
may be due to the influence of food intake on follicular atresia. Autopsies (Saint
Girons 1957b) and nuclear magnetic resonance imaging data (unpublished) show
that undersized follicles, as well as atretic follicles, are common during vitellogenesis.
Well-nourished females may proceed with vitellogenesis rather than resorption of
some of these smaller follicles.
Importantly, the impact of additional food (= energy) on offspring mass
necessarily depends on litter size also, because a larger litter means that a given
amount of energy is divided among a greater number of offspring. Thus, offspring
size in the aspic viper is affected by the mother’s feeding success over the preceding
few years (because her accumulated energy stores will determine litter size) as well
as her feeding success in the weeks immediately preceding ovulation. Similarly, litter
sizes are affected by feeding rates in this vitellogenic period as well as in previous
years. The end result is that the link between food supply and reproductive output is
complex. We must understand short-term as well as long-term rates of food intake
before we can interpret variation in reproductive output in aspic vipers.
Trade-offs between litter size and offspring size occur in several snake
species (Ford & Seigel 1989a; Madsen & Shine 1992a), including the aspic viper
76

(see above). Our study also shows that this trade-off (competition among growing
follicles during the initial partitioning of energy) can be obscured by subsequent
variation in food intake (Lessells 1991). Thus, although large litters should result in
small follicles, this trade-off may be obscured by variation among females in the
magnitude of additional energy reserves gathered during vitellogenesis (see van
Noordwijk and de Jong 1986; Doughty and Shine 1997). Follicles "compete" for
resources not only at the initiation of vitellogenesis, but throughout the vitellogenic
period (when additional energy from recently-consumed prey becomes available).
Thus, sibling ova compete for resources at two distinct levels and for two distinct
energetic sources: first at the onset of vitellogenesis for the energy stored by the
mother prior to reproduction, and second during vitellogenesis for the prey caught by
the mother.
In combination with previous work on this species (Naulleau 1965; Naulleau &
Bidaut 1981; Naulleau et al. 1996), our data clarify the complex relationship between
maternal foraging success and reproductive activity. Non-reproductive individuals
accumulate large body reserves over an interval of one to three years, by storing a
large proportion of the energy they assimilate from prey consumed over that period.
During that time, they remain very secretive (Bonnet and Naulleau 1996) until they
exceed the body-condition threshold necessary for reproduction (Naulleau and
Bonnet 1996). After this long period of abstinence, vitellogenesis takes the form of
an "explosive" investment: almost all maternal reserves are allocated to reproduction
to produce the greatest possible number of young. We have recorded several litter
masses greater than the mass of the post-parturient female (maximum = 112%).
Energy opportunistically obtained by feeding during vitellogenesis is also directed to
the developing follicles, and significantly increases offspring size. After reproduction,
females are very emaciated and many do not survive until the next year (Bonnet
77

1996; Bonnet et al. 2000a). Thus, many female aspic vipers are semelparous,
producing only one litter in their lifetimes (unpublished data). Given the low
probability of breeding again, the “costs” of additional investments to an already high
reproductive effort will be low in terms of decrements in future reproductive
opportunities (Williams 1966a). Consequently, females should be under strong
selection to maximise the effective output from any reproductive attempt, rather than
conserve resources to invest in subsequent litters. This situation may be widespread
in viperid snakes (Madsen & Shine 1993; W. S. Brown, pers. com.).
Reptiles have increasingly been used as "model organisms" for research in
this field (Seigel 1993; Shine & Bonnet 2000). However, both our methods and our
results provide a contrast to most previous work on these animals. Firstly, we focus
on events during vitellogenesis (the period during which most of the variation in litter
size and offspring mass is generated) rather than during pregnancy (e.g., Stewart
1989; Shine & Harlow 1993; Baron et al. 1996; Gregory & Skebo 1998).
Vitellogenesis may extend for periods as long as pregnancy in many reptile species
(including the aspic viper), and is a crucial phase in terms of maternal reproductive
"decisions". Vitellogenesis can be viewed as the “explosive” phase of the energetic
investment of reproductive females (Bonnet et al. 1994). Secondly, maternal body
size has little effect on reproductive output in the aspic viper (Bonnet et al. 2000b), in
strong contrast to most other snake species for which similar data are available (e.g.,
Seigel and Ford 1987). This result is not an artefact of low statistical power; our
sample size is larger than in most previous analyses (mean sample size was 32.8 ±
51.3 in 61 studies reviewed by Seigel and Ford [1987]). Other studies on the aspic
viper from the same region revealed a positive correlation between maternal size and
litter size (Naulleau & Saint Girons 1981). The difference between these studies is
probably explained by the particular methodology we employed (collection of all
78

animals that were potentially reproductive). If we had searched only for obviously
gravid snakes (e.g. snakes with particularly large litters), we would have ignored
many females with small litters relative to their body size.
Another difference between our study and previous analyses is that we have
used a very strict criterion as to when in the reproductive cycle we estimate body
reserves from overall body condition. Thus if measurements are taken on females
during pregnancy or close to the time of ovulation, the reproductive products
(follicles, eggs or embryos) constitute a significant proportion of maternal body mass.
Hence, body condition at this time is likely to be a direct measure of reproductive
output, not an indicator of the magnitude of body reserves such as the fat bodies and
liver (Bonnet & Naulleau 1994, 1995). In such a case, a positive correlation between
"body condition" and reproductive output is almost inevitable, since the same variable
occurs in both sides of the equation. In addition, the origin of the energy invested
into the clutch cannot be determined by such a methodology, and the respective
influences of food intake and maternal reserves become indistinguishable. Our data
constitute a basis for developing hypotheses that can be tested by rearing snakes in
controlled diets from the onset of vitellogenesis to the production of the offspring (e.g.
laying or parturition).
Despite these methodological problems, data from other snake species are
sufficient to document substantial interspecific differences in the relationship between
energy acquisition and reproductive expenditure. For example, experimental studies
show that food intake during vitellogenesis affects clutch sizes but not offspring sizes
in two American snake taxa (Thamnophis marcianus and Elaphe guttata: Ford and
Seigel 1989b; Seigel and Ford 1991). In both of these "income-breeding" genera,
maternal energy reserves may play only a small role in fuelling reproductive
expenditure (Whittier & Crews 1990; Naulleau & Bonnet 1995). In contrast, field data
79

on a "capital-breeding" viperid snake, Vipera berus suggest that offspring size may
be affected by prey availability in this species (Andren & Nilson 1983). Our own data
suggest that reproductive frequency in the aspic viper is driven by long-term energy
stores, but that litter sizes and offspring sizes respond in complex ways to maternal
feeding rates over both the short and the long term.
More generally, we predict that broad patterns will be apparent across
species, depending on the extent of their reliance upon stored energy for breeding.
In "capital" breeders, early body condition will determine reproductive status (Diller &
Wallace 1984; Brown 1991; Naulleau & Bonnet 1996) and will largely determine
offspring number (present study). In such species, breeding frequencies will often be
low because long periods of time are necessary to accumulate energy reserves
(Martin 1993; Bonnet & Naulleau 1996). The mass of the clutch relative to maternal
mass will often be high, because there is a massive investment of body reserves to
reproduction. In contrast, we predict that the reproductive decisions of "income"
breeders (which can store only limited body reserves) will be less dependent on
maternal body condition prior to vitellogenesis (Plummer 1983; Whittier and Crews
1990; Naulleau & Bonnet 1995). Breeding frequency should be higher, with the
number of offspring dependent on maternal foraging success immediately prior to
breeding (Ford & Seigel 1989b). Relative clutch (litter) mass should be relatively low.
There is undoubtedly a continuum between capital and income breeders in snakes,
as in other organisms (Chastel et al. 1995), so that there is ample opportunity for
robust empirical tests of these predictions. This interspecific diversity in the role of
energy reserves for reproduction, and the partial decoupling of control systems that
regulate different aspects of reproductive output (offspring size versus number) mean
that snakes offer exceptional opportunities to answer Fisher’s (1930) challenge about
the ways in which animals allocate energy to reproduction.
80

Acknowledgements
We thank L. Schmidlin, M. A. L. Zuffi, and C. Thiburce for helping to measure, weigh,
and mark several hundred neonates. We also thank S. Duret and M. Vacher, who
spent many months to catch the snakes. T. Madsen, P. Doughty, D. Reznick, S. J.
Hall and S. D. Bradshaw provided helpful comments on the manuscript. Financial
support was provided by the Conseil Général des Deux Sèvres, the Centre National
de la Recherche Scientifique (France) and the Australian Research Council. Rex
Cambag, J. T. Wilmslow and Gisèle Gaboune helped to solve many technical
problems.
81

C. Article 2
When does a reproducing female viper
"decide" on her litter size?
Olivier Lourdais 1 2 3 , Xavier Bonnet 1 , Richard Shine 4 and Emily N. Taylor 5
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 Biological Sciences A08, University of Sydney, NSW 2006 Australia
5 Department of Biology, Arizona State University, Tempe, AZ, 85287-1501,USA
Accepted for publication in Journal of Zoology (London)
82

Abstract
Some organisms rely on stored energy to fuel reproductive expenditure (capital
breeders) whereas others use energy gained during the reproductive bout itself
(income breeders). Most species occupy intermediate positions on this continuum,
but few experimental data are available on the timescale over which food intake can
affect fecundity. Mark-recapture studies of free-ranging female aspic vipers have
suggested that reproductive output relies not only on the energy in fat bodies
accumulated in previous years, but also on food intake immediately prior to ovulation.
We conducted a simple experiment to test this hypothesis, by maintaining females in
captivity throughout the vitellogenic period and controlling their food intake. A
female's energy input strongly influenced the amount of mass that she gained and
the number of ova that she ovulated. Multiple regression showed that litter size in
these animals was affected both by maternal body condition in early spring (an
indicator of foraging success over previous years) and by food intake in the spring
prior to ovulation. Our experimental data thus reinforce the results of descriptive
studies on free-ranging snakes, and emphasise the flexibility of energy allocation
patterns among vipers. Reproducing female vipers may combine energy from
"capital" and "income" to maximise their litter sizes in the face of fluctuating levels of
prey abundance.
Key words: capital breeding, vitellogenesis, fecundity, snakes
83

Introduction
Reproduction requires a considerable expenditure of energy, especially in female
organisms that produce large clutches or litters relative to their own body mass.
Because food availability fluctuates through time for many species, coupling energy
acquisition (feeding) with expenditure (reproduction) is not a trivial problem. The
notion of "capital" versus "income" breeding offers a useful conceptual framework in
which to explore such issues (Drent & Daan 1980). Capital breeders gather the
energy to fuel reproduction long before the actual reproductive event, whereas
income breeders simultaneously gain and expend energy. However, these
definitions probably describe the extremes of a continuum. Most kinds of organisms
probably depend to some degree on both kinds of resources to support reproductive
expenditure. Indeed, different facets of reproductive output within the same
reproductive bout by a single female (such as offspring size versus number) may
depend upon different timescales of energy acquisition (Bonnet et al. 2001b).
Squamate reptiles provide good models for studies on this topic, because they
display a diversity in systems of energy allocation. Some of the most extreme
examples of capital breeding systems are found among viperid snakes, with the aspic
viper (Vipera aspis) perhaps the most intensively studied capital-breeder (Saint
Girons 1949, 1957a,b; Saint Girons & Duguy 1992; Bonnet 1996; Naulleau et al.
1999; Bonnet et al. 1999a, 2000a,b, 2001b, 2002b). Female vipers typically
reproduce only once every two to three years and sometimes less often (Saint Girons
& Naulleau 1981). They delay reproduction until they attain a threshold value for
body condition (Naulleau & Bonnet 1996) and females can reproduce successfully
even if they do not feed during the entire year in which the litter is produced (i.e.,
throughout vitellogenesis plus gestation). Thus, "capital" stored prior to reproduction
84

can support the entire reproductive effort of female aspic vipers. Nonetheless, field
data indicate that female vipers often feed during vitellogenesis, and suggest that
food acquired at this time can influence some components of reproductive output
(Bonnet et al. 2001b, Lourdais et al. 2002b). These results suggest flexibility in the
system of energy allocation, whereby food intake in the current year, as well as long-
term storage of energy gained during previous years, can influence a female's
reproductive output.
This scenario concerning the sources of energy for litter production in aspic
vipers is, however, based largely on descriptive studies of free-ranging snakes. In
these studies, maternal feeding rates have been inferred from maternal mass gain
(Bonnet et al. 2001b). Although logic and indirect evidence support the assumption
that these two traits are linked, it remains possible that other factors (such as
maternal disease or metabolic expenditure) could also modify rates of gain in body
mass. If so, correlations between mass gain and reproductive output might reflect
such additional factors, rather than a straightforward effect of enhanced feeding rates
on litter sizes. Experimental manipulation of food supplies offers a direct and
powerful approach to resolving such uncertainties (Ford & Seigel 1989b; Seigel &
Ford 1991,1992; Gregory & Skebo 1998). To investigate the relative influences of a
female viper's initial body stores and subsequent energy intake during vitellogenesis
on her reproductive output, we maintained vipers in captivity, directly modified their
rates of prey consumption, and examined the effects of this manipulation on the
numbers of offspring that they produced.
85

Material and Methods
Study species
The aspic viper (V. aspis) is a small viviparous snake that is abundant in central
western France. In this area, females typically reproduce on a less-than-annual
schedule (Saint Girons 1957a,b; Bonnet & Naulleau 1996; Naulleau & Bonnet 1996;
Naulleau et al. 1999). In reproductive females, the recruitment of ovarian follicles is
controlled by endocrine factors at the onset of vitellogenesis in March (Bonnet et al.
1994). After this initial phase, follicular atresia (i.e. death of ovarian follicles prior to
ovulation, Méndez-De la Cruz et al. 1993) is the proximate factor controlling litter size
(Saint Girons 1957b; Saint Girons & Naulleau 1981). That is, females initially begin
to enlarge more follicles than they eventually ovulate, and hence are potentially able
to adjust their eventual (ovulated) litter size depending on conditions that they
experience prior to ovulation. Ovulation occurs during the first two weeks of June
(Naulleau 1981), and parturition occurs two to three months later, from late August to
late September.
Capture and housing
A total of 108 snakes (48 males and 60 females) was captured in spring 2000 in
three adjacent localities in West-central France (Château d’Olonnes, Les Sables
d’Olonnes and Rochefort). The snakes were collected from late February to mid-
April, when they first emerged to bask after the winter hibernation period. Individuals
were given identification marks by scale-clipping, and were measured (snout-vent
length, SVL + 0.5 cm) and weighed (+ 0.1 g). Mating occurred in captivity, with each
female given a 10-day period of contact with numerous males in an indoor enclosure
(2.5 X 1.5m) with a heat source and water. Copulation was frequently observed in
86

this enclosure. After mating (late March), the snakes were examined by abdominal
palpation to detect vitellogenic follicles. This procedure revealed that thirty-nine
females were reproductive and twenty-one were non-reproductive.
The thirty-nine vitellogenic females were placed in eight outdoor enclosures (5 X 3m,
mean density: 5 snakes/enclosure) recreating the natural habitat and exposed to the
climatic conditions of the field research station of Chizé (Forêt de Chizé, Deux-
Sèvres, 46°07’ N, 00°25’ W). The enclosures were located side by side and did not
differ in term of orientation and sunlight exposure. Each enclosure was equipped
with the same number of external dens to serve as hiding-places. Water was
provided ad libitum and vegetation (mainly annual grasses, Poacae) was kept high
(20 - 40 cm) to provide shade and shelter.
Experimental design
To examine the effects of absolute energy intake during vitellogenesis on subsequent
litter sizes, the 39 females were randomly allocated to one of two diet treatments:
high-intake group (n = 19, enclosures 1 to 4): one large mouse (mean mass: 25 ± 5g)
per snake per feeding in each enclosure, provided on four occasions (every two
weeks from mid-April to early June).
low-intake group (n = 20, enclosures 4 to 8): one small mouse (14 ± 4g) per snake
per enclosure, offered on two occasions only (mid-April and mid-May).
Snakes of both treatment groups were fed by placing recently killed mice close their
dens, early in the afternoon when climatic conditions were favourable. On each
feeding occasion the mass of prey offered was the same (± 1g) for each replicate
enclosure within each treatment. This method allowed us to calculate the total mass
of prey consumed (g) for each snake during the experiment. Uneaten prey items
were removed the next morning. Prey consumption was recorded by direct
87

observation of feeding, or by less direct means if feeding was not observed (by
palpation of mice inside the snake and by increase in body mass).
Snakes were weighed three times during the study: early vitellogenesis (early April),
mid-vitellogenesis (mid-May), and close to the time of ovulation (mid-June). Enlarged
ovarian follicles were counted by palpation (Fitch 1987) at mid-vitellogenesis (mid
May) and ovulation (mid June). This method enables us to detect objects as small as
2 g (Bonnet et al. 2001b). One female from the low-intake group was killed by a feral
cat in early June, and hence data from this individual were not used in most of our
analyses.
Our two treatment groups differed both in prey size and number in order to
mimic the natural situation where snakes may sometimes encounter relatively few,
small prey and in other years may encounter prey items that are larger and also more
abundant. Thus, the treatments were designed to span the normal range of variation
in food supply in terms of both prey size and prey number. Feeding frequency and
relative prey sizes are important parameters of snake biology that may influence
conversion efficiency (Secor & Diamonds 1995). However, the aim of our experiment
was simply to modify the level of energy available for follicular growth and thus, the
important concern was to generate variations in feeding opportunities comparable to
annual variations in food consumption that occur in the field (Lourdais et al. 2002b).
Statistics
Data were analysed using Statistica 5.1. To provide an index of body condition
(mass relative to length), we calculated residual values from the regression of log-
transformed body mass against log-transformed body length (Jayne and Bennett
1990). We compared snout vent length (SVL) of the two groups of females at the
beginning of the experiment in terms of size-frequency distributions as well as mean
88

values. To do so, the data for SVL were standardised (Z = (X - mean value)/SD) so
that the distribution had a mean of zero and standard deviation (SD) of one. Three
size classes were identified: small (Z

Reproductive females in the two food-intake groups did not differ in mean
body mass (ANOVA, F(1.38) = 0.66; p = 0.42), mean SVL (ANOVA, F(1.38)= 0.50; p
= 0.48), or initial body condition (ANOVA, F(1.38)= 0.25; p = 0.62). The two diet
groups were also similar in term of size-frequency distributions (χ 2 = 0.025; df = 2; p
= 0.98). For thirty-seven of the thirty-nine reproductive females, ovarian follicles were
detected each time that we palpated the animals. For the remaining two animals,
however (one in each diet treatment), palpation in mid-vitellogenesis revealed no
detectable eggs. Thus, these animals commenced vitellogenesis but terminated the
process prior to ovulation. These individuals were in lower body condition (residual
values < -0.14 ) that were the other females belonging to the same size class at the
onset of the experiment (residual values > -0.07). Also, neither of these snakes ate
the first prey item they had been offered. After deleting these two cancellations,
females in the two treatments still did not differ significantly in either mean body mass
(p = 0.4), mean SVL (p = 0.43) or more importantly, body condition at the beginning
of the experiment (p = 0.69).
Food intake and changes in body mass
Among the 19 individuals of the low-intake group, eleven females ate one prey item
and eight females ate two prey items. Among the 17 females of the high-intake
group, four females consumed only one mouse, seven females consumed two mice
and six females ate three mice. Unsurprisingly, the total amount of prey ingested (g)
differed significantly between the two groups (Kruskal-Wallis Test: H (1, n = 37) =
15.4, p = 0.0001), with mean values (+SD) of 49 ± 19 of prey consumed by females
in the high-intake group compared to 22 ± 7g for the low-intake group. The variance
in food intake was also higher for the high food group (χ 2 = 0.88, df =1, p < 0.0001),
reflecting the higher number of feeding opportunities within this group.
90

We pooled data from the two groups to examine the relationship between absolute
food intake and subsequent changes in body mass. The two variables were
significantly correlated (r = 0.89, r 2 = 0.79, n = 36, F(1.35)= 89.79; p < 0.0001, see
Figure 2), with energy intake explaining 79% of the observed variance in mass gain.
Food intake was not related to female SVL (r = 0.22; n = 36; F(1.35)= 1.76; p = 0.20).
Changes in body mass (g)
40
30
20
10
0
-10
r 2 = 0.79, n = 36, p < 0.0001
-20
10 20 30 40 50 60 70 80
Cumulative food intake (g)
Figure 2. The relationship between total food intake (g) of 36 female aspic vipers and change in body
mass (g) during the experiment. Each point represents an individual snake. The two treatment groups
were pooled for this analysis.
The difference in average food consumption between the two treatments generated
significant differences in the amount of mass gained by females. The increment in
mass midway through vitellogenesis was greater for the high-intake group (ANCOVA
using mass gain as dependent variable, treatment as factor and initial body mass as
covariate, F(1.34)= 4.29; p = 0.046). The magnitude of this difference was enhanced
by the time of ovulation (F(1.33)= 8.00; p = 0.007).
91

The influence of feeding treatment on change in maternal body mass was examined
with a repeated measure ANOVA. Using SVL-adjusted body mass as the dependent
variable, treatment as factor and the three consecutive records of masses as
repeated measures, revealed a significant interaction (Wilk’s lambda = 0.75; F(3.29)=
3.18; p = 0.038). The mass gain was significantly more marked in the high-intake
group (treatment effect, F (2.62)= 5.77; p < 0.005; see Figure 3).
SVL- adjusted body mass
125
120
115
110
105
100
Early
vitellogenesis
Mid
vitellogenesis
high food intake
low food intake
Ovulation
Figure 3. Influence of experimental treatment (diet) on changes in body mass of female vipers at three
times during the experiment. Black circle, continuous line: high food intake group; black triangle,
dotted line: low food intake group.
Determinants of fecundity
The number of ova palpated in mid-vitellogenesis was greater in the high-intake
group than in the low-intake group (one-way ANCOVA with number of palpated ova
as the dependent variable, treatment as the factor and SVL as the covariate;
92

F(1.34)= 9.04; p < 0.005). The divergence was even greater at ovulation (F(1.33)=
10.68; p < 0.0025, see Figure 4). The SVL-adjusted numbers of palpated ova were
respectively 7.1 ± 2.46 (high-intake group) and 5.14 ± 1.89 (low-intake group) in mid-
vitellogenesis and 6.67 ± 1.82 and 5.03 ± 1.47 at the time of ovulation.
number of ovulated eggs
10
8
6
4
2
0
non restricted
p < 0.0025
Treatment
restricted
Figure 4. Effects of experimental manipulation of food intake on the number of ova ovulated by aspic
vipers. Error bar represent the standard deviation.
A regression analysis pooling data from the two treatments confirmed that
food intake during vitellogenesis influenced the number of ova at ovulation (r = 0.42,
r 2 = 0.17, n = 36, p = 0.01). In addition to food intake, female SVL and early body
condition are known to influence fecundity in aspic vipers (Bonnet et al. 2001b). We
thus included all three of these variables in a multiple regression analysis. As
predicted, the highest proportion of fecundity variation in our data set was explained
by a model including all three predictor variables (r 2 = 0.41, see Table 1). When
univariate regression analyses were conducted separately on data from the two
93

treatment groups, however, strong differences were evident. Early body condition
significantly affected fecundity in the high-intake females (r = 0.71, r 2 = 0.50, n = 17,
p < 0.02) but not in the low-intake group (r = 0.34, n = 19, p = 0.56). Similarly,
fecundity increased with maternal SVL in the high-intake animals (r = 0.56, n = 17, p
= 0.018), but not in the low-intake group (r = 0.33, r 2 = 0.31, n = 19, p = 0.16). As a
consequence, the combination of these two factors in a multiple regression explained
a significant fraction of fecundity variation among the high-intake females (r = 0.69, r 2
= 0.46, n = 17, p = 0.01), but not among the low-intake snakes (r = 0.34, r 2 = 0.11 n
= 19, p = 0.56).
Table 1. The effects of maternal body size (SVL), food intake during vitellogenesis (FOOD), and
maternal body condition (BC) prior to vitellogenesis (residual scores from log-transformed mass
versus SVL) on litter size of female aspic vipers. Results shown are from a multiple regression carried
out on the entire data set (i.e., including female vipers from both treatment groups). The highest
proportion of fecundity variance was explained by a model that included all three independent
variables.
Multiple Regression r = 0.64; r² = 0.41; n = 36; F(3.32)= 7.3265; p = 0.00071
BETA Partial correlation Semi partial p-value
SVL 0.36 0.36 0.42 0.012
FOOD 0.39 0.38 0.44 0.008
BC 0.30 0.30 0.36 0.033
94

Discussion
The experimental data from this study strongly support the conclusions of a recent
descriptive study by Bonnet et al. (2001b) on free-ranging snakes. That is, the
number of ova (litter size) produced by a female aspic viper is influenced not only by
her body condition at the beginning of the year in which she will reproduce, but also
by the amount of food that she consumes during the period of vitellogenesis. The
body size of the female snake also plays a role in determining fecundity, probably
because a larger female has more abdominal space in which to accommodate the
litter (Saint Girons 1957a). Thus, litter size in a female V. aspis is affected in a
complex way by her body size, her pre-existing energy stores, and her food intake in
the weeks immediately prior to ovulation.
Experiments on the influence of energy input on reproductive output have
provided valuable information in many vertebrate species (Arcese & Smith 1988,
Bolton et al. 1993, Monaghan, P. & Nager. R.N. 1997 and references therein)
including snakes (Ford & Seigel 1989b, 1994, Seigel & Ford 1991, 1992, 2001,
Gregory & Skebo 1998). However, in a capital breeding species, the levels of pre-
vitellogenic maternal reserves largely determine reproductive output (Bonnet et al.
2001b). The possible effects of additional income energy (i.e. through manipulation
of the diet) should be framed within this context. Unfortunately, quantitative data on
the processes involved in of mobilisation of body reserves for follicular growth are
available in only one snake species: Vipera aspis. Both empirical and experimental
works on this species have shown that a peak in 17- β oestradiaol level triggers body
reserve mobilisation and yolk deposition (Bonnet et al. 1994, Bonnet 1996). In the
absence of equivalent data on the hormonal control of vitellogenesis in other species,
interspecific comparisons would be premature.
95

The female's "decision" as to whether or not to reproduce appears to be
determined primarily by her initial body condition, making V. aspis a typical "capital
breeder" in this respect (Naulleau & Bonnet 1996). However, the two "cancellations"
(females that initiated but did not maintain vitellogenesis) suggest that food intake
during vitellogenesis might also play a role in this early "decision". That is, a female
close to the energy-store threshold for reproduction may begin vitellogenesis, but
abandon the process unless she obtains prey relatively soon. Resorption of follicles
is probably adaptive, enabling the animal to recover resources if reproduction does
not proceed (Blackburn 1998).
Our experimental manipulation of food intake significantly modified not only
the amount of maternal mass that was gained, but also the number of ova that were
ovulated (and hence, litter size). Female body size and initial body condition also
affected fecundity in aspic vipers, as they do in other species of snakes (Seigel &
Ford 1987). Pooling data from the two treatment groups allowed us to detect
significant influences of maternal SVL, fat stores and food intake on fecundity.
However, conducting the analyses separately revealed that the effects of body
reserves and body length were only statistically significant in the high-intake group.
The differing importance of body length as an influence on fecundity may reflect the
fact that total abdominal space (and thus, female body length) became a significant
constraint on fecundity in the high-intake snakes. If so, we would expect to see
larger litters in larger females than in smaller animals. In contrast, because snakes in
the low-intake treatment group produced small litters fitting easily within the females'
bodies, maternal body size was not a constraint (nor a correlate) of the variation in
fecundity among these animals.
The differing role of pre-existing energy reserves is less easy to explain, but
may simply reflect the fact that the variation in food intake (and thus, the number of
96

developing ova) among the high-intake females was higher than in the low-intake
snakes. Data collected by Saint Girons & Naulleau (1981) demonstrate that in female
aspic vipers the mass of abdominal fat bodies is correlated with the number of
growing follicles. Our results show that, after this initial phase, an exogenous source
of energy will affect the number of ovulated eggs. The greater variation in food intake
and reproductive traits might explain why we detected correlates of reproductive
output more easily in the high-intake group than in the low-intake snakes.
Our study confirms that V. aspis is a typical capital breeder in some respects,
notably in the "decision" whether or not to reproduce. Thus, capital stores (which
reflect energy intake over long time periods) will ultimately determine reproductive
frequencies in aspic vipers. In contrast, litter sizes will be determined not only by
those pre-existing stores, but also by the female's foraging success in the weeks
immediately prior to ovulation (Bonnet et al. 2001b). This flexibility in energy
allocation enable the animals to adjust reproductive investment relative to local
resource levels and is consistent with field data (Lourdais et al. 2002b). Capital
breeding is widespread among ectotherms (Doughty & Shine 1997; Bonnet et al.
1998) and may be particularly advantageous in situations of strong inter-annual
fluctuations in food availability (Calow 1979). Vipera aspis is a sit-and-wait predator
which feeds mainly on rodents, especially voles (Microtus arvalis) that show dramatic
fluctuations in population density (Delatre et al. 1992, Lourdais et al. 2002b). In a
situation where prey densities are unpredictable from one year to the next, a female
aspic viper directly benefit from being able to:
(1) reproduce successfully without having to depend upon food intake during the
reproductive year. In a year when prey is scarce, a female relying upon "income"
might fail to breed successfully, and either waste resources already invested or
97

threaten her own survival. "Capital breeding" provides a mechanism for a temporal
dissociation between feeding and breeding.
(2) modify her reproductive output in a flexible fashion depending upon current levels
of prey availability. Essentially, this flexibility allows the female to manipulate energy
investment into reproduction and hence take advantage of "good" years by producing
more offspring than would have been expected from the long-term average levels of
prey availability in her habitat. Thus, although many aspects of reproduction in aspic
vipers are driven by "capital" stores, some degree of reliance on “income” may help
to fine-tune reproductive expenditure to food intake during the critical phase of egg
production.
Acknowledgements
We thank Gwénael Beauplet for helpful comments on the manuscript. Financial
support was provided by the Conseil Régional de Poitou-Charentes, Conseil Général
des Deux Sèvres, the Centre National de la Recherche Scientifique (France).
Manuscript preparation was supported by the Australian Research Council. We
thank Pr. Oumid Popoye and Mr Celestin Le khobrato for helping in snake collecting
and the construction of the electrical fence to impede feral cat predation. Finally, Mr
Dujardin helped to solve many technical problems.
98

D. Article 3
Capital-breeding and reproductive effort in a
variable environment: a longitudinal study of a
viviparous snake
Olivier Lourdais 1 2 3 , Xavier Bonnet 1 , Richard Shine 4 , Dale DeNardo 5 , Guy
Naulleau 1 and Michael Guillon 1
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 Biological Sciences A08, University of Sydney, NSW 2006, Australia
5 Department of Biology, Arizona State University, Tempe, AZ, 85287-1501,USA
Published in Journal of Animal Ecology 71 :470-479
(2002)
99

Summary
1. We examined the ways that fluctuations in prey abundance and weather
conditions can affect reproductive output in a "capital breeding" ectotherm, the aspic
viper (Vipera aspis).
2. Our longitudinal study confirms that female aspic vipers adjust reproductive
investment by integrating allocations of energy from stores (“capital”) and facultative
feeding (“income”). Thus, long-term energy storage enabled females to reproduce
successfully even in years when prey were scarce.
3. Not surprisingly, temporal changes in body reserves of female vipers preparing for
reproduction depended upon current feeding rates. However, the mean
environmental temperature during the active season also affected mass gain.
4. Allometric patterns suggest that reproductive output was limited by energy
availability in 8 out of the 9 years of our study. In the other year, high prey availability
in the preceding season meant that reproductive output was maximised within the
constraints set by maternal body size (and thus, abdominal volume).
5. High summer temperatures increased basking opportunities of gravid vipers and
thus accelerated gestation. However, maternal metabolic costs also increased in
such situations, resulting in low post-partum body condition.
Key words: ectothermy, capital breeding, environmental fluctuations, temperature,
food availability
100

Introduction
Variation in reproductive success is a central theme in evolutionary biology.
Theoretical models predict that variation in quality among individuals provides the
basic substrate for natural selection, resulting in a differential contribution to the
number of descendants produced. Intrinsic sources of variation (including genetic
effects), however, are only one source of phenotypic variation. Proximate
(environmental) factors also affect reproductive performance. In many biological
systems, year-to-year variation in environmental characteristics such as food supply
or weather conditions can have dramatic repercussions on reproductive traits such as
clutch size or reproductive frequency. Such influences have been documented in a
variety of taxa (Lack 1954; Ballinger 1977, 1983; Brand & Keith 1979; Todd, Keith &
Fisher 1981; Seigel & Fitch 1985).
To cope with limitations and/or fluctuations of food resources, organisms have
evolved a wide range of strategies for energy acquisition and allocation to
reproduction. One fundamental dichotomy is between species in which reproduction
is fuelled by recently acquired energy ("income breeders") and species where
storage constitutes the primary energy source for reproduction ("capital breeders",
Drent & Daan 1980). For income breeders, reproductive output should be closely
linked to current resource availability, while in capital breeders a temporal separation
should exist between the phase of energy acquisition and investment in reproduction.
Capital breeding may be especially advantageous to buffer resource fluctuations
among years, or if annual food levels are stable but insufficient to permit successful
reproduction (Calow 1979). However, the acquisition and storage of large amounts
of energy requires time and is also potentially costly (Jönsson 1997). Therefore, the
costs and benefits of alternative tactics of resource use (i.e., capital versus income
101

eeding) will vary among species. For instance, many features associated with
ectothermy pre-adapt organisms to store large reserves and to use them for
reproduction (Bonnet, Bradshaw & Shine 1998). The duration of energy gathering
may sometimes cover long periods (years) and therefore often results in a low
frequency of reproduction (Bull & Shine 1979).
In vertebrates, capital breeding systems coupled with infrequent (less-than-
annual) reproduction have been observed in many reptiles (Saint Girons & Naulleau
1981; Brown 1991; Brana, Gonzales & Barahona 1992; Doughty & Shine 1997;
Madsen & Shine 1999). Some of the best examples are among viperid snakes
(Madsen & Shine 1992a; Brown 1993; Martin 1993), with some species showing very
low reproductive rates. For example, female aspic vipers (Vipera aspis Linné) in
western France do not initiate vitellogenesis until they have accumulated sufficient
energy stores to exceed a high body condition threshold (Naulleau & Bonnet 1996).
The time necessary to accumulate body reserves entails a delayed maturity (2.5 to
3.5 years of age with an average lifespan of 5 years; Bonnet et al. 1999a;
unpublished data) and a low breeding frequency (once every 2 to 3 years). Due to
high costs of reproduction, most females will produce only one or two litters within
their lifetimes (Bonnet et al. 2000a; 2002a), and the same may be true for many
temperate-zone viperid species (Madsen & Shine 1992a; Brown 1993).
In west-central France, the habitat of the aspic viper is characterised by strong
annual fluctuations in availability of rodents that are the snakes' main prey (Naulleau
1997). Thermal conditions also vary significantly among years in this area. For an
ectothermic species, variations in the thermal environment may affect the rate of
processes such as digestion, metabolism and reproduction (Huey 1982; Naulleau
1983a, b). In the present paper, we use data from a longitudinal study of a
population of vipers, to examine how annual variation in both food availability and
102

temperature affect reproductive output in a capital breeding ectotherm species.
Specifically, we predicted that:
The long duration of energy gathering prior to reproduction combined with the
female's ability to store large amounts of energy within her body should result in a
high investment per reproductive bout. Nonetheless, because "income" is also
allocated to reproduction (Bonnet et al. 2001b), we expect that litter mass will be
influenced by prey abundance in the year preceding reproduction as well as the
current year.
Capital breeding coupled with a fixed body condition threshold (Naulleau & Bonnet
1996) means that all females initiating reproduction do so with substantial energy
reserves and hence, the success of the litter should not be compromised by an
unanticipated energy shortage. The proportion of healthy offspring versus non-viable
components in a litter should be high, and independent of resource fluctuations.
Among non-reproducing females (i.e., individuals preparing for reproduction), body
reserves accumulated at the end of the activity period should depend upon current
food levels and also be influenced by thermal conditions that determine digestion
rate.
Thermal conditions should directly affect the rate of embryogenesis and thereby
gestation length.
Material and methods
Study Animals
The aspic viper (Vipera aspis Linné) is a small viviparous snake of the western-
Paleartic region and is locally abundant at the northern limit of its distribution in
France. Females mature at 40 cm snout-vent length (SVL), which is attained in 2.5
103

to 3.5 years (Bonnet et al. 1999a). In this area, females typically reproduce at a less-
than-annual frequency (Saint Girons 1957a,b; Bonnet & Naulleau 1996; Naulleau &
Bonnet 1996; Naulleau et al. 1999). Ovulation typically occurs during the first two
weeks of June with limited geographical or altitudinal variations (Saint Girons 1957b,
1973; Naulleau 1981). Parturition occurs two to three months later, from late August
until late September (Bonnet et al. 2001b).
Study site and methods
The study site is near the village of Les Moutiers en Retz in west-central France
(47 o 03N'; 02 o 00W'). It is a 33-hectare grove with a mosaic of meadows and
regenerating scrubland. The site is characterised by a temperate oceanic climate.
From 1992 to 2000, one to four people patrolled the site using a standardised
searching method on almost every favourable day during the vipers’ annual activity
period (late February to late October). The total searching effort exceeded 4,000
hours. Snakes were caught by hand, sexed by eversion of the hemipenes, weighed
to the nearest g with an electronic scale, and measured (SVL and total length) to the
nearest 5 mm. Over 400 adult and sub-adult female vipers were marked using
passive integrated transponder (PIT) tags (Sterile transponder TX1400L, Rhône
Mérieux, 69002 LYON France, product of Destron/IDI Inc). Each snake was then
released at its exact place of capture. Because the study site is surrounded by
habitat unsuitable for vipers (Vacher-Vallas, Bonnet & Naulleau 1999) and this
species is highly philopatric (Naulleau, Bonnet & Duret 1996), any snake not
captured over a long period (> 2 years) had almost certainly died rather than
emigrated or avoided capture.
104

Initial total body length and body mass were measured in March-April (before
any significant food intake). Changes in body mass and body length were calculated
from March-April to August-November within a given year and from March-April to the
next March-April between years. Since our data indicate that vipers did not show any
significant change in either body mass or body length over hibernation, we excluded
that time period from our calculations. To provide an index of body condition (mass
relative to length), we calculated residual values from the regression of log-
transformed body mass against log-transformed body length (Jayne & Bennett 1990).
Initial body condition (calculated in March-April) provides an accurate indicator of the
level of fat stores in female vipers (Bonnet 1996).
Reproductive status was determined using two methods. First, any female
whose initial body condition value was greater than the threshold at this time (March-
April) was considered reproductive (see Bonnet, Naulleau & Mauget. 1994; Naulleau
& Bonnet 1996). Second, from mid-vitellogenesis (May) to the end of gestation (late
August) reproductive status was easily determined either by palpation of ova and/or
embryos or by records of parturition (Fitch 1987; Naulleau & Bonnet 1996). Gravid
females were captured and maintained in captivity after the first parturition of the year
was recorded (generally in late August). For each year, mean annual changes in
maternal mass prior to parturition were calculated as pre-partum body mass minus
initial body mass. Date of capture of pregnant females had no significant effect on
this parameter for two main reasons. First, all pregnant females were re-captured
over a short period (one to two weeks) at the end of gestation, by which time there
was little further mass change before parturition. Second, during the active season,
food intake occurs mainly during vitellogenesis in spring and almost cases after
ovulation (Saint Girons & Naulleau 1981, Bonnet et al. 2000a; Lourdais, Bonnet &
105

Doughty 2002a). Overall, then, pre-partum mass change provides a robust indicator
of food intake during reproduction.
Pregnant females were maintained in separate cages in the laboratory until they
gave birth one to 21 days later. Mass was recorded every two days and immediately
after parturition. For each female, we calculated pre-partum and post-partum body
condition. Body condition of pre-partum females includes both the litter mass and
female carcass mass, whereas post-partum body condition offers an indicator of the
degree of female emaciation (Bonnet et al. 2000a).
Reproductive data were obtained on 173 litters from 149 different females. For
most individuals (129) only a single litter was obtained, but 16 females produced two
litters and 4 individuals produced three litters. Treating these successive litters by
the same female as independent data may introduce problems with
pseudoreplication. However, none of our conclusions from statistical tests differed
depending on whether these ‘repeat” litters were included or excluded. Thus, the
following analysis presents calculations based on the total data set. The components
of the litter were characterised (yolked eggs, dead offspring, healthy offspring),
counted, and weighed (± 0.1g). Additionally, young and stillborn were measured (±
0.5 cm) and sexed. For one individual, total litter mass was not available.
We could not distinguish unfertilised ova from ova that had been fertilised but had
died early in embryogenesis. Hence, eggs where only yolk was visible (fertilised or
not), underdeveloped embryos and stillborn were all grouped in the same “non-
viable” category. Litter mass and litter size included all of these elements along with
healthy offspring, whereas "fit litter mass" and "fit litter size" included healthy
offspring only (Gregory, Larsen & Farr 1992). Relative litter mass (RLM) was defined
as the residual score from the general linear regression of litter mass against post-
106

parturient mass of the mother. Gestation periods were calculated from parturition
dates, assuming a fixed ovulation date in mid June (Naulleau 1981).
Environmental factors
Food levels - Vipers feed mainly on voles (Microtus arvalis Pallas) whose populations
typically fluctuate in a three to four year cycle (Delatre et al. 1992). Variations in vole
abundance directly influence rates of feeding by the snakes, as revealed by the
proportion of snakes captured with a prey item in the stomach (Bonnet et al. 2000a;
2001b). Data on the annual proportions of adult snakes containing prey at the time
of capture were available from 1993 to 2000. In the following analysis we used this
parameter as an index of food abundance. This estimator of food levels provides
results that are consistent with those from line trapping; for example, the same
annual peaks in vole density are detected by both methods (Salamolard et al. 2000;
Bonnet et al. 2000a). However, feeding rates of the snakes provide a more sensitive
measure of prey abundance, because trapping can fail to detect voles at low
population densities. We acknowledge that feeding rates may also be sensitive to
environmental temperature (due to thermal effects on snake activity and foraging
success) and hence, our measures of annual variation in temperatures and in food
supply would not be independent. However, we found no significant correlation
between mean annual temperature and the proportion of snakes captured with a prey
item in the stomach (see below). Because the aspic viper is a very sedentary animal
occurring at a high density in our population, feeding rates are likely to be tightly
linked to prey abundance.
Temperatures - the study site is near the northern limit of the species' range and,
therefore, climatic conditions may constrain the animals' reproductive biology. This
inference is supported by the fact that parturition in southern populations occurs in
107

mid July, two months earlier than in West-central France (Zuffi, Giudici & Iolae,
1999). Throughout the nine years of our study, daily thermal maxima (°C) under
shelter were recorded. Two biological periods were considered: the active season
(March to October) and the gestation period (mid June to August).
Results
Fluctuation in environmental factors
Food levels - feeding rates varied markedly during the course of our study (χ ² =
25.53, df =7, p < 0.0006). Two years were distinguishable, with one year of very low
(1994) and one of particularly high (1996) food levels (see Figure 1).
Food index
0.5
0.4
0.3
0.2
0.1
0.0
1992 1993 1994 1995 1996 1997 1998 1999 2000
Year
Figure 1. Annual variation in the proportion of aspic vipers containing prey items in the stomach at the
time of capture. Analyses in this paper use this proportion as an index of availability of prey for the
snakes. Most of these prey were voles (Microtus arvalis). See text for details on the method.
108

Temperature - during the course of the study, mean daily temperature during the
snakes' active season (and during the gestation period) varied significantly among
years (ANOVA, F(8, 2196) = 3.88; p < 0.0001; F(8, 819) = 2.89; p < 0.0035,
respectively, see Fig 2). We found no significant relationship between feeding rates
and mean temperature during the active season (r = 0.27, n = 8, F(1, 6) = 0.5, p =
0.50).
Mean temperatures (active season)
25
24
23
22
21
20
19
1992 1993 1994 1995 1996 1997 1998 1999 2000
Year
Active season
Gestation
Figure 2. Annual variation in thermal conditions over the course of our study. Mean ambient
temperatures (± S.E.) under cover items were calculated separately for the active season (March to
October; black circles, continuous line) and for the gestation period (Mid-June to August; black
triangles, dashed line).
Influences on viper reproduction
1 - Annual fluctuations in food levels and reproductive output
a) Variation in litter mass and reproductive effort
Reproductive investment was generally high (mean litter mass = 34 ± 15g),
representing on average 52% of female post-partum body mass. Mean litter mass
109
28
27
26
25
24
23
22
21
20
Mean temperatures (gestation)

showed significant annual fluctuations (ANCOVA, F(8, 160) = 2.26, p < 0.02; using
litter mass as the dependent variable, female body length as a co-factor). Among the
173 litters obtained, 15 individuals produced entirely non-viable litters (e.g., only
undeveloped ova and still-born offspring) and thereby had very low litter mass values.
Even when these non-viable litters were excluded from analyses, mean litter mass
displayed significant annual variation (ANCOVA, F(7, 142) = 3.76, p < 0.0009, Figure
3). The variation was mainly due to two consecutive years of especially high litter
mass values: 1996 (a high food year) and 1997 (an intermediate food year).
Size adjusted total litter mass (g)
55
50
45
40
35
30
25
5
0
1992 1993 1994 1995 1996 1997 1998 1999 2000
Year
Figure 3. Annual fluctuations in mean total litter mass (± S.E.) of female aspic vipers. Values have
been scaled with female body length. Fifteen litters that were entirely non-viable were excluded from
this analysis (see text for statistics).
Substantial year-to-year variations in maternal characteristics were also
detected (see Figure 4). Mean initial body condition of reproducing females varied
110

significantly (ANOVA, F(8, 257) = 3.53, p < 0.0007) with the highest values observed
in 1997 (after a year of high food abundance, see Figure 2). Pre-partum body
condition also showed significant variation (ANOVA, F(8, 160) = 6.27, p

91% of the variance in mean annual litter mass (r = 0.95, n = 8, F(2, 4) = 19.74, p <
0.009; Table 1).
Table 1. Effects of food level in the year of (n) and the year prior to (n-1) reproduction on absolute litter
mass of aspic vipers in south-western France.
Multiple Regression r = 0.95; r² = 0.91; n = 7; F(2, 4) =19.745; p = 0.008
Bêta Partial correlation Semi-partial p value
Food year n 0.83 0.94 0.82 0.005
Food year n-1 0.57 0.88 0.57 0.019
b) Influence of maternal body size
Because maternal body size is highly correlated with litter size or litter mass in many
snakes (Seigel & Ford 1987), we looked for correlations between maternal length and
reproductive output (litter size and litter mass) in our population. Combining data for
all years, we detected a significant but weak influence of maternal body size on
reproductive output (r = 0.20, n = 173, F(1,171) = 7.1, p

Table 2. Correlation between maternal body length and reproductive output (litter size and litter mass)
for each year of the study. The power (1-β) of the analysis and the sample size required for α

was high during the study. Fit litter mass averaged 86% of total litter mass and, not
surprisingly, these two values were strongly correlated (r = 0.95, n = 171, F(1, 169) =
1747.0, p < 0.0001). To avoid statistical problems associated with ratios (Seigel and
Ford 1987), we calculated residual values from the regression of fit litter mass
against litter mass to provide an index of reproductive “efficiency”. This adjusted
measure of fit litter mass peaked in 1994 (ANOVA, F(8, 160) = 3.21, p < 0.0021, see
Figure 5; considering only years where data on more than 10 litters were obtained,
F(3, 118) = 5.94; p < 0.0008).
Adjusted fit litter mass (g)
40
35
30
25
20
5
0
1992 1993 1994 1995 1996 1997 1998 1999 2000
Year
Figure 5. Annual fluctuations in our measure of reproductive “efficiency” based on the regression of fit
litter mass against litter mass for aspic vipers. For simplicity, the graph shows fit litter mass values
adjusted for total litter mass (± S.E.).
Paradoxically, reproductive “efficiency” was highest in 1994 (i.e., during a year of
food scarcity) and lowest in 1996 (high food year). Considering fit litter size rather
than fit litter mass did not change those results. This annual variation mostly reflected
changes in the proportions of litters with at least one non-viable component (χ² =
114

9.92; dl = 3; p = 0.02) rather than annual variation in proportional viability among
litters with at least one non-viable component (ANOVA, F(4, 67) = 0.64, p = 0.64).
Females with low reproductive success (i.e., > 60% of the litter non-viable) did not
have decreased survival (χ² = 0.13, dl = 1, n = 154, p = 0.71), but they did have
longer gestation periods (ANOVA, F(1, 170) =14.83, p

Annual mass change (g)
80
70
60
50
40
30
20
10
0
1992 1993 1994 1995 1996 1997 1998 1999 2000
Year
Growth
Mass change
Figure 6. Fluctuations in annual growth rates (open hexagons, dashed line) and annual mass
changes (open square, dotted line) for non-reproducing female V. aspis. Values were scaled with
female initial (spring) body size. Error bars represent standard errors.
b) Interaction between food levels and thermal conditions
A multiple regression analysis revealed a significant combined influence of food
levels and mean temperature during the active season on annual mass gain of non-
reproducing females. These two factors explained 96% of the variance in mean
annual mass gain (see Table 3). Interestingly, we were not able to detect any
significant interaction between food levels and thermal conditions in the
determination of annual rates of growth in body length (partial correlation: r = 0.14, n
= 8, F(1, 6) = 0.5, p = 0.76).
116
8
7
6
5
4
3
2
1
0
Annual growth (cm)

Table 3. Combined influences of mean ambient temperature during the active season (Mean temp)
and food levels on adjusted annual mass change of non-reproducing female vipers.
Multiple Regression r = 0.96; r² = 0.92; n = 8; F (2, 5) = 30.9; p < 0.0015
Bêta Partial correlation Semi-partial p value
Mean temp 0.56 0.89 0.54 0.007
Food level 0.95 0.96 0.92 0.0007
3 - Influence of thermal conditions during the gestation period
a) Reproductive success
Mean daily temperature during pregnancy did not affect the proportion of viable
neonates (r = 0.08, n=173, F(1, 171) = 1.12, p< 0.29).
b) Gestation length
Mean daily temperature during pregnancy strongly affected the length of gestation (r
=- 0.47, n = 173, F(1, 171) = 49.36, p < 0.00001, Figure 7).
Gestation length (days)
105
100
95
90
85
80
75
23.0 23.5 24.0 24.5 25.0 25.5
Mean temperature during gestation (°C)
Figure 7. Influence of mean environmental temperature during the gestation period (summer) on the
duration of gestation in free-ranging aspic vipers. For simplicity, the graph shows mean annual
gestation length. See text for detailed statistical analyses of these data.
117

Because low reproductive success (i.e., > 60% of litter non-viable) also affected
gestation length, we re-analysed the data excluding these litters, thus enabling us to
focus on the effect of temperature (r = - 0.61, n = 138, F(1, 135) = 83.01, p <
0.00001). Mean temperatures during pregnancy explained 40% of the variance in
gestation length.
c) Maternal condition post-partum
We found a negative correlation between mean summer temperature and mean post-
partum body condition (r = 0.24, n = 172, F(1, 170) =10.7, p < 0.00125). Post-partum
body condition in this species is also positively influenced by energy acquired during
the year of reproduction (Bonnet et al. 2001b). We found a positive correlation
between mean mass changes prior to parturition and post-partum body condition, (r =
0.84, n = 8, F(1, 7) =17.56, p < 0.004). We reanalysed the data holding pre-partum
body condition constant (i.e., an indirect measure of energy acquired during the year
of reproduction). The partial correlation analysis supported our initial finding (Table
4).
Table 4. Combined influences of temperature during gestation (Ges temp) and food intake during the
year of reproduction (indirectly measured using pre-partum body condition value) on the post-partum
body condition of female aspic vipers.
Multiple Regression r = 0.45; r² = 0.21; n = 169; F(2, 166) =22.05;

Discussion
Previous studies on this population have shown that the reproductive “decisions” of
female aspic vipers largely rely on stored energy reserves ("capital breeders":
Naulleau & Bonnet 1996; Bonnet et al. 2001b). Demographic data clearly identify
prey-driven variations in the proportion of reproducing females in this population, with
food levels in a given year influencing the number of reproductive females the
following year (Bonnet et al. 2001b; Lourdais et al. 2002d). In the present work, we
have focussed on annual variation in reproductive output to further understand the
system of energy acquisition and allocation in this species. As predicted, capital
breeding allowed for a high reproductive investment regardless of food availability in
the current year. However, reproductive investment varied among years, with mean
litter mass higher in 1996 and 1997 than in other years (Figure 3). Reproducing
females gained substantially in body condition during the year of highest prey
availability (1996), reflecting high energy intake during that year. High feeding rates
at this time also resulted in high initial body mass of reproducing females early in the
following season (1997). That is, the high prey availability of a single year (1996)
increased litter masses not only that year but also in the following year as well (by
increasing energy storage of non-reproducing females).
This complex system of energy allocation involving both capital breeding and
facultative income breeding is well illustrated in the multiple regression analysis
combining food levels both in the year of reproduction and in the preceding year
(Table 1). A female viper’s reproductive strategy combines both “rigid” and “flexible”
components. Firstly, female vipers have to reach a fixed body condition threshold to
engage in reproduction (Naulleau & Bonnet 1996). A certain level of flexibility is
evident however, with some pre-reproductive females eating rapidly enough to
119

“overshoot” the body condition threshold, and hence accumulating body reserves
above the fixed threshold (Naulleau & Bonnet 1996). These “extra” reserves are
invested into reproduction and they positively influence litter size (Bonnet et al.
2001b). After this initial phase that determines both the female’s reproductive
decision and the number of follicles she recruits, facultative income breeding enables
her to adjust her reproductive effort to current food levels during vitellogenesis. This
composite system of energy allocation is advantageous for at least two reasons.
First, reliance upon stored reserves secures a high reproductive “efficiency”
independently of current prey availability. The proportion of undeveloped ova or
stillborn offspring produced was low and was not related to food levels. Second,
instead of a completely “rigid” capital breeding system, facultative income breeding
enabled individuals to adjust reproductive investment to local resource levels
reproductive and notably to take advantage of occasional periods when prey are
abundant. A similar flexibility may occur in the closely related adder, Vipera berus
Linné (Andren & Nilson 1983).
The effects of marked fluctuations in food levels were also evident in non-
reproducing females (i.e., individuals preparing for reproduction). Among these
females, change in body mass and growth in body length were both highly
dependent on food levels. However, variance in mass gain was linked not only to
feeding rates, but also to thermal conditions during the active season. This result
may explain the high rates of mass gain observed in 1997, a year when climatic
conditions were particularly favourable but prey abundance was only intermediate
(confirmed by trapping, Salamolard et al. 2000). Thermal conditions may influence
average mass change in several ways. Favourable thermal conditions may influence
average mass change by accelerating digestion rates whereas low temperatures
prolong digestion and may even stimulate regurgitation of the meal (Naulleau 1983a,
120

1983b). While our data show no correlation between feeding rates and thermal
conditions, we can not exclude the possibility of undetected complex interactions
between concurrent fluctuations in thermal factors and prey availability, and therefore
we encourage further study of this relationship.
Several results from our study support an energy limitation model for
reproduction in this population. First, elevated rates of body growth and mass gain
were observed during a year of high food levels (1996). Second, maternal body size
correlated only weakly and inconsistently with reproductive output (litter size and litter
mass) in our aspic vipers. This correlation is high in most other snakes (Seigel &
Ford 1987), perhaps because maternal abdominal volume (rather than energy
supply) generally constrains reproductive output (Shine 1988). In a capital breeder,
however, body stores may provide a greater constraint on reproductive output, and
may often be below the level at which the litter mass would be constrained by (and
hence, correlated with) maternal body size. Under this scenario, we would expect a
positive influence of female body size on fecundity to be more easily detected when
energy intake is sufficient to allow maximisation of body reserves. This prediction is
consistent with our results: the correlation between maternal body size and
reproductive output was significant only in 1997, the "best" year in terms of energy
availability.
While the system of energy allocation allows for adjustment to fluctuations in
food availability, some aspects of reproduction were directly dependent upon
variations in thermal conditions. High midsummer temperature accelerated
gestation, as has been reported previously in captive snakes (Blanchard & Blanchard
1941). Our data also suggest that summer temperatures influence "costs of
reproduction" for females. The maintenance of a higher body temperature and thus
higher metabolic rate (Saint Girons, Naulleau & Célérier 1985) during pregnancy
121

translated into negative effects on female post-partum body condition. During the
course of gestation, female aspic vipers are virtually anorexic and feed only
opportunistically. Embryo maintenance generates substantial fecundity-independent
mass loss in females and constitutes an important component of energy expenditure
(Lourdais et al. 2002a). Our data suggest that the magnitude of this metabolic cost
varied from year to year, depending upon thermal conditions.
This result demonstrates difficulties associated with accurate estimation of
reproductive effort in viviparous ectotherms. Classically, "reproductive effort" has
been quantified using simple measures of reproductive output, such as the ratio of
absolute litter mass to post-partum female body mass (Seigel & Fitch 1984).
However, our results, in combination with related work (Bonnet 2001b, unpublished
data) suggest that the post-partum body condition of a female aspic viper is affected
in complex ways by several factors including her direct investment in the litter
(absolute litter mass), her food intake during the year of reproduction, and her
metabolic expenditure during gestation. Although all of these factors affect female
emaciation (and thus both current and future reproductive effort), they are controlled
by very different variables: (1) investment in the litter depends on energy stores
combined with current food intake; (2) independent of reproductive investment, food
intake during the year of reproduction enhances female post-partum body condition;
and (3) female emaciation is also influenced by climatic conditions. Complex
interactions between the two varying environmental factors (prey abundance and
thermal conditions) thus are likely to affect the degree of female emaciation.
Integrative approaches will be needed to identify the different means by which energy
is expended during reproduction in ectotherms.
In conclusion, our long-term study clarifies the influence of two major
environmental factors (food availability and thermal conditions) on the reproductive
122

iology of female vipers. The snakes' responses to annual fluctuations in food
availability support the hypothesis that reproductive output in V. aspis is determined
by a combination of capital breeding (i.e., the use of energy stores) and income
breeding (i.e., the use of energy from current food intake). This mixed (capital plus
income) system masks but does not eliminate a trend for correlated fluctuations in
prey abundance and reproductive output. The reliance on stored energy also
enables a high and relatively invariant “efficiency” of reproduction (proportion of
viable embryos).
Annual fluctuations in thermal conditions also entailed both direct and indirect
repercussions on viper reproduction. Among females preparing for reproduction,
thermal conditions, in combination with food levels, significantly influenced the
acquisition of energy stores. Additionally, thermal conditions have at least two direct
effects on reproducing females during gestation period. First, gestation length is
influenced by ambient temperatures that determine the thermal regimes experienced
by pregnant females (Naulleau 1986). Second, high midsummer temperatures not
only shorten gestation, but they also increase maternal metabolism and thus
decrease the female's post-partum body condition. Such a decrease may well
translate into a lower probability of survival, or into a delay in production of the
eventual next litter (Bonnet et al. 2002a).
Acknowledgements
We thank Gwenaël Beauplet for comments on the manuscript. Financial support
was provided by the Conseil Régional de Poitou-Charentes, Conseil Général des
Deux Sèvres, the Centre National de la Recherche Scientifique (France). Special
thanks to Melle, notably for being her.
123

III. Les coûts de la
reproduction: amplitude et
degré de dépendance avec
PROBABOLITY OF SURVIVAL (%)
100
90
80
70
60
50
40
30
20
10
0
la fécondité
Reproductive
124
Non
Reproductive

A. Résumé du chapitre:
Au cours de chaque épisode reproducteur, la vipère aspic femelle déploie un effort
particulièrement élevé, tout au moins en comparaison avec les autres vertébrés
amniotes. Ainsi, le rapport entre la masse de la portée et celle de la mère après la
parturition est très élevé, parfois supérieur à 1. Un tel investissement énergétique
traduit la mobilisation massive d’importants stocks de réserves corporelles parfois
associée à l’effet de prises alimentaires facultatives. Les modèles théoriques
prévoient que les relations entre l’effort de reproduction, le succès reproducteur et les
coûts qui les accompagnent sont des éléments primordiaux pour comprendre
l’évolution des traits d’histoire de vie. Dans ce chapitre, les principales composantes
des coûts de la reproduction, mortalité, fréquence de reproduction, dégradation de la
condition corporelle, ralentissement de la croissance, ainsi que leurs valeurs relatives
sont examinés.
Dans un premier temps, le suivi de la population des Moutiers nous a permis
de mesurer l’influence d’une reproduction donnée sur les probabilités de
reproduction futures. Cette étude révèle l’existence de coûts écologiques très élevés
(article 4) avec une très forte mortalité chez les femelles reproductrices en
comparaison avec les non-reproductrices. Cette mortalité réduit dramatiquement le
nombre d’épisode reproducteur dans la vie de l’animal avec une majorité d’individus
“semélipares”. Outre les coûts en survie, nos résultats soulignent plusieurs sources
de coût énergétiques potentiels avec notamment un amaigrissement marqué et une
croissance très réduite (souvent nulle) pendant l’année de reproduction. L’essentiel
des réserves corporelles est investit dans la reproduction courante et une femelle
bénéficiant d’un important stock initial investira d’autant plus d’énergie dans la
reproduction. A la différence des coûts en survie qui se traduisent directement par
125

une baisse du succès reproducteur à vie, les conséquences des coûts énergétiques
potentiels sur le succès des reproductions futures (individus itéropares) sont plus
difficiles à cerner et à quantifier. Cette situation reflète une forte hétérogénéité inter
individuelle dans la cinétique de reconstitution des stocks. Ainsi, indépendamment
de l’effort reproducteur, ce sont les femelles qui vont reconstituer rapidement leurs
réserves l’année suivante qui vont bénéficier d’une meilleure survie et d’une plus
grande probabilité de reproductions futures. Ce résultat souligne l’influence
déterminante de la cinétique de la reconstitution des réserves sur la vie reproductrice
des femelles.
Dans un second temps, nous avons cherché à préciser la nature des
contraintes énergétiques imposées par les activités de reproduction. Si la
vitellogénèse constitue la phase clé de l’allocation énergétique pour la production
des follicules, la gestation, qui peut être considérée comme une forme de soin
parental prénatal , est la période capitale du développement embryonnaire. Comme
la vitellogenèse, elle entraîne de profonds changements éco-éthologiques chez les
femelles reproductrices. Le second article de ce chapitre (article 5) révèle d’ailleurs
l’existence de coûts énergétiques élevés et très spécifiques de la gestation. Pendant
cette phase, le déplacement des préférences thermique vers des températures
élevées est associé à une perte de masse marquée qui résulte de l’épuisement des
réserves corporelles résiduelles à la vitellogenèse. De façon remarquable, la
dépense énergétique n’est pas influencée par le nombre de jeunes en
développement. Cette indépendance de la fécondité suggère clairement une
amplitude “ fixe” des contraintes métaboliques de la gestation, probablement parce
que les températures optimales de développement des embryons ne changent pas
quel que soit leur nombre. En effet, le déplacement des préférences thermiques
pendant la gestation est lié au maintien d’un optimum physiologique pour le
126

développement embryonnaire qui se manifeste sous la forme d’un palier. Pendant
cette période, la thermorégulation des femelles maternelles est donc directement
déterminée par le statut reproducteur et non le nombre de jeunes en développement.
Nos observations de terrain et en captivité confirment les publications
antérieures suggérant que les femelles gestantes réduisent fortement leur prise
alimentaire. Toutefois, nos analyses révèlent que ces pertes en opportunités
alimentaires sont indépendantes de la fécondité. La gestation génère donc des
pertes d’opportunités énergétiques qui s’ajoutent aux dépenses métaboliques
maternelles. En captivité, l’alimentation pendant cette période limite l’impact
métabolique de la gestation et améliore l’état des femelles post-parturientes. Notre
étude sur la consommation en oxygène pendant le gestation (article 6) vient
confirmer l’impact du régime thermique maintenues par la femelle reproductrice par
rapport aux non-reproductrices sur le métabolisme et indique à nouveau une
influence très réduite de la fécondité sur la consommation en oxygène. Par exemple,
l’effet du nombre d’embryons vivants sur la consommation d’oxygène n’est détecté
qu’en fin de gestation, lorsque le développement des vipéreaux s’achève et leur
consommation en oxygène devient détectable .
Enfin, dans un dernier volet (article 7), nous avons cherché à identifier
l’échelle temporelle sur laquelle se manifestent les coûts énergétiques de la
reproduction. Notamment si certains coûts s’expriment pendant la reproduction
(prédation par exemple), d’autres composantes, vont pouvoir intervenir avec un
certain décalage temporel étendu. Un tel décalage peut être particulièrement
important pour des espèces comme la vipère aspic chez lesquelles de longues
périodes sont nécessaires pour la reconstitution des réserves pour la reproduction.
Nos résultats indiquent qu’une fraction importante de la mortalité s’exprime un an
après la reproduction. Cette mortalité post-reproductrice est liée à l’épuisement
127

physiologique des femelles après la mise bas. La vipère aspic est donc affectée par
une combinaison complexe de coûts se manifestant à la fois l’année de la
reproduction (“breeding cost”) mais aussi l’année suivante (“post-breeding cost”)
alors que les coûts associés à la phase de capitalisation des réserves avant la
reproduction (“pre-breeding cost”) semblent plutôt réduits.
128

B. Article 4
Reproduction in a typical capital breeder:
costs, currencies and complications in the
aspic viper (Vipera aspis)
Xavier Bonnet 1 , Olivier Lourdais 1 2 3 , Richard Shine 4 & Guy Naulleau 1
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 Biological Sciences A08, University of Sydney, NSW 2006, Australia
Published in Ecology 83, 2124-2135.
(2002)
129

Abstract
Female aspic vipers (Vipera aspis) are "capital breeders", and delay reproduction
until they have amassed large energy reserves. Data from an eight-year mark-
recapture study on free-ranging vipers suggest that potential costs of reproduction
were high for these animals, in terms of survival as well as growth and energy
storage. Females that reproduced experienced higher mortality rates than non-
reproductive females, and hence exhibited a tendency toward semelparity, grew less,
and devoted most of their energy stores to reproduction. Both the depletion of body
reserves and the low survival of reproductive females translated into significant costs
(decrements of LRS). However, the cessation of growth during pregnancy had no
detectable effect on LRS. Most females produced only a single litter during their
lifetimes. A female’s “costs” in energy terms were not negatively correlated with her
future reproductive output, probably because female vipers vary considerably in the
rate at which they can accumulate energy. This notion is supported by the
observations that (1) females with higher initial body reserves expended more energy
during reproduction, and (2) females that accumulated energy more rapidly after
parturition were more likely to survive and to breed again. This kind of variation
among females masks any underlying trade-off between current reproductive effort
and probable future reproductive success. Despite this complication, a strong link
between rates of survival and post-reproductive mass recovery suggests that
changes in body reserves govern reproductive effort in this species.
Key words: breeding frequency, cost of reproduction, energy storage, reproductive
effort, semelparity, snakes.
130

Introduction
The extensive scientific literature on "costs of reproduction" falls into two main
categories, with relatively little overlap. This dichotomy involves theory-based
(primarily mathematical) explorations on the one hand, and empirical studies of living
animals on the other. Many mathematical models in life-history theory incorporate
some causal link between an animal's current reproductive expenditure and its
probable future reproductive output. These models suggest that the exact nature of
that link has profound implications for the kinds of life-history strategies that will
maximize lifetime reproductive success and hence, are expected to evolve under the
conditions posited in the model (e.g., Williams 1966a,b; Schaffer 1974; Winkler &
Wallin 1987; Shine & Schwarzkopf 1992).
Unfortunately, it is difficult to translate these apparently simple notions into
practicable measures of reproductive costs (Reznick 1992; Jönsson & Tuomi 1992;
Reznick et al. 2000). Thus, much of the empirical literature on "costs of reproduction"
relies on measuring variables that are linked only indirectly to the potential costs
experienced by reproducing organisms (Stearns 1992). Such variables include
measures of reproductive output (i.e. clutch sizes), reproductive output relative to
maternal size (e.g., Relative Clutch Mass; Cuellar 1984; Seigel & Fitch 1984), or
effects of reproduction on maternal traits (e.g., metabolism, locomotor speeds, post-
parturition maternal body condition: e.g., Shine1980; Birchard et al; 1984; Seigel et
al. 1987; Lee et al. 1996). However, it is not easy to determine whether or not these
measures correlate with the degree to which current expenditure decreases future
probable reproductive success (i.e., decrements in lifetime reproductive success [=
LRS]: Williams 1966a,b). First, LRS is extraordinarily difficult to measure in mobile or
long-lived animals (Clutton-Brock 1988). Second, effects of reproductive output on
131

LRS can be mediated via several different processes. The most obvious dichotomy
is between survival costs and energy costs (e.g., Calow 1979), but there are many
subtleties even within these two broad categories. For example, higher energy
expenditure on current reproduction may reduce LRS via decreased energy stores
and/or by decreasing subsequent growth rates (and hence fecundity if body size
enhances reproductive success). Third, these currencies are not independent
(Bauwens & Thoen 1981; Brodie 1989). Fourth, the magnitude of various "costs" is
likely to shift among habitats and years (Festa-Bianchet et al. 1998). Fifth, variation
in levels of resource availability among individuals may generate a positive rather
than negative correlation between current reproductive output and future
reproductive success, thereby masking a trade-off between these two traits (Bell &
Koufopanou 1986; Van Noordwijk & de Jong 1986).
In order to overcome some of these difficulties, such studies should focus on
various species that offer logistical advantages for measuring the relevant traits (i.e.
reproductive expenditure and its consequences) (Reznick 1992; Seigel 1993; Shine
& Bonnet 2000). In the current paper, we describe an eight-year study on such a
system, the aspic viper (Vipera aspis). Aspic vipers are abundant, sedentary (hence,
easily recaptured), and live in a relatively cool climate (so that thermoregulatory
needs during vitellogenesis and gestation substantially modify patterns of movement
and feeding). Perhaps more importantly, females reproduce on a less-than-annual
basis, so that we can readily compare females that are in the reproductive versus
non-reproductive years of their cycles (Bonnet et al. 2000b). This species is a typical
capital breeder (sensu Stearns 1992). Large body reserves must be accumulated
during long periods (years) before reproduction, both for the induction of
vitellogenesis and to fuel most of the reproductive effort (Saint Girons 1957; Bonnet
et al. 1994, 2001b; Naulleau & Bonnet 1996). The massive depletion of body
132

eserves in the course of reproduction results in a low breeding frequency and hence
can be considered as a typical energy cost (Naulleau & Bonnet 1996; Bonnet et al.
2001b). Like other ectotherms, female aspic vipers are pre-adapted to capital
breeding (Pough 1980; Bonnet et al. 1998). In contrast, endotherms are less well-
suited to long-term storage of body reserves (Jönsson 1997), and these animals tend
to rely on "income" rather than "capital" to fuel reproduction (Else & Hulbert 1981;
Bonnet et al. 1998; Bronson 1998; Schneider et al. 2000). As a result, changes in
body mass during reproduction may be a poor indicator of energy costs of
reproduction in mammals and birds. First, such changes may reflect fluctuations in
food availability independently of reproductive effort. Second, the high basal
metabolic rate of these animals means that body reserves can be depleted rapidly
during short periods (days) of starvation in non-reproductive individuals (Nagy 1987).
This situation seriously complicates direct comparisons between reproductive and
non-reproductive individuals. By contrast, most ectotherms can survive for long
periods of time (months to years) during total starvation with minor changes in body
mass (Pough 1980). Hence, any massive decrease in body mass that is temporally
and physiologically associated with reproductive effort is likely to be a direct
consequence of reproduction. The comparison between reproductive and non-
reproductive females is thus straightforward: both survival rates and changes in body
mass are likely useful candidates for measuring potential costs of reproduction in this
snake species. Body reserves can influence LRS through major components such
as breeding frequency (Naulleau & Bonnet 1996), current fecundity (Bonnet et al.
2001a), and survival (Bonnet et al. 2000a).
133

We set out to answer three main questions:
(1) What form do “costs of reproduction” take in aspic vipers? We explore three
potential currencies in this respect: probabilities of survival, decreases in energy
reserves, and rates of growth in body length (because fecundity is generally
associated with size in snakes).
(2) Because we have long-term data, we can assess whether or not these kinds
of “cost” indices (such as reductions in growth and body condition), or the magnitude
of expenditure on current reproduction, actually translate into lower reproductive
success in the future. This must be true for survival costs, although even here it is
possible that the effect is trivial (e.g., if subsequent post-reproductive survival rates
are so low that few females live long enough to reproduce again anyway). For
energy-storage costs, is it true that an unusually emaciated female will delay the
production of her next litter? For growth costs, will females forfeit fecundity
increments in later litters if they grow less after their first reproduction? Are females
with high reproductive output in their first litter, less likely to reproduce again or
produce a small litter if they do?
(3) What attributes of a female in the year after she reproduces (e.g., her rate of
growth in body length, or her rate of replenishment of body condition) offer the best
predictors of her subsequent reproductive output (i.e., determine the time she
reproduces, how many offspring she produces, and their size)? Data on this issue
can help us to identify which of these traits may offer the best currency in which to
measure “costs” of reproduction in a typical ectothermic capital breeder.
134

Materials and methods
Animals and study site
The aspic viper (Vipera aspis) is a stocky, medium-sized venomous snake species
widely distributed through western Europe (Naulleau 1997). Adults in our population
average 47.7 ± 3.4-cm snout-vent length (SVL), 54.3 ± 3.8cm total length. Female
vipers mature at an age of approximately three years (Bonnet et al. 1999a). Mating
occurs in spring (March-April: Saint Girons 1952, 1957a,b; Vacher-Vallas 1997;
Naulleau et al. 1999). In our population, females are gravid over summer, and give
birth to a litter of 1 to 13 large (20.7 ± 1.2 cm TL, 6.1 ± 1 g) offspring in autumn (late
August-September). Most female vipers do not reproduce every year (Bonnet &
Naulleau 1996). The exact frequency of reproduction depends upon thermal
conditions (especially, length of the activity season) and food supply, so that
reproductive frequencies differ among areas and among years (Saint Girons 1952,
1957a,b, 1996). This less-than-annual frequency of reproduction results from the
time taken to replenish energy stores for the next litter: females delay reproduction
until they exceed a minimum body-condition threshold (Naulleau & Bonnet 1996).
We studied the aspic viper in a closed population in western central France
(Les Moutiers en Retz, 47 o 03N'; 02 o 00W'; Bonnet & Naulleau 1996). The study site
is 33 ha in extent. It is bordered to the north and east by roads, to the south by the
Atlantic Ocean, and to the west by a camping site (Vacher-Vallas et al. 1999). It is a
typical parkland habitat that has not been intensively managed for 15 years. Thus,
the hedges form a dense network, and bushes (especially brambles) have invaded
the meadows to varying degrees. In some meadows, oak and pine plantations have
recently (1993 to 1995) been established. The climate is a temperate oceanic one
(see Bonnet & Naulleau 1993 for average temperatures).
135

Procedure
One to three people checked the area almost every sunny day from the time the
snakes emerged (late February for the females) to the end of their reproductive
period (September), and less frequently in late September-October. Searching effort
averaged 95.3 days per annual activity season (sd = 32.2, range = 51 to 124 days)
and 523.7 hours per annual activity season (sd = 198.4, range = 232 to 614 hours).
Over the period 1992 to 1999, we hand-captured 469 different adult female vipers.
Classification of these animals as adults is based on the minimum size we have
recorded for parturition in this population (41.5 cm SVL, 47 cm total length). Each
female was individually marked for future identification by scale-clipping in 1992, and
fitted with a Passive Integrated Transponder (PIT) tag since 1993, measured (to the
nearest 0.5 cm, SVL and TL), weighed (nearest 1 g), palpated for prey, eggs or
embryos, and released at her exact place of capture. Reproductive status was
determined by palpation of eggs or embryos, by records of parturition or by obviously
post-parturient body condition. Immediately after giving birth, females are very
emaciated, with a flaccid abdomen and extensive skin folds. Our analyses exclude
body-mass data taken from individuals containing prey items or oviductal embryos.
Recapture probabilities were high (see Bonnet & Naulleau 1996), but few females (N
= 5) were recaptured in eight consecutive years due to the low survival rate of
reproductive females.
Survival
We scored a female as having died if we failed to locate her on >250 days’ searching
over a period of > 2 years. Given the very low vagility (5m/day on average: Naulleau
et al. 1996) and high recapture rates within this closed population, we can be
confident that such animals had died rather than moved away. Only one female
136

escaped capture during three consecutive years (marked in 1993 and not recaptured
until 1997), and only eight animals were “missed” during two consecutive years.
Changes in body mass and body size
We measured changes in body mass and body length from the onset of
vitellogenesis to the post-parturition period. This covers the entire activity season (6-
8 months per year). Our criteria for inclusion of data in the analyses were as follows:
“early vitellogenesis” was defined as the period from March to April (Bonnet et al.
1994), and data from snakes captured after this period were not included for
analyses of change in body mass or length. However, we did include these later
captures for analyses of rates of survival and future reproduction. Changes in body
mass and rates of growth in body length were calculated from March-April to August-
November within a given year, and (ignoring hibernation) from March-April to the next
March-April between years. Our extensive data indicate that vipers did not show any
significant change in either body mass or body length over the hibernation period.
Reproductive output
As soon as we recorded the first parturition of the year (generally in the second half
of August), we collected all of the gravid females that we could locate, and held them
in captivity in individual cages (for periods of up to one month) so that we could
count, measure, weigh, sex and mark the offspring. Captive females were weighed
every two days, and immediately after parturition. We defined Relative Clutch Mass
(RCM) as the total mass of the litter (including stillborn offspring, etc.) divided by the
post-parturient mass of the mother. Data were obtained on 195 litters from 157
different females. Several females were captured shortly before parturition in two
different years, so that we were able to quantify reproductive output on each
137

occasion. Data on these animals allowed us to explore the relationship between
initial reproductive output and subsequent changes in body size, body condition, and
reproductive output. However, for most analyses, data were not available from all of
the females (e.g., data for mass change during the reproductive period were
available on 301 females).
Activities associated with reproduction likely to be costly in aspic vipers
In combination with extensive studies in other parts of France (e.g., Saint Girons
1952, 1957a,b, 1996), our studies reveal that reproduction imposes marked changes
on several aspects of the biology of female vipers. The major modifications are as
follows:
(1) Relative to males and non-reproductive females, reproductive female vipers
become more sedentary in the course of gestation: mean home ranges decrease
sharply from 3,000 m 2 to 300 m 2 (Naulleau et al. 1996). (2) Gravid vipers spend
more of their time in behavioral thermoregulation than do other animals within the
population (Bonnet & Naulleau 1996), and hence could be more exposed to
predation (mainly birds; Naulleau et al. 1997). (3) Pregnant females progressively
reduce their rate of feeding, and may cease feeding in the latter stages of gestation
(unpublished data, and see Saint Girons 1952, 1957a,b; 1996). (4) Female vipers
show a consistent pattern of change in body mass over the course of the
reproductive cycle. Body condition (mass relative to length) increases during the
non-reproductive years, until it exceeds the threshold level required to initiate
vitellogenesis (Naulleau & Bonnet 1996). The female's mass drops dramatically at
parturition. The magnitude of this decrease in maternal mass (i.e., from the
beginning to the end of the reproductive bout) offers a measure of her net energy
138

expenditure over that period: (body reserves invested into the litter + metabolic
expenditure) - food intake.
Analyses
Body condition was calculated as residual values from the regression of body mass
(Log) against body size (Log) (Jayne & Benett 1990). We randomly selected a single
record per female to avoid pseudo-replication bias in this analysis. However,
ignoring such bias, and including 753 females where reproductive status is known
(among a total of 853 “female-year” data, reproductive status was unknown on 100
occasions) in the analyses did not change any results significantly . Importantly, the
mean body sizes of reproductive females and non-reproductive females were similar
(ANOVA with reproductive status as the factor and SVL as the dependent variable;
F1, 267 = 0.90, P = 0.34), allowing us to compare these two categories of females
without having to take into account possible effects of body size on rates of survival
or growth (Bonnet et al. 2000b).
The snakes' tendency toward semelparity greatly reduced our sample sizes for
tests comparing successive reproductive events by the same female. Conclusions
from such tests are problematic because of their low power to reject the null
hypothesis. Statistical conventions are much more rigid with respect to α (the
decision to reject null hypotheses with an error α < 0.05) than with respect to β (the
type II error). Statistical textbooks generally recommend that the power of a test (1-
β) should be > 0.80. Because statistical tests (especially correlation analyses) have
very low power when sample size is small, we performed power analysis to estimate
the ability of our statistical tests to detect “significant” effects. In all our ANOVAs,
power was close to 1.0 and has not been reported. Correlation analyses are more
sensitive; a low correlation between two variables inevitably (due to the structural
139

trade-off between α and β error rates) leads to a low power of the analysis, even with
a large sample size. Such a low power does not invalidate the analysis, but means
that caution is needed in interpretation. In such cases, we calculated the sample
sizes that would be required to detect a "significant" result at low α and β error rates
(0.05 and 0.10). Power analyses and required sample size estimates are not a
panacea to low sample sizes, but no conflict arose among our different tests. We
used Statistica 5.1 and 6.0 to perform the statistical analyses.
Results
How high are the potential costs of reproduction for female aspic vipers?
We can estimate these potential costs by comparing females in reproductive years
versus non-reproductive years of their cycles. Our data provide four separate
indices: (1) whether or not a female survived to produce her litter; (2) how much her
energy stores (as measured by changes in body mass) decreased over the
reproductive period (vitellogenesis plus gestation); (3) how much she grew in body
length over this period; and (4) her body condition (mass relative to length) after
giving birth. The first three variables are self-explanatory; the fourth is relevant
because previous studies on two species of viviparous snakes sympatric with V.
aspis have suggested that this trait (maternal post-parturition condition) may be a
significant predictor of maternal survival rates (Madsen & Shine 1993; Luiselli et al.
1996). Comparisons between reproductive and non-reproductive females suggest
that potential costs are high in each of these currencies.
1. Survival rates. We have data for 381 randomly sampled females caught
between 1992 and 1997 (females caught in 1998 and 1999 are ignored), for which
140

eproductive status and survival are known. Adult female vipers experienced much
higher survival during non-reproductive years (123 of 145 records, = 85%) than
during reproductive years (106 of 236 records, = 45%; χ 2 = 59.7, 1 df, p< 0.0001;
see Figure 1). The total potential cost of reproduction in terms of survival is actually
higher and more complex than this, because a female viper’s food-gathering activities
during her non-reproductive years also constitute a component of reproduction
(Bonnet et al. 2000a). However, the clear result is that activities directly associated
with the production of the litter cause a substantial decline in annual survival. The
major component of the low survival of reproductive females seems to occur mainly
before parturition (i.e. during vitellogenesis and gestation) with 88 of 142 females
dying before parturition versus 31 of 78 females dying later, between parturition and
the following spring (χ 2 = 10.0, df = 1, p = 0.0016; the sample size was reduced for
this test because we selected females caught several times during a given year).
Considering the 236 reproductive females, the studied population exhibited a
tendency toward semelparity: 182 females became vitellogenic only once, 43 twice,
and 11 were vitellogenic on three different occasions; leading to a mean number of
1.28 reproductions per female during their life. Importantly, many (roughly 50%) of
these females died before giving birth, and the average total number of litter per
“reproductive female” was actually less than 1. For reliability, we excluded from this
analysis the few females that survived over long periods but that we failed to
recapture in some years (for example, caught in 1992 and in 1995 but missed in
1993 and 1994), because of uncertainty about their number of reproductive episodes.
2. Changes in maternal body mass. All reproductive females (regardless of whether
we used the full data set or only one randomly-selected data point per female) lost
body mass over the period from spring (vitellogenesis) to autumn (post-parturition).
141

In contrast, non-reproductive females generally gained substantially in mass (ANOVA
with reproductive status as the factor: F(1.299) = 616.8, p < 0.0001; see Figure 1).
This decrease corresponds to the depletion of body reserves that are transferred in
the embryos plus the metabolic costs of vitellogenesis and gestation (6 months).
Factoring out the potential effect of size through ANCOVAs (reproductive status as
the factor, changes in body mass as the dependent variable and body size as the
covariate) leads to similar results (F(1.298)=606, P < 0.0001).
3. Growth rates in body length. Most (57 of 68 = 84%) reproductive females
showed no detectable growth in body length during the reproductive period, whereas
non-reproductive females showed significant growth (ANOVA with reproductive
status as the factor: F(1.375)= 37.2, p< 0.0001; see Figure 1).
4. Body condition post-parturition. Females that reproduced were in much lower
body condition after giving birth than were non-reproductive females at the same time
of year (ANOVA with reproductive status as the factor: F(1.255)=229.7, p< 0.0001;
see Figure 1). In fact, almost all post-parturient females were in poor body condition,
as indicated by their abundant skin folds, typical of snakes with minimal body
reserves (Bonnet 1996).
142

PROBABOLITY OF SURVIVAL (%)
GROWTH RATE
(cm/year)
CHANGES IN BODY MASS (g)
100
90
80
70
60
50
40
30
20
10
0
2.5
2.0
1.5
1.0
0.5
0.0
50
45
40
35
30
25
20
15
10
5
0
Reproductive
Reproductive
Will
Reproductive
Again
Non
Reproductive
Non
Reproductive
Will not
CHANGES IN BODY MASS
(g/year)
BODY MASS ADJUSTED
FOR SVL (g)
CHANGES IN BODY MASS (g)
30
20
10
0
-10
-20
-30
-40
110
100
90
80
70
60
50
40
30
20
10
0
50
45
40
35
30
25
20
15
10
5
0
Reproductive
Reproductive
Biennial
Non
Reproductive
Non
Reproductive
Triennial
Figure 1. Comparisons of survival rates (top left), change in body mass (top right), rate of growth in
body length (middle left) and maternal body condition (residual scores from linear regression of ln-
transformed mass versus snout-vent length) immediately after the time of parturition (middle right) in
reproductive versus non-reproductive female vipers (for ease of interpretation, this graph shows body
mass adjusted for maternal SVL rather than the body condition index). Female vipers that regained
mass more rapidly after reproduction were more likely to reproduce again (bottom left) and did so after
a briefer delay (bottom right: mean mass changes of females that had an inter-litter interval of 2
(biennial) or 3 (triennial) years.
143

Do these measures of “cost” translate into lower future reproductive output?
Although it seems plausible that future reproductive output will be compromised by
high levels of current expenditure (an index of energetic reproductive effort), this
assumption requires empirical verification. To do this, we can examine the
relationship between our measures of potential cost (survival rate, mass loss, growth
rate, and maternal body condition post-partum) and reproductive output (Relative
Clutch Mass) on the one hand, and future reproductive output on the other. That is,
do high levels of potential cost or reproductive output correlate with lower levels of
future output, as predicted by the “costs” hypothesis? In other words, do our
measures of potential costs reveal real costs of reproduction?
1. Survival rates. Obviously, females who die during reproduction are less likely
to breed again than are surviving females. Our data clearly show that reproductive
females have a low probability of survival. However, the degree to which this is a
significant cost to LRS depends not only upon survival probabilities in the absence of
reproduction, but also upon a postpartum female's probability of reproducing again
even if she survives until the following active season. If most "surviving" post-
parturient females in this population actually die before a second reproductive
opportunity (regardless of their reproductive output in the first litter), then there may
be little difference in the incidence of second litters between animals that survive to
their first parturition versus those that die during their first reproductive year. If so,
the high mortality associated with reproduction will no longer be a real cost. This is
not the case in our population. Almost 50% (59 of 123) of the reproductive females
that survived initiated reproduction a second time. Although a significant proportion
of these animals undoubtedly died during their second pregnancy and thus did not
actually produce two litters, females that survive to reproduction have significant
144

opportunity to reproduce again. Overall, the low survival rate of reproductive
females, regardless of the period during which mortality peaks (i.e. before or after
parturition), entails significant costs by strongly decreasing future probabilities of
reproduction.
2. Changes in maternal body mass. Were females that lost less body mass over
the reproductive period more likely to breed again? Our data show that this was not
the case (logistic regression; χ 2 = 2.16, df = 1, n = 71, p = 0.14; Table 1). Similarly,
we might expect females that lost more mass when producing their first litter to delay
subsequent reproduction for a longer period in order to recoup their energy stores.
No such effect was apparent in our data (χ 2 = 0.77, df = 1, N = 64, p = 0.38). Lastly,
we might also predict that females who lost more mass would produce smaller-than-
average litters and/or smaller-than-average neonates at their next reproductive
episode. Neither of these patterns appeared in our data set (for litter size, r = 0.24, n
= 20, p = 0.29; for offspring size, r = 0.18, n = 10, p = 0.64). However, we note that
the sample sizes were small for these later tests, and hence that the power of these
analyses was low (0.27 and 0.13 respectively), perhaps reflecting the non-
significance of the results. Nonetheless, the correlations were positive rather than
negative, and the sample sizes that would be required to obtain a significant effect
were high (178 and 320 respectively). Thus, any “undetected” effects were probably
weak or negligible.
3. Growth rates in body length. The notion that growth costs decrease future
reproductive output depends upon the assumptions that (i) a decrease in growth rate
during reproduction will influence body size at the next reproduction; and (ii) a larger
body size will allow a larger reproductive output. Neither of these assumptions is well
supported by our data. In our population, body size influences reproductive output
145

only slightly (see Bonnet et al. 2000b for a detailed discussion of this issue). In
addition, the extent of a female's growth during reproduction did not correlate with her
body size at the next reproduction (r = 0.24, n = 15, p = 0.39). The power (0.22) of
this analysis was low, and we may have failed to detect a slight effect. Nonetheless,
the weak influence of maternal size on reproductive output (Bonnet et al. 2000b)
suggests that growth effects of reproduction probably have little effect on future
fecundity. Thus, the almost total inhibition of growth by reproducing female vipers did
not translate into a significant cost of reproduction.
4. Relative Clutch Mass. Female vipers that produced small litters relative to
their own body size were no more likely to reproduce again than were conspecifics
producing larger litters (χ 2 = 0.78, df = 1, n = 137, p < 0.37; Table 1). Females that
produced large first litters (high RCM) did not exhibit low rather than high RCMs in
their second litters (r = 0.45, n = 14, p = 0.11). The power of this analysis was only
0.51. Regardless, the positive rather than negative correlation strongly suggests
(counter-intuitively) that high initial RCM does not translate into a reduced
subsequent RCM. Females with high RCMs in their first litter did not produce smaller
offspring (relative to other females) in their second litters (r = -0.20, n = 12, p = 0.54;
but the power was low = 0.15) and did not delay their subsequent breeding attempts
relative to other females (χ 2 = 0.04, df = 1, n = 63, p = 0.82). A ratio measure such
as RCM facilitates intuitive understanding of reproductive output relative to maternal
size, but may introduce statistical artifacts into analyses (e.g. Seigel and Ford 1987).
To overcome this problem, we repeated all of these analyses using alternative
measures of reproductive output: either absolute mass of the litter, or residual scores
from the general linear regression of litter mass to body mass. We obtained similar
results in each case.
146

5. Maternal body condition post-partum : Female vipers are emaciated
immediately after parturition, and have relatively low energy reserves at this time.
Nonetheless, the degree of maternal emaciation did not correlate with a female’s
probability of survival to the next season (logistic regression χ 2 = 1.88, 1 df, n= 119,
P = 0.17), or with her probability of breeding again (analysis restricted to the females
that survived, logistic regression χ 2 = 1.15, 1 df, n=71, p= 0.28; Table 1). Similarly,
her litter size at the next reproduction (r = -0.25, n=14, p= 0.36), or offspring size at
the next reproduction (r = 0.18, n=12, p=0.58), was not strongly influenced by a post-
parturition female’s body reserves. The low power of these two later analyses (0.22
and 0.14) require caution in interpretation; but the required sample sizes to obtain an
α level < 0.05 were relatively high (164 and 320), suggesting that any “missed” effect
was weak. Post-partum maternal body condition tended to affect the number of
years’ delay until her next reproduction, but this trend did not attain the conventional
level of statistical significance (χ 2 = 3.39, 1 df, n= 63, p= 0.065).
Table 1. Mean values (± SD) of mass loss, growth rate, relative litter mass, and post-partum body
condition recorded in the course of reproduction in wild female aspic vipers. The changes in mass and
growth rate were calculated from the onset of vitellogenesis to parturition. All of the females used in
this analyses survived to produce their first litter and were classified based upon whether or not they
also reproduced again in the future. We found no significant differences between the two groups of
females (See text for statistics; except for growth rate: ANCOVA with SVL as the co-variable and
growth rate the dependent variable; F(1.70) = 2.53, p = 0.12).
Trait Will breed again Will not N
Mass loss (g) -37.2 ± 17 -35.1 ± 16 71
Growth rate (cm year –1 ) 0.32 ± 1.46 0.87 ± 1.46 73
Relative litter mass (%) -0.55 ± 0.22 -0.51 ± 0.24 137
Post-partum body
condition
0.005 ± 0.1 -0.002 ± 0.1 119
147

MASS RECOVERY
g year-1
What currencies of potential costs influence future reproductive output?
Another way to identify the appropriate currency in which to assess “costs” is to
ignore reproductive output per se, and focus instead on the long-term consequences
of rates of change in the attributes that we know to be affected by reproductive
expenditure. Body reserves and growth rates can readily be examined in this way.
1. Maternal change in body mass after the first litter. If body reserves are
important, we expect that females that recoup their energy (body mass) reserves
rapidly will be more likely to breed again (and will breed sooner) than females that
regain mass only slowly. Analysis supports this proposition for the female’s
probability of reproducing again (χ 2 = 10.53, df = 1, n = 55, p < 0.001) and for the
duration of the delay to her next reproduction (χ 2 = 27.70, df = 1, n = 42, p< 0.0001:
see Figure 2). The female’s litter size and neonate size at her second reproduction
showed no significant relationship with her rates of mass increase after the first litter
(for litter size, r = 0.45, n = 14, p= 0.10; and r = 0.29, n = 11, p= 0.38 for neonate
size). Although our sample sizes were small and associated statistical power low
(0.51 and 0.22 respectively), these correlations were positive rather than negative.
100
80
60
40
20
0
a b
-20
-30 -20 -10 0 10 20 30
RELATIVE LITTER MASS
residuals
148
GROWTH RATE
(residuals)
6
4
2
0
-2
-4
-30 -20 -10 0 10 20 30
RELATIVE LITTER MASS
(residuals)

Figure. 2. Relationships between a female viper's reproductive output in her first litter and her
subsequent rate of recovery of body reserves and growth rate in body length. Female vipers that
invested more into reproduction exhibited higher rather than lower rates of mass recovery or growth, in
contradiction to predictions from the "costs" hypothesis. Females with higher relative clutch mass
(RCM, residuals of litter mass on post-parturient female’s mass) recovered their body reserves more
rapidly during the following year (r = 0.43, n = 36, p = 0.01; the power of this analysis was 0.85; Figure
2a); without trading this body reserve replenishment against growth rate the following year (r = 0.20, n
= 36, p = 0.20; power = 0.32; Figure 2b). To control for the effect of size on growth rate, growth rate
was calculated as the residual values of the regression of absolute gain in size (cm/year) on initial
body size (SVL in cm) (Fig2b). Because mass recovery was not affected by body size, we used
absolute values (Figure 2a).
2. Rate of growth in body length after the first litter. Faster-growing females were
more likely to breed again (χ 2 = 6.10, df = 1, n = 89, p= 0.01), and bred sooner than
slower-growing animals (χ 2 = 4.1, df = 1, n = 58, p= 0.04). Taking into account the
effect of a female's mean body size on her growth rate (by using residuals of the
regression between absolute growth rate and initial SVL: Bonnet et al. 2000b), does
not alter this conclusion (χ 2 = 8.20, df = 1, n = 67, p= 0.004; and χ 2 = 4.90, df = 1, n
= 47, p= 0.003). Reproductive output in the second litter was not associated with
rates of body growth in the period following production of the first litter (for litter size, r
= 0.05, n = 18, p= 0.84; for offspring size, r = 0.12, n = 15, p= 0.65). Despite a low
power of these analyses (0.05 and 0.07), the very weak correlations would require
sample sizes of 4,198 and 725 to attain statistical "significance", and hence suggest
an absence of effect.
Discussion
Our analyses of costs of reproduction differ from most previous studies in this field,
by incorporating two steps into our analyses. First, we have measured the ‘potential
costs of reproduction’ for female aspic vipers in terms of energy and survival.
149

Second, we have examined the consequences of these ‘potential costs’ in terms of
lifetime reproductive success. That is, we have tried to determine whether or not
‘potential costs’ actually translate into a significant decrease in future reproductive
success. By adopting this method, we can better tease apart two components of the
life-history: reproductive effort per se, and costs of reproduction (Niewiarowski &
Dunham 1994). Our data support four main conclusions:
1) Reproducing female vipers commit themselves to a considerable effort in
vitellogenesis and gestation, to the degree that most females produce only a single
litter during their lifetime. Such reproductive effort is reflected in high rates of mass
loss, a virtual cessation of growth, and a decrease in the probability of survival.
2) The different components of reproductive effort do not systematically, and equally,
translate into real ‘costs of reproduction’. The cessation of growth rate during
reproductive years had no measurable influence on future reproduction. By contrast,
the strong mobilization of maternal reserves necessary to fuel reproductive effort had
a considerable impact on breeding frequency and lifetime reproductive success
through distinct, but interconnected mechanisms. First, vitellogenesis and gestation
entail a strong increase in basking frequency and expose females to predation
(Bonnet & Naulleau 1996; Naulleau 1997). Second, many post-parturient females
die from starvation; and even if a female survives, she delays reproduction for a long
period of time (1-3 years) until she has restored her body reserves (Naulleau &
Bonnet 1996; Bonnet et al. 2000a). The lower survival caused by such thermal and
energetic requirements of reproduction, along with the long period between
reproductive bouts (that automatically increase mortality from other causes, Bonnet
et al. 2000a) constitute a high cost. Reproductive females that die, thereby lose
150

significant opportunities to reproduce again and to increase the total number of
offspring they could produce during their life.
3) The magnitude of reproductive effort in a given reproductive bout does not affect
future reproductive success. Levels of reproductive output and our measures of
reproductive effort (such as lowered survival, lower maternal body condition, loss in
mass, or decrements in growth) were not negatively correlated with future
reproductive success.
4) Changes in maternal body reserves over the complex alternation of non-
reproductive and reproductive phases play a key role in the reproductive biology of
female aspic vipers. Notably, the emaciation of post-parturient females determines
the low breeding frequency and a high proportion of the mortality experienced by
reproducing females. Despite the difficulty of detecting trade-off between
reproductive investment and future reproductive success, the strong link between
rate of mass recovery in post-parturient females, versus the duration of delay to
production of the next litter (and probability of survival to this time) offers strong
evidence that energy stores are a crucial currency. Thus, the rate that a female viper
can replenish her energy stores after reproduction may strongly affect her LRS.
These conclusions have several implications for studies on “costs of
reproduction”. For example, they offer a challenge to simplistic attempts to use
measures of reproductive output such as RCM, increased metabolism or cessation of
growth as a shorthand index of “costs of reproduction” (i.e. Shine 1980; Vitt & Price
1982; Birchard 1984; Seigel et al. 1987). The reality is far more complex: even if
reproductive effort is high, its magnitude may not correlate with future reproductive
output in simple phenotypic comparisons (Bauwens & Thoen 1981; Brodie 1989;
Dunham et al. 1994; Olsson et al. 2000). Nonetheless, our study is encouraging in
151

that a relatively easily measured trait (changes in maternal mass) may offer a
reasonable currency in which to estimate several of the major “costs”, at least in
ectothermic animals. Because they entail significant costs, the physiological
mechanisms that control the allocation of body reserves during reproduction should
be under strong selection (Sinervo & Svensson 1998; Bonnet et al. 2002b). Such a
situation enables us to identify more precisely the ecological context that can favor
the emergence of capital breeding instead of income breeding as alternative
reproductive strategies (Stearns 1992; Jönsson 1997; Bonnet et al. 1998).
Although a comparative approach will be needed to examine the evolution of
such traits, at present we cannot compare the absolute magnitude or form of
reproductive “costs” in V. aspis with that in other reptiles because most previous
studies have relied on indirect measures of “cost”. Indeed, “costs of reproduction”
may be manifested differently. For example, there is apparently no significant
survival cost of reproduction in Orsini’s Viper, whereas such costs are high in both
adders and aspic vipers (Madsen & Shine 1992, 1993; Capula et al. 1992; Baron et
al. 1996; Luiselli et al. 1996). Nonetheless, it may often be true that female
viviparous snakes living in cool climates experience such high mortality during
reproduction that many females produce only a single litter in their lifetimes, and
hence exhibit a strong tendency toward semelparity (Brown 1991). In the case of V.
aspis, the mortality comes not only from starvation per se (because some females
maintain sufficient reserves to avoid this threat) but also from vulnerability to
predation during pregnancy (probably due to increased basking) and associated
dangers such as occlusion of the oviducts by inviable embryos (Naulleau 1997).
Thus, although energy stores are a crucial currency that limits a female’s
reproductive output, energy limitation is not the only proximate determinant of
reproduction-associated mortality in female vipers. When they engage in
152

eproduction, females shift from a very secretive to a conspicuous way of life (Bonnet
& Naulleau 1996). The thermal requirements of vitellogenesis and gestation result in
high rates of basking in reproductive females (Shine & Harlow 1993), regardless the
number of eggs/embryos they carry (unpublished). Hence, reproductive female
vipers are disproportionately exposed to avian predation. Because such costs of
reproduction can be independent of fecundity (Bull & Shine 1979), high energy and
survival costs are often associated with the extreme reproductive effort in capital
breeders such as viperid snakes. The low survival rates of reproductive female aspic
vipers are not affected by fecundity in natural conditions (Bonnet et al. 2002b), and
females may optimize their reproductive effort by producing the greatest number of
offspring per litter in order to minimize the cost paid per neonate. Capitalizing large
amount of body reserves prior to reproduction is an elegant way to produce a
massive reproductive effort when the probability of experiencing more than a single
reproductive bout is low.
Our study revealed another classical complication in studies of costs of
reproduction. The expected underlying trade-off between current versus future
reproduction (as evidenced by negative correlations between energy “costs” and
future reproduction) was masked. The comparisons between reproductive and non-
reproductive females, and the link between rates of mass recovery and future
reproduction, provide strong evidence for the existence of costs. Why, then, are they
not manifested in negative correlations between output and costs on the one hand,
and future reproduction on the other?
The answer almost certainly lies in substantial differences among females
within our population in their ability (opportunities) to gather resources (Glazier 2000).
For example, females with large initial body reserves produce larger litters and have
a greater output relative to their own body sizes (Bonnet et al. 2001b), but they are
153

nonetheless no less likely to breed again (this study) and they do not produce a
relatively smaller litter at their second bout (this study). Despite their high
reproductive expenditure, they do not recoup energy reserves more slowly after
parturition, and thus, they eventually reproduce again without additional delay. That
is, females that invest more (higher RCM) in their first litter were not less likely to
survive (logistic regression with survival as the dependent variable and RCM
[residuals] as the independent variable: χ² = 1.03, df = 1, n = 141, p = 0.31), and
even showed a tendency to regain mass more rapidly (Figure 2), than the “less lucky”
females who exhibited lower reproductive output (and thus, who superficially appear
to have paid lower “potential costs”). This situation may reflect strong differences in
female “quality” (as manifested in traits such as energy reserves prior to
reproduction) as has been documented in other species (e.g., Van Noordwijk and de
Jong 1986; Doughty & Shine 1997; Reznick et al. 2000). It may often be true that
any given level of reproductive output is a greater “cost” (e.g., to survival) for a
female in poor body condition (e.g., Cichon et al. 1998), and for variation in maternal
quality to generate positive rather than negative correlations between reproductive
output and survival (e.g., Bell & Koufopanou 1986; Winkel & Winkel 1995). We have
no data on the determinants of female “quality”, but some correlates suggest that
females with the highest reproductive effort (i.e. RCM) also exhibited the highest
abilities to recover after parturition (Figure 2), and hence to reproduce again. This
variation might reflect underlying genetic factors, or processes acting during
ontogeny (e.g., developmental temperatures; feeding opportunities early in life; low
parasite numbers). Alternatively, these inter-individual variations may simply reflect
the fact that some females have been luckier than others during foraging prior to,
during and after reproduction. Food intake can affect reproductive body reserves,
154

eproductive success and recovery at each of these phases, and these effects can
interact strongly (Bonnet et al. 2001b).
Despite the masking of phenotypic trade-off by variations among females, our
data nonetheless provide support for the use of changes in maternal body mass as a
realistic currency in which to estimate “costs of reproduction” in female vipers. We
base this conclusion on several facts. First, reproduction is expensive energetically:
a female viper’s reproductive output is tightly linked to her energy stores prior to
vitellogenesis (Bonnet et al. 2001b). Second, the rate at which a female can recoup
her energy stores (expended during the previous reproduction) is a significant
predictor of her future reproductive output (this study). The high growth rate
observed in females that recouped their body reserves rapidly reflects the fact that
females with high food availability invested both in body-reserve recovery and
growth, and changes in body mass integrate these two effects. Third, female vipers
postpone reproduction until they have achieved a critical threshold level of body
condition (Naulleau & Bonnet 1996). Thus, the substantial variation among post-
parturient females in their rates of recovery of body condition is convincing evidence
that rates of gain in body mass offer a biologically meaningful currency in which to
gauge a female’s ability to survive and reproduce. Moreover, the absolute mass that
post-parturient females must regain to reach the threshold for reproduction fits well
with the absolute mass loss during reproduction (Figure 1b and 2b). The other
currencies that we examined appear to be less useful for measuring “costs of
reproduction”. Survival is obviously important and this parameter must be included in
analyses, but it is difficult to integrate with energetic measures in any single currency.
Also, the real survival “cost” of producing a litter involves survival rates over the entire
reproductive cycle (not just the year in which the litter is produced), whereas energy
155

allocation can be calculated from the “reproductive” year only. Litter sizes are
predictable from energy stores, as manifested in body condition.
One major advantage of using changes in maternal body mass as a currency
for reproductive effort and “potential costs,” is that such changes are easy to
measure under field conditions. Also, reproductive output can be quantified in the
same currency, so that the two measures of investment (litter mass plus maternal
loss in body loss) can simply be added together to calculate a female’s total
expenditure on reproduction. This cannot be done if investment in aspects other
than the litter is measured in other currencies. Thus, our study is encouraging in that
a logistically feasible currency can be used to quantify a female’s investment into
current reproduction, in terms that can be directly translated into effects on future
reproductive success. This currency may well prove to be useful for other species as
well, especially those in which maternal energy reserves fuel most of the reproductive
expenditure (“capital breeders”: Drent & Daan 1980).
Acknowledgements
M. Vacher-Vallas, S. Duret, L. Patard, and M. Pedrono assisted in field work. For
comments on the manuscript, we thank R. Cambag, S. Cucullatus, B. S. White and
B. Wilmslow. Financial support was provided by the Conseil Général des deux
Sèvres (XB), the Centre National de la Recherche Scientifique and the Australian
Research Council.
156

C. Article 5
Costs of anorexia during pregnancy in a
viviparous snake (Vipera aspis)
Olivier Lourdais 1 2 3 , Xavier Bonnet 1 , Paul Doughty 4
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 Univertisty of Canberra, ACT, Australia
Published in Journal of Experimental Zoology 292, 487-493.
(2002)
157

Summary
Spontaneous anorexia has been documented in various animal species and is
usually associated with activities competing with food intake. In natural conditions,
most female aspic vipers (Vipera aspis) stop feeding during the two months of
pregnancy. We carried out a simple experiment on 40 pregnant females to determine
whether anorexia was obligatory or facultative, and to investigate the energetic
consequence of fasting on post-partum body condition and litter traits. Three diet
treatments were applied during gestation: no food, one feeding occasion, and two
feeding occasions. Twelve non-pregnant un-fed females were used as a control
group. Most gravid females accepted captive mice during gestation, suggesting that
anorexia reported in the field was a side effect of the tremendous changes in activity
pattern associated with pregnancy. Mass loss was high for un-fed reproductive
females, indicating high-energy expenditure associated with embryo maintenance.
Prey consumption allowed compensation for metabolic expenditure and enhanced
post-partum female body condition, but had no effects on litter characteristics. The
magnitude of the metabolic expenditure during gestation appeared to be independent
of fecundity.
Keywords: snakes, costs of reproduction, life history trade-offs
158

Introduction
Life history theory has been largely influenced by the concept of cost of reproduction
based on a possible trade-off between current reproduction and future reproductive
success (Fischer 1930; Williams 1966b). This notion is supported by a substantial
amount of theoretical study suggesting that the form of the relationship between
reproductive investment and the magnitude of associated costs influences the
evolution of reproductive strategies (Williams 1966b; Bull & Shine 1979; Shine &
Schwarzkopf 1992).
Identification of proximate mechanisms by which costs are mediated is an
important empirical challenge and two major categories of reproductive costs are
classically distinguished (Calow 1979). The first component is ecological and linked
with a reduction of survival probabilities associated with reproduction. The second
component recognised as a “fecundity” cost involves an energy allocation trade-off:
the investment in current reproduction affects the residual reproductive value through
a depletion of body reserves or growth rate reduction (Williams 1966b; Shine 1980).
Such a classification is somewhat artificial as the two major forms of costs are
interconnected; for example, low body reserves may also affect survival. In addition,
there is growing evidence that organisms can change the relative magnitude of the
different components through behavioural modifications (Bauwens & Thoen 1981;
Brodie 1989). Costs based on allocation trade-offs can take a diversity of forms.
Most studies have been carried out on the effects of direct energy investment into
reproduction (Anguiletta & Sears 2000). However, substantial expenditure may also
arise indirectly from cessation or reduction of feeding during all or part of
reproduction. Empirical studies indicate that feeding cessation or reduction during
reproduction is a widespread phenomenon (Engelmann & Rau 1965; Batholomew
159

1970; Mrosovsky & Sherry 1980; Weeks 1996). Cessation of feeding when food is
available is superficially paradoxical and needs explanation. Furthermore, this
indirect component of reproductive effort may translate into a substantial energetic
cost of reproduction.
Anorexia is associated with hibernation, migration or incubation in various
endotherm vertebrates (Mrosovsky & Sherry 1980). Although less studied,
ectotherm vertebrates may also serve as a a good model group for the study of
feeding cessation. Endothermy is tightly linked with parental cares (Farmer 2000),
and the lack of such cares among numerous ectotherm species lead to a substantial
simplification as reproductive effort is sealed prior to oviposition or parturition. Such
situation will facilitate the assessment of the relationship between reproductive effort,
reproductive output (fecundity) and associated costs. Among squamate reptiles
(lizards and snakes), reproduction entails major behavioural changes, notably
decrease or cessation of food intake during gestation (Shine 1980; Madsen and
Shine 1993; Gregory & Skebo 1998; Gregory et al. 1999). Such phenomena are
particularly obvious for “capital breeding” species in which long term energy storage
constitutes the primary source of energy for reproduction (Bonnet et al. 1998).
The aspic viper (Vipera aspis) is a medium size (50 cm) viperid snake that
displays those characteristics. In females, energy stores permit them to fuel the
entire energetic requirements of reproduction. Depending upon prey availability, food
intake may occur during the egg production phase in spring (Saint Girons & Naulleau
1981; Bonnet et al. 2001b). However, field data clearly indicate that many females
virtually stop feeding during the two months of pregnancy. Three hypotheses about
the proximate factors involved in gestational anorexia can be proposed: 1) anorexia
could be a consequence of abdominal space limitation to accomodate both embryos
and prey items (Saint Girons 1979). 2) cessation of feeding may be related to a loss
160

of appetite intrinsically associated with gestation (i. e. due to changes in hormonal
balance; Bonnet et al. 2001b). 3) fasting may simply be the result of low foraging
success due to behavioural changes in gravid individuals (thermal needs, predator
avoidance).
For hypothesis 1 and 2, gestational anorexia is supposed to be obligatory. In the
case of hypothesis 3, this phenomenon is expected to be facultative. In the present
study, we conducted a simple experiment to test if gestational anorexia is obligatory
or facultative and to examine to what extent feeding cessation during pregnancy
translated into energetic costs.
Materials and Methods
Study species
The aspic viper (V. aspis) is a small viviparous snake, abundant in central western
France. In this area, females typically reproduce on a less-than-annual schedule
(Saint Girons 1957a,b; Bonnet & Naulleau 1996; Naulleau & Bonnet 1996; Naulleau
et al. 1999). Ovulation occurs during the first two weeks of June (Naulleau 1981),
and parturition occurs two to three months later, from late August to late September.
Captures and housing
Forty reproductive females were collected in June 2000 from three localities: Château
d’Olonnes, Les Sables d’Olonnes (both in Vendée district) and Rochefort (Charentes
Maritimes district). Individuals were given a unique scale clip number, measured to
the nearest 0.5 cm and weighed to the nearest 1g. Females were placed in six
outdoor enclosures (5 X 3 m, mean density: 5 snakes/enclosure) broadly recreating
the natural habitat and exposed to the climatic conditions of the field research station
of Chizé (Forêt de Chizé, Deux-Sèvres, 46°07’ N, 00°25’ W). Each enclosure was
161

equipped with numerous external dens to serve as hiding-places. Water was
provided ad libitum and vegetation mainly composed of annual Poacae was kept high
(20 - 40 cm) to provide shade and shelter.
Experimental design
Females were randomly assigned to one of the three feeding treatments during
gestation: Group 1 (9 individuals): never fed; Group 2 (17 individuals): one prey item
offered in July ; Group 3 (14 individuals): two prey were presented in July and early
August. Snakes of both feeding treatment groups were fed by placing a recently
killed mouse (average mass 20 g) close to their dens. Prey consumption was
recorded by direct observation of feeding, or by less direct means if feeding was not
observed (by palpation of mice inside the snake and by a sudden increase in body
mass).
As a control group, 12 non-pregnant females (un-fed during one month in the same
conditions to reproductive females) were weighed to measure mass loss during
fasting, independently of gestation.
Records of body mass and reproductive output
The snakes were all weighed at the onset of the experiment in early July (i.e. after
ovulation). At this time the number of eggs was assessed via abdominal palpation
(Fitch 1987, Bonnet et al. 2001b for further details on the acuracy of the method).
Females were all recaptured at the end of gestation (late August); and weighed again
to determine absolute mass changes between early July and late August. Daily mass
change during gestation was calculated using absolute mass change (g) and time
elapsed (days) between the two mass records. Snakes were then brought in the
laboratory until parturition. We recorded post-partum female body mass and the
162

number, mass (± 0.1 g) and length (± 0.5 cm) of healthy offspring. The number of
unfertilised eggs and stillborn were also recorded. We made a distinction between
the total litter size including healthy neonates, still born offspring and undeveloped
eggs as well (Farr & Gregory 1991; Gregory et al. 1992), and “fit” litter size where
only viable neonates were considered. Five females (one in group 2 and two each in
group 1 and 3) produced only unfertile eggs. Because snakes were caught after the
mating season (Naulleau 1997), those individuals were removed from analysis of
reproductive output.
In this species, post-partum female body condition (mass adjusted by size)
positively influence survival the year following reproduction and hence residual
reproductive success (Bonnet et al. 2000a). We calculated body condition as the
residual score from the general linear regression of log-transformed body mass value
versus log-transformed snout-vent length value for all females (Jayne & Benett 1990,
Bonnet et al. 2000a). Such index provides an accurate estimation of body reserves
(Bonnet 1996).
Results
Prey consumption
Pregnant females accepted prey most of the time (41 of 45 feeding occasions).
Among the 17 snakes fed once, 15 (88%) ate the prey offered, and among the 14
snakes fed twice, 12 (86%) ate both prey offered; two females ate only one mouse
each. Hence, at the end of gestation it was possible to classify females by the actual
number of prey eaten: no prey (11 individuals); one prey consumed (17 individuals);
two prey consumed (12 individuals). In the following analysis, we considered both
group treatments and actual number of prey eaten.
163

Changes in body mass during gestation
The three groups did not differ in snout-vent length (one factor ANOVA,
F(2,37)=1.74; p=0.20, size-adjusted initial body mass (ANCOVA F(2,36)=1.39;
p=0.26), or size-adjusted number of eggs (ANCOVA, F(2,36)=0.45; p=0.63; Table
1). Group 1 females showed a significantly higher daily mass loss during gestation in
comparison to the control group (0. 22 g/d versus 0.11 g/d; F(1,18)=12.58;
p

Hence, using either initial group treatments or the number of prey actually consumed
led to similar results. In the following analysis, we only present calculations based on
the number of prey ingested during pregnancy, because it should more directly bear
upon the influence of energy intake on reproductive output and post partum female
body condition.
Changes in body mass (grams per days)
0.1
0.0
-0.1
-0.2
-0.3
-0.4
control group
un-fed NR
females
p < 0.002
no prey one prey
two prey
Figure 1. Effects of diet on female daily mass change (in grams per day scaled with initial body mass)
during the course of gestation. Error bars represent standard error. See text for statistics.
Litter characteristics and females post partum condition
The number of prey consumed during pregnancy did not influence litter size
(ANCOVA, F(2,32)= 0.21; p=0.81), litter mass (F(2,32)=1.67; p=0.23), fit litter size
(F(2,32)=1.39; p=0.26) and fit litter mass (F(2,32)=1.84; p=0.21). Similarly, no
difference in mean offspring snout vent length (ANCOVA, F(2,32)=0.57; p=0.57) or
mean offspring mass (ANCOVA, F(2,32)=0.55; p=0.58) were detected (Table 2).
165

However, females eating twice were in higher body condition than females fasting or
eating only one prey (ANOVA, F(2,32)=5.54; p

Body mass (grams)
160
140
120
100
80
60
40
O
July 1
(early gestation)
G P
August 31
(late gestation)
no prey
one prey
two prey
September 15
(post parturition)
Figure 3. Influence of diet on pattern of mass changes during the experiment. Symbols represent
size-adjusted body mass ± 1 standard error. (O: ovulation; G: gestation; P : parturition).
Discussion
Comparisons were made in the feeding behavior among three groups of pregnant
vipers and with a control group of non pregnant snakes. Pregnant females accepted
prey most of the time (41 of 45 feeding occasions). This observation clearly
invalidates the hypothesis of an intrinsic (physiological or anatomical) origin of
anorexia during pregnancy. Dramatic reduction of food intake reported in the field is
thus a facultative phenomenon and appears to be a consequence of modification in
the activity pattern. In female viviparous snakes, important behavioural changes are
associated with gestation (increase in basking rate, physical burden, Birchard et al.
167

1984; Seigel et al. 1987; Brodie 1989). In female aspic vipers, gestation is
accompanied by a drastic reduction of the home range of monitored females from
thirty to a few square meters (Naulleau et al. 1996). Pregnant females also adopt
higher thermal preferences and substantially increase basking time (Saint Girons
1952; Bonnet & Naulleau 1996). Thermal conditions are important to optimise
embryonic developmental speed along with many offspring traits in squamate reptiles
(Fox et al. 1961; Packard & Packard 1988; Shine et al. 1997). If changes in gravid
female thermal preferences optimise developmental rates of the embryos,
interference with other activities may also occur. Most notably, the time devoted to
thermoregulation will trade off with the time spent foraging. In this study, we
facilitated feeding behaviour by placing prey very close to the snakes. In support of
this, anecdotal cases of gravid females with a prey in the stomach even in late
gestation have been reported in the field (Naulleau 1997). Those females were
extremely sedentary (i.e. not engaged in foraging activities) as indicated by
extensive radio tracking data (Naulleau et al. 1996) and thus were probably “lucky” in
catching voles passing very close to the basking site. The same situation has been
documented in a closely related species, the adder (Vipera berus, Madsen & Shine
1992a).
As food was accepted, it was possible to explore the effects of the different
diets. Food intake significantly affected the pattern of mass change during gestation
(Fig 1). The higher body mass loss detected among un-fed pregnant females in
comparison with the control group suggests that metabolic expenditure associated
with embryonic development is high. Data on fasting pregnant females allowed us to
assess the relationship between daily mass loss during pregnancy (i.e. metabolic
expenditure) and reproductive output characteristics. Interestingly, the magnitude of
mass loss was not correlated with total litter size (n=9, r=0.06, F(1,7)=0.02, p

or more importantly fit litter size, (n=9, r=0.27, F(1,7)=0.56, p

current level of food availability notably by marked years to years variations in prey
(voles) density reported in the field (Delattre et al. 1992; Bonnet et al. 2001b).
Despite the limited sample size (n=9), the fecundity-independent nature of
metabolic cost associated with pregnancy agrees with results gathered by Madsen &
Shine (1993), who reported significant fixed costs in the adder (Vipera berus). Such
fecundity-independent costs are probably a widespread phenomenon influencing the
evolution of life history strategies (Bull & Shine 1979). In the case of the aspic viper,
the integration of those particular energetic costs will help to further understand the
“all or nothing” system of energy allocation displayed by this species (Bonnet and
Naulleau 1996; Bonnet et al. 2002a).
Acknowledgements
We thank Gwenaël Beauplet, Dale DeNardo, Olivier Dehorther and Yves Cherel for
constructive comments on the manuscript. We are grateful to Patrice Quistinic and
Michaël Guillon for help in snake collecting, and Hélène Blanchard, Günter and Felicy
for picnic organisation. Manuscript preparation was supported by the Conseil
Régional de Poitou-Charentes. Special thanks to Melle, notably for the numerous
Italian coffees. Finally, Rex Cambag provided his own inimitable commentary.
170

D. Article 6
Gestation, thermoregulation and metabolism in
a viviparous snake, Vipera aspis: evidence for
fecundity-independent costs
Mitchell Ladyman 1 , Xavier Bonnet 2 , Olivier Lourdais 2 3 4 , Don Bradshaw 1 and
Guy Naulleau 2
1 Department of Zoology, University of Western Australia, Perth, WA 6009
2 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
3 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
4 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
Accepted for publication in Physiological and Biochemical Zoology
171

Abstract
Oxygen consumption of gestating Aspic vipers, Viper aspis (L.), was strongly
dependent on body temperature and mass. Temperature-controlled, mass-
independent oxygen consumption did not differ between pregnant and non-pregnant
females. Maternal metabolism was not influenced during early gestation by the
number of embryos carried, but was weakly influenced during late gestation. These
results differ from previous investigations that show an increase in mass-independent
oxygen consumption in reproductive females relative to non-reproductive females
and a positive relationship between metabolism and litter size. These data also
conflict with published field data on Vipera aspis that show a strong metabolic cost
associated with reproduction. We propose that, under controlled conditions (i.e.
females exposed to precise ambient temperatures), following the mobilisation of
resources to create follicles (i.e. vitellogenesis), early gestation per se may not be an
energetically expensive period in reproduction. Under natural conditions, however,
the metabolic rate of reproductive females is strongly increased by a shift in thermal
ecology (higher body temperature and longer basking periods), enabling pregnant
females to accelerate the process of gestation. Combining both laboratory and field
investigation in a viviparous snake, we suggest that reproduction entails discrete
changes in the thermal ecology of females to provide optimal temperatures to the
embryos, whatever their number. This results in the counterintuitive notion that
metabolism may well be largely independent of fecundity during gestation, at least in
an ectothermic reptile.
172

Introduction
Reproduction represents a major disruption to the typical day-to-day life of any
female organism. The decision to reproduce, or to do so successfully, is often
strongly resource orientated; mediated by food availability (Nagy et al. 1984, 1995;
Boggs 1992; Meijer & Drent 1999), by body condition (Larsen 1980; Anderson &
Karasov 1981; Nagy 1987; Brown 1991; Naulleau & Bonnet 1996; Meijer & Drent
1999), or both (Meijer & Drent 1999; Bonnet et al. 2001b). For example, endogenous
reserve levels can determine the initiation of vitellogenesis (Bonnet et al. 1994), or
reproductive effort and associated costs (Madsen & Shine 1993; Erikstad et al. 1997;
Festa-Bianchet et al. 1998; Bonnet et al. 2001b). The amounts of body reserves can
also influence the maintenance of brooding (Chastel et al. 1995; Dearbon 2001),
gestation (Boyd 1984; Cresswell et al. 1992), the intensity of parental care (Olsson
1997), and hence offspring number; or can even affect the sex of the offspring.
In many vertebrates, reproduction occurs when an animal is in positive energy
balance as folliculogenesis maintains a low priority in the ongoing competition for
energy allocation (Bronson 1998; Schneider et al. 2000). Any attempt to reproduce
during periods of resource deficiency may potentially decrease both current and
future reproductive success (Stearns 1992). The association between reproduction
and resources is very intimate because reproduction is an energetically demanding
process. An analysis of the available literature suggests that there is a clear
difference between the metabolic rate of reproductive and non-reproductive
individuals (Birchard et al. 1984; Speakman & McQueenie 1996; Mauget et al. 1997;
Angilletta & Sears 2000). This conclusion is based on comparisons of rates of
metabolism measured during discrete periods of the reproductive cycle (see Guillette
1982). However, to generalise such a broad divergence in metabolism across the
173

whole reproductive process could be slightly hazardous given that this process is
clearly not static. For example, the physiological and behavioural mechanisms that
underlie folliculogenesis, gestation, and lactation are extremely different (Thibault &
Levasseur, 1991; Speakman & McQueenie 1996). Even among males, where the
energetic contribution to gametic production is almost negligible in most species,
metabolic rate varies with pubertal maturity (Brown et al. 1996) or with the alternation
of mating and non-mating periods (Olsson et al. 1997). Moreover, the ecological
implications of changes in metabolism over the reproductive period are important
considering the fact that food resource and other environmental conditions
(temperature, rainfall etc.), are often limited and/or often fluctuate.
Whatever the taxon or reproductive mode (i.e. viviparity versus oviparity), the
total metabolic effort expended over time is generally higher in reproductive females
than non-reproductive females during the reproductive period. Limiting the focus to
gestation, and hence considering viviparous species only, the extra metabolic
demands associated with the developing foetus are higher than the normal metabolic
demands for growth and repair (Hahn & Tinkle 1965; Mauget et al. 1997). Such a
difference is detectable even in the late luteal phase of the menstrual cycle (Curtis
1996). In mammals, this greater metabolic expenditure is often driven by a change in
the hormonal balance of the reproducing individual that, in turn, provokes an increase
in their basal metabolic rate through increased levels of cellular work (protein
synthesis, mitosis, ion pumping etc.) over the reproductive period (Thibault and
Levasseur 1991; Howe et al. 1993; Butte et al. 1999).
Physiological changes also occur in viviparous ectotherms to cope with the
demand of developing embryos. Birchard et al. (1982), and other authors, describe
an increase in mass-independent oxygen consumption in reproductive females
suggesting a modification of metabolic regulation set point, and this may also be
174

associated with hormonal changes (Highfill & Mead 1975; Fergusson and Bradshaw
1991; Callard et al. 1992). However, ectothermic vertebrates are limited in the extent
to which the can elevate the rate of cellular work and therefore their metabolic rate
through purely physiological mechanisms (Harlow & Grigg 1984; Shine et al. 1997;
Cadenas et al. 2000; Wang et al. 2001). Instead, ectotherms may depend on the
environment to compensate for such a physiological constraint (Bradshaw 1997),
using available thermal energy (heat) to increase metabolic reaction rate, and
increase development of embryos. It is a common observation that in viviparous
reptiles living in temperate areas, gravid females bask for longer periods than non-
gravid females (Gregory et al. 1987; Seigel & Ford 1987; Bonnet & Naulleau 1996;
Shine 1998) with the accelerated development of embryos occurring as a result
(Naulleau 1986). Changes in body temperature of pregnant female ectotherms is the
most effective way to increase the rates of cellular activity involved with production of
young and we may expect that this increase is proportional to the load of active
tissues represented by the embryos. Such changes will involve a change in
thermoregulatory behaviour. In addition, if we admit that variation in female metabolic
rate is an integrative effect of the cellular work of the tissues of the mother plus those
of developing embryos, we may expect a positive relationship between the size (and
mass) of the litter and maternal mass-independent oxygen consumption. Such a
relationship has been previously documented in viviparous reptiles, but the number of
studies has been limited (Birchard et al. 1984; Demarco & Guillette 1992).
Importantly, the form of the relationship between fecundity (litter size),
reproductive effort (mass-independent oxygen consumption, materials invested into
follicles, loss of feeding opportunities etc...), and the potential costs (decreased
reproductive value, here simply viewed as a combination of lower survival and
depletion of energy stores) has immense consequences on the evolution of life
175

history traits, and on the underlying physiological regulations (Shine & Schwarzkopf
1992; Stearns 1992; Niewiarowski & Dunham 1994; Sinervo & Svensson 1998).
Clearly both empirical and experimental data are required in this field. For instance,
increasing our knowledge in the costs/benefits relationship of pregnancy is a major
prerequisite to better understanding the oviparity versus viviparity transition that has
evolved independently several hundreds of times in many vertebrate taxa (fishes,
amphibians, squamate reptiles; Shine 1985).
Snakes are suitable models for investigations such as this, as there are
oviparous and viviparous species in the same family group (Shine 1985). Because
they are ectothermic, we can also experimentally control body temperature and
assess the impact of this factor on metabolism in both reproductive and non-
reproductive individuals (Beaupré & Duvall 1998). In addition, the broad range of
fecundity among conspecific females allows us to identify any potential
fecundity/reproductive effort/cost relationships. Finally, many snake species have a
less than annual breeding frequency (Saint Girons 1957b; Seigel & Ford 1987).
Therefore within any given year, both reproductive and non-reproductive individuals
are influenced by the same variation of climate and resource availability making the
two groups more comparable.
In this investigation we examined the influence of the reproductive status and
fecundity on the metabolic rate of viviparous female snakes. Notably, under the same
thermal conditions, are pregnant females more metabolically active than non-
reproductive females over the gestative period? As an aside, we demonstrate that
broad generalisations on the effects of reproductive status on metabolism, based on
laboratory experiments over discrete periods, may potentially limit the value of
conclusions in field conditions.
176

Materials and Methods
Origin and Description of Study Animals
In this study we used wild caught gravid female Aspic vipers (Vipera aspis). We used
this species because this snake is abundant in our study area, tolerant to captivity
and manipulation, and previous work provides a useful baseline on the reproductive
ecology and physiology of this species (Saint Girons 1957b; Saint Girons & Duguy
1992; Bonnet et al. 1994, 2000a,b, 2001a, b, 2002a, b; Aubret et al. 2002; Lourdais
et al. 2002a, b; and references therein). The Aspic viper is a stocky, venomous snake
species widely distributed throughout Western Europe (Naulleau 1997). Average
snout-vent length (SVL) is 48.5 cm and body mass (BM) is 85.5 g. Females mature at
approximately three years (Bonnet et al. 1999a), mating occurs in spring (March –
April)(Saint Girons 1957b; Naulleau et al. 1999), and parturition occurs in autumn
(September; Saint Girons 1957b). Clutch size varies from 1 -13 individuals (mean SVL
= 17.9±1.2 cm, mean BM = 6.3±1.1 g)(Bonnet et al. 2000b). Females in western
France do not breed every year (Saint Girons 1957b; Bonnet & Naulleau 1996) and
delay reproduction until they exceed a minimum body condition threshold through the
accumulation of large body reserves such as abdominal fat bodies (Naulleau &
Bonnet 1996; Aubret et al. 2002).
Individuals used in this investigation were collected by hand from Les
Moutiers-en-Retz and Les Sables d’Olonne, central western France (47 o 03’ N; 02 o
00’ W and 46° 30’ N; 01° 44’ W respectively). Both study sites lie within 60 km of
each other. Habitats were consistent between the two study sites and the climate for
the region is a temperate oceanic one (see Bonnet & Naulleau, 1993, for average
temperatures).
177

Specimens used for the metabolic tests were collected during early spring,
after the mating period and during vitellogenesis (Saint Girons 1957b). At this time,
reproductive females had already committed to the development of a number of
follicles. Individuals were housed in outdoor terraria (8 – 16 m²) in our laboratory (46°
08’ N; 00° 25’ W), from June to late September, and exposed to a similar climate to
the field site from where they were collected. Within each enclosure snakes were
provided with a mosaic of microhabitats, including many shelters and well-exposed
sites to facilitate optimal thermoregulation. A unique code of ventral scale clipping
identified individuals and these marks are permanent as the regenerated tissues
exhibit a different colour.
Resting Oxygen Consumption
The oxygen consumption of 50 reproductive females and 19 non-reproductive
females was measured in the laboratory under controlled environmental conditions.
Subjects were tested for rates of oxygen consumption over two phases during the
survey period: early gestation and late gestation. A total of 59 snakes was tested
during early gestation (40 reproductive and 19 non-reproductive) and a subset of 17
reproductive snakes (among the 40 reproductive females) was used for repeated-
measures analysis in late gestation. An additional 12 reproductive snakes were
measured in late gestation at 32 o C to increase the statistical power of the data set
used to derive relationship with fecundity. Table 1 summarises the number of females
sampled and re-sampled at different temperature regimes and/or at the two periods
of gestation.
178

Table 1: Details on the number of female Aspic vipers assayed for O2 consumption. The first number
provides the total number of females used in each treatment group. The number in brackets refers to
females used in repeated measures experiment over early and late gestation.
Reproductive Status Gestation period 17.5 o C 25 o C 32.5 o C
Reproductive Early 14 13 13
Non-reproductive Early 6 5 8
Reproductive Late 7 (7) None 22 (10)
Rates of oxygen consumption (VO2 - mL O2 g -1 h -1 ) were measured using a
flow-through respirometry system. Dry incurrent air was drawn through a small, clear
Perspex metabolic chamber at a rate of 204+4 mL min -1 by a Byoblock Scientific 6 L
air pump and flow was controlled using a Platon mass flow controller. The chamber,
specifically built to accommodate Aspic vipers, was large enough (internal diameter:
15 x 15 x 5 cm) to accommodate snakes up to 200 g, without preventing voluntary
activity. Oxygen concentration was maintained at approximately 20.1 %. The
metabolic chamber was located within a sealed Cryosystem temperature-controlled
chamber and positioned such that snakes could be observed during the trial through
a small viewing port. During early-gestation the body temperature of each individual
was checked prior to commencement of the test using a quick-registering
thermometer (Novo) inserted 2 cm into the cloaca until stabilisation of the measure.
Excurrent air was passed through two column desiccators containing drierite,
then through a paramagnetic O2 transducer (Servomex Series 1100). The differential
output of the oxygen analyser (ambient air minus excurrent air) was recorded,
adjusted to standard temperature and pressure conditions and plotted on a standard
desktop PC. The metabolic chamber was calibrated to the outside atmosphere
(Pressure Indicator Druck DPI 260) and set at zero oxygen consumption by running
179

an empty chamber for one hour prior to each snake being tested. Differential output
was presented graphically as it was acquired and snakes were run for as long as
necessary to obtain stable oxygen consumption for a period greater than half of an
hour. Monitoring of snakes ensured that fluctuations in VO2 were attributed to activity
and not technical perturbations. Oxygen consumption was calculated following the
equation of Decopas & Hart (1957).
Reproductive and non-reproductive snakes were run at 17.5 o C, 25 o C and
32.5 o C degrees during early gestation, and reproductive snakes only were run at
17.5 o C and 32.5 o C during late gestation. This corresponds to the range of
temperatures we have recorded, using radio telemetry, in the field in females
(reproductive and non-reproductive) during the pregnancy period (Naulleau et al.
1996; unpublished data). Before measurements of oxygen consumption, animals
were fasted for a minimum of four days and introduced to the system on the night
prior to testing. Because the Aspic viper is diurnal, all experiments were conducted
during the normal light phase appropriate for the time of the year and location
(approximately 0600 to 1900 hours). Snakes of both reproductive statuses were
tested randomly throughout the day to remove any possible influence of diel cycle.
Body mass of snakes was recorded on a top-loading electronic scale (±0.1 g) before
each run, and the measurements were used to derive mass-independent volume of
oxygen consumed (mass-independent VO2).
Measurements of Fecundity
Fecundity was determined by palpation in early gestation and, in this species, follicles
as small as 2.0 g can be detected. Fecundity was confirmed at parturition. Live and
stillborn neonates were included in the analysis as metabolically active tissues during
pregnancy. Unfertile and/or undeveloped eggs (i.e. where only yolk was identifiable)
180

that are frequent in Aspic viper’s litters (Bonnet et al. 2001b) were not considered as
metabolically active tissues during pregnancy as these undeveloped eggs are formed
during vitellogenesis, but not pregnancy.
Field Body Temperatures
We collected field data to compare the temperature regimes imposed on snakes in
our laboratory experiment with the far more complex situation experienced by wild
vipers. Using internal temperature radio transmitters, we determined the body
temperature of reproductive and non-reproductive females in the field during most of
the period of reproduction (15 th April 1996 to the 15 th July 1996), encompassing the
45 last days of vitellogenesis and the 45 first days of gestation. The methodology has
been described elsewhere (Naulleau et al. 1996), and has also been used
successfully in a closely related species, the adder (Vipera berus)(Madsen & Shine
1992a; 1993).
Radio tracking enabled the collection of over 1896 temperature records from
21 female Aspic vipers (11 reproductive and 10 non-reproductive). The temperatures
were collected, without disturbing the snake, during the day. Each animal was
sampled one to three times per day: in the morning, at mid-day and in the afternoon
(mean number of records per snake per day was 2.33±0.84, range 1-6 with 99.9 % of
the cases comprised between 1 and 4 and 97 % below 4). Because we sampled
each snake randomly irrespective of the reproductive status, the exact time elapsed
between two consecutive records was also random. However, it was always greater
than 3 hours and often greater than 12 hours. The duration between samples
ensures that consecutive records will reflect the thermal behaviour adopted by each
snake rather than the thermal inertia of the body of the snake. As body temperature
is affected by ambient temperature, and ambient temperature increased over our
181

sampling period, we took into account the effect of date in most analysis involving
field body temperatures.
Statistical Analysis
Body mass (BM-grams) and oxygen consumption (VO2 – mL O2 h -1 ) were log10
transformed to meet the normality assumption (Shapiro-Wilk W = 0.986, P = 0.731 for
Log – BM and Shapiro-Wilk W = 0.972, P = 0.158 for Log – VO2)(Zar 1984). Oxygen
consumption data were adjusted for body mass by regressing Log - VO2 on Log -
BM. The residuals from this regression yielded mass-independent oxygen
consumption (mass-independent VO2), which was used for several analysis (see
results). We did not use ratios to scale oxygen consumption (Atchley et al. 1976;
Packard and Boardman 1988). Instead we performed ANCOVAs, for example, to
compare reproductive versus non-reproductive females using BM as a covariate
(Garcia-Berthou 2001). Statistical analysis of VO2 was performed on the mean rate of
oxygen consumption measured over a stable half hour period for each individual.
Analyses were performed using Statistica 5.1 and 6.0 (Statsoft 1995, 2001).
Results
Reproductive Status, Body Size and Body Condition in Early Gestation
At the beginning of gestation, there was no difference in SVL between reproductive
and non-reproductive females (one factor ANOVA with SVL as the dependent
variable and reproductive status as the factor: F(1,56)=2.74, p=0.10; Table 2).
However, reproductive females were significantly heavier than non-reproductive
females (same design ANOVA with BM as the dependent variable: F(1,56) = 17.16,
p< 0.001; Table 2), and, logically were in better body condition than non-reproductive
182

snakes (ANCOVA with Log - BM as the dependent variable and Log - SVL as the
covariate: F(1,55) = 34.12, p< 0.001; Table 2). Importantly, at the beginning of
gestation, the greater body condition of reproductive females relative to non-
reproductive females does not mean that they possess larger body reserves. In fact,
most of the extra-mass of reproductive females is represented by the litter and,
despite their external appearance, pregnant females are relatively emaciated during
early gestation (Bonnet et al. 2002a).
Table 2: Morphometrics of 40 reproductive and 19 non-reproductive female Aspic vipers used during
early-gestation. Means are expressed ± SD. Comparing reproductive versus non-reproductive
females, adjusted body mass (scaled by size using SVL as a covariate) was greater in pregnant
females; but not necessarily the amounts of body reserves (see text).
Reproductive Status Snout-vent Length
(cm)
Body Mass (g) Adjusted Body
Mass (g)
Reproductive 48.80±6.14 99.84±38.01 93.59±3.25
Non-reproductive 46.17±4.11 60.73±18.16 66.98±4.80
Effect of Body Mass, Ambient Temperature and Reproductive Status on Oxygen
Consumption in Early Gestation
Body mass and body temperature are the two most important determinates of any
ectotherm’s metabolic rate, therefore, these variables must be controlled before
looking for an effect of reproductive status. We performed an ANCOVA with Log –
VO2 (ml -1 h -1 ) as the dependent variable, reproductive status as the factor, Log - BM
and body temperature as the covariates. The whole model (test of the sum of
squares [SS] of the whole model versus SS residuals) explained a large proportion of
the variance in oxygen consumption (r² = 0.60; F(3,54) = 26.59, p < 0.0001; Levene’s
183

test for homogeneity of the variance, degrees of freedom for all F’s - 1, 56; for Log –
VO2: F = 1.14, p= 0.29; for Log - BM: F = 0.47, p= 0.50; for Body Temperature: F =
1.46, p = 0.23), reinforcing the notion that body mass and body temperatures play a
major role in the oxygen consumption of ectotherms (Saint Girons et al. 1985;
Beaupré & Zaidan III 2001). Although both covariates strongly and positively
influenced oxygen consumption (respectively Log - BM: F(1,54) = 12.29, p < 0.001;
and body temperature F(1,54)= 51.76, p < 0.0001), we did not find any significant
effect caused by the reproductive status (F(1.54)= 2.82, p= 0.10)(Figure 1). Using
ambient temperature (17.5 °C, 25 °C and 32.5 °C) as a second factor (due to its
discontinuous nature in the experiment), instead of body temperature as a covariate,
did not change the result (whole model: r² = 0.61; specific effect of ambient
temperature: F(2,51)=25.07, p

Mass Adjusted Oxygen Consumption
Log-VO2 (ml -1 h -1 )
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
Reproductive
Non reproductive
6
14
13
17.5°C 25.0°C 32.5°C
Figure 1. Effect of reproductive status and temperature on oxygen consumption of female Aspic vipers
during early-gestation. Each point represents the adjusted mean (least-squares using Log - VO2 ( mL -1
h -1 ) as the dependent variable, reproductive status and ambient temperature as the factor and Log -
BM as the covariate)(±SE and sample size) for each group of snakes in each category.
Table 3. Comparison of the current mass-independent VO2 (ml g 1 h -1 ) data against data previously
collected an ecologically similar species and an ecologically different species tested at three imposed
temperature regimes (Secor & Nagy 1994). Vipera aspis and Crotalus cerastes are similar, sit-and-
wait predators (capital breeders), while Masticophis flagellum is a typical active forager (income
breeder).
Species Temperature Category
5
Cold (17.5 o C) Medium (25.0 o C) Hot (32.5 o C)
Vipera aspis 0.018 0.031 0.05
Masticophis flagellum 0.012 0.028 0.06
Crotalus cerastes 0.009 0.020 0.04
185
8
13

Effect of Gestation Period on Oxygen Consumption
In pregnant females, there was no change between the mean mass-independent VO2
(residuals) for the two periods, early and late gestation. To control for possible inter-
individual differences in reproductive stage, or body mass, we used the same
females on the two occasions under the same temperature conditions (seven
females at 17.5 °C and 10 females at 32.5 °C). The seventeen females were
sampled during early gestation and again, two months later, during late gestation.
Despite being sensitive enough to detect any potential difference in mass-
independent VO2 between gestation periods the T-test for dependent samples
provided a non-significant result (t = 1.56, df = 16, p= 0.14; Figure 2).
Mass Independent Oxygen
Consumption (residuals)
0.3
0.2
0.1
0.0
-0.1
-0.2
-0.3
-0.4
EARLY LATE
GESTATION
Figure 2. The effect of early (shortly after ovulation) versus late (shortly before parturition) gestation,
on oxygen consumption in 17 pregnant female vipers. Each female was plotted two times (early and
late gestation) with a line connecting the two values. See text for statistics and details.
186

Effect of Fecundity on Oxygen Consumption
For the data collected in early gestation, a partial correlation analysis was performed
to take into account the effect of body temperature, body mass and fecundity on
oxygen consumption (r² = 0.51, F(3,34) = 11.97, p < 0.0001; specific effect of body
temperature, F(3,34) = 4.62; p < 0.001; Log - BM, F(3,34)=2.68, p = 0.011; fecundity,
F(3,34)= -0.31, p = 0.758). Overall, in early gestation, fecundity per se did not
influence significantly oxygen consumption of reproductive individuals (Figure 3a).
For this analysis the sample size was reduced and the three ambient temperatures
used (17.5 °C, 25 °C and 32.5 °C) created discontinous data between groups of
females, making the use of body temperature as an independent variable suspect.
For this reason, we adjusted the oxygen consumption to the mean value obtained at
32.5 °C by multiplying the values obtained under low and medium temperature by 2.8
and 1.6 respectively; the factor by which oxygen consumption increased as
temperature increased. Even after adjusting oxygen consumption, the result
remained unchanged (r² = 0.18, F(2,35) = 3.90, p < 0.03; specific effect of fecundity,
F(2,35) = 1.32, p = 0.195).
However, in late gestation the data show a positive and significant relationship
between oxygen consumption and fecundity, measured as the number of offspring
produced at parturition (Figure 3b). A partial correlation analysis was performed to
take into account the effect of body mass on oxygen consumption. For this analysis
the sample size was further reduced (n = 19 females) and only two ambient
temperatures were used (17.5 °C and 32.5 °C), thus we used the temperature
adjusted oxygen consumption (r² = 0.38, F(2,16) = 4.96, p = 0.02; specific of
fecundity, F(2,16) = 2.97, p = 0.008) and, despite a small sample size, the power [1-
β] of this analysis was relatively high (0.83).
187

Mass Specific Temperature-Adjusted
Oxygen Consumption (residuals)
Mass Specific Temperature-Adjusted
Oxygen Consumption (residuals)
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
1 2 3 4 5 6 7 8 9 10 11 12 13
0.5
0.4
0.3
0.2
0.1
0.0
-0.1
-0.2
-0.3
-0.4
Fecundity at ovulation
-0.5
1 2 3 4 5 6 7 8
Number of viable offspring
Figure 3. Relationship between oxygen consumption (residuals from the regression of Log – BM
against Log - VO2, adjusted to an ambient temperature of 32.5°C, see text) and fecundity of
reproductive female vipers during early gestation (N = 38, upper graph [Figure 3a], Y = 0.02X - 0.413,
P = 0.227), and the number of viable neonates during late gestation (N = 19, lower graph [Figure 3b],
Y = 0.610X - 0.501, P < 0.006). Fecundity at ovulation was determined by palpation (see text), and
was measured as the number of viable neonates or fully developed stillborn young at parturition. A
number of ovulated follicles do not develop and lead to unfertile eggs (where only yolk is identifiable),
explaining the difference between the ranges of values between the two X axis (Bonnet et al. 2000b).
A similar relationship was observed regressing the mass of live young
produced against temperature adjusted VO2 (r 2 = 0.38, p = 0.029, n = 19, power =
0.83). Interestingly, we found that post-parturient body mass of the females had no
influence on the oxygen consumption measured before parturition (r² = 0.004, n = 19,
188

p = 0.80), suggesting that the contribution of the fatigued maternal organisms in the
total variance was minor in comparison to the contribution of the offspring they carry.
Field Body Temperatures
We found a strong effect of the sampling date on the mean body temperature
selected by snakes in the field (ANOVA, F(89,1806) = 7.44, p < 0.0001). As
expected, the mean body temperature of the snakes increased from mid-spring to
mid-summer (correlation between mean body temperature of the females and date: r²
= 0.16, F(1,88) = 17.23, p < 0.0001), with important daily fluctuations as expected
under temperate-oceanic climate (see Figure 4). Field data showed that in the course
of the reproductive period, reproductive females maintain higher body temperature
(25.01±0.26 °C, N = 989) than non-reproductive females (22.03±0.27 °C, n =
907)(ANCOVA with reproductive status as the factor, temperature records as the
dependent variable and date as the covariate; reproductive status, F(1,1893) = 44.05,
P < 0.0001; date, F(1,1893) = 59.86, p < 0.0001; Levene’s test for homogeneity of the
variance; F(1,851) = 1.01, p = 0.31 and F(1,851) = 0.93, p = 0.34 for body
temperature and date respectively). Such a difference was more pronounced during
gestation where the body temperature of reproductive snakes was approximately
3.79 °C higher (3.27 °C for date-adjusted means); the mean body temperature of
reproductive and non-reproductive females being 26.39±0.29 °C (±SD, N = 605) and
22.60±0.37 o C (±SD, n = 438), respectively (same design ANCOVA; reproductive
status, F(1,1040) = 47.10, p < 0.0001; date, F(1,1040) = 15.75, p < 0.0001, Figure 4).
Despite the time elapsed between two consecutive temperature records, each
individual was represented more than once, and this may lead to spurious pseudo-
replication effects. We checked for the possibility that pseudo-replication generated
the observed difference in body temperature between the two classes. We used the
189

Mean field-body temperature
of female asp vipers (°C)
mean body temperature calculated over the reproductive period for each female
(generating independent data but weakening the sensitivity of the analysis) and
performed an ANOVA with reproductive status as the factor and mean body
temperature as the dependent variable. The above results were unchanged (ANOVA,
F(1,19) = 7.06, p = 0.015), with mean body temperature of reproductive and non-
reproductive females being 24.54±1.23 o C (±SD, N = 11) and 19.80±1.29 °C (±SD, N
= 10), respectively.
35
30
25
20
15
5
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46
Days of Gestation
Figure 4. During gestation, reproductive females (grey symbols) maintain higher body temperatures
than non-reproductive females (white symbols). Each symbol represent the mean body temperature
(±SE) calculated for several females (n = 11.85±5.29 records on average, range 2 - 25). Despite
marked daily fluctuations (due to environmental variations: i.e. passage of very cold fronts in late June,
days 26 and 30), the mean body temperature of reproductive females was often several degrees
higher compared to non-reproductive females. Body temperature was taken at the point of capture
irrespective of behavioural status. See results for statistics.
190

Discussion
The data from this investigation suggest that, throughout the reproductive period,
metabolism in both reproductive and non-reproductive snakes was strongly affected
by both temperature and body mass. An important body of published data have
already reported these effects in ectotherms, for temperature (Bennett & Dawson
1976; Naulleau et al. 1984; Andrews & Pough 1985; Loumbourdis & Hailey 1985;
Bradshaw et al. 1991; Zari 1991; Thompson & Withers 1992) and body mass
(Bennett & Dawson 1976; Dmi'el 1986; Zari 1991; Thompson & Withers 1992;
Beaupré & Zaidan III 2001). The expected difference in the metabolic rate between
reproductive and non-reproductive females, however, was not apparent in our data
when the two groups of snakes were placed under the same temperature regime.
Furthermore, in reproductive females, metabolism was largely independent of
fecundity. With regard to these latter outcomes, our results differ from previous
investigations in this field (Guillette 1982; Birchard et al. 1984; Beaupré & Duvall
1998; Kunkele 2000).
Why do our data conflict with other investigations and fail to support some of
our initial hypotheses? Firstly, it is necessary to consider the methodology. For
example, the period over which oxygen consumption was measured was different
between the current and previous studies. In the live bearing lizard Sceloporus
aeneus and the garter snake, Thamnophis s. sirtalis comparisons were made with
reproductive females less than two weeks prior to parturition (Guillette 1982; Birchard
et al. 1984). Beaupré & Duvall (1998) focused on vitellogenesis and they showed that
reproductive female western diamondback rattlesnakes, Crotalus atrox, consume 1.4
times the amount of oxygen as non-reproductive females; however, it is not made
clear exactly when during vitellogenesis (a prolonged period in snakes; two to four
191

months, or more in viperids) the data were collected. Because the metabolic process
from the beginning of vitellogenesis to the end of pregnancy would be very dynamic,
it is difficult to compare directly these data and previous investigations. In addition,
the few snake species studied belong to very different lineages (colubrids versus
viperids for instance). It may also be that our methodology lacked the capacity to
detect subtle differences that may affect the outcome. Important inter-individual
differences, such as sensitivity to manipulation, hormonal levels related to stress, and
state of metabolically active tissue, such as the intestine epithelium, would certainly
increase the variation in our data set.
Nevertheless, the consistency between published and current data suggests
that the system employed during this investigation lead to reasonable and
interpretable results (Table 3) and important effects such as those linked to body
mass or body temperature were clearly visible in our data. For instance our Q10 of
2.05 is similar to the typical Q10 of between two and three reported for most reptiles
(Bennett & Dawson 1976; Andrews & Pough 1985; Thompson & Withers 1992). The
experimental process undertaken, augmented by long term monitoring of large
number of female Aspic vipers in the field, enables us to propose a number of
falsifiable hypotheses to explain our unexpected results. For that, a simple approach
is to follow the chronology of reproduction in female viviparous snakes: from
vitellogenesis to the end of gestation.
Vitellogenesis
Clear differences in the metabolic effort between reproductive and non-reproductive
females during vitellogenesis may be expected in capital breeders, such as female
rattlesnakes or vipers. In capital breeders, non-reproductive females accumulate
energy, sometimes over very prolonged periods (years in vertebrates; Bull & Shine
192

1979), until a sufficient store (the capital) has been constituted following which
vitellogenesis can be engaged (Drent & Daan 1980; Stearns 1992; Naulleau &
Bonnet 1996; Bonnet et al. 1998). During vitellogenesis there is an extensive
mobilisation of maternal body reserves to produce a large number of offspring (Bull &
Shine 1979). In snakes, the vitellogenic process is a very intensive physiological
event (Bonnet et al. 1994) where the energetic expenditure of reproductive females
increases substantially over that of non-reproductive females, as was demonstrated
by Beaupré & Duvall (1998) in the viperid Crotalus atrox. Although we have no data
on the metabolic rate of Aspic vipers during vitellogenesis, indirect information
supports the notion that this period corresponds to an increase of oxygen
consumption for reproductive females over non-reproductive females. Firstly, large
amounts of body resources are mobilised to develop follicles whilst non-vitellogenic
females do not exhibit any sign of such a mobilisation at that time (Bonnet et al.
1994). Second, vitellogenic females move and forage intensively at that time, mostly
to supplement the energy available for follicular development which, in turn, improves
their reproductive success; at the same time, non-vitellogenic females are more
sedentary, probably to save energy and to minimise predation risk (Naulleau et al.
1996; Bonnet et al. 2001a, 2002a). Third, vitellogenic females engage in
energetically demanding acts of sexual behaviour (courtship, mating) whilst non-
vitellogenic females do not (Naulleau et al. 1999; Aubret et al. 2002). Fourth,
vitellogenic females bask in the sun much more frequently than non-vitellogenic
females (Bonnet & Naulleau 1996) with a concomitant increase in metabolism due to
the fundamental effects of increased temperature on reaction rates (see Results).
Overall, development of follicles during vitellogenesis is dependant on the increased
activity of metabolically active tissues such as the liver, the ovaries, the intestine
epithelium, and the locomotor muscles, whilst these tissues are relatively less active
193

in non-vitellogenic females (Secor et al. 1994; Singh & Ramachandran 2000;
O’Sullivan et al. 2001). The completion of vitellogenesis occurs in early June
(Naulleau & Bidaut 1981), and this period corresponds to the beginning of summer in
our study area.
Pregnancy
Pregnancy immediately follows vitellogenesis. At this time follicular development and
the mobilisation of maternal reserves are complete (Bonnet et al. 1994, 2001a) and
the extra-metabolic cost of being pregnant in Aspic vipers may only be a small effort
associated with the maintenance of follicles. Our data support this notion with the
decrease in reproductive effort apparent in the lack of difference between the
oxygen consumption of reproductive and non-reproductive females. However, during
the early stages of pregnancy a precise comparison of oxygen consumption between
the two classes of female snakes is difficult due to the potentially confounding effect
of the mass of the yolk follicles. Yolk is comprised mainly of fat (50%) and water,
components likely to contribute little to organismal metabolic effort (Darken et al.
1998). In early pregnancy developing embryos are very small (less than 5mm in total
length; unpublished data) and the yolk component is relatively large. This is a
methodological difficulty that can only be resolved using techniques such as
ultrasound (Beaupré & Duvall 1998) and Nuclear Magnetic Imaging and Proton
Spectroscopy (unpublished) to measure the exact volume and chemical composition
(i.e. water versus lipids using NMPS) of the follicles, from which mass and energy
can be calculated in a non-destructive manner.
Field data further support the idea that early gestation is not metabolically
demanding. Gravid females in early gestation are often already emaciated at this
stage (i.e. fat bodies represent less than 3% of the mass of the post-ovulatory
194

females; Bonnet et al. 2002b). In this situation any substantial increases in
metabolism would lead to a rapid depletion of these reserves before parturition and
the possible failure to successfully produce viable offspring. Pregnant females could
forage at this time to compensate for their low energy stores. However, field data
show that this is not often the case. Instead they sharply reduce both locomotor and
hunting activities during the two to three months of pregnancy, seldom capturing prey
(Naulleau et al. 1996; Lourdais et al. 2002a). The limited body reserves of pregnant
females, sometimes supplemented by energy from prey, provide sufficient energy to
sustain the metabolism of the females until parturition (Lourdais et al. 2002a). These
experimental, field, and anatomical (comparative metabolic activity of muscles, liver,
or intestinal epithelium between vitellogenesis and gestation) data suggest that early
pregnancy is not a costly metabolic phenomenon per se. As a consequence the
absence of a difference in mass-independent VO2 between pregnant and non-
reproductive female vipers in early gestation is not surprising when temperature is
controlled by the experimenters.
In late gestation embryonic development was mostly complete, and each
foetus could potentially contribute to the overall metabolism of the reproductive
female. However, the current investigation suggests that reproductive females
maintain a similar rate of mass-independent VO2 throughout pregnancy. Such a result
is partly logical because VO2 was systematically scaled by maternal mass, which
includes the mass of the embryo and extra-embryonic fluids (the total water content
of embryo and fluids being close to 80 %; unpublished data) in late gestation. But this
is not entirely satisfactory as we may expect a slight increase in the mass-
independent oxygen consumption of the females in the course of pregnancy. Firstly,
because the overall body composition (mother and embryos) now includes a larger
proportion of active tissues (muscles of the embryos etc. versus yolk) at the end of
195

gestation. Second, the oxygen consumption by developing embryos is not the only
factor of pregnancy that will affect metabolism. Other energetic components, such as
supplying oxygen to the foetuses and handling foetal nitrogenous waste, are also
paid by the mother (Clark & Sisken 1956). Birchard et al. (1984) suggest that the
contribution of foetal oxygen consumption to the overall oxygen consumption of the
mother is the factor responsible for the observed difference in metabolic rate of
pregnant versus non-pregnant garter snakes. Perhaps inter-individual variance
rendered our data too noisy to detect such a difference that is likely to be subtle
anyway. In late gestation there was a significant positive relationship between
fecundity and oxygen consumption, though the relationship was weak, indicating that
our measurements enabled to detect the fact that the intra-uterine embryos were
metabolically active. An alternative explanation may be that any detectable difference
generated by the embryos may have been masked by a decrease in maternal
metabolism at the end of gestation due to the fatigued body condition of the
reproductive females in the later stages of pregnancy (Bonnet et al. 2002b).
Field versus laboratory data
Overall, when ambient temperature was imposed in the laboratory, reproductive
status and fecundity did not influence, or only weakly influenced, maternal
metabolism. However, field body temperatures show that a commitment to
reproduction means a dramatic shift in behavioural thermoregulation with pregnant
female vipers maintaining, on average, a 4 o C higher body temperature than non-
reproductive females. This result is consistent with the observation that, in the field,
pregnant females bask more often than non-reproductive females (Bonnet &
Naulleau 1996). Given the fact that reproduction (at least vitellogenesis) is an
energetically expensive process, and considering the difficulties of reproductive
196

females to sustain metabolism at a higher rate than non-reproductive females
independently from a shift in thermoregulatory behaviour (as they are ectothermic), it
seems intuitive that the maintenance of higher preferred body temperature over
prolonged time periods accelerates metabolism to the rate where development of
young is completed over an appropriate time scale. Embryos are sensitive to
temperature, both in terms of extremes and variations (Burger 1998; Rhen & Lang
1999), and development is usually optimised by precise temperatures (Shine &
Harlow 1996; Downes & Shine 1999; Elphick & Shine; Shine 1999; Shine & Downes
1999; Andrews et al 2000). It is very likely that the optimal temperatures for follicular
growth and for embryonic development are the same whatever the number of
offspring. This may explain the absence of any relationship, in the current
investigation, between mean field body temperature measured in pregnant females
and fecundity.
Conclusion
Taken together these results suggest that, in female Asp vipers, the change of
reproductive status corresponds to an all-or-nothing system in terms of energetic
metabolism. Vitellogenesis and pregnancy impose a discrete modification in the day-
to-day life of the females rather than progressive adjustments to the size of the litter
(from zero to 13 offspring in our study zone). Capture-recapture data have shown
that survival is independent of fecundity in our population (Bonnet et al. 2002b). We
hypothesise that increased thermoregulatory behaviour linked to vitellogenesis and
gestation substantially increase maternal metabolism, provoke a strong emaciation
and result in increased exposure of the females to avian predation. As a result, the
majority of the females will not survive a single reproductive event whatever their
current fecundity (Bonnet et al. 2002a, b). Our field and laboratory data support the
197

notion proposed by Bull & Shine (1979) that in systems where the costs of
reproduction independent of fecundity are significant, the emergence of reproductive
strategies based on a low breeding frequency should be favoured (which is the case
for the many viperid snakes; Saint Girons 1957b; Brown 1991; Martin 1993; Bonnet &
Naulleau 1996). The notion that costs of reproduction may be partly disconnected to
fecundity provides immense potential for better understanding reproductive strategies
such as capital breeding and or semelparity (Bull & Shine 1979; Bonnet et al. 1998;
Olsson, et al. 2000). Although incomplete, our data clearly show that the true value of
understanding the relationships between fecundity, the energetics of reproduction
and reproductive effort lie not in a simple comparison with non-reproductive
conspecifics tested under laboratory conditions, but in the synthesis of laboratory
data and field observation.
Aknowledgements
A special thanks to the CNRS, French Embassy in Canberra (A. Moulet, A. Littardi,
and J. Mordeck), the Faculty of Science and the Zoology Department (UWA) for the
financial support that made this project possible. Laurence Pastout taught the use of
the metabolic chamber, Guy Merlet solved several technical problems, and Christian
Thiburce cared for the animals between measurements.
198

E. Article 7
What is the Appropriate Timescale for
Measuring Costs of Reproduction in a “Capital
Breeder” such as the aspic viper?
X. Bonnet 1 , G Naulleau 1 , R Shine 4 1 2 3
, & O Lourdais
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 Biological Sciences A08, University of Sydney, NSW 2006, Australia
Published in Evolutionary Ecology 13: 485-497
(2000)
199

Abstract
Before we can quantify the degree to which reproductive activities constitute a cost
(i.e., depress an organism's probable future reproductive output), we need to
determine the timescale over which such costs are paid. This is straightforward for
species that acquire and expend resources simultaneously (income breeders), but
more problematical for organisms that gather resources over a long period and then
expend them in a brief reproductive phase (capital breeders). Most snakes are
capital breeders; for example, female aspic vipers (Vipera aspis) in central western
France exhibit a two- to three-year reproductive cycle, with females amassing energy
reserves for one or more years prior to the year in which they become pregnant. We
use long-term mark-recapture data on free-living vipers to quantify the appropriate
timescale for studies of reproductive costs. Annual survival rates of female vipers
varied significantly during their cycle, such that estimates of survival costs based only
on years when the females were "reproductive" (i.e., produced offspring) substantially
underestimated the true costs of reproduction. High mortality in the year after
reproducing was apparently linked to reproductive output; low energy reserves (poor
body condition) after parturition were associated with low survival rates in the
following year. Thus, measures of cost need to consider the timescale over which
resources are gathered as well as that over which they are expended in reproductive
activities. Also, the timescale of measurement needs to continue for long enough
into the post-reproductive period to detect delayed effects of reproductive "decisions".
Key-words: body condition; capital breeder; energy stores; foraging; snake; Vipera
aspis
200

Introduction
One of the primary aims of life-history theory is to facilitate comparisons among
different kinds of organisms. Only in this way, by examining diverse taxa in a single
conceptual framework, can we hope to derive general insights. One of the great
successes in this respect has been the development of theory concerning costs of
reproduction and the allied concept of Reproductive Effort. The field is based upon
an original idea by George Williams (1966a,b): the realisation that an iteroparous
organism will maximise its lifetime reproductive success not by maximising effort at
the first reproductive opportunity, but by reproducing at a level that does not too
greatly reduce it’s probable future output. This notion has been incorporated into
elegant mathematical models(e.g., Schaffer 1974; Winkler & Wallin 198; Sibly &
Calow 1984; Jönsson et al. 1995a,b), and has stimulated extensive empirical studies
on a diverse array of species (e.g., Bell 1980; Bell & Koufopanou 1986; Clutton-Brock
1991).
In the present paper, we point out an important complication in measuring the
costs of reproduction: we have to be careful about specifying what activities are
included under our definition of reproduction. The significance of this superficially
trivial caveat is that organisms differ in the timescale over which various components
of “reproductive” activities are carried out. In particular, some animals concentrate all
of the activities associated with reproduction (i.e., acquisition as well as expenditure
of energy) in a single time period. For these income breeders (Drent & Daan 1980;
Jönsson 1997), we can evaluate costs of reproduction by comparing energy balance
and survival rates between reproductive and non-reproductive animals (e.g., Bell &
Koufopanou 1986) or by documenting the effects of manipulating reproductive
expenditure (e.g., Tombre & Erikstad 1996; Cichon et al. 1998). The situation is
201

more complex with capital breeders that rely upon stored energy reserves to support
reproductive output. For such taxa, energy acquisition and expenditure are
temporally dissociated, so that measurements of cost taken during the period of
reproductive expenditure will fail to include the components of cost that accrue during
the energy-gathering phase.
Timescales of Reproductive Costs
Much of the scientific literature on costs of reproduction is based on studies of birds,
and one technique that is often used is to manipulate reproductive expenditure by
adding or removing eggs from the nest at the beginning of the period of parental care
(e.g., Nur 1988; Tombre & Erikstad 1996; Cichon et al. 1998). Importantly, the facet
of reproductive effort that is being affected by this manipulation is the level of the
parents’ foraging effort to provision the nestling. If we compare this situation to that
of a viviparous snake such as our study organism, the aspic viper, the contrast
becomes obvious. Female aspic vipers produce litters only once every two or three
years (or less often), with the intervening (“non-reproductive”) years being used to
amass energy reserves that will fuel the eventual litter (Naulleau & Bonnet 1996).
Females may eat relatively little during gestation; indeed, some may become
completely anorexic (Saint Girons 1952, 1957a,b). For such an animal, most of the
foraging effort occurs over a period of years prior to the year of reproductive output.
If we evaluate costs of reproduction in such an organism by comparing females in
“reproductive” versus “non-reproductive” years of their cycles, then we completely fail
to assess the kind of reproductive cost (risk, etc., due to additional foraging) that has
been the primary focus of studies on income-breeding species such as most birds.
Any comparison of such costs between a bird and a snake – exactly the kind of
202

comparison that is an aim of life-history theory - would be invalidated if we measured
different components of cost in the two taxa.
One interesting aspect of the bird-snake comparison is that the validity of such
a comparison will depend on the currency in which costs are to be measured. If the
currency is energetics, then it may be meaningful to measure total energy allocation
to the reproductive event in both taxa, and to measure this trait over the period of
expenditure only. Despite the fact that the bird may have gathered the energy over a
period of weeks and the snake over a period of years, their allocation of energy to
reproduction is still directly comparable. Thus, for example, one could compare total
reproductive allocation to body size between these two taxa. Unfortunately, the
comparison breaks down as soon as we try to compare energy allocation to
reproduction versus to other activities (maintenance, growth, etc.). At this point, the
timescale becomes important – and for the reptile, the appropriate timescale is surely
that over which these resources have been gathered, not just the “reproductive” year.
The problem is even worse for survival rates, the other main potential currency in
which costs can be assessed (and probably, the most important such currency for
many kinds of animals –Shine & Schwarzkopf 1992). By analogy with the bird, the
real survival cost of reproduction for a female viper involves the risks that she takes
throughout the “non-reproductive” (energy acquisition) years as well as her risk
during reproduction itself.
Unfortunately for the logistics of assessing costs of reproduction, the vast
majority of living species probably depend to a significant degree on stored reserves
to fuel reproductive expenditure. This characteristic is particularly common in
ectothermic species, for several reasons (Pough 1980; Bonnet et al. 1998).
Ectotherms may also differ from endotherms in the timescale over which costs are
expressed after the overt reproductive expenditure. Because endotherms (especially
203

irds) are under strong energetic constraints, even a brief period of unfavourable
energy balance may be fatal (e.g., Pough 1980). In contrast, the low metabolic
requirements of ectotherms mean that any effect of reproductive expenditure on
survival may not be apparent for a much longer period. For example, a bird that
compromises its energy reserves or thermoregulatory efficiency due to reproductive
costs may thereby die the following winter (e.g., McCleery et al. 1996; Daan et al.
1996; Nilsson & Svensson 1996) whereas a reptile in the same situation can simply
hibernate (which requires very little energy expenditure: Gregory 1982) over the
entire winter period, and not have to face the consequences of its reduced energy
reserves until the following spring. We stress, however that there is a continuum of
timescales, and that some birds will resemble some reptiles in important respects.
The difference is one of degree, but nonetheless may often be so substantial that we
need studies on “costs” experienced by both kind of organisms. Even superficially
similar phenomena in different types of organisms may differ in important respects.
For example, high mortality in some passerines and small mammals in the year
following breeding (i.e. Gustafsson & Sutherland 1988) is not directly comparable to
the delayed post-reproductive decrease of survival in snakes. The “delayed survival
costs” paid by small endotherms occur after several reproductive episodes (3 to 5
clutches [litters] per year on average); but after a single reproductive episode in the
snakes.
Previous research on costs of reproduction in reptiles has concentrated
primarily on events during the actual period of reproductive expenditure: for example,
the decrease in survival rates, food intake and mobility of gravid females (e.g., Shine
1980; Seigel et al. 1987; Madsen & Shine 1993). Our six-year mark-recapture study
of free-ranging vipers allows us to document these costs over a longer timespan (i.e.,
the female’s entire reproductive cycle, not simply the year in which she produces
204

offspring). We examined the data to calculate the proportion of the total survival
costs of reproduction that accrue during different phases of the female cycle. If the
probability of survival is very high during “non-reproductive” years, and very low
during “reproductive” years, then estimates based only on the latter timeframe may
nonetheless provide a reasonable index of overall survival costs. However, if survival
rates are low during other phases of the cycle, then measurements restricted to
"reproductive" years may substantially underestimate the true costs of reproduction
(Jönsson et al. 1995a,b).
If we can quantify survival rates at each stage of the female's cycle, we can
compare the magnitude of pre-breeding, breeding and post-breeding components of
the overall cost of reproduction. This comparison would be impossible in an income
breeder, because the costs of energy acquisition and offspring production occur
simultaneously, whereas they are separated temporally in a capital breeder. The
relative magnitude of post-breeding costs is also likely to differ in consistent ways
between ectotherms and endotherms, because the high metabolic rates of the latter
group mean that over-depletion of energy reserves during reproduction is likely to
cause death rapidly. Such an effect may well be postponed for a very long period in
an organism with lower metabolic needs, such as a viperid snake. To evaluate
whether or not reproducing vipers pay long-term costs in survival, we can use the
comparison among years of a female’s cycle (above). A higher rate of mortality in
the year immediately after parturition would support the notion of “delayed” survival
costs. Also, we can compare survival rates in that post-partum year to a female’s
body condition immediately after she has reproduced in the preceding year.
205

Materials and Methods
Aspic vipers (Vipera aspis) are medium-sized (average adult = 48.5 cm snout-vent
length, 85.5 g) venomous snakes widely distributed through Europe. We studied a
population in central western France (Les Moutiers en Retz), in a mosaic of
meadowland and thicker vegetation. The snakes were hand-captured and
individually-marked (scale clipping or, later in the study [1993] with electronic tags,
sterile transponder TX 1400L, Rhône Mérieux, Destron/IDI INC). Recapture rates
were high and emigration was extremely rare, because the snakes are very
sedentary (Naulleau et al. 1996) and the 33-hectare study area is bounded by habitat
unsuitable for this species. Thus, snakes that disappeared had almost certainly died
rather than emigrated. To ensure that the lower catchability of non-reproductive
females relative to reproductive females did not falsify our results, we waited at least
two years to classify a given female as dead or not (and thus we did not score
survival of females caught in 1996 or 1997).
Further details on the study area and our methods are given elsewhere
(Bonnet & Naulleau 1996; Naulleau & Bonnet 1996; Naulleau et al. 1996; Bonnet et
al. 2000b). Female vipers in this population typically reproduce with a two to three
year cycle, although many females do not live long enough to produce more than a
single litter (Naulleau et al. in prep.). Litters consist of 1 to 13 large (17.9 ± 1.2 cm
SVL, 6.3 ± 1.1 g) neonates. Females with a body size greater than the minimal size
at which parturition has been recorded (41.5 cm SVL, 47 cm total length) were
considered as adult.
206

For the analysis of survival rates in each year of the reproductive cycle, we
had to classify females with respect to their stage of the cycle. This procedure was
straightforward for reproductive females, and for non-reproductive females one to
four years after reproduction (post-reproductive), but more problematical for females
that were pre-reproductive – i.e., those that were in the years prior to their first
“intended” litter. The individuals allocated to this category were those that we caught
one to four years before they first reproduced. In order to qualify as pre-reproductive,
the animals had to be in relatively good body condition at the first capture (indicating
that they were not post-parturient: e.g. absence of flaccid abdomen or extensive skin
folds), and they needed to have been regularly recaptured (so that we were sure that
they did not produce a litter during this period). Some of them were first caught as
juveniles (based on their small size) and later recaptured after they had attained adult
body size but before their first reproduction. In practice, the maternal body-condition
threshold for breeding in this species is so consistent (Naulleau & Bonnet 1996) that
it was possible to classify such females with confidence. Females increase steadily
in condition throughout the “non-reproductive” years of their cycle (Figure 1). Each
female was represented only once in the analyses. In total, data on 527 adult
females was used in the following analyses.
207

MATERNAL BODY CONDITION
0.3
0.2
0.1
0.0
-0.1
-0.2
-0.3
VIT
1
GEST
PART
2 2
YEAR 1 YEAR 2 YEAR 3 YEAR 4
VIT2
GEST2
3 - YEARS CYCLE
2 - YEARS CYCLE
Figure 1. Patterns of change in maternal body condition (residual score from the general linear
regression of ln-transformed mass versus snout-vent length) over the course of the reproductive cycle
in female vipers, Vipera aspis. Females lose considerable body condition at parturition (because of
the mass of the litter) and must gradually recover this condition over a period of two to four years
before they can reproduce again. The Figure provides data from females reproducing on a two-year
(circles and black lines) or three-year (triangles and grey lines) cycle length. “Vit” = during the period
of vitellogenesis; “gest” = during gestation; “part” = after parturition. Arrows indicate the onset of
vitellogenesis in a given year (1 = first reproduction, 2 = second reproduction).
Results
Figure 2 shows that survival rates varied significantly over the course of the female’s
reproductive cycle (χ 2 = 35.2, 3 df, p< 0.0001, all sample sizes indicated in Figure 2).
Female vipers experienced very high mortality (46%) in years when they
"reproduced" (i.e., initiated vitellogenesis and [if they survived] produced offspring).
This rate of mortality was significantly higher than that exhibited by females at other
208
PART2
VIT3
GEST3
PART3

stages of their reproductive cycles (versus pre-reproductive females: χ 2 = 30.7, 1 df,
p < 0.0001; versus females one year after parturition: χ 2 = 4.2, 1 df, p= 0.041; versus
females two years after reproduction: χ 2 = 6.9 [Yates correction], 1 df, p= 0.01).
Mortality rates were also high in the year following parturition (32.5% of snakes died),
but were relatively low in other years of the cycle (e.g., annual mortality rate was only
16% two years after parturition). Survival was particularly high (80%) for females in
the year immediately preceding reproduction, when they were in very good body
condition (Figure 2).
PROBABILITY
OF SURVIVAL
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
188
234
80
BEFORE DURING + 1 YEAR + 2 YEARS
REPRODUCTIVE STAGE
Figure 2. Survival rates of adult female vipers (Vipera aspis) in each year of their reproductive cycle.
“Before” = pre-reproductive females; “during” refers to the year when offspring are produced; “+ 1
year” means the year following litter production; “+ 2 years” means the subsequent 12-month period.
Sample size are indicated above each bar. See text for explanation of criteria used, and statistical
tests on these data.
209
25

We can use these data to estimate the relative magnitude of each component of cost.
The total survival cost of reproduction for a female aspic viper can be divided into
three components:
(1) Pre-breeding cost. - This is the decrease in survival probability caused by the
female delaying reproduction past the time when she has attained adult body size.
This delay is clearly used to build up energy reserves (Figure 1); a female that was
an income breeder would not need to delay for this additional year, and so would not
pay this cost. Females in this phase comprise two of the groups in Figure 2: pre-
reproductive animals, and females two years post-partum. For both groups, annual
survival rates were approximately 82%. Assuming for simplicity that females differ
only in the length of their cycle, the pre-breeding cost averaged an additional 18%
probability of mortality for a female viper with a three-year reproductive cycle. For a
female with a four-year cycle (also common in our study population), this component
of cost is paid in two successive years as energy stores are laid down. The total
additional risk for such a female is thus 33% (= 1.0 - [0.82 X 0.82]). Much of this
additional mortality may not be a direct consequence of reproductive-related activities
such as increased foraging effort, but may be due to random mortality that also
affects immature individuals and adult males. Nonetheless, the mortality is
experienced because of the need to delay reproduction until females reach the
reproductive threshold, and thus can legitimately be considered as a “cost” of this
delay.
(2) Breeding cost. - In the year that they initiated vitellogenesis, females experienced
an annual survival rate of only 54%. Thus, the cost of the activities directly
associated with offspring production (e.g., mating, gestation, parturition) averaged
46%. As above, we note that some component of this mortality risk may be unrelated
210

to reproductive activities, but nonetheless comprises part of the “costs” that are paid
during the reproductive year.
(3) Post-breeding cost. - Females experienced high mortality in the year immediately
following parturition (Figure 2; 67% survival, = 33% cost: note above caveat).
Survival rates of female vipers were lower in the immediately post-parturient year
than in other “non-reproductive” years (Figure 2; comparing survival rates in the
years immediately preceding versus following the “reproductive” year: χ 2 = 4.7, 1 df,
p=0.031). This difference supports the notion that there are mortality risks
associated with reproducing, that are not manifested until long after the litter is
produced.
To further test this proposition, we can compare a female’s probability of
survival in that post-parturient year, to her body condition (residual score from the
linear regression of ln-transformed mass to SVL) immediately following parturition in
the preceding year. We used logistic regression for this analysis, because the
dependent variable (survived versus died) is a dichotomous trait. As predicted, a
female’s probability of survival in the year after she reproduced was significantly
higher if she was in relatively better body condition immediately after giving birth the
year before (χ 2 = 5.8, 1 df, p= 0.016; Figure 3). This significant association supports
the interpretation that mortality in the year after litter production constitutes a delayed
(post-breeding) cost of reproduction for female vipers.
211

SURVIVAL
1
0
- 0.4 - 0.3 - 0.2 - 0.1 0.0
BODY CONDITION
Figure 3. Logistic regression of a female viper’s probability of survival in the twelve months following
her production of a litter, as a function of her post-parturient body condition. Body condition was
calculated as the residual score from the general linear regression of ln-transformed mass versus
snout-vent length for all females within the population; almost all values are negative because post-
parturient females are always in much poorer body condition than are other (pre-reproductive)
females. See text for explanation and statistical results.
Discussion
The central result from our analysis is a very straightforward one: the timescale over
which we measure costs of reproduction needs to reflect the timescale over which an
animal engages in activities that support that reproductive bout. In capital-breeding
species, that timescale may well be very much greater than the actual period over
which overt “reproduction” (production of offspring) occurs. The extended timescale
reflects two factors: a longer pre-reproductive period of energy-gathering, and a
longer post-reproductive period when effects of reproductive activities are
manifested. Interspecific comparisons based on shorter timescales are likely to be
212

misleading if they compare different components of reproductive cost of one kind of
organism (e.g., an avian income-breeder) versus the other (a reptilian capital-
breeder).
Previous theoretical treatments have identified the importance of this
distinction between pre- and post-breeding costs of reproduction for understanding
the evolution of reproductive tactics (Sibly & Callow 1984; Stearns 1992; Jönsson et
al. 1995a,b; Jönsson 1997). Our data provide strong empirical support for the
assumptions that underpin these models, especially those proposed by Jönsson et
al. (1995a,b). Thus, our data support the idea that optimal reproductive investment
should increase when costs experienced late in the reproductive cycle (breeding plus
post-breeding costs) are higher than pre-breeding costs. This situation is exactly the
one that we have found in the asp viper (see above results and Bonnet et al. 1994).
Our results also allow us to develop this idea further. The evolution of semelparity
can be viewed as a consequence of extreme capital breeding tactics, where most of
the maternal somatic resources are invested during a single reproductive bout
(Bonnet et al. 1998). Semelparity may be favoured when the sum of survival costs
measured over a long timescale (3 to 4 years on average in our study model) are
particularly high. This extreme reproductive tactic, observed almost exclusively in
ecthotherms, may also be associated with components of reproductive costs that are
independent of fecundity (Bull & Shine 1979; Olson et al. 2000).
In the case of female aspic vipers, costs of reproduction are so high that most
females produce only a single litter during their lifetimes. Especially if energy
acquisition is required over a period of two years rather than one, the annual survival
rates of females are so low (Figure 2) that few females survive long enough to
produce a second litter. Although the survival cost in the year of litter production is <
50%, the additional risks due to pre-breeding mortality (18 to 33%, depending on
213

cycle length) and post-breeding mortality (33%) combine to make semelparity the
norm for female vipers in our study population (Naulleau et al., in prep.). This result
emphasises the importance of understanding all components of reproductive costs,
not simply those that are paid during the actual "reproductive" bout. In our study
animals, these additional components (mostly post-breeding costs) sum to at least as
high a cost as the overt mortality risk experienced by a female in the year in which
she produces offspring.
More generally, methodologies for measurement of costs need to be
evaluated carefully before comparisons can be made. This caveat extends to
particular techniques as well as to timescales. For example, the popular technique of
assessing avian costs through clutch-size manipulation after laying does not
incorporate any effects of the additional clutch size on maternal mobility prior to
laying. Such effects may well occur in birds (e.g., Lee et al. 1996), and are believed
to be an important component of the total costs experienced by some reptiles (e.g.,
Shine 1980; Seigel et al. 1987; Sinervo & DeNardo 1996).
Similarly, our logistic regression detected a significant mortality cost of
reproduction associated with maternal body condition after parturition (see above,
and Figure 3). Post-parturient condition also affects maternal survival in two other
species of snakes that are partly sympatric with aspic vipers, but in both cases the
correlation is apparent in the few months following parturition (Vipera berus –
Madsen & Shine 1993; Coronella austriaca – Luiselli et al. 1996). No such link is
apparent within aspic vipers over that period (post-partum body condition did not
affect a female viper’s probability of survival to the next season; N = 93, p = 0.31 -
Naulleau et al. in prep.): the survival cost of lowered maternal condition is only seen
over the ensuing 12 months. This contrast suggests that even when species display
214

similar relationships between reproductive effort and cost, the taxa may differ in the
timescale over which such effects are manifested.
These kinds of complications do not invalidate broad-scale comparisons of
costs: indeed, we enthusiastically endorse attempts to do so. There will be many
species that are phylogenetically distant from each other, but for which the form and
timescale of costs are sufficiently similar that comparisons are relatively
straightforward. For example, although ectothermy predisposes animals to capital-
breeding, there are many income-breeders within this group also (e.g., short-lived
lizards, James & Whitford 1994). Similarly, some endotherms rely upon stored
“capital” for reproduction (e.g., Cherel et al. 1993; Cherel 1995) and some
endotherms experience relatively delayed, long-term survival costs (McCleery et al.
1996, Daan et al. 1996). Thus, there are many opportunities to carry out appropriate
comparisons among suitably-matched groups of species. Many of the comparisons
that we have made (such as capital versus income, or timescales for energy
acquisition in birds versus snakes) are clearly continuous rather than dichotomies.
Future research could usefully quantify such timescales.
Given the logistical difficulties of the kind discussed here, however, it may also
be worth investigating the massive potential of intrageneric (and even, intraspecific)
comparisons to clarify costs of alternative life-history traits. For example, many
reptile lineages display variation in traits (such as mean body sizes, degrees of
sexual size dimorphism, reproductive mode) at these levels (Fitch 1981; Blackburn
1982, 1985). This diversity, among taxa that are otherwise very similar, offers a
particularly powerful opportunity to characterise and quantify the costs associated
with phylogenetic shifts in traits of interest. Ultimately, such comparisons may be
more revealing than those made between taxa that differ so substantially in the form
215

and timescale of costs that it is difficult to overcome the confounding variables
involved.
Acknowledgements
We thank S. Duret, M. Vacher-Vallas and L. Patard for help during field work.
Financial support was provided by the Conseil Général des Deux Sèvres, the Centre
National de la Recherche Scientifique (France) and the Australian Research Council.
We also thank Rex Cambag for maintenance of the electronic material.
216

IV. Les déterminants de la
tendance semélipare:
description et implications
démographiques
217

A. Résumé du Chapitre:
Chez la vipère aspic, l’effort de reproduction est associé à des coûts très élevés
s’exprimant sur un échelle de temps complexe. Dans notre population d’étude les
contraintes climatiques et énergétiques vont ralentir les possibilités de reconstitution
des réserves corporelles. Ces contraintes vont affecter la survie des reproductrices et
limiter le nombre d’opportunités de reproduction avec une majorité de femelles
semélipares. Toutefois certaines femelles réussissent à se reproduire plusieurs fois.
La population des Moutiers est donc constituée une fraction d’individus
“ semélipares ” en coexistence avec une minorité d’individus “ itéropares ”. Une telle
situation offre une excellente opportunité pour examiner les déterminants de ces
stratégies démographique contrastées.
Dans un premier temps, nous avons cherché à tester les avantages sélectifs
des modes de reproduction semélipares versus itéropares dans cette population.
Nous avons voulu examiner si la “stratégie semélipare” était associée à un
investissement reproducteur plus élevé illustrant ainsi la concentration de l’effort
reproducteur maternelle sur une opportunité unique de reproduction (article 8). Nos
résultats vont clairement à l’encontre de cette hypothèse : les femelles semélipares
produisent des portées moins lourdes constituées de vipéreaux moins nombreux et
plus légers que les femelles itéropares. La “stratégie” itéropare est donc la plus
avantageuse en assurant un meilleur succès reproducteur à vie. Cependant elle
reste minoritaire ce qui suggère l’existence de fortes contraintes sur l’expression de
l’itéroparité. La comparaison des caractéristiques des femelles semélipares et
itéropares suggère une origine environnementale de la seméliparité. En effet, les
individus des deux stratégies ne diffèrent pas en taille, masse ou condition corporelle
initiale. En revanche on détecte des différences significatives de condition pré/post
218

parturition en faveur des itéropares. Le meilleur succès de ces femelles semblent
donc directement lié à des conditions énergétiques favorables (prises alimentaires
facultatives) qui vont influencer les caractéristiques de la portée. Cet apport
d’énergie va aussi améliorer la condition de la mère après la mise bas et favoriser les
reproductions futures.
De plus, nos données soulignent l’existence de forte variations interannuelles
dans la proportion d’individus semélipares. Ces variations sont étroitement corrélées
aux fluctuations en nourriture observées dans l’année faisant suite à la reproduction
et donc au conditions trophiques pendant la phase cruciale de récupération. Une
dernière analyse nous permet de modéliser la probabilité d’être semélipare dans la
population. L’expression de la seméliparité dépend ainsi de contraintes énergétiques
intimement liées que sont l’état d’émaciation post mise bas (qui reflète le niveau
d’alimentation et les conditions thermiques l’année de la reproduction) et les
conditions trophiques l’année faisant suite à la reproduction (phase de récupération).
Ces résultats invalident donc l’existence d’une “stratégie” semélipare au sens propre
dans la population. La tendance semélipare reflète en fait un système
d’investissement reproducteur extrême associé à des coûts post-reproducteurs dont
l’amplitude est élevée et largement contrôlée par des facteurs environnementaux
fluctuants. Les variations d’amplitude des coûts illustrent un système où les
possibilités de récupération énergétiques sont très précaires et déterminantes de
l’espérance de vie reproductrice des femelles.
Dans un dernier volet, nous avons examiné l’impact populationnel de la
stratégie reproductrice femelle en comparant la démographie des deux sexes (article
9). La production de vipéreaux impose des contraintes et des contributions très
différentes selon les sexes. L’investissement des mâles est réduit dans le temps et
réclame moins d’énergie que celui des femelles. Mâles et femelles partagent la
219

même niche écologique (même type de proie, habitat similaire), cette situation offre
un bonne opportunité pour examiner les conséquences démographiques de
divergences éco-physiologiques entre les sexes sans avoir à tenir compte des
autres facteurs.
Nos résultats montrent l’existence d’un système démographique hautement
dimorphe avec d’importantes fluctuations dans la taille de la population femelle qui
contrastent fortement avec une certaine stabilité des mâles. Les variations dans la
population femelle reflètent les contraintes reproductrices maternelles (vitellogénèse,
gestation) et illustrent le pas de temps pertinent sur le lequel les coûts de la
reproduction se manifestent. Ainsi, la dynamique de la population femelle est
intimement liée aux fluctuations en proies qui influencent positivement le recrutement
de nouveaux adultes et qui affecte l’amplitude des coûts payés pendant les phases
de récupération. Chez les mâles le système d’allocation de l’énergie est graduel et
on n’observe pas de fluctuation majeure dans la taille de la population. Les mâles
peuvent ajuster leur effort reproducteur en fonction de leur état physiologique, par
exemple en arrêtant de chercher les femelles si leur condition corporelle est trop
faible. Par contre, une fois engagées, les femelles doivent poursuivre leur effort de
reproduction jusqu’aux mises bas au risque de tout perdre. Cette dichotomie dans les
contraintes reproductrices se traduit par des patterns démographiques contrastés.
220

B. Article 8
Y-a-t’il un avantage à être semélipare?
Comparaisons des tactiques démographiques
de la vipère aspic (Vipera aspis)
O. Lourdais 1 2 3 , X. Bonnet 1 , R. Shine 4 , & G Naulleau 1
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 Biological Sciences A08, University of Sydney, NSW 2006, Australia
Manuscrit en préparation...
221

Résumé
Chez la vipère aspic (Vipera aspis) les populations situées au Nord de l’aire de
répartition présentent une majorité (72.5%) de femelles semélipares. Toutefois, une
minorité de femelles (27.5%) sont itéropares et se reproduisent entre 2 et 4 fois. De
façon surprenante l’effort de reproduction n’est pas supérieur chez ces individus
semélipares. Au contraire, ces femelles semblent moins bien réussir leur
reproduction que les femelles itéropares qui produisent des portées plus lourdes
constituées de vipéreaux dont le poid moyen est plus élevé. En se reproduisant de
façon répétée, les individus itéropares produisent un nombre supérieur de jeunes et
bénéficient d’un meilleur succès reproducteur.
Nos données indiquent qu’il n’existe pas deux “stratégies” au sens propre
dans la population : la minorité d’itéropares est en fait composée d’indivividus ayant
bénéficié d’un apport énergétique supérieur pendant la reproduction. Ces individus
jouissent d’une meilleure condition coprorelle post-partum, une variable déterminante
des probabilités de survie et de reproductions ultérieures. L’origine environnementale
de la seméliparité est par ailleurs confirmée par l’étude des variations inter-annuelles
dans la proportions de semélipares. Ainsi, pour une année donnée, la fraction
d’individus semélipares est fortement dépendante des conditions d’alimentation de
l’année faisant suite à la reproduction. Le fort degré de seméliparité observé dans
cette population illustre les faibles possibilités de récupération énergetique après la
mise bas et s’explique donc par des facteurs environnementaux dévaforables et/ou
fluctuants.
Mots clés: seméliparité, itéroparité, succès reproducteur
222

Introduction
L’extraordinaire diversité dans les stratégies reproductrices déployées par les êtres
vivants reflète l’existence de contraintes environnementales. Dans un contexte
limitant, les êtres vivants doivent ainsi faire face à des compromis ou “trade-off” entre
des activités ou voies concurentielles que sont croissance, reproduction et survie
(Stearns 92). Les différentes combinaisons entre les traits d’histoire de vie sont
généralement interprétées comme autant de réponses évolutives permettant aux
organismes d’optimiser leur succès reproducteur à vie et donc de “répandre”
efficacement leurs gènes au sein des populations. Les organismes sont ainsi soumis
à des pressions de sélection qui favorisent l’évolution d’un effort reproducteur
optimum (Stearns 1992, Roff 1992). Cette répartition optimale de l’effort reproducteur
est le résultat d’un compromis entre le succès attendu dans la reproduction courante
et celui des reproduction futures.
L’existence d’une dichotomie entre les organismes semélipares et d’autres
itéropares constitue un des contrastes les plus surprenants dans l’orientation des
stratégies reproductrices (Cole 1954). Alors que chez les premiers la vie
reproductrice est constituée d’une succession d’épisodes reproducteurs, chez les
seconds la reproduction est suivie de la mort de l’organisme. Les avantages de ces
différentes stratégies ont été largement discutés par le passé (Cole 1954, Stearns
1992). De nombreux modèles mathématiques ont été développés pour comprendre
l’évolution de la seméliparité (Cole 1954; Gadgil & Bossert 1971; Charnov & Schaffer
1973). Les travaux les plus récents reposent sur la confrontation de stratégies
semélipares et itéropares définies par des paramètres démographiques (Young
1990; Young & Auspurger 1991; Ranta et al. 2000a,b). Dans tous les cas ces
modèles permettent l’existence d’un mode de reproduction caracterisé par une mort
223

programmée. Seméliparité et itéroparité sont alors considerées comme des
stratégies pré-existentes, entrant en compétition au sein de populations théoriques
(Ranta et al. 2000a,b). Les résultats obtenus suggèrent l’influence de paramètres
démographiques et environnementaux sur la sélection de l’une ou l’autre des
stratégies. Ces travaux ont apporté des informations précises sur l’influence relative
des taux de mortalités adulte et juvénile sur les avantages de l’une ou l’autre des
stratégies. Ainsi, dans certaines circonstances bien définies, un “mutant” semélipare
pourrait envahir la population et vice-versa (Ranta et al. 2000a,b). Cependant, les
modèles élaborés reposent tous sur un déterminisme très simple de la partition de
l’effort reproducteur (itéropare versus semélipare) en considérant l’existence d’une
stratégie “résidente” confrontée à l’invasion d’une stratégie mutante. Il est important
de signaler l’absence de support empirique de telle situation dans la nature.
Notamment, la détermination de l’effort reproducteur et de sa partition dans la vie de
l’organisme fait intervenir des soubassements physiologiques complexes reposant
sur des actions génétiques variées (Sinervo & Svensson 1998). Un tel contexte rend
donc peu réaliste les postulats d’un codage simplifié du mode de parité, implicite à la
construction des modèles.
L’étude de l’évolution du mode de parité et les traits d’histoire de vie associés
a donc fait l’objet de nombreuses approches théoriques, en dehors de tout support
phylogénétique. Si cette approche a permis de mieux comprendre les avantages et
inconvénients de la seméliparité, la nature des pressions évolutives responsables de
transition entre l’itéroparité et la seméliparité demeurent très largement inconnue
(Crespi & Teo 2002). Plus récemment, quelques travaux de comparaisons
interspécifiques ont été entrepris dans un cadre phylogénétique robuste. Ces
approches ont été très fructueuses en révélant à la fois chez les plantes et les
animaux l’existence d’un continuum, et non d’une dichotomie, entre seméliparité et
224

itéroparité (Hautekèete et al. 2001; Crespi & Teo 2002). De plus, en comparant des
espèces réparties sur un gradient entre des itéropares longévifs et des semélipares
strictes, ces travaux apportent des informations précises sur les changements dans
les traits d’histoire de vie associés à une transition vers la seméliparité. Au sein d’un
tel gradient, les espèces itéropares présentant des populations à tendance
semélipares, constituent des modèles d’études particulièrement intéressants. Par
tendance semélipare on considère les espèces dont le mode de reproduction est
l’itéroparité mais où de nombreux individus ne se reproduirent qu’une seule fois et
présentent ainsi une trajectoire semélipare. En occupant une position charnière, ces
situations rendent possible le test d’hypothèses sur les déterminants de la répartition
de l’effort reproducteur et sur les possibles contraintes environnementales impliquées
dans la transition vers des systèmes semélipares. En effet, une telle approche n’est
pas envisageable chez les espèces semélipares strictes chez qui la mort des
organismes est programmée après la reproduction.
De telles opportunités existent chez certaines espèces de reptiles et
notamment parmi les squamates (Madsen & Shine 1992b, Brown 1993, Bonnet et al.
2002a). Par exemple, chez la vipère aspic, il existe une tendance semélipare
marquée dans les populations située au Nord de l’aire de répartition (Bonnet et al.
2002b) alors que dans les populations du Sud, les femelles sont longévives et
itéropares (Zuffi et al. 1999). Nous avons exploité les données d’un suivi à long terme
d’une population de vipères à tendance semélipare en examinant les questions
suivantes :
1) Les individus semélipares diffèrent-ils des itéropares en terme d’effort
reproducteur. Notamment, la seméliparité est-elle associée à un
investissement plus massif dans la reproduction ?
225

2) Existe-t’il une variabilité inter-annuelle dans la proportion d’individus
semélipares. Si oui, quelle est la nature des facteurs déterminants de
l’expression de la tendance semélipare au sein de cette population.
Matériels et méthodes
Zone et espèce d’étude
La vipère aspic est un serpent venimeux de taille moyenne (55 cm). Cette espèce
présente un système d’allocation de l’énergie basé sur l’accumulation d’importantes
quantités de réserves lipidiques qui servent de support à la reproduction (Bonnet
1996). Nous avons étudié une population de vipères aspic dans l’Ouest de la France
(Les Moutiers en Retz, 47 o 03N'; 02 o 00W') à proximité de la limite Nord de l’aire de
distribution de l’espèce. Dans cette région, une à quatre années sont nécessaires
pour accumuler des stocks de réserves suffisants pour la reproduction et les femelles
présentent typiquement un cycle reproducteur supérieur à un an. L’investissement
dans la reproduction, très élevé pour un vertébré, est associé à de forts coûts
écologiques et énergétiques. Le nombre d’épisodes reproducteurs est réduit et il
existe une tendance marquée vers la seméliparité (Bonnet et al 2002a). Cette
situation contraste fortement avec les populations situées plus au Sud où les
femelles sont longévives et présentent un cycle reproducteur annuel (Zuffi et al.
1999).
Les adultes se nourrissent de micromammifères et essentiellement de
campagnol (Microtus arvalis Pallas) dont les populations présentent d’importantes
fluctuations inter-annuelles. Plusieurs travaux suggèrent une influence complexe de
ces variations trophiques sur la dynamique des réserves et l’investissement
reproducteur (Bonnet et al 2001b, Lourdais et al. 2002b). Le climat est de type
226

océanique tempéré et avec des variations des conditions thermiques significatives
(Lourdais et al. 2002b) En combinaison avec la ressource trophique les variations
climatiques affectent de nombreux aspect du cycle de vie des vipères. Notamment
les conditions thermiques pendant la gestation semblent contraignantes et les mises-
bas sont observées fin septembre soit deux mois plus tard que dans les populations
méridionales. De 1992 à 2000 la zone d’étude a été patrouillée par 1 à 4 personnes
pendant la saison active (Mars à Avril). Les serpents observés pendant les phases
de thermorégulation ont été capturés à vue, sexés, marqués individuellement et
relachés sur le lieu exact de capture. Les indices de consommation de proies ont été
relevés pour permettre l’estimation de l’abondance en campagnols. Des informations
détaillées sur les méthodes de marquage et l’effort de recherche sont diponibles
dans des différents travaux publiés sur cette même population (Bonnet et al. 2001b,
2002a, Lourdais et al. 2002b)
Mesures realisées
Les analyses de cet article reposent sur un jeu de données de 2200 captures
provenant de 524 femelles adultes différentes. Tous les serpents capturés ont été
pesés au gramme près et mesurées à 0.5 cm près. Le statut reproducteur a été
déterminé en utilisant plusieurs méthodes complémentaires : au début du printemps,
les femelles présentant une condition corporelle supérieure au seuil sont considérées
comme reproductrices (voir Bonnet et Naulleau 1994 pour les détails sur la
méthode). Plus tard dans l’année, la palpation de l’abdomen a permis de détecter et
compter des follicules en croissance (vitellogénèse) ou des embryons (gestation).
Les femelles reproductrices capturées peu de temps avant la mise bas (fin de l’été)
ont été ramenées en captivité afin d’obtenir des informations précises sur la
reproduction (voir section Méthodes d’étude p 37).
227

Estimation de la survie
Les individus non-recapturés au-delà d’une période de deux ans sont considérés
comme morts. Cette méthode repose sur une série d’arguments complémentaires :
i) Le site d’étude est entouré par des zones défavorables aux vipères qui réduisent
les possibilités d’émigration (Vacher-Vallas, Bonnet & Naulleau 1999). ii) Cette
espèce est très philopatrique (Naulleau,et al. 1996) et les données de radiotracking
ont révélé des déplacements très réduits par rapport au lieu de capture initiale
(toujours

du suivi (9 ans d’études) et la forte tendance semélipare dans la population. Le
succès reproducteur à vie des vipères a été estimé en condidérant les individus dont
l’histoire reproductrice est connue et dont l’ensemble des mises bas a été obtenu en
captivité. Enfin, la proportion annuelle d’individus semélipares et itéropares a été
calculée de 1992 à 1998.
Une fois la classification des femelles réalisée, nous avons cherché à
comparer les deux stratégies en examinant notamment les traits morphologiques
maternels (taille, condition corporelle) et les caractéristiques des portées (pour les
femelles recapturées avant la mise bas). Pour réaliser ces comparaisons, nous
avons uniquement considéré les données de première reproduction des itéropares et
il n’y a donc pas de réplication de données pour un même individu. Nous avons par
ailleurs comparé les dynamiques pondérales chez les femelles itéropares et
semélipares depuis le printemps jusqu’à la mise bas. La population d’étude est
affectée par de nombreuses variables environnementales (Bonnet et al. 2001b
Lourdais et al. 2002b) et les comparaisons des stratégies reproductrices ont été
réalisées en tenant compte de la variabilité inter-annuelle. Ainsi l’étude d’un trait
donné (variable dépendante) a été effectuée en considérant comme facteurs 1)
l’année de comparaison et 2) la nature de la stratégie (itéropare ou semélipare).
Lorsque le trait étudié est influencé pas la taille maternelle, ce paramètre a été
introduit comme co-facteur. Enfin nous avons testé l’influence de la nourriture sur
l’expression de la stratégie reproductrice en utilisant l’indice d’abondance en
campagnol déjà décrit par ailleurs (section Méthodes d’étude p 37, Lourdais et al.
2002b).
Résultats
1) Estimation de la tendance semélipare
229

Sur les 9 ans d’étude, nous avons pu dénombrer 241 individus semélipares et 91
iteropares. La classe des semélipares comprend majoritairement des animaux se
reproduisant l’année n et jamais revus ensuite (181). Une minorité d’individus (60) a
cependant été observée l’année suivante (année n+1), principalement au printemps
et plus jamais par la suite. Considérant notre incertitude sur l ‘estimation de la survie
(période de 2 ans sans observation), il n’est pas possible d’identifier le moment où
s’exerce la mortalité et donc de définir deux classes d’individus semélipares. Ainsi,
parmis les 181 individus jamais recapturés, de nombreuses femelles ont
probablement survécu jusqu’au printemps suivant en échappant à l’observation (la
capturabilité des non-reproductrices étant toujours inférieure à celle des
reproductrices, Lourdais et al. 2002d). Nos résultats indiquent clairement un forte
tendance semélipare dans la population (72.5 % des individus). De plus, nous avons
détecté des variations inter-annuelles significatives dans la proportion des individus
semélipares et itéropares ( χ² de Pearson=25.16, dl=6, p=0.0003, voir Table 1).
Morphométrie:
2) Les individus semélipares sont-ils différents des itéropares?
Nous avons tout d’abord examiné si les individus itéropares et semélipares diffèrent
dans leur morphologie au sortir de l’hivernage (printemps). Nos analyses suggèrent
l’absence de différences significatives en longueur corporelle initiale et en masse
corporelle initiale (Table 2). En conséquence, il n’existe pas de différence
significative de condition corporelle entre les deux catégories de femelles (analyse de
covariance avec la masse ajustée par la taille, voir Table 2). Les femelles itéropares
et semélipares sont donc indistinguables en début de reproduction en terme de
morphométrie ou d’état des réserves lipidiques (condition corporelle).
230

Table 1. Proportion d’individus itéropares et semélipares en fonction de l’année. Les données
s’arrêtent en 1998 en raison du pas de 2 ans nécessaire pour statuer de la survie de vipères. Dans
cette analyse nous avons considéré la seméliparité et l’itéroparité comme deux strategies sensu
stricto et un même individu itéropare peut contribuer à plusieurs années différentes.
Année Semélipares Itéropares Total Proportion
1992 12 9 21 0.57
1993 59 16 75 0.78
1994 61 39 100 0.61
1995 8 18 26 0.30
1996 26 30 56 0.46
1997 63 42 105 0.60
1998
10
Investissement reproducteur
10
Nous avons ensuite étudié les paramètres de l’effort reproducteur selon la stratégie
reproductrice. En accord avec les précédents travaux réalisés sur cette population,
nos analyses indiquent l’existence de variations inter-annuelles dans la plupart des
paramètres mesurés (voir Table 2). De facon suprenante, les femelles semélipares
produisent des portées significativement moins lourdes que les itéropares (Figure 1,
Table 2). Des résultats similaires sont obtenus en considérant la masse de vipéreaux
viable produite ou la masse relative de la portée (Analyse de covariance avec
comme variable dépendante la masse de la portée et comme covariable la masse
post-partum de la mère). En revanche, nous ne détectons pas de différence
significative dans les tailles de portées ou bien dans le nombre de vipereaux viables
produits (voir Table 2). Enfin l’analyse des caractéristiques des jeunes viables
produits indique que les portées des femelles itéropares sont constituées de jeunes
significativement plus lourds (6.7g) que ceux des femelles semélipares (6.2g, voir
231
20
0.50

Table 2) alors que l’on ne détecte pas de différence dans la taille des nouveau-nés
produits (F(1,79)=1.68; p

Table 2. Suite
Traits facteurs co-facteur F n p Valeurs
iteropare semélipare
Masse AN SVL 8.30 88

Succès reproducteur
Les femelles semélipares ne produisent pas plus de vipéreaux que les femelles
itéropares lors de leur unique mise bas. Les femelles itéropares en se reproduisant
plusieurs fois dans leur vie vont donc avoir un nombre plus élevé de descendants
que les semélipares (10.2 versus 4.3, ANOVA F(1, 73)=40.47, p

masses pre-partum avec des femelles itéropares plus lourdes (masse ajustée :
119.8, versus 107.4; voir Table 2). De même, les femelles itéropares bénéficient
d’une masse corporelle supérieure après la mise bas (masse ajustée: 68.9g versus
63.6 Table 2). Les itéropares présentent donc une condition corporelle post-partum
supérieure aux semélipares (résidus : 0.064 versus -0.015)
Considérant ces éléments, nous avons voulu déterminer l’influence de
l’abondance en proies sur les variations annuelles dans les proportions d’individus
semélipares. Nos résultats suggèrent une absence de relation entre la proportion de
semélipares et l’abondance en proies dans l’année de la reproduction (r=0.40,
r 2 =0.16, n=7, p=0.37). Par contre, il existe une relation négative significative entre la
proportion d’individus semélipares dans une année et l’abondance en nourriture
l’année suivante (r= 0.84, r²=0.71, F(1,5)=12.672 p

Proportion de semelipares (année n)
0.95
0.85
0.75
0.65
0.55
0.45
0.35
0.06 0.12 0.18 0.24 0.30 0.36 0.42
Abondance en proies dans l'année n+1
Figure 3. Relation entre la proportion annuelle de femelles semélipares et l’abondance en proies
l’année suivante.
Discussion
Cette étude apporte des informations paradoxales sur la tendance semélipare de la
vipère aspic. Notre suivi indique que sur 9 ans d’études, cette “stratégie“ est
largement dominante dans la population (72.5% des individus). Pourtant plusieurs
de nos analyses indiquent que ce mode de reproduction est bien moins bénéfique
que l’itéroparité et ce pour plusieurs raisons. Tout d’abord, les femelles semélipares
n’investissent pas plus d’énergie dans la production d’un grand nombre de
vipéreaux. Ainsi il n’existe pas de différence de taille de portée selon les stratégies.
De façon surprenante, la masse des portées est supérieure chez les femelles
itéropares et, par conséquent, ces dernières produisent des vipereaux
significativement plus lourds et donc en meilleur condition corporelle à la naissance.
Ces différences sont importantes car, chez les serpents, la condition corporelle à la
naissance est un indice de l’état des réserves qui influence ultérieurement la
croissance post-natale, le type de proies ingérables et peut-être la survie des
236

jeunes. En dépit de l’absence d’informations sur la survie des nouveaux-nés, nos
résultats suggèrent néanmoins des différences dans la qualité des jeunes produits.
En parallèle avec ces variations de qualité, les femelles itéropares produisent un
nombre de jeunes toujours plus élevé que les semélipares. De plus, il existe une
relation positive très nette entre le nombre de reproduction et le nombre total de
vipéreaux produits.
Dans un tel contexte, il est légitime de s’interroger sur les avantages de la
stratégie semélipare qui demeure pourtant majoritaire dans la population. Nos
analyses indiquent que les individus semélipares ne diffèrent pas des itéropares à
l’entrée dans la reproduction (pas de différence de condition corporelle ou de taille
des portées). Le fait que l’on détecte des différences dans les masses pre et post
partum ainsi que dans les masses des portées et des vipéreaux (Table 2)
suggèrent des prises alimentaires chez les itéropares. Ainsi ces individus semblent
bénéficier d’un apport énergétique supérieur pendant la reproduction (prises de
nourriture facultatives) et par conséquent d’une meilleure condition corporelle post -
partum. La condition corporelle est un très bon indice des possibilités de
récupération ultérieure des femelles vipères et nos données soulignent donc le rôle
majeur des prises alimentaires sur l’état d’émaciation des femelles et les possibilités
de reproduction. En plus de l’état d’émaciation, il existe une relation très forte entre
la proportion annuelle d’individus semélipares et la quantité de nourriture disponible
l’année suivante (i.e. pendant la phase de récupération). Ces résultats démontrent
donc une origine profondément environnementale de la tendance semélipare. Ainsi,
il n’existe pas de stratégies semélipare et itéropare pures et la seméliparité apparaît
plutôt comme le reflet des fortes contraintes énergétiques de la reproduction. Chez
la vipère aspic l’investissement reproducteur est très élevé et basé sur la
mobilisation des réserves lipidiques et protéiques. Les changements
237

comportementaux et morphologiques (thermorégulation accentuée, déplacements
réduits) associés à la reproduction aboutissent à une réduction des prises
alimentaires, notamment pendant la gestation (Lourdais et al. 2002a). Dans cette
population, il existe de très fortes variations inter-annuelles dans l’abondance des
proies qui vont directement influencer le degré de prises alimentaires facultatives
l’année de la reproduction (Lourdais et al. 2002b). La fraction d’individus itéropare
correspond donc à une minorité d’individus (chanceux) qui ont pu bénéficier de
prises alimentaires un peu plus nombreuses. Les différences de masse pre-partum
(107.4g vs 119.4g) suggèrent un apport énergétique “réduit” probablement lié à la
consommation de une, peut-être deux proies en plus. Ces quelques prises
alimentaires semblent néanmoins exercer une influence déterminante sur l’état
d’amaigrissement et la survie post mise bas. Outre les prises alimentaires l’année
de la reproduction, les conditions trophiques l’année suivante vont avoir une
influence majeure sur la proportion d’individus semélipares. Ainsi cette proportion
sera élevée lorsque l’année de la reproduction est suivi par une année à faible
abondance en proies. En revanche la fraction de semélipares sera plus réduite
lorsque la nourriture est abondante l’année suivante. La probabilité d‘être
semélipare une année est donc affectée par la combinaison de facteurs : la
condition corporelle post-partum et les conditions trophiques de l’année suivant la
reproduction. Des travaux récents (lourdais et al. 2002b) montrent que la condition
corporelle post-partum est elle même une variable complexe qui intègre les prises
alimentaires facultatives mais aussi les conditions thermiques pendant la gestation.
Notamment les coûts métaboliques sont plus élevés lorsque les conditions
climatiques offrent des opportunités de thermorégulation supérieures aux femelles
gestantes. Ainsi, pendant les étés chauds, la durée de gestation est plus courte
mais le niveau de catabolisme protéique supérieur. La probabilité d’être semélipare,
238

intimement liée au niveau d’amaigrissement post-partum, est donc influencée par
des interactions complexes entre trois variables environnementales :
i) Les prises alimentaires pendant la reproduction en combinaison avec ii) les
conditions thermiques de la gestation vont déterminer la condition corporelle post-
partum. iii) L’abondance en proies l’année suivant la reproduction va directement
influencer les possibilités de récupération des femelles.
Ces résultats sont très informatifs pour bien comprendre la stratégie
reproductrice de la vipère aspic. Tout d’abord, la proportion d’individus semélipares
et itéropares est calculée à partir du nombre d’individus observés au printemps
(période principale des captures, Lourdais et al. 2002d). La capturabilité des vipères
est ensuite plus réduite dans l’année et la fraction d’individus recapturés pour les
mises bas ne représente qu’une minorité des individus reproducteurs observés dans
l’année. La mortalité des vipères reproductrices implique des coûts de la
reproduction direct (prédation pendant la vitellogénèse, la gestation) combinés avec
des coûts de post-reproduction (mort liée à une situation énergétique post-partum
critique) (Bonnet et al. 2000a, 2002a). En l’absence d’estimation du niveau de
prédation, ces deux composantes semblent a priori difficiles à départager.
Cependant nos résultats indiquent clairement que les contraintes de la reproduction
génèrent principalement des coûts de post-reproduction. En effet, la vitellogénèse
entraine la mobilisation des réserves. La gestation, en imposant des régimes
thermiques élevés, va entrainer l’épuisement des réserves et un important
catabolisme protéique. Ces contraintes énergétiques s’illustrent par un fort niveau
d’émaciation après la parturition. La relation très forte entre la proportion d’individus
semélipares (une variable démographique) et l’abondance en proies l’année
suivante indique donc que la plupart des individus survivent jusqu’à la mise bas et
239

que la mortalité va s’exprimer pendant la phase de recupération au printemps
suivant (la mortalité pendant l’hivernage étant très réduite, données non publiées).
Notre étude suggère donc qu’il n’existe pas de stratégie semélipare ou
itéropare au sens propre dans la population des Moutiers. La reproduction est
associée à des coûts énergétiques relativement indépendants de la taille des
portées produites (Lourdais et al. 2002a, Ladyman et al. soumis). Le fort
amaigrissement qui en résulte et les faibles possibilités de récupération vont aboutir
à une survie très faible après la mise bas. La tendance semélipare de cette
population reflète donc l’existence de coûts énergétiques et écologiques à la fois
élevés et fixes s’exprimant dans un contexte environnemental fluctuant et
contraignant.
240

C. Article 9
Do sex divergences in reproductive eco-
physiology translate into dimorphic
demographic patterns?
Olivier Lourdais 1 2 3 , Xavier Bonnet 1 , Dale DeNardo 4 , Guy Naulleau 1
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général Des Deux Sèvres, Rue de L’abreuvoir, 79021, Niort, France
3 University of Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
4 Department of Biology, Arizona State University, Tempe, AZ, 85287-1501,USA
Accepted for publication in Population Ecology
241

Abstract
We examined the influence of sex divergences in reproductive role and physiology on
catchability and demographic patterns in a closed population of aspic viper (Vipera
aspis Linné). During eight years, there were 4800 captures of 988 adults. In both
sexes, captures were more frequent in spring when climatic conditions and
reproductive activities impose extended basking periods that make animals more
detectable. On average, males were captured more than females, reflecting intense
sexual activity (i.e., mate searching) in spring. Reproductive females were more
catchable than non-reproductive females, illustrating a major increase in basking
behaviour associated with reproduction.
Estimates of population size revealed a sexually dimorphic demographic
system with marked year-to-year fluctuations in females contrasting with a more
stable male population. This sex difference in population dynamic reflects sex
divergences in the acquisition and allocation of energy for reproduction. In both
sexes reproduction is fuelled by body reserves. Females, however, need to
accumulate substantial body reserves to reach a high body condition threshold prior
to reproduction, while the male pattern of energy allocation is more gradual (i.e., no
fixed threshold). In addition, reproduction entails major survival cost in females (i.e.,
most females reproduce just once), whereas males are generally annual breeders.
As a consequence of this sex divergence, food abundance, through its direct effect
on body store dynamic, influenced major demographic parameters of females (e.g.,
proportion of reproducing individuals, annual changes in population size) but not
males.
Keywords: catchability; population size; capital-breeding; snake.
242

Introduction
Descriptions of variations in life history traits and their causations constitute a
fundamental theme in the study of evolution and adaptation (Stearns 1992). Accurate
quantification of variation expressed in wild populations often requires capture and
recapture of marked individuals over periods that are biologically relevant to the
species’ generation time. Therefore, estimates of population parameters such as
growth rate, survival, and population size necessitate long-term mark-recapture
techniques (Southwood 1988). While requiring extensive effort and time
commitment, longitudinal approaches provide a multitude of benefits that have been
extensively illustrated in different fields of ecology (Tinkle 1979). Data gathered in
long-term studies permit the examination of individual reproductive success, survival
and possibly life-time reproductive success (e.g., the basic raw material for natural
selection, Clutton-Brock 1988). Additionally, longitudinal work provides an opportunity
to correlate population characteristics with environmental factors such as food
availability, predator abundance, or climatic fluctuation that affect life history traits
(Ballinger 1977; Seigel & Fitch 1985).
Unfortunately, for logistical reasons, long-term mark-recapture studies are
usually not feasible. Individuals of some species are too small to permit marking or
too secretive to catch a reasonably large subset of a population. Similarly, recapture
avoidance and long-distance displacement invalidate or at least seriously complicate
any estimate of population size or mortality (Nichols et al. 1987; Brodie 1989; Cooper
et al. 1990; Lebreton et al. 1992; Shine & Schwarzkopf 1992; Houston & Shine
1994). Due to their cryptic behaviour, snakes are superficially poor models for
recapture studies (Seigel 1993; St Girons 1996). This impression, however, is not
necessarily valid as some species fit within the requirements outlined above.
243

Temperate viperid snakes, for example, are typical sit and wait predators that can be
locally abundant, and a number of field studies have been successfully conducted on
these animals (Saint Girons 1949, 1952, 1957a,b, 1975; Fitch 1960; Klauber 1972;
Brown 1991; Madsen & Shine 1993; Martin 1993). Snakes from temperate climates
may be especially conducive to studies based on recaptures. While snakes in
general are extremely secretive animals and frequently go undetected, species in
cooler climates often must bask in the sun to meet the thermal requirements of major
physiological processes such as digestion, ecdysis, and particularly reproduction
(Huey 1982, Peterson et al. 1993). In males and females, reproduction entails
alterations in activity pattern (e.g., mate searching, increased basking activities) and
this may render the snakes more visible (Bonnet et al. 1999b). Therefore, we
hypothesise that any shift observed in catchability will be directly linked to underlying
physiological changes imposed by reproduction and/or foraging activities. That is,
capture rates not only provide raw data for estimating demographic parameters such
as population size, but also provide information on the thermal and reproductive
biology of the species under study.
Another interesting feature of studying temperate viperid snakes is that they
offer extreme examples of capital-breeding systems in that body reserves often
constitute the primary fuel for reproductive activities (Seigel & Ford 1987; Naulleau
and Bonnet 1996). Though fat stores are probably important both for males and
females (Saint Girons 1957a,b; Olsson et al. 1997), proximate divergences exist in
term of patterns of energy allocation towards reproduction. Most notable is the length
of the reproductive cycle where females are generally pluriannual and males annual
breeders (Saint Girons 1957b; Andren & Nilson 1983; Seigel & Ford 1987; Naulleau
et al. 1999). Such a situation is particularly well described for the aspic viper (Vipera
244

aspis Linné), probably the most intensively studied European snake (Saint Girons
1952, 1996; Naulleau 1997, Zuffi et al. 1999).
Marked sexual divergences in reproductive biology are likely to entail major
ecological repercussions in term of sex specific catchability and possibly
demographic patterns. We tested for sex-differences in catchability patterns and
population dynamics of the aspic vipers using a data set from an eight-year mark-
recapture study of a closed population in west-central France that is characterised by
strong annual fluctuation in prey abundance (Bonnet et al. 2001b, Lourdais et al.
2002b). We also examined whether any detected differences were congruent with
pre-existing knowledge of the ecophysiology of this species.
Material and methods
Study species
The aspic viper is a small viviparous snake of the western-Palearctic region and
locally abundant in west-central France at the northern limit of its distribution.
Females mature at 40 cm snout-vent length (SVL), which is attained in 2.5 to 3.5
years (Bonnet et al. 1999a). Females are typical capital breeders that delay
reproduction until they have amassed enough energy-reserves to reach a high body
condition threshold (Naulleau & Bonnet 1996). In this area, the female reproductive
cycle is longer than annual (Saint Girons 1957a,b; Bonnet & Naulleau 1996; Naulleau
et al. 1999), leading to the coexistence of sub-populations of reproducing females
and non-reproducing ones. Reproductive activity in females imposes marked
behavioural changes that distinguish them from males and non-reproducing females.
Notably, females spend more time basking from the onset of follicle production
245

(March) through parturition (late August) to meet the metabolic requirements of
vitellogenesis and gestation. In addition, females are more sedentary during
gestation (Naulleau et al. 1996). These changes infer substantial survival costs of
reproduction and therefore most (>75%) female vipers reproduce no more than one
time in their lifetimes (Bonnet et al. 2000a, 2002a).
Male aspic vipers are also capital breeders in that fat store are the sole source
of energy during the sexual vernal anorexia. In contrast with females, reproductive
investment is temporally reduced (broadly six to eight weeks in spring) and mainly
concentrated on mate searching activities. Males do not have to reach a fixed body
condition to engage in reproductive activities, as reproductive effort is adjusted to
their body reserves (Aubret et al. 2002). As a consequence, the reproductive cycle is
generally annual in this sex (Vacher-Vallas et al. 1999). Since the smallest male
found copulating (with sperm transmission) was 36.5 cm SVL, all individuals longer
than this were considered adults.
Study site and methods
The study site is in west-central France near the village of Les Moutiers en Retz
47 o 03N'; 02 o 00W'). It is a 33-hectare grove with a mosaic of meadows and
regenerating scrubland. The site is characterised by a temperate oceanic climate.
From 1992 to 1999, one to four people patrolled on almost every favourable day
(sunny and partially cloudy, or cloudy with air temperature above 15°C)
encompassing the vipers’ annual activity period: late February to late October.
Although variations in searching effort occurred between years due to climatic
fluctuations, effectiveness in locating vipers and the searching method were largely
homogenous through the study period (except in 1992 due to a searching effort
biased toward females and large males) thanks to the extended searching period. As
246

a result, the catchability pattern observed pooling all the years is highly consistent
with the catchability pattern observed within each year (unpublished). Total searching
effort exceeded 4,000 hours and represented more than 670 “searching-days”.
Snakes were caught by hand, sexed by eversion of the hemipenes, weighed to the
nearest g with an electronic scale, and measured (SVL, and total length) to the
nearest 5 mm. Over 1,000 adult and sub-adult vipers have been marked using
passive integrated transponder (PIT) tags (Sterile transponder TX1400L, Rhônes
Mérieux, 69002 LYON France, product of Destron/IDI Inc). Upon capture, each snake
was color-marked on the back to avoid short-term recaptures and thus minimise
disturbance. All snakes were released at their point of capture.
Female reproductive status was determined using two methods. First, at the
beginning of vitellogenesis, a female with a body condition (mass scaled by size)
greater than a pre-determined threshold was considered reproductive (see Bonnet et
al. 1994; Naulleau & Bonnet 1996, for validity of the method). Second, from mid-
vitellogenesis (May) to the end of gestation (late August) reproductive status was
easily determined either by palpation of follicles and / or embryos or by records of
parturition (Fitch 1987; Naulleau & Bonnet 1996).
Catchability and population size
Different measurements of catchability were used in this study. First, for individuals
marked at the onset of the study (1992 and 1993), we estimated long-term
catchability by determining the number of consecutive years that an individual was
observed. Secondly, we examined intra-annual catchability by defining eighteen
successive two-week capture-recapture sessions from March to November. These
sessions were equally divided into three consecutive periods (broadly the spring,
summer, and fall seasons) that match with major events in the reproductive cycle
247

(see Table 1). For each individual, we calculated the mean number of captures per
session for each season (seasonal capture rate) and for the entire year (annual
capture rate).
Table 1. Biological cycle of the aspic viper (from Bonnet 1996) and organisation of the eighteen
capture sessions (RF, reproducing females; NRF, non-reproducing females; M, males).
PERIOD
RF
NRF
M
Capture
sessions
1 SPRING 2 SUMMER 3 FALL
vitellogenesis
fat store recovery
sexual anorexia
ovulation, gestation
fat store recovery
fat store recovery
parturition
fat store recovery
fat store recovery
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Month Mar Apr May Jun Jul Aug Sept Oct Nov
Because capture occurrence typically follows a Poisson distribution, we tested
the effect of different explanatory variables (body size, sex, and reproductive status)
using multiple Poisson regression. For descriptive purpose, the effect of body size
was also examined using three size classes. The variable of interest was
standardised (Z = (X - mean value) / SD) so that the distribution had a mean of zero
and a SD of one. The three size classes were: small (Z

Population estimates were calculated using CAPTURE (Otis et al. 1978). Data
from 1992 were excluded from the analysis because of low searching effort. The
model used assumes a closed population (e.g., no birth, death, or migration) and is
appropriate for use in a study covering a short time (Otis et al. 1978). Each two-week
period was considered a capture session. The analysis was restricted to spring
(March – May), since the survival rate is high (>0.8, unpublished data). We found no
evidence of emigration and any snakes not captured over a long period (> two years)
was considered dead (see Naulleau et al. 1996, and Vacher-Vallas et al. 1999). Birth
did not influence our analysis because only adults were considered and neonates
require at least 2.5 years to reach maturity (Bonnet et al. 1999a). Finally, CAPTURE
provides the opportunity to test different models including heterogeneity of capture
probabilities in populations (Mh), time-specific variation in probabilities of recapture
(Mt), behavioural response after initial capture (Mb), and combinations of these
models. In every case, the first model suggested by goodness-of-fit tests was
adopted (Chao et al. 1992). Annual changes in population size (year n) were
calculated from spring population size estimate in year n+1 minus spring population
size estimate in year n. Estimates were performed in spring, just after hibernation.
Because mortality during hibernation is extremely low in our population
(unpublished), the difference between “year-n+1” and “year-n” estimates corresponds
to the changes that occur over the active season (spring to hibernation) of the year n.
We analysed the influences of demographic changes occurring in a year n on the
operational sex ratio of the population in the following year. In the aspic vipers, sex is
genetically determined, and analysis were not confounded by the effects of
environmental variables on primary sex ratio. Estimates of sex ratio (SR) and
operational sex ratio (OSR) were calculated as follow: SR = total number of adult
males / (total number of adult males + total number of adult females); OSR = total
249

number of adult males / (total number of adult males + total number of reproducing
females). All statistics were performed using Statistica 6.0.
Results
Recapture rates and long-term catchability
During the course of the study, 988 adult snakes were marked (463 females and 525
males). The recapture rate (e.g., the percentage of snakes contacted at least once
after the initial capture) was 76.6% and the cumulative number of captures and
recaptures was 4,723. The number of individuals captured varied from year to year
(Table 2) with the highest number of animals observed in 1994. Considering only
snakes marked at the onset of the study (1992-1993), most individuals (70.4%) were
observed during a single or two consecutive years and only a limited number (

individual respectively – Marti 1990). This effect held true when sex was taken into
account (interaction between sex and body size : Wald χ²=3.2, df=2, p=0.21). Annual
capture rates were significantly higher in males than females (2.51 versus 2.16; Wald
χ²=5.79, df=1, p

Capture rate
2.0
1.5
1.0
0.5
0.0
***
***
Spring Summer Fall
Season
NS
Male
Female
Figure 1. Sex divergence in capture rate over the three seasons. Data for seasonal capture rate (e.g.,
mean number of capture per individual) are pooled from all years and error bars represent the
standard error (S.E.). Statistical analyses compares males versus females during the same season
(NS, non-significant; ***, p

Demographic patterns
In the spring, from 1993 to 1999, high recapture rates enabled us to estimate male,
female, and total population size. In all cases, goodness-of-fit tests indicated that a
time variation and individual heterogeneity in capture probabilities model was the
best fit for our data (Mth; Chao et al. 1992). Population size estimates are illustrated
in Figure 3.
Estimated population size
500
400
300
200
100
93 94 95 96 97 98 99
Year
Male
Female
Total
Figure 3. Annual fluctuation in total number of adult snakes (open squares, dashed line), males (open
circles, continuous line) and females (open triangles, dotted line) in spring. Population estimates (±
S.E.) were performed using the program CAPTURE (see text for statistics).
Over the seven years, the average estimated adult population size in spring
was 365 ± 65 individuals (coefficient of variation 0.17) with a mean density of 11
snakes/ha (total of 33 ha) and a biomass of 1.1 kg/ha (given a mean adult body mass
253

of 100g). Females exhibited greater fluctuation in population size than did males
(coefficient of variations 0.31 versus 0.08; Figure 3). Fluctuation in population size
was greatest when reproducing females were considered alone (coefficient of
variation 0.45). Using a proportion of 44% of the total female population (the long-
term mean proportion in the population, Table 2), the proportion of reproducing
females / non-reproducing females was significantly greater in 1996 and 1997, and
lower in 1995 (χ² = 60.9, dl = 6, p < 0.0001; Table 2, Figure 4).
Table 2. Annual variation in captures, proportion of reproducing females and the operational sex ratio.
Calculations were made as follows: Initial Captures = number of different individuals captured each
year; Total Captures = initial captures + recaptures; Proportion of reproducing females = estimate of
reproductive females number/ estimate of total female population size; Sex Ratio = estimate of adult
males population size /(estimate of adult males population size + estimate of adult females population
size); Operational Sex Ratio = estimate of adult males population size / (estimate of adult males
population size + estimate of reproductive females number).
Year
1993
1994
1995
1996
1997
1998
1999
mean
Initial
Captures
284
385
273
297
341
192
167
277
Total
Captures
640
1079
649
732
834
321
279
565
Proportion
of reproducing
females
0.38
0.41
0.20
0.64
0.59
0.33
0.53
0.44
Sex Ratio
0.40
0.41
0.53
0.54
0.40
0.53
0.57
0.48
Operational
Sex Ratio
0.64
0.63
0.85
0.65
0.53
0.77
0.80
0.69
Prey availability at the site varied annually, with 1996 having high rodent
abundance, 1994 having low abundance, and all other years being intermediate
(Bonnet et al. 2001b, Lourdais et al. 2002b). Food abundance in a given year was
independent of food abundance in the preceding year (r=0.16, n=7, F(1,5)=0.14,
254

p

Table 3. Combined influences of food levels in year n (Food n) and food levels in year n-1 (Food n-1)
on the proportion of reproducing females (RF).
Multiple Regression r = 0.97; r² = 0.88; n = 7; F(2,4)=14.51 p < 0.014
Bêta±SE Partial correlation Semi-partial p value
Food n 0.74±0.17 0.90 0.73 0.013
Food n-1 0.71±0.17 0.89 0.70 0.015
Annual changes in total female population size appeared closely related to
current food levels (r = 0.92, F(1,4) = 23.57, n = 6, p < 0.008). However, changes in
female population size were better explained in a multiple regression combining
current food abundance with food abundance in preceding year (model 1, Table 4).
While food level in year n positively influenced changes in female population size
during this year, a negative influence was detected for food level in year n-1.
A similar influence was detected when replacing food level in year n-1 by the
proportion of reproducing females in year n (model 2, Table 4). Hence, the annual
proportion of reproducing females (year n) negatively influenced the change in
female population size over this year n. For males, variations in population size were
limited and no relationship was found between changes in population size and food
level in year n, year n-1 or a combination of both (respective p values: 0.17, 0.29 and
0.27).
256

Table 4. Examination of annual changes in female population size. Population changes (during a year
n) were calculated from spring population size in year n+1 minus spring population size in year n.
Model 1 was obtained by combining food levels in year n (Food n) and food levels in year n-1 (Food
n-1) in the multiple regression. Model 2 was obtained by replacing food levels in year n-1 by the
proportion of reproducing females in year n (% RF n).
Model 1 r = 0.99; r² = 0.98; n = 6; F(2,3)=112.92 p

Figure 5. Relationship between annual changes in female population size (calculated as spring
population size in year n+1 minus spring population size in year n) and operational sex ratio in year
n+1 (see text for statistics).
Operationnal sex ratio in year n+1
0.90
0.85
0.80
0.75
0.70
0.65
0.60
0.55
0.50
-140 -115 -90 -65 -40 -15 10 35 60 85 110 135
Change in female population size (year n)
Discussion
Recently, it has been experimentally shown that female aspic vipers need to reach a
high and precise body condition threshold to initiate reproduction and sexual
behaviours, whereas males exhibit a more gradual relationship between body
reserves and reproduction (Aubret et al. 2002). It has also been shown that this
dichotomy is underlay by an “all-or-nothing” versus a “gradual” hormonal regulation of
reproduction in females and males, respectively (Aubret et al. 2002). The present
field study shows that such sex differences in the energy budget and regulation of
reproduction translate into marked sex divergences of demographic characteristics.
Sex differences in annual catchability patterns reflecting differences in reproductive
258

activities are somewhat a classical result (Saint Girons 1949, 1952, 1957b).
However, the juxtaposition of our extensive data set (>4500 captures) with published
eco-physiological data provides for the first time a sequence of functional links
between sex-specific reproductive roles, energy investment to reproduction,
catchability characteristics, and resulting demographics patterns. The possibility to
reveal the interdependence between these various traits (usually considered
separately), results from the major significance attributable to the marked shifts in
catchability of ectotherms.
In both sexes capture rate was influenced by a combination of environmental
and biological factors. Body size affected catchability with smaller-sized adults being
less catchable than large adults. Body size positively affects catchability in the asp
viper (Naulleau & Bonnet 1996; this study) as well as other snakes (Bonnet et al.
2002c). Perhaps, small snakes adopt a more secretive behaviour in response to a
size-dependent vulnerability to predators (Lima & Dill 1990; Houston & Shine 1994).
Additionally, smaller-sized snakes may be less catchable due to a higher body
surface to volume ratio that shortens heating times and thus reduces basking.
Alternately, the thermal requirements of small snakes may be lower than in larger
individuals. Whatever the case, in snakes, the major effect of body size on
catchability is associated with sexual maturity. After birth, snakes remain extremely
secretive until they reach maturity (unpublished data on more than 600 marked
neonates; Madsen et al. 1999). The dramatic increase in catchability with maturity
provides strong support to the notion that temperate snakes provide the opportunity
to connect reproduction to survival costs associated with vulnerability and
consequently to demographic patterns (Bonnet et al. 1999b).
Though snakes from all size and sex categories were caught throughout the
active season, most of the captures occurred in spring when cool ambient
259

temperatures make it necessary for animals to bask in the sun to achieve and
maintain optimal temperatures. Additionally, mate searching and male-male combat
lead to a strong increase in the catchability of males at that time. After the relatively
short mating season (March-April), males adopt more secretive behaviours and were
observed only occasionally, usually during shedding episodes or digestion
(unpublished obs). In females also, reproduction strongly and positively influenced
catchability with increased exposure of reproducing females over prolonged time
periods (March to September). The higher catchability rates of reproducing females
are linked to the high thermal/metabolic requirements of vitellogenesis and gestation
(Bonnet et al. 1994; Bonnet & Naulleau 1996). Each year, most of the adult males
undertake reproductive activities whilst only a fraction of the females are
reproductive. As a result, male captures outnumbered female captures in spring, but
not later in the year when males were no longer involved in reproduction whereas the
female reproductive activities continued through the end of summer. Overall, when
comparing males and females, both the difference in the absolute values and the
temporal shift in catchability are explainable in the light of their respective system of
allocation of energy to reproduction (gradual versus threshold dependent).
In the course of the study, total population size fluctuated widely and year-to-year
variations in rodent abundance (voles) appeared to be an important regulator. A sex-
specific analysis revealed that inter-annual fluctuations in population size were mainly
driven by the female population, notably reproductive females. These variations are
linked to the annual recruitment rate of reproductive females among non-reproductive
adults and sub-adults, and to the annual survival cost of reproduction. As
reproduction requires a female to have attained a high body condition threshold, and
since food availability influences the accumulation of body stores, food availability
thus influences the proportion of reproducing females in the following year (Bonnet et
260

al. 2001b). However, the proportion of reproducing females was also elevated the
year of particularly high food availability (1996, Figure 4), and this absence of a
temporal delay between food availability and reproduction suggests a more direct
influence of prey availability on reproductive status. We hypothesise that some
females (e.g., ones close to the threshold) may respond positively to high food levels
(1996) and reproduce under such favourable condition. This complex reproductive
decision process that involves both long-term storage and facultative food intake is
well illustrated in the multiple regression analysis (Table 3).
In this population, survival costs of reproduction are high and most females
die during or shortly after reproduction (Bonnet et al. 2000a, 2002a). Part of this
mortality is attributable to exposure to avian predation (Naulleau 1997) while part is
due to the extreme emaciation of post-parturient females (Bonnet et al. 2000a). A
similar effect of reproduction on survival through predation and body reserves
depletion has been reported in a similar adder, Vipera berus (Madsen & Shine 1993).
A direct consequence of this high mortality of reproducing females is that a
fluctuation in the proportion of reproducing females will generate substantial variation
in absolute annual mortality. This was clearly confirmed in our data set as the
proportion of reproducing females in a given year n negatively influenced the annual
changes in total female population over this year n (Table 4). However, current food
levels also positively affected annual changes in female population size (Table 4). As
most female vipers die after reproduction, positive changes in adult population size
can be attributable to the recruitment of new adult females. Therefore, this positive
relationship between food availability and adult population size suggests a strong
linkage between food levels and maturation, and this is largely confirmed by field
data. Growth rate and reserve storage are directly dependent on fluctuation in food
levels. In 1996, the best year in terms of food availability, we recorded the highest
261

growth rates, up to 20 cm in juveniles (Bonnet et al. 1999a) and several females
reached maturity in 2.5 years (instead of the typical 3.5) and reproduced in 1997
(Bonnet et al. 1999a). Hence, demographic patterns in female vipers were clearly
understandable in the frame of their particular reproductive biology, combining
delayed reproduction (precise body condition) and a short reproductive life (tendency
toward semelparity).
In strong contrast to females, annual changes in male population size were
limited and apparently not affected by food levels. Though food levels influence
growth in this sex as well (Bonnet et al. 1999a), they did not directly effect population
dynamic. In males, the gradual system of energy allocation to reproduction does not
impose extended phases of energy gathering, hence most males can initiate
reproductive activities even with limited body reserves. Furthermore, reproductive
activities are concentrated in spring and do not entail prolonged exposure to
predation as observed in reproductive females and, as a result, most males
reproduce repeatedly during their life span. This gradual system of energy allocation
(no body condition threshold, iteroparity) offers a likely explanation for the stable
male population size.
The differential effects of food abundance on each sex within a population led to
marked fluctuations in both absolute and operational sex ratio (Table 3). To our
knowledge, such a strong effect has never been documented in snakes. Since OSR
influences sexual selection gradients and the reproductive strategy of males (Duvall
et al. 1992, Madsen & Shine 1992b), it deserves further study.
In conclusion, while differential catchability is classically viewed as a
confounding factor complicating population size analysis; in ectotherms it rather
offers opportunities to explore complex relationships between the energy budget of
reproduction, seasonal catchability, and population dynamic. Our study of aspic
262

vipers shows that within the array of capital breeders and within a given population,
two sub-populations (males versus females) nonetheless exhibit marked differential
sequences in the links between various life-history traits. Sexual differences in
ecology (i.e., diet, behaviour, and habitat selection) or demography in relation to
climatic or resource fluctuations have been reported in endotherm species, notably
ungulates (Clutton-Brock et al. 1987; du Toit 1995; Myseterud 2000; Oakes et al.
1992; Owen-Smith 1993). The clear dichotomous system revealed in the present
work, however, contrasts sharply with results gathered on endotherms. The temporal
dissociation between the phases of energy acquisition and allocation to reproduction
exhibited by ectotherms provides a unique opportunity to better unravel the complex
effects of resource-fluctuating environments on reproductive strategies (Pough 1980;
Shine & Bonnet 2000a). Finally, the present study also emphasizes the
complementary aspects of ecological and physiological approaches to interpret
capture-recapture data (Bonnet et al. 2002a).
Acknowledgements
We thank Gwénael Beauplet, Hervé Fritz, and Emily Taylor for comments on the
manuscript. Financial support was provided by the Conseil Régional de Poitou-
Charentes, Conseil Général des Deux-Sèvres, Centre National de la Recherche
Scientifique (France). Thanks to Melle for enthralling debates on multiple regression
analysis. Finally, Jean De Riboulin solved many technical problems.
263

V Discussion-conclusion :
reproduction sur réserves,
coûts de la reproduction et
évolution vers la
seméliparité
264

Les données obtenues dans ce travail rendent possible l’identification des
contraintes liées à la reproduction chez la vipère aspic. L’étude de la contribution
respective de ces contraintes constitue une approche puissante pour la
compréhension de la stratégie reproductrice particulière déployée par cette espèce et
basée sur un investissement reproducteur massif et peu fréquent. Plusieurs de nos
résultats peuvent être replacés dans un contexte évolutif général et permettent
d’envisager un scénario explicatif à l’évolution des systèmes de gestion des
ressources et des stratégies démographiques extrêmes.
A. Stratégie de reproduction de la vipère aspic
L’investissement reproducteur
La formation des œufs (vitellogénèse) représente une étape clé de la reproduction
chez la vipère aspic. La synthèse et le dépôt du jaune dans les follicules va mobiliser
de grandes quantités d’énergie. Les importants stocks de réserves lipidiques en
combinaison avec des prises alimentaires facultatives vont fournir le support de base
de l’investissement reproducteur. Une fois cette première étape d’allocation achevée,
la gestation va être associée à de nouveaux types d’investissements maternel. En
effet, Pendant la phase de développement embryonnaire, le budget temps de la
femelle aspic va être consacré à la thermorégulation et au maintien de conditions
optimales pour le développement embryonnaire.
Relation entre effort reproducteur, coûts de la reproduction et fécondité
Si l’investissement gamétique est dépendant du nombre de follicules, nous avons
montré que les dépenses métaboliques de la gestation présentent une nature fixe et
265

indépendante de la fécondité. Ainsi, quelque soit le nombre de jeunes, la femelle doit
maintenir un profil thermique optimum pour assurer le développement embryonnaire.
En outre, les activités de thermorégulation et la réduction des déplacements vont
limiter les possibilités d’alimentation. Ces pertes en opportunités alimentaires sont
par leur nature indépendantes du nombre de jeunes produits. La succession des
phases d’investissement énergétique dans la vitellogénèse (investissement direct)
d’une part et dans la gestation (effort métabolique indirect lié à des modifications du
profil d’activité) ensuite vont avoir de profondes conséquences sur l’état
physiologique des femelles après la mise bas (amaigrissement). Les faibles valeurs
de condition corporelle observées chez les femelles post parturientes illustrent ces
phases successives d’investissement et limitent les possibilités de récupération
après la reproduction. Les contraintes énergétiques de la gestation sont donc
composites en impliquant un volet métabolique (changement des préférences
thermiques) accentué par les conséquences écologiques de la thermorégulation
(pertes en opportunités alimentaires). Le niveau d’amaigrissement post partum des
femelles et les coûts en survie seront donc déterminés par ces contraintes et peu ou
pas affectés par le nombre de jeunes en développement. Du fait de leur nature fixe,
les coûts énergétiques rapportés au nombre de vipéreaux produits seront d’autant
plus élevés que la portée sera de taille réduite. En plus de ces aspects
énergétiques, les changements éthologiques pendant la vitellogénèse et la gestation
vont entraîner une augmentation brutale de l’exposition et par conséquent du risque
de prédation. L’engagement dans la reproduction va donc être à l’origine de coûts
écologiques directs (prédation) déterminés par le statut reproducteur et peu affectés
par le nombre de jeunes.
L’examen de ces différentes composantes de l’effort reproducteur et des coûts
associés suggère donc l’existence d’une relation non linéaire entre ces deux
266

variables : la reproduction va être associée à des coûts en survie directs (prédation)
et indirects (amaigrissement et coûts de post reproduction) peu dépendants du
nombre de jeunes produits. Nos données montrent, par ailleurs, que l’amplitude très
élevée de ces coûts vient limiter directement le nombre d’opportunités
reproductrices.
Gestion de la ressource et fréquence de reproduction
Face à de tels coûts fixes, Bull et Shine (1979) ont suggéré l’avantage de systèmes
reproducteurs basés sur la faible fréquence des reproductions. En se reproduisant
de façon alternée, l’organisme peut ainsi tirer un avantage en préparant son
investissement pour la reproduction suivante. Nos données permettent de confirmer
l’existence de telles situations et de connecter la proposition de Bull et Shine avec la
sélection de systèmes particuliers de gestion de la ressource. Il existe, chez la
vipère aspic, un ensemble de traits d’histoire de vie co-adaptés qui permettent de
préparer très efficacement la reproduction. Ainsi, la maturité tardive et l’accumulation
de réserves lipidiques jusqu’à un seuil élevé de condition corporelle sont des
éléments qui garantissent un investissement reproducteur massif et donc une
fécondité élevée, même en l’absence de source énergétique (alimentation) pendant
la reproduction. Face à des coûts de reproduction élevés (directs et post
reproduction), un tel système va assurer un succès reproducteur élevé et donc
l’amortissement des coûts payés par vipéreaux. Cette orientation particulière de
l’effort reproducteur constitue une réponse évolutive qui va influencer l’ensemble de
la stratégie reproductrice (voir schéma récapitulatif page suivante).
267

80% 40% 24% 12%
Préparation
Système de
capitalisation
de l’énergie
+ + +
Reproduction 1
Investissement
reproducteur
élevé
Contraintes environnementales :
Prédation, climat, nourriture
Récupération
Reproduction 2
Faibles possibilités de
récupération : nombre
de reproduction réduit
Schéma récapitulatif : chez la vipère aspic femelle les coûts de reproduction et de
post-reproduction sont élevés et affectés par plusieurs variables environnementales
(en haut). Nos données indiquent que la probabilité de survivre va s’effondrer
(pourcentages) dès la première reproduction. Ceci illustre l’importance des
contraintes écologiques/énergétiques de la reproduction qui se manifestent
indépendamment de la fécondité (prédation, pertes en opportunités alimentaires,
dépense métabolique pendant la gestation). Dans ce contexte où les possibilités de
reproduction ultérieure sont précaires, la capitalisation de l’énergie est très
avantageuse. En effet, cette stratégie d’allocation permet un investissement
reproducteur élevé et l’amortissement des coûts en optimisant le succès à chaque
reproduction.
268

Pour les vipères qui survivent jusqu’à la mise bas, la phase de récupération
sera déterminée par les fluctuations d’abondance en proies qui vont influencer les
possibilités de reconstitution des stocks de réserves. L’interaction entre les
contraintes énergétiques de la reproduction et les limitations trophiques va rendre
peu avantageuse la réalisation de compromis entre l’investissement dans la
reproduction courante et dans les reproductions futures. Dans ce contexte, les
pressions de sélection vont favoriser la maximisation de l’investissement dans
chaque reproduction. La stratégie d’allocation de l’énergie de la vipère aspic reflète
fidèlement cette orientation des compromis d’allocation. Le stockage de réserves
corporelles avant la reproduction et l’existence d’un seuil de condition corporelle
élevé et fixe, garantissent le succès reproducteur même lorsque les conditions
trophiques sont limitantes pendant la reproduction. Cette stratégie d’allocation de
l’énergie constitue potentiellement une réponse évolutive adaptée à de fortes
contraintes reproductrices indépendantes de la fécondité (voir schéma récapitulatif).
Nos résultats soutiennent fortement la proposition initiale de Bull et Shine (1979) et
soulignent l’importance proximale des systèmes de gestion de la ressource
(capitalisation) dans la concentration de l’effort reproducteur.
Origine de la seméliparité
Si le système d’allocation de l’énergie est, chez la vipère aspic, fortement orienté
vers la capitalisation de la ressource, il est important de noter que cette espèce n’est
pas un reproducteur sur réserves “rigide”. Ainsi lorsque la nourriture est abondante,
les prises alimentaires facultatives seront observées pendant la reproduction. Cette
entrée d’énergie va permettre l’augmentation de l’investissement reproducteur et la
production de vipéreaux plus lourds (alimentation pendant la vitellogénèse). La
femelle va aussi bénéficier d’une partie de cet apport qui va amortir des coûts
269

énergétiques en améliorant sa condition corporelle post partum. Nos données
suggèrent que dans cette population, l’apport énergétique facultatif est généralement
réduit et fluctuant. Il en est de même, après la reproduction, où l’abondance en proie
va limiter les possibilités de récupération.
Dans notre population d’étude, l’impact de ces facteurs prend une dimension
très particulière. Les mises bas tardives (contraintes climatiques) vont limiter, voir
annuler, les possibilités de récupération avant l’hiver. Une fois l’hivernage achevé,
les limitations trophiques dans l’année qui suit la reproduction vont également limiter
les possibilités de récupération. En conséquence, les femelles ne vont se reproduire
qu’une seule fois et il en résultera une forte tendance semélipare.
Une approche comparative avec le cycle des mâles dans cette région où
avec des populations plus méridionales apporte des informations intéressantes sur
l’ampleur des contraintes reproductrices maternelles. Notamment, les mâles sont
beaucoup moins contraints dans leur reproduction que les femelles et
l’investissement, chez ce sexe, est beaucoup plus graduel. Cette situation est
clairement illustrée par une dynamique populationnelle profondément divergente
entre les sexes avec de très fortes fluctuations dans la population femelle qui
contrastent avec la stabilité populationnelle des mâles. Ces fluctuations qui affectent
la population dans son ensemble sont directement liées aux coûts de la reproduction
spécifiques des femelles et aux influences environnementales sur l’amplitude de
ces coûts.
La comparaison avec des populations dans des régions climatiques plus
favorables apporte des éléments de reflexion intéressant sur les facteurs influençant
l’expression des coûts de la reproduction et leurs conséquences démographiques.
Ainsi, le suivi de populations Italiennes (Zuffi et al. 1999 ; données non publiées)
indique des reproductions annuelles et une répétition des reproductions (souvent
270

supérieure à 4-5) au cours de la vie des femelles. Dans ces populations
méridionales, les conditions climatiques sont favorables et n’imposent pas de
longues phases d’exposition. Des conditions plus favorables vont permettre un cycle
plus court, des expositions beaucoup plus limitées et des possibilités de
récupération élevées l’année même de la reproduction. Ces informations sont
importantes et soulignent l’influence majeure des variables environnementales
(climat) qui vont moduler l’expression des coûts énergétiques et écologiques.
B. Proposition d’un scénario évolutif de la
transition vers la seméliparité
Les populations de vipères aspic de l’Ouest de la France sont donc très
intéressantes en occupant une position charnière dans le continuum entre itéroparité
et seméliparité. A l’image des populations méridionales, le mode de reproduction
dans ces zones périphériques est l’itéroparité. Pourtant, les contraintes
environnementales sont telles que les femelles suivent des trajectoires
reproductrices semélipares. Une telle situation offre une opportunité de retracer les
étapes possibles de la transition vers la seméliparité.
Comme chez beaucoup d’autres ectothermes (Bull and Shine 1979), la
reproduction chez cette espèce est à l’origine de contraintes brutales générant des
coûts élevés, indépendants de la fécondité. Le système d’allocation de l’énergie,
basé sur une longue préparation de la reproduction, permet d’amortir l’amplitude des
coûts payés par jeune produit. Dans les populations en limite de l’aire de répartition,
les contraintes environnementales (climatiques et trophiques) renforcent l’amplitude
de ces impacts sur la vie reproductrice ultérieure (la plupart des femelles meurent
après la reproduction). Cependant, la situation n’est pas fixée et lorsque les
271

conditions sont favorables, la récupération énergétique et l’investissement dans
d’autres reproductions est possible. Cette situation peut être considérée comme le
contexte initial de la transition vers un mode de reproduction semélipare. Si on
considère une légère accentuation des contraintes environnementales (dégradation
des conditions climatiques, limites trophiques), les chances de reproductions
ultérieures vont se réduire et s’annuler complètement si l’amplitude des coûts fixes
dépasse les possibilités de récupération de l’organisme. Le maintien ou non de ce
type de conditions, à une échelle macro-évolutive, pourra alors avoir des
conséquences majeures sur l’orientation de la stratégie reproductrice. Notamment, si
les contraintes environnementales rendent extrêmement improbables de futures
reproductions, les pressions de sélection vont favoriser la maximisation de
l’investissement dans la reproduction, en capitalisant par exemple, de l’énergie sur
de longues périodes. Une telle situation correspond en fait à une relaxation du
compromis classique entre reproduction courante et future (Williams 1966b)
favorisant l’orientation de la physiologie de l’organisme dans la réalisation d’un effort
reproducteur ”explosif”. Une telle sélection directionnelle sur les processus
d’allocation de l’énergie rend encore moins probable les chances de survie ultérieure
de l’organisme. Le résultat final possible va être la fixation du système (voir scénario
page suivante) avec une mort qui devient inévitable suite à l’épuisement de
l’organisme ou, à l’extrême, de façon génétiquement programmée (en dehors de
causes environnementales) comme chez de nombreux céphalopodes (Boyle 1983,
1987) ou certains saumons (Crespi & Teo 2002).
272

Scénario de transition vers la seméliparité
1) La reproduction draine parfois une telle quantité d’énergie que cela entraîne la
mort de organisme
2) Si un tel phénomène a lieu très régulièrement, il n’existe plus de bénéfices à
réduire l’effort reproducteur courant en dessous du maximum
physiologiquement envisageable car il n’existe pas d’autres possibilités de
reproduction.
3) Ce contexte favorise l’accumulation de traits qui vont augmenter le succès
reproducteur et en même temps réduire de plus en plus les chances de survie
ultérieures.
4) Le résultat évolutif final est l’accumulation de mutation et la fixation génétique
du système avec une mort inévitable et systématique après la reproduction.
Dans un tel scénario, les populations de vipère à tendance semélipare se
situent à l’échelon 1. Les systèmes à mort programmée rencontrés chez les
céphalopodes et les saumons se trouvent au stade 4. Un tel schéma offre une piste
cohérente pour la transition évolutive vers les systèmes semélipares. Alors que
seméliparité et itéroparité sont généralement étudiées et modélisées comme deux
stratégies en compétition, notre schéma propose l’existence d’une seule stratégie de
base : l’itéroparité. Dans certaines situations environnementales, la sélection
favoriserait la concentration de l’investissement sur la première reproduction. La
seméliparité constitue alors un état dérivé de l’itéroparité et résulte d’une orientation
273

de l’organisme vers la réalisation d’un effort reproducteur explosif et unique et non
d’une programmation rigide de la mort.
Notre étude apporte des éléments de réflexion sur l’évolution des stratégies
reproductrices. Dans le monde vivant, il existe en fait de nombreuses activités qui
peuvent générer des coûts indépendants de la fécondité et ces composantes
méritent donc une attention particulière pour la compréhension des systèmes
d’allocation de l’énergie. Ce travail s’est principalement concentré sur des aspects
énergétiques et quantitatifs de l’investissement reproducteur. Pourtant, si le nombre
de jeunes produits est une variable clé, il est important de souligner que la qualité de
la progéniture est un paramètre majeur qu’il est aussi nécessaire d’intégrer. En effet,
la thermorégulation accentuée des femelles pendant la gestation explique en partie
les coûts énergétiques élevés. L’origine de ces choix thermiques est directement liée
à la nécessité d’assurer un développement embryonnaire optimal. Dans un travail
récent (voir annexe), j’ai pu mettre en évidence de profondes influences des
conditions thermiques de la gestation sur le phénotype des jeunes (longueur,
écaillure) ainsi que sur la mortalité embryonnaire. Si l’allocation de l’énergie et
l’optimisation de l’effort reproducteur sont des éléments importants des stratégies
reproductrices, la qualité de la descendance va aussi exercer une influence majeure
sur le succès reproducteur. La diversité des soins parentaux prodigués par les
vertébrés ectothermes (notamment les soins pré-nataux pendant la gestation) offre
un champ d’étude très fertile pour examiner l’importance de l’optimisation de la
qualité des jeunes dans la compréhension des stratégies reproductrices. Je vais
désormais orienter mes travaux de recherche dans ce domaine particulier.
274

Bibliographie
A.
Anderson RA & Karasov WH (1981) Contrasts in energy intake and expenditure
in sit-and-wait and widely foraging lizards. Oecologia 49 : 67-72.
Andren C & Nilson G (1983) Reproductive tactics in an island population of
adders, Vipera berus (L.), with a fluctuating food resource. Amphibia-Reptilia 3 :
63-79.
Andrews RM (1982) Patterns of growth in reptiles. In : Gans C & Pough H (eds)
Biology of Reptilia 13, pp 273-312. Academic Press, London, United Kingdom,.
Andrews RM & Pough FH (1985) Metabolism of squamate reptiles: allometric
and ecological relationships. Physiological Zoology 58 : 214-231.
Andrews RM, Mathies T & Warner D (2000) Effect of incubation temperature on
morphology, growth and survival of juvenile Sceloporus undulatus. Herpetological
Monographs 14 : 420-431.
Andries JC (2001) Endocrine and environmental control of reproduction in
polychaeta. Canadian Journal of Zoology 79 : 254-270.
Anguiletta MJ & Sears MW (2000) The metabolic cost of reproduction in an
oviparous lizard. Functional Ecology 14 : 39-45.
Arcese CD & Smith JNM (1988) Effects of population density and supplemental
food on reproduction in song sparrow. Journal of Animal Ecology 57 : 119-136.
Atchley WR, Gaskins CT & Anderson D (1976) Statistical properties of ratios I.
Empirical results. Systematic Zoology 25 : 137-148.
275

Aubret F, Bonnet X, Shine R & Lourdais O (2002) Fat is sexy for females but not
males : the influence of body reserves on reproduction in snakes (Vipera aspis).
Hormones and Behavior 42 : 135-147.
B.
Ballinger RE (1977) Reproductive strategies : food availability as a source of
proximal variation in a lizard. Ecology 58 : 628-635.
Ballinger RE (1983) Lizard Ecology : Studies on a model organism. In : Huey R,
Pianka ER & Schoener TW (eds) Life history variations, pp. 241-260. Harvard
University press, Cambridge, Massachusetts, USA.
Baron JP, Ferrière R, Clobert J & Saint Girons H (1996) Stratégie
démographique de Vipera ursinii ursinii au Mont-Ventoux (France). Comptes
Rendus de l’Académie des Sciences, Série III 319 : 57-69.
Bartholomew GA (1970) A model for the evolution of pinniped polygyny.
Evolution 24 : 546-559.
Bauwens D & Thoen C (1981) Escape tactics and vulnerability to predation
associated with reproduction in the lizard Lacerta vivipara. Journal of Animal
Ecology 50 : 733-743.
Beaupré SJ & Duvall D (1998) Variation in oxygen consumption of the western
diamondback rattlesnake (Crotalus atrox) : implications for sexual size
dimorphism. Journal of Comparative Physiology 168(b) : 497-506.
Beaupré SJ & Zaidan F (2001) Scaling CO2 production in the timber rattlesnake
(Crotalus horridus), with comments on cost of growth in neonates and
comparative patterns. Physiological and Biochemical Physiology 74 : 757-768.
Bell G (1980) The costs of reproduction and their consequences. American
Naturalist 116 : 45-76.
Bell G & Koufopanou V (1986) The cost of reproduction. In : Dawkins R & Ridley
M (eds) Oxford Surveys in Evolutionary Biology, pp. 83-131. Oxford University
Press, Oxford.
276

Bennett AF & Dawson WR (1976) Metabolism. In : Gans C & Dawson WR (eds)
Biology of the Reptilia. Vol. 5 Physiology A, pp. 127-223. New York: Academic
Press.
Birchard GF, Black CP, Schuett GW & Black V (1984) Influence of pregnancy on
oxygen consumption, heart rate and haematology in the Garter Snake :
implications for the cost of reproduction in live bearing reptiles. Comparative
Biochemistry and Physiology 77(a) : 519-523.
Blackburn DG (1982) Evolutionary origins of viviparity in the Reptilia I. Sauria.
Amphibia-Reptilia 3 : 185-205.
Blackburn DG (1985) Evolutionary origins of viviparity in the Reptilia II.
Serpentes, Amphisbaenia, and Icthyosauria. Amphibia-Reptilia 6 : 259-291.
Blackburn DG (1998) Resorption of oviductal eggs and embryos in squamates.
Herpetological Journal 8 : 65-71.
Blanchard FN & Blanchard FC (1941) Factors determining time of birth in the
garter snake Thamnophis sirtalis sirtalis (Linnaeus). Pap, Mich, Acad, Sci, Arts,
Lett 26 : 161-176.
Blem CR & Blem LB (1990) Lipid reserves of the brown water snake Nerodia
taxispilota. Comparative Biochemistry and Physiology 97A : 367-372.
Boggs CL (1992) Resource allocation : exploring connections between foraging
and life history. Functional Ecology 6 : 508-518.
Bolton M, Monaghan P & Houston CD (1993) Proximate determination of clutch
size in lesser black-backed gulls : the roles of food supply and body condition.
Canadian Journal of Zoology 71 : 273-279.
Bonnet X (1996) Gestion des réserves corporelles et stratégie de reproduction
chez Vipera aspis. Thèse de Doctorat, Université Claude Bernard, Lyon I.
Bonnet X & Naulleau G (1993) Relationship between glycemia and seasonal
activity in Vipera aspis. Amphibia Reptilia 14 : 295-306.
Bonnet X & Naulleau G (1994) Utilisation d'un indice de condition corporelle
(BCI) pour l'étude de la reproduction chez les serpents. Comptes Rendus de
l’Académie des Sciences 317 : 34-41.
277

Bonnet X & Naulleau G (1995) Estimation of body reserves in living snakes using
a Body Condition Index (BCI). In : Llorente GA, Montori A, Santos X & Carretero
MA (eds) Scientia Herpetologica, pp 237-240. Barcelona, Spain.
Bonnet X & Naulleau G (1996) Catchability in snakes : consequences on
breeding frequency estimates. Canadian Journal of Zoology 74 : 233-239.
Bonnet X, Naulleau G & Mauget R (1994) The influence of body condition on 17-
β Estradiol levels in relation to vitellogenesis in female Vipera aspis (Reptilia
viperidae). General & Comparative Endocrinology 93 : 424-437.
Bonnet X, Bradshaw SD, & Shine R (1998) Capital versus income breeding: an
ectothemic perspective. Oikos 83 : 333-341.
Bonnet X, Naulleau G & Shine R (1999a) The dangers of leaving home :
dispersal and mortality in snakes. Biological Conservation 89 : 39-50.
Bonnet X, Naulleau G, Lourdais O & Vacher-Vallas M (1999b) Growth in the asp
viper (Vipera aspis L.) : insights from long term field study. In : Miaud C &
Guyetant R (eds) Current Studies in Herpetology, pp 63-69. SEH, Le Bourget du
Lac.
Bonnet X, Naulleau G, Shine R & Lourdais O (2000a) What is the appropriate
time scale for measuring costs of reproduction in a capital breeder? Evolutionary
Ecology 13 : 485-497.
Bonnet X, Naulleau G, Shine R & Lourdais O (2000b) Reproductive versus
ecological advantages to larger body size in female Vipera aspis. Oikos 89 : 509-
518
Bonnet X, Naulleau G, Bradshaw D & Shine R (2001a) Changes in plasma
progesterone in relation to vitellogenesis and gestation in the viviparous snake
Vipera aspis. General and Comparative Endocrinology 121 : 84-94.
Bonnet X, Naulleau G, Shine R & Lourdais O (2001b) Short-term versus long-
term effects of food intake on reproductive output in a viviparous snake (Vipera
aspis). Oikos 92 : 297-308.
Bonnet X, Lourdais O, Shine R & Naulleau G (2002a) Reproduction in snakes
(Vipera aspis): costs, currencies and complications. Ecology 83 : 2124-2135.
278

Bonnet X, Naulleau G & Lourdais O (2002b) The benefits of complementary
techniques : using capture-recapture and physiological approaches to understand
costs of reproduction in the asp viper. In : Schuett GW, Hoggren M & Greene HW
(eds) Biology of the Vipers. Biological Science Press, Traverse City, MI (USA) : In
press.
Bonnet X, Pearson D, Ladyman M, Lourdais O & Bradshaw D (2002c) Heaven
for serpents? A mark-recapture study of Tiger Snakes (Notechis scutatus) on
Carnac Island, Western Australia. Australian Ecology : In press.
Boyd IL (1984) The relationship between body condition and the timing of
implantation in pregnant Grey seals (Halichoerus grypus). Journal of Zoology 203
: 113-123.
Boyle PR (1983) Cephalopods life cycles vol I. In : Boyle PR (ed) Species
accounts. Academic press. London
Boyle PR (1987) Cephalopods life cycles vol II. In : Boyle PR (ed) Comparative
reviews. Academic press. London
Bradshaw SD (1997) Homeostasis in Desert Reptiles. Cloudsey-Thompson JL
(ed), 213 pp. Springer, Heidelberg.
Bradshaw SD, Saint Girons H & Bradshaw FJ (1991) Seasonal changes in
material and energy balance associated with reproduction in the green lizard,
Lacerta viridis, in western France. Amphibia Reptilia 12 : 21-32.
Brana F, Gonzalez F & Barahona A (1992) Relationship between ovarian and fat
body weights during vitellogenesis for three species of lacertid lizards. Journal of
Herpetology 26 : 515-518.
Brand CJ & Keith LB (1979) Lynx demography during a snowshoe hare decline
in Alberta. Journal of Wildlife Management 43 : 827-839.
Brodie EDI (1989) Behavioral modification as a means of reducing the cost of
reproduction. American Naturalist 134 : 225-238.
Brodie EDI & Ducey PK (1989) Allocation of reproductive investment in the
redbelly snake Storeria occipitomaculata. American Midland Naturalist 122 : 51-
58.
279

Bronson FH (1998) Energy balance and ovulation: small cages versus natural
habitats. Reproduction, Fertility and Development 10 : 127-137.
Brown WS (1991) Female reproductive ecology in a northern population of timber
rattlesnake (Crotalus horridus). Herpetologica 47 : 101-115.
Brown WS (1993) Biology, status and management of the timber rattlesnake
(Crotalus horridus) : a guide for conservation. Herpetological Circulars 22 : 1-78.
Brown DC, Kelnar CJH & Wu FCW (1996) Energy, metabolism during male
human puberty II. Use of testicular size in predictive equations for basal metabolic
rate. Annals of Human Biology 23(4) : 281-284.
Bryant EH (1971) Life history consequences of natural selection : Cole’s result.
American naturalist 105 : 75-76.
Bull JJ & Shine R (1979) Iteroparous animals that skip opportunities for
reproduction. American Naturalist 114 : 296-316.
Burger J (1998) Effects of incubation temperature on hatchling pine snakes :
Implications for survival. Behavioural Ecology and Sociobiology 43 : 11-18.
Butte NF, Hopkinson JM, Mehta N, Moon JK & Smith E (1999) Adjustments in
energy expenditure and substrate utilisation during late pregnancy and lactation.
American Journal of Clinical Nutrition 69(2) : 299-307.
C.
Cadenas S, Buckingham JA, St-Pierrie J, Dickinson K, Jones RB & Brand MD
(2000) AMP decreases the efficiency of skeletal-muscle mitochondria.
Biochemical Journal 351(2) : 307-311.
Callard IP, Filete LA, Perez FE, Sorbera LA, Giannoukos G, Klosterman LL,
Tsang P & McCracken JA (1992) Role of the corpus luteum and progesterone in
the evolution of vertebrate viviparity. American Zoologist 32 : 264-275.
Calow P. (1979) The cost of reproduction - a physiological approach. Biological
Reviews 54 : 23-40.
Capula M, Luiselli L & Anibaldi C (1992) Complementary study on the
reproductive biology in female adder, Vipera berus (L.), from eastern Italian Alps.
Vie Milieu 42 : 327-336.
280

Chao A, Lee SM & Jeng SL (1992) Estimating population size for captures-
recaptures data when captures probabilities vary by time and individual animal.
Biometrics 48 : 201-216.
Charnov EL (1982) The Theory of Sex Allocation. Princeton University Press,
Princeton.
Charnov EL & Schaffer WM (1973) Life history consequences of natural selection
: Cole’s results revisited. American Naturalist 107 : 791-793.
Chastel O, Weimerskirch H & Jouventin P (1995) The influence of body condition
on reproductive decisions and reproductive success in the blue petrel. Ecology
76 : 2240-2246.
Cherel Y (1995) Nutrient reserve storage, energetics, and food consumption
during the prebreeding and premoulting foraging periods of king penguins. Polar
Biology 15 : 209-214.
Cherel Y, Charassin JB & Handrich Y (1993) Comparison of body reserves build-
up in prefasting chicks and adults of king penguins (Aptenodytes patagonicus).
Physiological Zoology 66 : 750-770.
Cichon M, Olejniczak P & Gustafsson L (1998) The effect of body condition on
the cost of reproduction in female collared flycatchers Ficedula albicollis. Ibis 140
: 128-130.
Clark H & Sisken BF (1956) Nitrogenous excretion by embryos of the viviparous
snake Thamnophis s. sirtalis. Journal of Experimental Biology 33 : 384-393.
Clutton-Brock TH (1988) Reproductive Success. Studies of individual variation in
contrasting breeding systems. University of Chicago Press, Chicago.
Clutton-Brock TH (1991) The evolution of parental care. Princeton University
Press, Princeton, New Jersey.
Clutton-Brock TH (1998) Intoduction : studying reproductive cost. Oikos 83 : 421-
423
Clutton-Brock TH, Iason GR & Guiness FE (1987) Sexual segregation and
density-related changes in habitat use in male and female red deer (Cervus
elaphus). Journal of Zoology 211 : 275-289.
281

Coates M & Ruta M (2000) Nice snakes, shame about the legs. Trends in
Ecology and Evolution 15 : 503-507.
Cole LC (1954) The population consequences of life history phenomena.
Quarterly review of Biology 29 : 103-137.
Congdon JD, Dunham AE & Tinkle DW (1982) Energy budgets and life histories
of reptiles. In : Gans C & Pough FH (eds) Biology of the Reptilia, pp 233-271.
Academic Press, New York.
Cooper WE Jr, Vitt LJ, Hedges R & Huey RB (1990) Locomotor impairment and
defense in gravid lizard (Eumeces laticeps): behavioral shift in activity may offset
costs of reproduction in an active forager. Behavioural Ecology and Sociobiology
27 : 153-157.
Corkum LD, Ciborowski JJH & Poulin RG (1997) Effects of emergence date and
maternal size on egg development and sizes of eggs and first-instar nymphs of a
semelparous aquatic insect. Oecologia 111 : 69-75.
Cossins AR & Roberts N (1996) The feast and famine. Nature 379 : 23.
Cresswell WJ, Harris S, Cheeseman CL & Mallinson PJ (1992) To breed or not to
breed: an analysis of the social and density-dependent constraints on the
fecundity of female badgers (Meles meles). Philosophical Transactions of the
Royal Society London B. 338 : 393-407.
Crews D & Gans C (1992) The interaction of hormones, brain, and behavior : an
emerging discipline in herpetology. In : Gans C & Crews D (eds) Biology of the
Reptilia 18, pp 1-23. The university of Chicago press.
Crespi JB & Teo R (2002) Comparative Phylogenetic Analysis of the evolution of
semelparity and life history in Salmonid fishes. Ecology 56 : 1008-1020.
Cuellar O (1984) Reproduction in a parthenogenetic lizard: with a discussion of
optimal clutch size and a critique of the clutch weight/body weight ratio. American
Midland Naturalist 111 : 242-258.
Curtis V, Henry CJK, Birch E & Ghusain-Choueiri A (1996) Intra-individual
variation in the basal metabolic rate of women : Effect of the menstrual cycle.
American Journal of Human Biology 8(5) : 631-639.
282

D.
Darwin C (1859) On the origin of species by means of natural selection or the
preservation favoured races in struggle for life. John Murray (ed). London
Daan S, Deerenberg C & Dijkstra C (1996) Increased daily work precipitates
natural death in the kestrel. Journal of Animal Ecology 65 : 539-544.
Darken RS, Martin KLM & Fisher MC (1998) Metabolism during delayed hatching
in terrestrial eggs of a marine fish, the Grunion, Leuresthes tenuis. Physiological
Zoology 71(4) : 400-406.
Dearbon DC (2001) Body condition and retaliation in the parental effort decisions
of incubating great frigate birds (Fregata minor). Behavioral Ecology 12 : 200-206.
Decopas F & Hart SS (1957) Use of a Pauling oxygen analyser for measurement
of oxygen consumption of animals in open-circuit systems and in a short-lag,
closed-circuit apparatus. Journal of Applied Physiology 10 : 388-392.
Delattre P, Giraudoux P, Baudry J, Truchetet D, Musard P, Toussaint M, Stahl P,
Poulle ML, Artois M, Damange JP, & Quere JP (1992) Land use patterns and
types of Common Vole (Microtus arvalis) population kinetics. Agriculture,
Ecosystem and Environment 3 : 153-169.
Demarco V & Guillette LJ jr (1992) Physiological cost of pregnancy in a
viviparous lizard Sceloporus jarrovi. Journal of Experimental Zoology 262(4) :
383-390.
Diller LV & Wallace RL (1984) Reproductive biology of the northern pacific
rattlesnake (Crotalus viridis oreganus). Herpetologica 40 : 182-193.
Dmi'el R (1986) Intra- and interspecific allometry of oxygen consumption in
snakes. Journal of Herpetology 20 : 265-267.
Doughty P & Shine R (1997) Detecting life history trade-offs : measuring energy
stores in “capital” breeders reveals costs of reproduction. Oecologia 110 : 508-
513.
Doughty P & Shine R (1998) Energy allocation to reproduction in a viviparous
lizard (Eulamprus tympanum) : the role of long-term energy stores. Ecology 79 :
1073-1083.
283

Downes SJ & Shine R (1999) Do incubation-induced changes in a lizard's
phenotype influence its vulnerability to predators? Oecologia 120(1) : 9-18.
Drent RH & Daan S (1980) The prudent parent: energetic adjustments in avian
breeding. Ardea 68 : 225-252.
Du Toit JT (1995) Sexual segregation in Kudu : sex differences in competitive
ability, predation risk or nutritional needs? South African Journal of Wildlife
Research 25 : 127-134.
Dunham AE, Miles DB & Reznick DN (1988) Life history patterns in squamate
reptiles. In : Gans C & Huey RB (eds), Biology of the Reptilia 16, pp 441-522.
Alan R Liss Inc, New York.
Duval D, Arnorld S & Schuett GW (1992) Pitviper mating system: Ecological
potential, sexual selection and microevolution. In : Campbell JA & Brodie ED Jr
(eds) Biology of the pitvipers, pp 321-336. Tyler, Texas.
Duval D, Schuett GW & Arnorld SJ (1993) Ecology and evolution of snake mating
systems. In : Seigel RA, Collins JT & Novak SS (eds) Snakes, Ecology and
Behavior, pp 165-200. McGraw-Hill, Inc, New York.
E.
Elphick MJ & Shine R (1999) Sex differences in optimal incubation temperatures
in a scincid lizard species. Oecologia 118(4) : 431-437.
Else PL & Hulbert AJ (1981) Comparison of the "mammal machine" and the
"reptile machine" : energy production. American Journal of Physiology 240 : R3-
R9.
Engelmann F & Rau I (1965) A correlation between feeding and the sexual cycle
in Leucophaea maderae (Blattaria). Journal of Insect Physiology 11 : 53-64.
Erikstad EK, Asheim M, Faucahld P, Dahlaug LT & Tveraa T (1997) Adjustment
of parental effort in the puffin ; the roles of adult body condition and chick size.
Behavioural Ecology and Sociobiology 40 : 95-100.
F.
Farmer CG (2000) Parental care : key to understanding endothermy and other
convergent features in Birds and Mammals. American Naturalist 155 : 326-334.
284

Farr DR & Gregory PT (1991) Sources of variation in estimating litter
characteristics of the garter snake, Thamnophis elegans. Journal of Herpetology
25 : 261-268.
Fergusson B & Bradshaw SD (1991) Plasma arginine vasotocin, progesterone,
and luteal development during pregnancy in the viviparous lizard Tiliqua rugosa.
General and Comparative Endocrinology 82 : 140-151.
Festa-Bianchet M, Gaillard JM & Jorgenson JT (1998) Mass- and density-
dependent reproductive success and reproductive costs in a capital breeder.
American Naturalist 152 : 367-379.
Fisher RA (1930) The Genetical Theory of Natural Selection. Oxford University
Press, Oxford.
Fitch HS (1960) Autoecology of the copperhead. University of Kansas
Publications of the Museum of Natural History 13 : 85-288.
Fitch HS (1981) Sexual size differences in reptiles. Miscellaneous Publications
of the Museum of Natural History, University of Kansas 70, 72.
Fitch HS (1987) Collecting and life-history techniques. In : Seigel RA, Collins JT
& Novak SS (eds) Snakes Ecology and Evolutionary Biology, pp 143-164.
Macmillan, New York.
Ford NB & Seigel RA (1989a) Relationships among body size, clutch size, and
egg size in three species of oviparous snakes. Herpetologica 45 : 75-83.
Ford NB & Seigel RA (1989b) Phenotypic plasticity in reproductive traits :
evidence from a viviparous snake. Ecology 70 : 1768-1774.
Ford NB & Seigel RA (1994) Phenotypic plasticity : implication for captive
breeding and conservation programs. In : Murphy JB, Collins JT & Adler K (eds)
Captive Managment and Conservation of Amphibians and Reptiles, pp 175-182.
Society for the Study of Amphibians and Reptiles, Oxford, OH.
Forsman A (1996) An experimental test for food effects on head size allometry in
juvenile snakes. Evolution 50 : 2536-2542.
Fox W, Gordon C & Fox MH (1961) Morphological effects of low temperatures
during the embryonic development of the garter snake, Thamnophis elegans.
Zoologica 46 : 57-71.
285

G.
Gadgil M & Bossert PW (1970) Life historical consequences of natural selection.
American Naturalist 104 : 1-24.
Garcia-Berthou E (2001) On the misuse of residuals in ecology : testing
regression residuals vs. the analysis of covariance. Journal of Animal Ecology 70
: 708-711.
Glazier DS (2000) Is fatter fitter? Body storage and reproduction in ten
populations of the freshwater amphipod Gammarus minus. Oecologia 122 : 335-
345.
Gregory PT (1982) Reptilian hibernation. In : Gans C & Pough FH (eds) Biology
of the Reptilia, Volume 13, pp. 53-154. Academic Press, London.
Gregory PT, Larsen KW & Farr DR (1992) Snake litter size = live young + dead
young + yolks. Herpetologica 2 : 145-146.
Gregory PT, Lisa HC & Skebo KM (1999) Conflicts and interactions among
reproduction, thermoregulation and feeding in viviparous reptiles : are gravid
snakes anorexic? Journal of Zoology London 248 : 231-241.
Gregory PT, Macartney JM & Larsen KW (1987) Spatial patterns and
movements. In : Seigel RA, Collins JT & Novak SS (eds.), Snakes : Ecology and
Evolutionary Biology, pp. 366-395. Macmillan Publishing Company, New York.
Gregory PT & Skebo KM (1998) Trade-offs between reproductive traits and the
influence of food intake during pregnancy in the garter snake, Thamnophis
elegans. American Naturalist 151 : 477-486.
Guillette LJ (1982) Effects of gravidity on the metabolism of the reproductively
bimodal lizard, Sceloporua avenues. The Journal of Experimental Zoology 223 :
33-36.
Gustafson L & Sutherland W (1988) The costs of reproduction in the collared
flycatcher Ficedula albicollis. Nature 335 : 813-815.
286

H.
Hahn WE & Tinkle DW (1965) Fat body cycling and experimental evidence for its
adaptive significance to ovarian follicle development in the lizard, Uta
stansburiana. Journal of Experimental Zoology 158 : 79-86.
Harlow P & Grigg G (1984) Shivering thermogenesis in the brooding diamond
python, Python spilotes spilotes. Copeia 1984 : 959-965.
Highfill DR & Mead RAM (1975) Function of corpora lutea of pregnancy in the
viviparous Garter snake Thamnophis elegans. General and Comparative
Endocrinology 27 : 401-407.
Hoglünd J & Sheldon BC (1998) The cost of reproduction and sexula selection.
Oïkos 83 : 478-483.
Houston D & Shine R (1994) Movement and activity pattern of Arafura file snakes
(Serpentes: Achrocordida) in tropical Australia. Herpetologica 50 : 349-357.
Howe JC, Rumpler WV & Seale JL (1993) Energy expenditure by indirect
calorimetry in premenopausal women : Variation with one menstrual cycle.
Journal of Nutritional Biochemistry 4(5) : 268-273.
Huey BR (1982) Temperature, Physiology, and the Ecology of Reptiles. In : Gans
C & Pough HF (eds) Biology of the Reptilia 12, Physiological Ecology, pp. 25-95.
Academic Press.
Hulbert AJ & Else PL (1989) Evolution of mammalian endothermic metabolism :
mitochondrial activity and cell composition. American Journal of Physiology 256 :
R63-R69.
I.
Ineich I (1995) Actual knowledge on venomous snakes systematics. Bulletin
de la société Herpetologique de France 75-76 : 7-24
J.
Jakob EM, Marshall SD & Uetz GW (1996) Estimating fitness : a comparison of
body condition indices. Oikos 77 : 61-67.
287

James CD & Whitford WG (1994) An experimental study of phenotypic plasticity
in the clutch size of a lizard. Oikos 70 : 49-56.
Jayne BC & Bennett AF (1990) Selection on locomotor performance capacity in a
natural population of garter snakes. Evolution 44 : 1204-1229.
Jönsson KI (1997) Capital and income breeding as alternative tactics of resource
use in reproduction. Oikos 78 : 57-66.
Jönsson KI & Tuomi J (1994) Costs of reproduction in a historical perspective.
Trends in Ecology and Evolution 9 : 304-307.
Jönsson KI, Tuomi J & Järemo J (1995a) Reproductive effort tactics : balancing
pre- and postbreeding costs of reproduction. Oikos 74 : 35-44.
Jönsson KI, Tuomi J & Järemo J (1995b) On the consequences of pre- and
postbreeding costs in the evolution of reproductive effort tactics. Ecoscience 2 :
311-320.
K.
King RB (1993) Determinants of offspring number and size in the brown snake,
Storeria dekayi. Journal of Herpetology 27 : 175-185.
Klauber LM (1972) Rattlesnakes : their habits, life histories and influence on
Mankind, 2nd ed., 2 vols., University of California Press. Bekerley.
Krebs CJ & Myers JH (1974) Population cycles in small mammals. Advances in
Ecological Research 8 : 267-399.
Kunkele J (2000) Energetics of gestation relative to lactation in a precocial
rodent, the guinea pig (Cavia porcellus). Journal of Zoology (London) 250 : 533-
539.
L.
Lack D (1954) The natural regulation of animal numbers. Oxford University
Press, Oxford.
Larsen LO (1980) Physiology of adult lampreys, with special regard to natural
starvation, reproduction, and death after spawning. Canadian Journal of Fisheries
and Aquatic Science 37 : 1762-1779.
288

Lebreton JD, Burnham KP, Clobert J & Anderson DR (1992) Modelling survival
and testing biological hypothesis using marked animals : a unified approach with
case studies. Ecological Monograph 62 : 67-118.
Lee SJ, Witter MS, Cuthill IC & Goldsmith AR (1996) Reduction in escape
performance as a cost of reproduction in gravid starlings, Sturnus vulgaris.
Proceedings of the Royal Society of London, Series B 263 : 619-623.
Lequette B & Weimerskirch H. (1990) Influence of parental experience on the
growth of wandering albatross chicks. Condor 92 : 726-731.
Lessells C (1991) The evolution of life histories. In : Krebs R & Davies ND (eds)
Behavioural Ecology, pp. 32-68. Blackwell, Oxford, London.
Levins R (1968) Evolution in changing environments. Princeton University press,
Princeton.
Lima SL & Dill LM (1990) Behavioral decisions made under the risk of predation:
a review and prospectus. Canadian Journal of Zoology 68 : 619-640.
Loumbourdis NS & Hailey A (1985) Activity metabolism of the lizard Agama
stellio stellio. Comparative Biochemistry and Physiology 82A : 687-691.
Lourdais O, Bonnet X & Doughty P (2002a) Costs of anorexia during pregnancy
in a viviparous snake (Vipera aspis). Journal of Experimental Zoology 292 : 487-
493.
Lourdais O, Bonnet X, Shine R, DeNardo D, Naulleau G & Guillon M (2002b)
Capital-breeding and reproductive effort in a variable environment: a longitudinal
study of a viviparous snake. Journal of Animal Ecology 71 : 470-479.
Lourdais O, Bonnet X, Shine R & Taylor E. (2002c) When does a reproducing
female viper (Vipera aspis) "decide" on her litter size? Journal of Zoology, in press
Lourdais O, Bonnet X, DeNardo D & Naulleau G. (2002d) Does sex differences in
reproductive eco-physiology translate in different demographic patterns ?
Population Ecology, in press.
Luiselli L, Capula M & Shine R (1996) Reproductive output, costs of reproduction,
and ecology of the smooth snake, Coronella austriaca, in the eastern Italian Alps.
Oecologia 106 : 100-110.
289

M.
Madsen T & Shine R (1992a) Determinants of reproductive success in female
adders Vipera berus. Oecologia 92 : 40-47.
Madsen T & Shine R (1992b) Temporal variability in sexual selection on
reproductive tactics and body size in male snakes. American Naturalist 141 : 161-
171.
Madsen T & Shine R (1993) Costs of reproduction in a population of european
adders. Oecologia 94 : 488-495.
Madsen T & Shine R (1996) Determinants of reproductive output in female water
pythons (Liasis fuscus: Pythonidae). Herpetologica 52 : 146-159.
Madsen T & Shine R (1999) The adjustment of reproductive threshold to prey
abundance in a capital breeder. Journal of Animal Ecology 68 : 571-580.
Madsen T, Shine R, Olsson M & Wittzell H (1999) Restoration of an inbred adder
population. Nature 402 : 34-35.
Marti DM (1990) Sex and dimorphism in the barn owl, a test of mate choice. Auk
107 : 246-254.
Martin WH (1993) Reproduction of the timber rattlesnake (Crotalus horridus) in
the Appalachian mountains. Journal of Herpetology 27 : 133-143.
Martin K (1995) Patterns and mechanisms for age-dependent reproduction and
survival in birds. American Zoologist 35 : 340-348.
Mauget C, Mauget R & Sempéré A (1997) Metabolic rate in female european roe
deer (Capreolus capreolus): Incidence of reproduction. Canadian Journal of
Zoology 75(5) : 731-739.
Mayr E (1963) Animal species and evolution. Harvard University press.
McCleery RH, Clobert J, Julliard R & Perrins CM (1996) Nest predation and
delayed cost of reproduction in the great tit. Journal of Animal Ecology 65 : 96-
104.
Meijer T & Drent S (1999) Re-examination of the capital and income dichotomy in
breeding birds. Ibis 141 : 399-414.
290

Méndez De La Cruz FR, Guillette LJ Jr & Villagran-Santa Cruz M (1993)
Differential atresia of ovarian follicles and its effects on the clutch size of two
populations of the viviparous lizard Sceloporus mucronatus. Functional Ecology 7
: 535-540.
Monaghan P & Nager RN (1997) Why don’t birds lay more eggs? Trends in
Ecology and Evolution 7 : 270-274.
Monney JC (1994) Note sur la reproduction et la taille des nouveau-nés chez la
vipère aspic (Vipera aspis) et la vipère péliade (Vipera berus) dans l’Oberland
bernois. Bulletin de la Société Fribourgeoise de Sciences Naturelles 83 : 61-74.
Mrosovsky N & Sherry DF (1980) Animal anorexias. Science 207 : 837- 842.
Mysterud A (2000) The relationship between ecological segregation and sexual
body size dimorphism in large herbivores. Oecologia 124 : 40-54.
N.
Nagy KA (1987) Field metabolic rate and food requirement scaling in mammals
and birds. Ecological Monograph 57 : 111-128.
Nagy KA, Huey RB & Bennett AF (1984) Field energetics and foraging mode of
Kalahari lacertid lizards. Ecology 65 : 588-596.
Nagy KA, Meienberger C, Bradshaw SD & Wooller RD (1995) Field metabolic
rate of a small marsupial mammal, the Honey Possum (Tarsipes rostratus).
Journal of Mammalogy 76 : 862-866.
Naulleau G (1965) La biologie et le comportement prédateur de Vipera aspis au
laboratoire et dans la nature. Bulletin Biologique franco-belge 99 : 395-524.
Naulleau G (1981) Détermination des périodes de l’ovulation chez Vipera aspis
et Vipera berus dans l’ouest de la France, étudiée par radiographie. Bulletin de la
Société des Sciences Naturelles de L’Ouest de la France 3 : 151-153.
Naulleau G (1983a) Action de la température sur la digestion chez cinq espèces
de vipères Européennes du genre Vipera. Bulletin de la Société Zoologique de
France 108 : 47-58.
Naulleau G (1983b) The effects of temperature on digestion in Vipera aspis.
Journal of Herpetology 17 : 166-170.
291

Naulleau G (1986) Effects of temperature on “gestation” in Vipera aspis and
Vipera berus (Reptilia: Serpentes) Studies in Herpetology (ed Röcek, Z), pp. 489-
494, Charles Univ., Prague.
Naulleau G (1997) La vipère aspic. Eveil Nature, Saint Yrieix.
Naulleau G & Bidaut C (1981) Intervalle entre l'accouplement, l'ovulation et la
parturition chez Vipera aspis L. (Reptiles, Ophidiens, Vipéridés), dans différentes
conditions expérimentales, étudié par radiographie. Bulletin de la Société
Zoologique de France 106 : 137-143.
Naulleau G & Saint Girons H (1981) Poids des nouveau-nés et reproduction de
Vipera aspis (Reptilia: Viperidae), dans des conditions naturelles et artificielles.
Amphibia-Reptilia 2 : 51-62.
Naulleau G & Bonnet X (1995) Reproductive ecology, body fat reserves and
foraging mode of two contrasted snakes species: Vipera aspis (terrestrial,
viviparous) and Elaphe longissima (semi-arboreal, oviparous). Amphibia-Reptilia
16 : 37-46.
Naulleau G & Bonnet X (1996) Body condition threshold for breeding in a
viviparous snake. Oecologia 107 : 301-306.
Naulleau G, Célérier ML & Saint Girons H (1984) Influence de la température
d’acclimatation sur la consommation d’oxygène de Vipera aspis (Reptilia,
Viperidae). Comptes Rendus de l’Académie des Sciences, Paris, Série III 12 :
511-514.
Naulleau G, Bonnet X & Duret S (1996) Déplacements et domaines vitaux des
femelles reproductrices de vipères aspic Vipera aspis (Reptilia, Viperidae) dans le
centre ouest de la France. Bulletin de la Société Herpétologique de France 78 :
5-18.
Naulleau G, Verheyden C & Bonnet X (1997) Prédation spécialisée sur la Vipère
aspic Vipera aspis par un couple de buses variables Buteo buteo. Alauda 65 :
155- 160.
Naulleau G, Bonnet X, Vacher-Vallas M, Shine R & Lourdais O (1999) Does less-
than-annual production of offspring by female vipers (Vipera aspis) mean less-
than-annual mating? Journal of Herpetology 33 : 688-691.
292

Nichols JD, Hepp GR, Pollock KH & Hines JE (1987) The husting dilemma : a
methodological note. Ecology 68 : 213-217.
Niewiarowski PH & Dunham AE (1994) The evolution of reproductive effort in
squamate reptiles : costs, trade-offs, and assumptions reconsidered. Evolution
48:137-145.
Niewiarowski PH & Dunham AE (1998) Effect of mortality risk and growth on a
model of reproductive effort : why the Shine & Schwarzkopf model is not general.
Evolution 52 : 1236-1241
Nilsson JA & Svensson E (1996) The cost of reproduction : a new link between
current reproductive effort and future reproductive success. Proceedings of the
Royal Society, London B 263 : 711-714.
Nur N (1988) The costs of reproduction in birds : an examination of the evidence.
Ardea 76: 155-168.
O.
Oakes EJ, Harmsen R & Erberl C (1992) Sex, age, and seasonal differences in
the diets and activity budgets of muskoxen (Ovibos moschatus). Canadian
Journal of Zoology 70 : 605-616.
Oakwood M, Bradley AJ & Cockburn A (2001) Semelparity in a large marsupial.
Proceedings of the Royal Society. Series B 268 : 407-411.
Olsson M, Madsen T & Shine R (1997) Is sperm really so cheap? Costs of
reproduction in male adders, Vipera berus. Proceedings of the Royal Society
séries B 264 : 455-459.
Olsson M, Shine R & Bak-Olsson E (2000) Locomotor impairment of gravid
lizards: is the burden physical or physiological? Journal of Evolutionary Biology 13
: 263-268.
Olsson O (1997) Clutch abandonment: a state-dependent decision in king
penguins. Journal of Avian Biology 28 : 264-267.
O'Sullivan A, Martin A & Brown MA (2001) Efficient fat storage in premenopausal
women and in early pregnancy: A role for estrogen. Journal of Clinical
Endocrinology and Metabolism 86(10) : 4951-4956.
293

Otis DL, Burnham KPCG & Anderson DR (1978) Statistical inference for capture
data on closed animal populations. Wildlife Monograph 62 : 1-135.
Owen-Smith N (1993) Comparative mortality rates of male and female Kudus: the
costs of sexual size dimorphism. Journal of Animal Ecology 62 : 428-440.
P.
Packard GC & Packard MJ (1988) The physiological ecology of reptilian eggs
and embryos. In : Gans C. Huey R. (eds). Biology of the Reptilia, pp 523-605.
Alan R Liss, New York, USA 16.
Packard GC & Boardman TJ (1988) The misuse of ratios, indices and
percentages in ecophysiological research. Physiological Zoology 61 : 1-9.
Peterson CR, Gibson AR & Dorcas ME (1993) Snake thermal ecology: the
causes and consequences of body-temperature variation. In : Seigel RA, Collins
JT & Novak SS (eds) Snakes, Ecology and Behavior, pp 241-314. McGraw-Hill,
Inc, New York.
Plummer MV (1983) Annual variation in stored lipids and reproduction in green
snakes (Opheodrys aestivus). Copeia 1983 : 741-745.
Pough FH (1980) The advantages of ectothermy for tetropods. American
Naturalist 115 : 92-112.
Price T, Kirkpatrick M & Arnold SJ (1988) Directionnal selection and the evolution
of breeding date in birds. Science 240 : 798-799.
R.
Ranta E, Kaitala V, Alaja S, Tesar D (2000a) Nonlinear dynamics and the
evolution of semelparous and iteropaous reproductive strategies. The American
Naturalist 155 : 294-300.
Ranta E, Tesar D, Alaja S & Kaitala V (2000b) Does evolution of iteroparous and
semelparous reproduction call for spatially structured systems? Evolution 54 :
145-150.
Reznick DN (1992) Measuring costs of reproduction. Trends in Ecology and
Evolution 7 : 42-45.
294

Reznick DN (1985) Costs of reproduction: an evaluation of the empirical
evidence. Oikos 44 : 257-267.
Reznick DN, Nunney L & Tessier A (2000) Big houses, big cars, superfleas and
the costs of reproduction. Trends in Ecology and Evolution 15 : 421-425.
Rhen T & Lang J (1999) Incubation temperature and sex affect mass and energy
reserves of hatchling snapping turtles, Chelydra serpentina. Oikos 86(2) : 311-
319.
Roff PA (1992) The evolution of life histories. Chapman et Hall, New York,
Routledge.
Rose MR & Bradley TJ (1998) Evolutionary physiolgy of the cost of reproduction.
Oikos 83 : 443-451.
S.
Saint Girons H (1949) Les moeurs nuptiales de la Vipère aspic. Terre et Vie 96 :
110-113.
Saint Girons H (1952) Ecologie et éthologie des vipères de France. Annales des
Sciences Naturelles, Zoologie 14 : 263-343.
Saint Girons H (1957a) Croissance et fécondité de Vipera aspis (L). Vie et Milieu
8 : 265-286.
Saint Girons H (1957b) Le cycle sexuel chez Vipera aspis (L) dans l'ouest de la
France. Bulletin Biologique de France et de Belgique 91 : 284-350.
Saint Girons H (1973) Le cycle sexuel de Vipera aspis (L) en montagne. Vie et
Milieu 23 : 309-328.
Saint Girons H (1975) Critères d’âge, structure et dynamique des populations de
reptiles. In : Bourlière F & Lamotte M (eds) Problèmes d’écologie, pp 233-252.
Masson, Paris.
Saint Girons H (1979) Les cycles alimentaires des vipères européennes dans
des conditions semi-naturelles. Annales de Biologie Animale, de Biochimie et de
Biophysique 19 : 125-134.
295

Saint Girons H (1996) Structure et évolution d'une petite population de Vipera
aspis (L.) dans une région de Bocage de l'ouest de la France. Terre et Vie 51 :
223-241.
Saint Girons H (1997) Vipera aspis. In : Gasc JP et al. (eds) Atlas of Amphibians
and Reptiles in Europe, pp 386-387. SEH, MNHN, Paris.
Saint Girons H & Naulleau G (1981) Poids des nouveau-nés et stratégies
reproductrices des vipères européennes. Terre et vie 35 : 597-616.
Saint Girons H & Duguy R (1992) Evolution de la masse corporelle et de la
masse relative des corps gras, des ovaires et des œufs au cours des cycles
reproducteurs chez Vipera aspis. Amphibia-Reptilia 13 : 351-364.
Saint Girons H, Naulleau G & Célérier ML (1985) Le métabolisme aérobie de
Vipera aspis et Vipera berus (Reptilia: Viperidae) acclimatées à des conditions
naturelles. Influence de la température et comparaison des deux espèces.
Amphibia-Reptilia 6 : 71-81.
Salamolard M, Butet A, Leroux A & Bretagnolle V ( 2000) Responses of an avian
predator to variations in prey density at a temperate latitude. Ecology 81 : 2428–
2441.
Schaffer WM (1974) Optimal reproductive effort in fluctuating environments.
American Naturalist 964 : 783-900.
Schaffer WM & Elson PF (1975) The adaptative significance of variations in life
history among local populations of Atlantic salmon in North America. Ecology 56 :
577-590.
Schneider JE, Zhou D & Blum RB (2000) Leptin and metabolic control of
reproduction. Hormones and Behavior 37 : 306-326.
Secor SM & Diamond J (1995) Adaptive responses to feeding in Burmese
pythons: pay before pumping. The Journal of Experimental Biology 198 : 1313-
1325.
Secor SM, Stein ED & Diamond JM (1994) Rapid up-regulation of snake intestine
in response to feeding : a new model of intestinal adaptation. American Journal of
Physiology 266 : G695-G705.
296

Seigel RA (1993) Summary : future research on snakes, or how to combat “lizard
envy”. In : Seigel RA, Collins JT & Novak SS (eds) Snakes, Ecology and
Behavior, pp 395-402. McGraw-Hill, Inc, New York
Seigel RA & Fitch HS (1984) Ecological patterns of relative clutch mass in
snakes, Oecologia 62 : 293-301.
Seigel RA & Fitch HS (1985) Annual variation in reproduction in snakes in a
fluctuating environment. Journal of Animal Ecology 54 : 497-505.
Seigel RA & Ford NB (1987) Reproductive Ecology. In : Seigel RA, Collins JT &
Novak SS (eds) Snakes: Ecology and Evolutionary Biology, pp 210-252.
Macmillan, New York.
Seigel RA & Ford NB (1991) Phenotypic plasticity in the reproductive
characteristics of an oviparous snake, Elaphe guttata: implication for life history
studies. Herpetologica 47 : 301-307.
Seigel RA & Ford NB (1992) Effect of energy input on variation in clutch size and
offspring size in a viviparous reptile. Functionnal Ecology 6 : 382-385.
Seigel RA & Ford NB (2001) Phenotypic plasticity in reproductive traits:
geographical variation in plasticity in a viviparous snake. Functionnal Ecology 15 :
16-42.
Seigel RA, Huggins MM & Ford NB (1987) Reduction in locomotor ability as a
cost of reproduction in gravid snake. Oecologia 73 : 481-485.
Shine R (1980) "Costs" of reproduction in reptiles. Oecologia 46 : 92-100.
Shine R (1985) The evolution of viviparity in reptiles : an ecological analysis. In :
Gans C & Billet F (eds.), Biology of the Reptilia Vol. 15, p. 605-694. John Wiley
and Sons, INC. New York.
Shine R (1988a) Parental care in reptiles. In : Gans C & Huey RB (eds) Biology
of the Reptilia 16, pp 275-329. Alan R. Liss, New York.
Shine R (1988b) Constraints on reproductive investment: a comparison between
aquatic and terrestrial snakes. Evolution 42 : 17-27.
Shine R (1991) Australian Snakes : A Natural History. Reed Publishing.
297

Shine R (1999) Why is sex determined by nest temperature in many reptiles?
Trends in Ecology and Evolution 14(5) : 186-189.
Shine R & Schwarzkopf L (1992) The evolution of reproductive effort in lizards
and snakes. Evolution 46: 62-75.
Shine R & Harlow P (1993) Maternal thermoregulation influences offspring
viability in a viviparous lizard. Oecologia 96 :122-127.
Shine R & Harlow PS (1996) Maternal manipulation of offspring phenotypes via
nest-site selection in an oviparous lizard. Ecology 77(6) : 1808-1817.
Shine R & Madsen T (1997) Prey abundance and predator reproduction : rats
and pythons on a tropical Australian floodplain. Ecology 78 : 1078-1086.
Shine R & Downes SJ (1999) Can pregnant lizards adjust their offspring
phenotypes to environmental conditions? Oecologia 119(1) : 1-8.
Shine R & Bonnet X (2000) Studying the snakes : a “new model organism” in
ecological research? Trends in Ecology and Evolution 15 : 221-222.
Shine R, Madsen T, Elphick M & Harlow PS (1997) The influence of nest
temperatures and maternal brooding on hatchling phenotypes in water pythons.
Ecology 78 : 1713-1721.
Sibly RM & Calow P (1984) Direct and absorption costing in the evolution of life
cycles. Journal of Theoretical Biology 111 : 463-473.
Sinervo B & Licht P (1991) Hormonal and physiological control of clutch size,
egg size, and egg shape in side-blotched lizards (Uta stansburiana) : constraints
on the evolution of lizard life histories. Journal of Experimental Zoology 257 :
252-264.
Sinervo B & DeNardo DF (1996) Costs of reproduction in the wild: path analysis
of natural selection and experimental tests of causation. Evolution 50 : 1299-
1313.
Sinervo B & Svensson E (1998) Mechanistic and selective causes of life history
trade-offs and plasticity. Oikos 83 : 432-444.
Singh D & Ramachandran AV (2000) Circannual variations in carbohydrate
metabolism in relation to testicular cyclicity and the influence of pinealectomy on
298

altered adrenocortical activity. Journal of Animal Morphology and Physiology
47(1-2) : 33-44.
Southwood TRE (1988) Tactics, strategies and templets. Oikos 52 : 3-18.
Speakman JR & McQueenie J (1996) Limits to sustained metabolic rate: The link
between food intake, basal metabolic rate and morphology in reproducing mice
Mus musculus. Physiological Zoology 69(4) : 746-769.
Statsoft (1995) Statistica. Statsoft, Tulsa.
Statsoft (2001) Statistica. Statsoft, Tulsa.
Stearns SC (1989) Trade-offs in life-history evolution. Functional Ecology 3 : 259-
268.
Stearns SC (1992) The evolution of life histories. Oxford University Press, New
York.
Stewart JR (1992) Placental structure and nutritional provision to embryos in
predominantly lecithotrophic viviparous reptiles. American zoologist 32 : 303-312.
T.
Thibault C & Levasseur MC (1991) La reproduction chez les mammifères et
l'homme. Ellipse Editions, INRA.
Thompson GG & Withers PC (1992) Effects of body mass and temperature on
standard metabolic rates for two Australian varanid lizards (Varanus gouldii and
V. panoptes). Copeia 1992 : 343-350.
Tinkle DW (1979) Long-term field studies. Biosciences 20 : 717.
Todd AW, Keith LB & Fisher CA (1981) Population ecology of coyotes during a
fluctuation of snowshoe hares. Journal of Wildlife Management 45 : 629-640.
Tombre IM & Erikstad KE (1996) An experimental study of incubation effort in
high-arctic barnacle geese. Journal of Animal Ecology 65 : 325-331.
Tuomi J, Hakala T & Haukioja E (1983) Alternative concepts of reproductive
effort, costs of reproduction and selection in life-history evolution. American
Zoologist 23 : 25-34.
V.
299

Vacher-Vallas M (1997) Comportement sexuel, condition corporelle et stratégie
de reproduction chez Vipera aspis: étude en milieu naturel. Unpublished Ph D
thesis, Université François Rabelais, Tours.
Vacher-Vallas M, Bonnet X & Naulleau G (1999) Relations entre les
comportements sexuels et les accouplements chez Vipera aspis: étude en milieu
naturel. Terre et Vie 54 : 375-391.
Van Noordwijk AJ & De Jong G (1986) Acquisition and allocation of resources:
their influence on variation in life-history tactics. American Naturalist 128 : 137-
142.
Villevieille S (1997) Application de la résonance magnétique nucléaire pour
localiser et quantifier les réserves lipidiques chez un reptile vivipare, Vipera aspis.
Unpublished D.E.A. manuscript, Université François Rabelais, Tours.
Vitt LJ & Congdon JD (1978) Body shape, reproductive effort, and relative clutch
mass in lizards: resolution of a paradox. American Naturalist 112 : 595-608.
Vitt LJ & Price HJ (1982) Ecological and evolutionary determinants of relative
clutch mass in lizards. Herpetologica 38 : 237-255.
W.
Weeks SC (1996) The hidden cost of reproduction : reduced food intake caused
by spatial constraints in the body cavity. Oikos 75 : 345-349.
Whittier JM & Crews D (1990) Body mass and reproduction in female red-sided
garter snakes (Thamnophis sirtalis parietalis). Herpetologica 46 : 219-226.
Wilbur HM & Morin PJ (1988) Life history evolution in turtles. In : Gans C & Huey
RB (eds) Biology of the Reptilia 16, pp 387-439. Alan R Liss, INC, New York.
Williams GC (1966a) Adaptation and Natural Selection : A Critique of Some
Current Evolutionary Thought. Princeton University Press, New Jersey, USA.
Williams GC (1966b) Natural selection, the costs of reproduction, and a
refinement of Lack's principle. American Naturalist 100 : 687-690.
Winkel W & Winkel D (1995) Costs and benefits of second broods in coal tits
(Parus ater). Journal of Ornithology 136 : 29-36.
300

Winkler DW & Wallin K (1987) Offspring size and number : a life history model
linking effort per offspring and total effort. American Naturalist 129 : 708-720.
Y.
Young TP (1990) Evolution of semelparity in Mount Kenya lobelias. Evolutionnary
Ecology 4 : 157-171.
Young TP & Auspurger CK (1991) Ecology and evolution of long-lived
semelparous plants. Trends in Ecology and Evolution 6 : 285-289.
Z.
Zar JH (1984) Biostatistical Analysis. Prentice-Hall, New Jersey.
Zari TA (1991) The influence of body mass and temperature on the standard
metabolic rate of the herbivourous desert lizard, Uromastyx microlepis. Journal of
Thermal Biology 16 : 129-133.
Zuffi MAL, Giudici F & Iolae P (1999) Frequency of reproduction and reproductive
effort in female Vipera aspis from a southern population. Acta Oecologica 20 :
633-638.
Zuffi MAL (2002) A critic of the systematic position of the aspic viper subspecies
Vipera aspis aspis (Linnaeus 1758), Vipera aspis atra Meisner, 1820, Vipera
aspis francisciredi Laurenti, 1768, Vipera aspis hugyi Schinz, 1833 and Vipera
aspis zinnikeri Kramer, 1958. Amphibia-Reptilia 23 : 191-213.
301

ANNEXE
Fluctuating climatic conditions affect embryonic
development in a viviparous snake (Vipera aspis)
OLIVIER LOURDAIS 1,2 , RICHARD SHINE 3 , XAVIER BONNET 1, 3 , MICHAËL
GUILLON 1 AND GUY NAULLEAU 1
1 Centre d'Etudes Biologiques de Chizé, CNRS, 79360, Villiers en Bois, France
2 Conseil Général des Deux Sèvres, rue de l‘abreuvoir, 79000 Niort, France
3 School of Biological Sciences, University of Sydney, Australia
Corresponding author: Olivier Lourdais,
CEBC-CNRS, 79360, Villiers en Bois, FRANCE
Tel: (33) 5 49 09 78 79
Fax: (33) 5 49 09 65 26
E-mail: lourdais@cebc.cnrs.fr
For consideration in: Ecology
Table: 3, Figures: 4
302

Abstract. Climatic conditions during embryonic development can exert profound and
long-term effects on many types of organisms, but most previous research on this
topic has focussed on endothermic vertebrates (birds and mammals). Although
viviparity in ectothermic taxa allows the reproducing female to buffer ambient thermal
variation for her developing offspring, even an actively thermoregulating female may
be unable to provide optimal incubation regimes in severe weather conditions. We
examined the extent to which fluctuations in natural thermal conditions during
pregnancy affect reproduction in a temperate viviparous snake, the aspic viper
(Vipera aspis). Data gathered from a long term field study demonstrated that
ambient thermal conditions influenced (1) female body temperatures and (2)
gestation length, embryo viability, and offspring phenotypes. Interestingly, thermal
conditions during each of the three months of gestation affected different aspects of
reproduction. Hotter weather early in gestation (June) increased ventral scale counts
(= number of body segments) of neonates; hotter weather mid-gestation (July)
hastened development and thus the date of parturition; and hotter weather late in
gestation (August) reduced the incidence of stillborn neonates. The population that
we studied is close to the northern limit of the species’ range, and embryonic thermal
requirements may prevent Vipera aspis from extending into cooler conditions further
north.
Key words: viviparity, snakes, ectothermy, temperature, development
303

INTRODUCTION
The environment affects living organisms in many ways. Influences that occur early
in an organism’s life – especially, while it is an embryo – typically have greater effects
on its subsequent development than do influences that occur later in life (Henry and
Ulijaszek 1996, Desai and Hales 1997). Recent research has documented many
strong and persistent effects of environmental factors acting during embryogenesis
on fitness-related traits (Lindström 1999, Lummaa and Clutton-Brock 2002). Most
studies on this topic have been based on birds and mammals, reflecting the
concentration of long-term individual-based studies on these taxa (Lindström 1999).
However, we also need data on other kinds of animals if we are to discern valid
generalities about effects of the early environment on subsequent phenotypic traits.
Ectothermic vertebrates are of particular interest in this respect. Because they
contain both oviparous and viviparous taxa, often without the confounding influence
of post-hatching parental care (Clutton-Brock 1991), they allow us to examine the
degree to which alternative modes of reproduction buffer the developing offspring
from environmental fluctuations.
Ambient temperatures not only fluctuate considerably through time, but they
also influence ectothermic vertebrates in many ways. The most obvious influences
concern variables such as the metabolic rates, activity levels and locomotor
performance of adult animals (Huey 1982, Hertz, Huey and Stevenson 1993).
However, temperature also influences the rates and trajectories of ontogenetic
development, and embryonic sensitivity to ambient temperature may be an important
(albeit, less-studied) component of ectothermic biology. For example, extensive
studies on oviparous reptiles show that the rate of embryonic development depends
upon thermal regimes inside the nest. High temperatures greatly accelerate
304

embryogenesis and thus, shorten the incubation period (Hubert 1985). Thermal
regimes during embryogenesis can also profoundly affect phenotypic traits of the
offspring, such as body size, scalation, and locomotor performance (Shine and
Harlow 1996, Shine et al. 1997, Downes and Shine 1999, Andrews and Mathies
2000, Flatt et al. 2001, Shine and Elphick 2001, Webb, Brown and Shine 2001). In
some taxa, incubation temperatures directly determine the sex of offspring (Bull
1980).
The sensitivity of reptilian embryogenesis to ambient temperature suggests
that thermal conditions may play a major role in the population ecology of these
animals. For example, thermal minima for embryonic development may constrain
oviparous species from reproducing in cold climates (Mell 1929, Weekes 1933, Tinkle
and Gibbons 1977). More generally, geographic distributions of oviparous species
may be set by the thermal requirements for embryogenesis (Shine 1987). Plausibly,
annual variation in climatic conditions may influence the attributes of offspring (size,
shape, time of hatching or birth) that enter the population each year, and thus
influence the relative size of different year-classes. However, we are not aware of
any data to show such an effect, apart from anecdotal reports of delayed oviposition,
hatching or parturition in unusually cold years (e.g., Saint Girons 1952, Pengilley
1972, Olsson and Shine 1997). In contrast, a wide range of studies on endothermic
vertebrates (especially mammals) have not only documented effects of climatic
variation on neonatal phenotypes, but also have shown that the resulting effects have
long-term consequences for survival and reproductive success of individuals from
those cohorts (e.g., Albon, Guinness and Clutton-Brock 1983, Post et al. 1997,
Lummaa and Clutton-Brock 2002).
The influence of ambient thermal fluctuations on hatching dates and offspring
phenotypes is likely to be obvious in oviparous species of reptiles, because females
305

in such taxa cannot control the thermal conditions experienced by their offspring
throughout post-oviposition development (except in cases of maternal egg-brooding:
e.g., Vinegar 1973). We might expect that this sensitivity to ambient thermal regimes
would be much lower in a viviparous (live-bearing) species. Viviparity has evolved
>100 times within squamate reptiles and this transition seems to have occurred
primarily in cold climates (Shine 1985, Blackburn 1985, 1999). The probable
selective force for these repeated transitions has been the egg-retaining female’s
ability to maintain high, relatively constant incubation temperatures for her developing
offspring (Shine 1985). Because the gravid female can regulate her temperature
behaviorally, moving among microhabitats to exploit thermal heterogeneity in the
environment, the temperatures experienced by an offspring developing in utero will
be buffered considerably from fluctuations in ambient temperature (Shine 1983a,
Burger and Zappalorti 1988, Charland and Gregory 1990, Schwarzkopf and Shine
1991, Peterson, Gibson and Dorcas 1993).
Nonetheless, if ambient thermal conditions fluctuate considerably, even a
carefully-thermoregulating viviparous female may be unable to maintain high,
constant temperatures for her developing offspring. In keeping with this inference,
laboratory studies that have manipulated basking opportunities for viviparous lizards
have found many of the same phenomena as described above for egg-layers. That
is, a viviparous female reptile’s access to basking opportunities not only determines
rates of embryogenesis (and thus, the duration of her pregnancy: Naulleau 1986,
Schwarzkopf and Shine 1991) but also affects many phenotypic traits of her offspring
(Shine and Harlow 1993, Swain and Jones 2000, Wapstra 2000, Arnold and
Peterson 2002). Unfortunately, the relevance of these results to free-ranging animals
remains unknown.
306

Although many authors have stressed the ability of gravid females in
viviparous reptile species to provide relatively high stable temperatures for their
developing offspring (see above references), less attention has been paid to
situations where they may be unable to do so. That is, do severely cold or rapidly
fluctuating weather conditions make it impossible for even an actively
thermoregulating viviparous female to provide an effective thermal buffer for her
offspring? In such a situation, we might expect to see the developmental rates and
phenotypic traits of offspring respond to annual variation in weather conditions,
despite the thermoregulatory efforts of their mothers. Such effects should be
especially important for individuals living at the altitudinal or latitudinal margin of the
geographic range of a species. We examined this possibility with data from a nine-
year study of a free-ranging population of viviparous snakes at the extreme northern
limit of the species’ range.
MATERIAL AND METHODS
Study Animals
The aspic viper, Vipera aspis Linné, is a small viviparous snake of the western-
Paleartic region and is locally abundant at the northern limit of its distribution in
France. Females mature at 40 cm snout-vent length (SVL), which is attained in 2.5
to 3.5 years (Bonnet et al. 1999a). Ovulation typically occurs during the first two
weeks of June with limited geographical variation (Saint 1957b, Naulleau 1981).
During gestation pregnant females display higher thermal preferenda and
substantially increase basking times (Saint Girons 1952, Naulleau 1979, Bonnet and
Naulleau 1996, Lourdais et al. 2003b, Ladyman et al. 2003). Parturition occurs two
to three months after ovulation, from late August through late September.
307

Study site and methods
The study site is near the village of Les Moutiers en Retz in west-central France
(47 o 03N'; 02 o 00W'). It is a 33-hectare grove with a mosaic of meadows and
regenerating scrubland. Details on the field site, methods and searching effort are
available in related works (Bonnet et al. 2000, 2001,2002). Previous results
suggest that climatic conditions in this area not only prolong gestation by one to two
months compared to warmer-climate (Mediterranean) populations, but also that the
magnitude of such effects varies among years (Lourdais et al. 2002a).
Gravid females were captured and maintained in captivity after the first
parturition of the year was recorded in the field (generally in late August).
Reproductive data were then obtained on 173 litters from 149 different females. For
most individuals (127) only a single litter was obtained, but 17 females produced two
litters and 4 individuals produced three litters.
The components of the litter were characterized (undeveloped ova, dead
embryo, fully-developed but stillborn, healthy offspring), counted, and weighed
(±0.1g). Young were measured (±0.5cm) and sexed. Stillborn offspring were
measured, weighed, and sexed when possible. Because we could not distinguish
unfertilized ova from ova that had been fertilized but had died early in
embryogenesis, these were grouped in the same category (undeveloped ova).
Using this method we gathered data on 817 healthy offspring, 132 undeveloped ova,
22 dead embryos and 78 stillborn offspring. From 1993 to 2000, ventral scales were
counted for 136 mothers and 681 healthy neonates. Gestation period was calculated
from parturition dates, assuming a fixed ovulation date of 10 June (Saint Girons
1980, Naulleau 1981).
308

Thermal conditions
In our study area, climatic conditions constrain many aspects of aspic viper ecology
(Lourdais et al 2002a). These snakes are diurnal with prolonged basking episodes.
Previous studies have revealed substantial daily variation in body temperatures, with
low overnight temperatures followed by basking to attain and maintain high body
temperatures during daylight hours. Pregnant females bask much more than do non-
pregnant animal, and thus are more often encountered and captured (Bonnet and
Naulleau 1996, Lourdais et al. 2002b). Thermal conditions fluctuate strongly from
one day to the next in this temperate-oceanic climate. Does such variation affect the
body temperatures of vipers, despite the buffering effects of behavioral
thermoregulation? To answer this question we need measures of both ambient
temperatures and viper body temperatures:
1) Ambient temperatures
As an index of ambient temperature we used daily maximum shaded air temperature,
as measured in a standard metereological shelter 1.8 m above the ground in Pornic
(47 o 06N'; 02 o 07W'), near our field site (47 o 03N'; 02 o 00W'). These daily maxima will
not necessarily have any close relationship to the actual temperatures experienced
by an embryonic viper inside its mother’s uterus, but instead should provide a rough
index of a viper’s opportunity for behavioral thermoregulation. Thermal maxima were
generally achieved in the afternoon, one to two hours after sun zenith.
Aspic vipers have a long period of embryonic development (up to three
months: Hubert and Dufaure 1968) and we distinguished three periods broadly
corresponding to major steps in embryogenesis (Hubert 1985, Hubert, Dufaure and
Collin 1966, Hubert and Dufaure 1968): (1) Early gestation (10 to 30 June), the onset
of embryogenesis (including blastulation, gastrulation, neurulation, somite
309

development, differentiation of the head and circulatory system); (2) Mid-gestation (1-
31 July), a period of organogenesis (development of the optic vesicles and olfactory
bulb, appearance of the jaws, cloacal split, genitalia and trunk/tail scales; rapid
growth of the embryo and development of spiral-coiling); and (3) Late-gestation (1-
31 August), a period of embryonic growth and completion of development (including
development of pigmentation, and differentiation of head scales).
In our analysis, we calculated mean daily temperature maxima during each of
these three periods of gestation for each year. We investigated the relative influence
of each period by regressing these mean temperatures against the duration of
gestation (as estimated from dates of birth: see above). Then, we examined the
effect of thermal conditions during development on offspring phenotype (scalation).
Finally, we examined the possibility that embryo mortality rates might be influenced
by thermal conditions during gestation.
2) Body temperatures of free-ranging vipers
Using internal temperature radiotransmitters (See Naulleau, Bonnet and Duret 1996
for the method), we monitored reproductive and non-reproductive female vipers
during the gestation period (1 June to 31 August 1996). Females were sampled one
to four times per day from 0800 to 2130 h, but to minimise temporal heterogeneity
and pseudoreplication our analyses were based only on a single late-afternoon (1700
– 20000 h) data point per female per day. Because of the 24-hour delay between
successive readings (during which time the snakes’ body temperatures dropped to
minimum levels overnight: Naulleau 1997, pers obs), we have treated successive
daily temperatures from the same individual as quasi-independent. Using this
procedure 241 temperature records from 16 female Aspic vipers (9 reproductive and
7 non-reproductive) were available (average number of records per individual
310

=16±7). The influence of daily maxima on female body temperature was examined
after accounting for female identity and reproductive status.
Statistics
All statistics were performed with Statistica 6.0. Influences of ambient thermal
conditions on female body temperature were examined using general linear
regression modeling (GLM). Reproductive status was considered as a fixed factor.
Female identity was treated as a random factor, nested within reproductive status.
Daily temperature maxima or sampling dates (Julian calendar) were then treated as
covariates (fixed effects). The influences of thermal conditions on offspring
phenotype were investigated using mixed-model ANCOVAS. To account for
correlated responses among offspring of individual litters and repeated female
contributions (17 females reproduced twice and 4 three times), maternal identity was
included as a random factor. Neonatal traits were the dependent variable, offspring
sex was a fixed factor and annual thermal conditions were treated as covariates
(fixed effects).
RESULTS
I) Determinants of female body temperatures
GLM analysis suggested that female body temperature was affected by at least three
factors (see table 1). First, our use of air temperature as an index of
thermoregulatory opportunities was validated by a significant influence of maximum
air temperature on female body temperature (table 1). Second, a significant
relationship between sampling date and daily thermal maxima reflected an increase
in ambient temperatures over the summer period. Third, reproductive status also
exerted a strong influence on female body temperature with pregnant females
311

maintaining body temperatures than did non-reproductive females (28.7±0.5°C, n
=133 versus 25.47±0.6°C, n =117; see table 1).
II) Fluctuations in thermal conditions
A repeated measure ANOVA (using year as the factor and gestation month within
each year as the repeated factor) first indicated that mean daily air temperature
significantly increased over the course of gestation (F(2, 522)=45.33; p < 0.00001)
and more importantly, that mean air temperature during gestation varied significantly
among years (F(8, 261)=3.33; p < 0.0012). We also detected a significant interaction
between gestation month and year (F(16, 738) = 4.53; p < 0.0001), reflecting
marked year to year fluctuations in thermal conditions over the three months of
gestation (Fig. 1). During the study period, thermal conditions during gestation fell
broadly into three patterns: parabolic (where the highest temperature was in July),
sigmoid (where the predominant change was a major increase in temperature
between June and July), and exponential (where the predominant change was a
major temperature increase between July and August).
As a consequence, we found no significant correlation between monthly mean
daily temperatures across years (F(1, 7) = 1.52, n = 9, p = 0.25 for June versus July;
F(1, 7) = 3.11, n = 9, p = 0.12 for June versus August; F(1, 7) = 0.01, n = 9, p = 0.93
for July versus August). Therefore, we consider mean temperatures during each of
the three months as independent variables for our subsequent analyses.
1) Duration of gestation period
III) Impact on reproduction
In a related study (Lourdais et al. 2002a), we showed that the duration of gestation
in this population was influenced by mean temperatures during the gestation period,
312

as well as by the frequencies of inviable elements in litters (i.e., undeveloped ova and
stillborn offspring). For the present study, we can examine this result more closely in
terms of the three phases of gestation defined above. We used stepwise multiple
regression analysis for this purpose, and restricted the analysis to the 80 females that
produced only viable offspring (i.e., no stillborns). Only mean daily temperatures
during July (mid-gestation) were retained in the model (see Table 1), accounting for
51.4% of the variance in overall gestation length.
2) Offspring scalation
First, we detected a strong influence of maternal identity on the number of ventral
scales in newborn vipers (ANOVA, F(101, 527) = 4.13, p=0.00001). This influence
was partially attributable to significant heritability of ventral scalation, as we detected
a significant relationship between maternal and new-born number of ventral scales (r
2 =0.11; F(1, 629) = 80.81; p < 0.0001 treating each offspring as an individual point
and r 2 =0.22; F(1, 114) = 34.64; p < 0.0001 when considering mean offspring number
of ventral scales per litter). The two sexes differed in mean numbers of ventral
scales, with neonatal females having more scales than their brothers (149.1±0.17,
n=329 versus 148.0±0.17, n=352, F(1, 563) = 11.37; p < 0.0007, mixed model
ANOVA using female identity as random factor and offspring sex as a fixed factor).
Interestingly, we detected significant year to year variation in the number of ventral
scales in neonatal snakes (ANOVA, F(7, 673)=12.400, p=0.00001, see figure 2). This
effect holds true even after accounting for maternal influence and offspring sex (F(7,
114)=3.08, 0.0004, mixed model ANOVA year using female identity as a random
factor, offspring sex and year as a fixed factors). Such annual variations appear to
be linked to the climatic fluctuations described above. For instance we detected a
significant influence of mean gestation thermal maxima on neonate number of ventral
313

scales (F(1, 114)=6.94, p=0.008, mixed model ANCOVA using female identity as
random factor, offspring sex as a fixed factor and mean gestation thermal maxima as
a fixed covariate). Because organizational effects of temperature are likely to occur
early in embryogenesis, we re-conducted the analysis by considering each of the
three components of the gestation period independently. Only mean daily
temperature maxima during the first period (i.e., the three weeks following ovulation)
led to significant results (Table 3, Fig. 3). This influence is reflected in a significant
relationship between mean thermal maxima in June and the mean number of ventral
scales in neonatal vipers (r 2 =0.49; F(1,6)=7.75 p

(χ 2 = 0.24; dl = 1; p = 0.62, pooling the nine years of the study), nor did sex ratios of
stillborn offspring vary significantly among years (χ 2 =10.27; dl = 8; p = 0.24).
Based on these results, we looked for a possible influence of mean
temperature during the latter part of gestation (July and August) on the probability of
producing stillborn offspring. Excluding females producing undeveloped ova, we
detected a significant negative influence of mean August temperature on the
probability of observing late embryonic death (Logistic regression, χ 2 = 8.18; n =
113; p = 0.0042). We also detected a significant negative relationship between mean
August temperatures and the proportion of stillborn offspring (r = 0.33; r 2 = 0.11; n =
113; F(1, 111) = 13.37; p < 0.0004). The same analysis using mean July or mean
June daily temperatures yielded non significant results.
DISCUSSION
Our relatively long-term field study demonstrates that natural climatic conditions
influence important aspects of embryogenesis in the aspic viper. Although gravid
vipers show distinctive thermoregulatory behaviors that result in relatively high, stable
body temperatures throughout pregnancy (Saint Girons 1952, Naulleau 1979, Bonnet
and Naulleau 1996, Ladyman et al. 2003), they are unable to completely buffer their
developing embryos from year-to-year thermal variations in this relatively northern,
cool-climate area. This result runs counter to the primary emphasis of published
studies on thermoregulation by gravid reptiles, which have stressed the
thermoregulatory precision of such animals (e.g., Shine 1983a, Charland and
Gregory 1990, Schwarzkopf and Shine 1991, Peterson, Gibson and Dorcas 1993).
Clearly, this stenothermy is relative: even if gravid females maintain higher
temperatures than do non-reproductive conspecifics, they may still vary enough in
body temperatures to impact on the embryos.
315

Unsurprisingly, high summer temperatures resulted in faster embryonic
development and thus, earlier parturition dates. Thermal dependence of embryonic
development is widespread in squamate reptiles (Blanchard and Blanchard 1941,
Hubert 1985) and our results in this respect are consistent with an earlier
experimental study conducted by Naulleau (1986) on this species and the closely
related adder (Vipera berus). However, our data extend previous knowledge of this
phenomenon in suggesting that the thermal sensitivity of gestation length seems to
be significant mostly (or only) in the middle part of gestation. In our multiple-
regression analyses, only July mean temperature significantly accelerated gestation.
This period coincides with major steps in embryogenesis (i.e., organogenesis and
active tissue synthesis). Recent studies on oviparous lizards reported that rates of
embryogenesis were most sensitive to incubation temperature relatively soon after
oviposition (Shine and Elphick 2001, Shine 2002). Given that the lizards involved in
that study lay their eggs when embryos are almost one-third through their total
developmental period (Shine 1983b), this result fits well with those from our own
study. The relationship between the thermosensitivity of gestation length and specific
embryonic stages warrants further study, to test this apparent generality.
Environmental conditions also had direct effects on the phenotypic traits of
offspring. In keeping with laboratory studies that have manipulated basking
opportunities for viviparous female reptiles and documented shifts in offspring
phenotype as a result (Shine & Harlow 1993, Shine & Downes 1999, Swain & Jones
2000, Wapstra 2000, Arnold and Peterson 2002), we found significant correlations
between ambient temperatures and phenotypic traits of the neonatal vipers. In
contrast with the experimental study of Arnold & Peterson (2002) that documented a
flat reaction norm for scale count in another viviparous species (Thamnophis
elegans), we detected a significant impact of temperature on ventral scalation in new
316

orn aspic vipers. Notably, mean temperature during early stages in embryogenesis
affected the number of ventral scales in offspring, with higher temperatures
increasing scale numbers. Similar influences of developmental temperature on
scalation have been reported from laboratory experiments on reptiles (Vinegar
1973,1974; Osypra & Arnold 2000), and inferred from climate-correlated geographic
shifts in scale counts (Klauber 1941). In most snakes, the number of ventral scales is
tightly correlated with the number of vertebrae, reflecting the number of pairs of
somites differentiated during early embryogenesis (Hubert 1985; Lindell 1996). The
number of ventral scales or body vertebrae shows considerable intraspecific variation
(Lindell 1996, Lindell, Forsman and Merilä 1993). If temperature influences
somitisation, we would expect that differences in scalation (reflecting vertebral
number) will be correlated with differences in body length. The proximal relationship
between those two traits was supported in our data set by a significant relationship
between ventral scalation and offspring snout-vent length (r=0.20; F(1,679)=27.978;
p

performance (Arnold 1988, Arnold and Bennett 1988). Hence, the influence of
natural climatic conditions on embryo development may affect the quality of offspring
produced by a female viper.
Finally, weather conditions affected neonatal fitness directly by influencing
rates of embryo mortality. Years with cool weather late in summer, close to the end
of gestation (August) resulted in a high incidence of stillborn offspring. This result
suggests that embryos may be particularly sensitive to low temperatures late in
development, a pattern previously reported in a field study on an oviparous species
(Burger and Zappalorti 1988). As our study population is close to the species’
northern range limit (Stewart 1971), the sensitivity of offspring development to
ambient weather conditions may be a direct result of climatic constraints on female
thermoregulation. In more favorable environments (e.g., southern populations),
viviparity may well allow female aspic vipers to selectively alter incubation conditions
and thus provide optimal incubation regimes for their offspring.
The ecological and evolutionary significance of such interactions among the
environment, female thermoregulatory behavior, and offspring phenotype is a
complex issue that requires further work, notably comparative studies of populations
facing different climatic conditions. The high levels of embryo mortality detected in
our study probably reflect the location of Les Moutiers at the northern limit of the
geographic range of the species. In this area, female aspic vipers experience high
survival costs of reproduction and most females reproduce only once in their lifetime
(Bonnet et al. 1999b, 2002). Further north the aspic viper is replaced by a sister
species, the adder (Vipera berus), with limited overlap in the distribution of the two
species (Saint Girons 1975). While similar in size and appearance, these two vipers
diverge in metabolic rates and in thermal requirements for digestion and gestation.
Both are lower in V. berus than in V. aspis, allowing the former species to penetrate
318

into cooler, more northern areas (Naulleau 1983, 1986 Saint Girons, Naulleau and
Célérier 1985). Our analyses suggest that the thermal optima for embryonic
development could also constrain the geographic distribution of V. aspis (as
suggested for oviparous squamates by Shine 1987 and Shine, Barrot and Elphick
2003).
In conclusion, we found that in this northern population of snakes, natural
thermal conditions significantly affected embryonic development despite active
maternal thermoregulation. Our results underline the importance and complexity of
ambient thermal influences on the lives of ectothermic vertebrates. Comparative
studies with southern populations facing a less constraining environment would be of
great interest. In addition, experimental examination of female thermoregulatory
behavior during particular thermosensitive phases (such as early embryogenesis) are
needed to clarify to the extent to which viviparity permits active maternal manipulation
of offspring phenotypes.
Acknowledgements
We thank Dale DeNardo for helpful comments on the manuscript. Financial support
was provided by the Conseil Général des Deux Sèvres, le Centre National de la
Recherche Scientifique (France).
Literature Cited
Albon, S.D., Guinness, F.E. and Clutton-Brock, T.H. 1983. The influence of climatic
variation on the birth weights of red deer (Cervus elephas). Journal of Zoology
(London), 200: 295-297.
Andrews, R.M. and Mathies, T. 2000. Natural history of reptilian development:
constraints on the evolution of viviparity. Bioscience, 50: 227-238.
Arnold, S.J. 1988. Quantitative genetics and selection in natural populations:
microevolution of vertebral numbers in the garter snake Thamnophis elegans.
Proceedings of the Second International Conference on Quantitative Genetics.
319

(Eds Weir, B.S., Goodman, M.M., Eisen, E.J. and Namkong, G.) pp. 619-636.
(Sinauer Assoc.: Sunderland, Mass.)
Arnold, S.J. and Bennett, A.F. 1988. Behavioural variation in natural populations. V.
Morphological correlates of locomotion in the garter snake Thamnophis radix.
Biological Journal of the Linnean Society, 34: 175-190.
Arnold, S.J. and Peterson, C.R. 2002. A model for optimal reaction norm : the case of
the pregnant garter snake and her temperature-sensitive embryos. The
American Naturalist 160: 307-316.
Blanchard, F.N. and Blanchard, F.C. 1941. Factors determining time of birth in
garter snake. Pap. Michigan Acad. Sci., 26: 161-176.
Blackburn, D.G. 1985. Evolutionary origins of viviparity in the Reptilia. II. Serpentes,
Amphisbaenia, and Icthyosauria. Amphibia-Reptilia, 6: 259-291.
Blackburn, D.G. 1999. Are viviparity and egg-guarding evolutionarily labile in
squamates? Herpetologica, 55: 556-573.
Bonnet, X. and Naulleau, G. 1996. Catchability in snakes : consequences on
breeding frequency estimates. Canadian Journal of Zoology 74 : 233-239.
Bonnet, X., Naulleau, G., Lourdais, O. and Vacher-Vallas, M. 1999a. Growth in the
asp viper (Vipera aspis L.) : insights from long term field study. In : Miaud C and
Guyetant R (eds) Current Studies in Herpetology, pp 63-69. SEH, Le Bourget
du Lac.
Bonnet, X., Naulleau, G., Shine, R. and Lourdais, O. 1999b. What is the appropriate
time scale for measuring costs of reproduction in a capital breeder?
Evolutionary Ecology 13 : 485-497.
Bonnet, X., Naulleau, G., Shine, R. and Lourdais, O. 2000. Reproductive versus
ecological advantages to larger body size in female Vipera aspis. Oikos 89 :
509-518
Bonnet, X., Naulleau, G., Shine, R. and Lourdais, O. 2001. Short-term versus long-
term effects of food intake on reproductive output in a viviparous snake, Vipera
aspis. Oikos 92: 297-308.
Bonnet, X., Lourdais, O., Shine, R. and Naulleau, G., 2002. Reproduction in snakes
(Vipera aspis): costs, currencies and complications. Ecology 83, 2124-2135.
Bull, J. J. 1980. Sex determination in reptiles. Quarterly Review in Biology, 55: 4-
21.
320

Burger, J. and Zappalorti R.T. 1988. Effects of incubation temperature on sex ratios
in pine snakes: differential vulnerability of males and females. The American
Naturalist, 132: 492-505.
Charland, M.B. and Gregory, P.T. 1990. The influence of female reproductive
status on thermoregulation in a viviparous snake, Crotalus viridis. Copeia,
1990: 1089-1098.
Clutton-Brock, T.H. 1991. The evolution of parental care. Princeton University Press,
Princeton, New Jersey.
Desai, M. and Hales, C.N. 1997. Role of fetal and infant growth in programming
metabolism in later life. Biological Reviews, 72: 329-342.
Downes, S.J. and Shine, R. 1999. Do incubation-induced changes in a lizard’s
phenotype influence its vulnerability to predators? Oecologia, 120: 9-18.
Flatt, T., Shine, R., Borges-Landaez, P.A. and Downes, S.J. 2001. Phenotypic
variation in an oviparous montane lizard (Bassiana duperreyi): the effects of
thermal and hydric incubation environments. Biological Journal of the Linnean
Society, 74: 339-350.
Forsman, A. 1993. Survival in relation to body size and growth rate in the adder,
Vipera berus. Journal of Animal Ecology, 62: 647-655.
Forsman, A. and Lindell, L. E. 1993. The advantage of a big head: swallowing
performance in adders, Vipera berus. Functional Ecology, 7: 183-189.
Forsman, A. and Shine, R. 1997. Rejection of non-adaptive hypotheses for
intraspecific variation in trophic morphology in gape-limited predators.
Biological Journal of the Linnean Society, 62: 209– 223.
Henry, C.J.K. and Ulijaszek, S.J. 1996. Long-term consequences of early
environment. Cambridge Univ. Press, Cambridge.
Hertz, P.E., Huey, R.B., and Stevenson, R.D. 1993. Evaluating temperature
regulation by field-active ectotherms: the fallacy of the inappropriate question.
American Naturalist, 142: 796-818.
Hubert. J. 1985. Embryology of the squamata. Biology of the Reptilia 15. (Eds Gans,
C. and Billet, F) pp. 1-55. (Wiley and sons: New York)
Hubert, J. and Dufaure, J. P. 1968. Table de développement de la vipère aspic:
Vipera aspis L. Bulletin de la Société Zoologique de France, 93: 135-148.
Hubert, J., Dufaure, J.P. and Collin, J.P. 1966. Matériaux pour une table de
321

développement de Vipera aspis L., I. La période de l’Organogénèse. Bulletin
de la Société Zoologique de France, 91: 779-788.
Huey, R.B. 1982. Temperature, physiology and the ecology of reptiles. Biology of the
Reptilia 12. (Eds Gans, C. and Pough, F.H.) pp. 25-91. (Academic Press:
London.)
Klauber, L.M. 1941. The correlation between scalation and life zones in San Diego
County snakes. Bull. Zool. Soc. San Diego, 17: 73-79.
Lindell, L. E. 1996. Vertebral number in adders, Vipera berus: direct and indirect
effects on growth. Biological journal of the Linnean society, 59: 69-85.
Lindell, L. E., Forsman, A. and Merilä, J. 1993. Variations in number of ventral
scales in snakes: effects on body size, growth rate and survival in the adder,
Vipera berus. Journal of Zoology, London, 230: 101-115.
Lindström, J. 1999. Early development and fitness in birds and mammals. Trends in
Ecology and Evolution, 14: 343-348.
Lourdais, O., Bonnet, X. and Doughty, P. 2002. Costs of anorexia during pregnancy
in a viviparous snake (Vipera aspis). Journal of Experimental Zoology, 292:
487-493.
Lourdais, O., Bonnet, X., Shine, R., DeNardo, D., Naulleau, G. and Guillon, M.
2002a. Capital-breeding and reproductive effort in a variable environment: a
longitudinal study of a viviparous snake, Journal of Animal Ecology, 71: 470-
479.
Lourdais, O., Bonnet, X., DeNardo, D. and Naulleau, G., 2002b. Do sex divergences
in reproductive ecophysiology translate into dimorphic demographic patterns?
Population Ecology 44 3: 241-249.
Lummaa, V. and Clutton-Brock, T. 2002. Early development, survival and
reproduction in humans. Trends in Ecology and Evolution, 17: 141-147.
Mell, R. 1929. Beitrage zur Fauna Sinica. IV. Grundzuge einer Okologie der
chinesischen Reptilien und einer herpetologischen Tiergeographie Chinas.
(Walter de Gruyter: Berlin.)
Mushinsky, H. R. 1987. Foraging ecology. Snakes – Ecology and Evolutionary
Biology (eds. Seigel, R. A., Collins, J. T., and Novak S. S.). pp. 302-334.
Mazmilan Publishing Company, New York.
Naulleau, G. 1971. Fertility of female Vipera aspis in captivity as a function of the
periods of mating. Herpetologica, 27: 385-389.
322

Naulleau, G. 1979. Etude biotélémétrique de la thermoregulation chez Vipera aspis
(L.) Elevée en conditions artificielle. Journal on Herpetology 13(2): 203-208.
Naulleau, G. 1981. Détermination des périodes d’ovulation chez Vipera aspis et
Vipera berus dans l’ouest de la France, étudiée par radiographie. Bulletin de la
Société des Sciences Naturelles de l’Ouest de la France, 3: 151-153.
Naulleau, G. 1983. Action de la température sur la digestion chez cinq espèces de
vipères européennes du genre Vipera. Bulletin de La Société Zoologique de
France, 108: 47-58.
Naulleau, G. 1986. Effect of temperature on gestation in Vipera aspic and V. berus
(Reptilia: Serpentes). Studies in herpetology, Roček Z. Prague: 489-494.
Naulleau, G. 1997. La vipère aspic. Eveil Nature, Saint Yrieix.
Olsson, M. and Shine, R. 1997. The seasonal timing of oviposition in sand lizards
(Lacerta agilis): why earlier clutches are better. Journal of Evolutionary Biology,
10: 369-381.
Osypka, N.M. and Arnold, S.J. 2000. The developmental effects of sex ratio on a
sexually dimorphic scale count in the garter snake Thamnophis elegans.
Journal of Herpetology, 34: 1-5.
Post, E., Stenseth, N.C., Langvaten, R. and Fromentin, J. 1997. Global climate
change and phenotypic variation among red deer cohorts. Proceedings of the
Royal Society of London, Series B, 264: 1317-1324.
Pengilley, R. 1972. Systematic relationships and ecology of some lygosomine lizards
from southeastern Australia. Ph.D. Dissertation. Unpubl. Thesis. Australian
National University.
Peterson, C.R., Gibson, A.R. and Dorcas, M.E. 1993. Snake thermal ecology: The
causes and consequences of body-temperature variation. Snakes: Ecology and
Behavior. (Eds Seigel, R.A. and Collins, J.T.) pp. 241-314. (McGraw-Hill: New
York.)
Saint Girons, H. 1952. Ecologie et éthologie des vipères de France. Annales des
Sciences Naturelles de Zoologie, 14: 263-343.
Saint Girons, H. 1957. Le cycle sexuelle chez Vipera aspis (L) dans l’ouest de la
France. Bulletin Biologique de France et de Belgique, 91: 284-350.
Saint Girons, H. 1975. Coexistence de Vipera aspis et de Vipera berus en Loire
Atlantique: Un problème de compétition interspécifique. Revue d’Ecologie
(Terre et Vie), 29: 590-613.
323

Saint Girons, H. 1980. Biogéographie et évolution des vipères européennes.
Compte Rendu des Scéances de la Société de biogéographie, 496: 146-172.
Saint Girons, H., Naulleau, G. and Célérier, M.L. 1985. Le métabolisme aérobie de
Vipera aspis et de la Vipera berus (Reptilia, Viperidae) acclimatées à des
conditions naturelles. Influence de la température et comparaison des deux
espèces. Amphibia-Reptilia, 6: 71-81.
Schwarzkopf, L. and Shine, R. 1991. Thermal biology of reproduction in viviparous
skinks, Eulamprus tympanum: Why do gravid females bask more? Oecologia,
88: 562-569.
Shine, R. 1983a. Reptilian viviparity in cold climates: testing the assumptions of an
evolutionary hypothesis. Oecologia, 57: 397-405.
Shine, R. 1983b. Reptilian reproductive modes: the oviparity-viviparity continuum.
Herpetologica, 39: 1-8.
Shine, R. 1985. The evolution of viviparity in reptiles: an ecological analysis. Biology
of the Reptilia, Volume 15. (Eds Gans, C. and Billett, F.) pp. 605-694. (John
Wiley and Sons: New York.)
Shine, R. 1987. Reproductive mode may determine geographic distributions in
Australian venomous snakes (Pseudechis, Elapidae). Oecologia, 71: 608-612.
Shine, R. 2002. Reconstructing an adaptationist scenario: what selective forces
favor the evolution of viviparity in montane reptiles? American Naturalist 160:
582-593.
Shine, R. and Downes, S.J. 1999. Can pregnant lizards adjust their offspring
phenotypes to environmental conditions? Oecologia, 119: 1-8.
Shine, R. and Elphick, M.J. 2001. The effect of short-term weather fluctuations on
temperatures inside lizard nests, and upon the phenotypic traits of hatchling
lizard. Biological Journal of the Linnean Society, 72: 555-565.
Shine, R. and Harlow, P. 1993. Maternal thermoregulation influences offspring
viability in a viviparous lizard. Oecologia, 96: 122-127.
Shine, R. and Harlow, P. 1996. Maternal manipulation of offspring phenotypes via
nest-site selection in an oviparous lizard. Ecology, 77: 1808-1817.
Shine, R., Barrot, E.G. and Elphick, M.J. 2003. Some like it hot: effects of forest
clearing on nest temperatures of montane reptiles. Ecology. In press
Shine, R., Madsen, T.R.L., Elphick, M. J. and Harlow, P.H. 1997. The influence of
nest temperatures and maternal brooding on hatchling phenotypes in water
324

pythons. Ecology, 78: 1713-1721.
Slip, D.J. and Shine, R. 1988. Reptilian endothermy: a field study of
thermoregulation by brooding diamond pythons. Journal of Zoology (London)
216: 367-378.
Stewart, J.W. 1971. The snakes of Europe. (Fairleigh Dickinson Univ. Press:
Rutherford).
Swain, R. and Jones, S.M. 2000. Maternal effects associated with gestation
conditions in a viviparous lizard, Niveoscincus metallicus. Herpetological
Monographs, 14: 432-440.
Tinkle, D. W. and Gibbons, J.W. 1977. The distribution and evolution of viviparity in
reptiles, Misc. Pub.Mus. Nat. Hist. Univ. Mich. 154:1-55.
Vinegar, A. 1973. The effects of temperature on the growth and development of
embryos of the Indian python, Python morulus (Reptilia: Serpentes: Boidae).
Copeia, 1973: 171-173.
Vinegar, A. 1974. Evolutionary implications of temperature induced anomalies of
development in snake embryos. Herpetologica, 30: 72-74.
Wapstra, E. 2000. Maternal basking opportunity affects juvenile phenotype in a
viviaprous lizard. Functional Ecology, 14: 345-352.
Webb, J.K., Brown, G.P. and Shine, R. 2001. Body size, locomotor speed and
antipredator behaviour in a tropical snake (Tropidonophis mairii, Colubridae):
the influence of incubation environments and genetics factors. Functional
Ecology, 15: 561-568.
Zuffi, M.A.L., Giudici, F. and Iolae, P. 1999. Frequency and reproductive effort of
reproduction in female Vipera aspis from a southern population. Acta
Oecologica, 20: 633-638.
325

Table 1. Determinants of female body temperature. Reproductive status
(STATUS) was considered as a fixed factor. Females identity (IDENTITY) was
considered as a random factor and was nested in the corresponding reproductive
status. Daily temperature maxima (MAX) and Julian sampling date (DATE) were then
treated as covariates (fixed effect).
Effect dl MS F p value
STATUS Fixed 1 831.82 29.36 0.0001
DATE Fixed 1 307.55 10.85 0.001
MAX Fixed 1 403.24 14.23 0.0002
IDENTITY Random 14 169.54 5.98 0.0001
Error 234
Table 2. Influence of mean temperatures during the three months of pregnancy on
the duration of gestation in female aspic vipers.
Multiple Regression r = 0.71; r² = 0.51; n = 80; F(3, 76) = 26.86; p < 0.0001
Bêta Partial correlation p value
June 0.23 0.18 0.11
July -0.77 0.62

Captions to figures
Figure 1. Annual variation in thermal conditions during the three months of
gestation. For simplicity, years were classified depending upon the thermal pattern
observed. (Jun:June; Jul:July; Aug: August)
Pattern 1 (parabolic): 1992 (open triangles down), 1994 (open circles), 1996 (open
squares), 1999 (open diamonds).
Pattern 2 (sigmoid): 1995 (open triangles up), 1997 (open triangles down).
Pattern 3 (exponential): 1993 (open squares), 1998 (open hexagons), 2000 (open
circles).
Figure 2. Annual variation in number of ventral scales in the offspring of aspic vipers
(± S.E.).
Figure 3. Relationship between mean June maxima and mean offspring number of
ventral scales over the course of the study.
Figure 4 Influence of mean June maxima (°C) on offspring number of ventral scales
(Offspring NVS) and body size (Snout vent length, cm).
327

Mean Temperatures
28
27
26
25
24
23
22
21
20
Figure 1
1 2 3
Jun Jul Aug Jun Jul Aug Jun Jul Aug
328

Number of Ventral scales
151
150
149
148
147
146
145
Figure 2
1993 1994 1995 1996 1997 1998 1999 2000
Year
329

Mean number of ventral scales
Figure 3
150
149
148
147
146
145
21.0 21.5 22.0 22.5 23.0 23.5 24.0
Mean June maxima
330

Figure 4
331