influence du climat et de la prédation sur l'utilisation de l'habitat et la ...

89
individual (using approximations described above) and the percentage of times the dominant
tree within 15 m was an aspen, fruit-bearing, deciduous, cedar or coni ferous tree.
Second, we tested whether juveniles differed in their use of coyer (at both scales) as weil
as in th eir use of den, ground and tree locations. F inally, when j uveniles showed significant
individual di fferences in use for a given variable, we utilized this variable of habi tat use in
survival analyses.
Sex, body mass, mobility and distance ta the mother
We sexedjuveniles by palpating the genital area (Dodge 1982) and weighted them once a
week from the date they were found to 15 August. We used body mass as a proxy fo r age of
the indi viduals as we expected earli er born individuals to be heavier than later born
indi viduals. We chose body mass on 10 June because ail j uveniles were born on or before
that date (range of birth date : 6 May-I 0 June), and because we measured body mass fo r 10 of
14 j uveniles on that date. For the remaining four individuals, we estimated body mass on 10
June by adding or removing 30 g/day (average dail y weight gain in June calculated over ail
individuals) from the body mass measured on the day closest to 10 June (27 May, 01 June, 0 1
July, 06 July for these four juveniles). We characterized the mobility ofjuveniles using the
mean di stance the y moved in a day. We calcul ated thi s distance by di viding the total di stance
(straight line) moved from one locati on to the next by the number of days elapsed between
successive locations. Finall y, when we kn ew the identity of the mother (n = Il ), we located it
twice a week, just before or after locating its juvenile, in order to calculate the di stance
between mother and juvenile (m).
Survival
We calculated survival time as the time from birth to death (exact time) or to the end of
the study period (1 5 August, censored time). We estimated date ofbirth from weight on first
capture considering a mean weight gain of30 g /day. Over 14 juvenil es, we knew the exact
death time fo r 10 indi viduals and death time was censored fo r the remaining fo ur indi viduals
(i .e. these indi viduals were still ali ve at the end of the study peri od). W hen we fo und a dead
j uvenile, we determined cause of mortality as starvation (not injured, intact carcass), predator
killed (traces of blood on radio co llar or wounds on the carcass), or unknown.