Evolutionary origins of novel conchologic growth patterns in tropical ...

Recommendations

Info

Goodwin et al.Evolution of novel conchologic growth patterns 649Fig. 5. Strict consensus of six most parsimonioustrees, using Juliacorbula scutata and Panamicorbulaventricosa as outgroup taxa.Bootstrap values and Bremmer Decay Indicesare shown above and below individual nodes,respectively. Growth form character state transitionsare shown in brackets on nodes, andgrowth form for each species is indicated in parenthesesafter each taxon name. See text fordiscussion.growth form. Their cross-sectional shape, however, is similarto other Caryocorbula, suggesting GF2 characterizes the entiresubclade.The distribution of growth forms in Clade B is more complex(Fig. 5). The basal node of the clade is characterized by atransition from GF1 to GF3. All the taxa in the Corbulaspp.1Notocorbula subclade are characterized by GF3. Thetwo basal-most taxa in the (Bothrocorbula spp.1Hexacorbula?1Hexacorbulaspp.) subclade, H. quirosana andH. hexachyma, also are characterized by GF3. The next mostderivedtaxon, Hexacorbula? sp. is characterized by GF2. Thischaracter state transition is somewhat surprising given that itis the only instance where a transition between GF2 and GF3occurs. Nevertheless, it may reflect an evolutionary propensityof this sublcade to modify patterns of growth. For example,the node at the base of the Bothrocorbula spp. clade reversesfrom GF3 back to the ancestral GF1. Four of the five speciesin this genus possess GF1. B. sp.cf.B. viminea, incontrast,possesses a well-defined nepioconch, and is best characterizedby GF3. This character state transition therefore representsanother reversal, from GF1 to GF3.Evolutionary mechanisms within BothrocorbulaIn documenting the phylogenetic distribution of growthforms, we focused on Bothrocorbula for several reasons. First,the genus is monophyletic and all known Bothrocorbula speciesare included in our phylogenetic analysis (Fig. 5). Second,we have robust support for the phylogenetic topology of thissubclade. These phylogenetic criteria are critical to identifyingthe process(es) responsible for observed evolutionary patterns(Fink 1982). Third, the character reversal at the base of theclade suggests that deviation from GF1 is a synapomorphiccharacter state within the Bothrocorbula, whereasGF3isanapomorphic character state in B. sp.cf.B. viminea. Thispatternis similar to that observed in tropical American corbulidsas a whole, where GF1 is symplesiomorphic and deviationsare synapomorphic. Thus, evaluation of evolutionary processesin Bothrocorbula serves as a model system in which toinvestigate evolutionary processes that may be operating inthe larger clade. Fourth, species of Bothrocorbula tend to berelatively large, allowing growth lines to be easily identifiedunder low magnification, and micromilled at high resolution(Goodwin et al. 2003). Finally, geographic and stratigraphicdistributions of Bothrocorbula span an interval of environmentalchange in the tropical western Atlantic associated withthe emergence of the Isthmus of Panama and the closure ofthe Panama Seaway (Jackson et al. 1996; Coates et al. 2004).Thus, the potential exists to link evolutionary patternsobserved in the fossil record with specific paleoenvironmentalchanges (e.g., Anderson 2001; O’Dea et al. 2007).To better understand the evolution of shell development inBothrocorbula, we examined the hypothesis that heterochronywas responsible for generating the distribution of growthforms within the genus. Heterochrony, changes in timing orrate of development through ontogeny from an ancestor to adescendant (de Beer 1958), seems a likely controlling mechanismbecause the initial phases of both derivative growthpatterns resemble that of GF1 (i.e., primary radial accretion)(see Alberch et al. 1979). Failure to support the null hypothesisFi.e.,heterochrony did not occurFpermitsustosuggestthat heterochronic processes were responsible for morphologicevolution using the time, size, and shape frameworkdeveloped by Alberch et al. (1979). Failure to reject the nullhypothesis suggests that a mode of ontogenetic modificationother than heterochrony (e.g., heterotopy) may have been responsiblefor the distribution of growth forms in this clade(Webster and Zelditch 2005).Two criteria must be met to evaluate the role heterochronyplays in the morphologic evolution of a clade
650 EVOLUTION&DEVELOPMENT Vol. 10, No. 5, September^October 2008Fig. 6. Cross-sections of two species of Bothrocorbula.(A)B. sp. cf.B. viminea. The shaded area is the nepioconch and the stippled areamarks shell material deposited after the nepioconch. (B) Crosssectionof the nepioconch from B. sp.cf.B. viminea. (C)Crosssectionof an adult B. synarmostes. Note the similar shape of thenepioconch of the B. sp. cf. B. viminea cross-section (nepoiconch)to that of the adult B. synarmostes.(i.e., heterochronic patterns of paedomorphosis vs. peramorphosis),whereas a third, if met, will identify thespecific heterochronic process responsible for that pattern(see Roopnarine 2001b for discussion). The first criterionrequires that a clear hypothesis of phylogenetic relationshipswithin a clade be available so that putative ancestor anddescendent taxa can be identified (Fink 1982). For Bothrocorbulawe have a hypothesis of evolutionary relationshipthat is robust and with low stratigraphic debt (Fig. 5; Fig.S1). In this hypothesis, B. synarmostes, with GF1, is thebasal-most and stratigraphically lowest taxon in Bothrocorbula,serving as a putative ancestor for other speciesin the genus.The second criterion, that ancestors and descendants havecoincident morphologic (shape) histories through their respectivedevelopment (Roopnarine 2001b), also can be demonstratedfor Bothrocorbula through three lines of evidence.First, conchologic features, with characters describing traitsthat range from valve ornament to dentition, strongly supportthe monophyly of Bothrocorbula (see also SupportingInformation: Bothrocorbula species descriptions). Second,combined internal and external landmarks from geometricmorphometric analysis indicate that Bothrocorbula is distinctmorphologically in both shape and size from closely relatedtaxa, including Hexacorbula, Hexacorbula?, and Caryocorbula(Anderson and Roopnarine 2005). Third, and most critical forour data on valve accretion patterns, the growth form presentin B. sp.cf.B. viminea (GF3) recapitulates growth of theancestral B. synarmostes (GF1), as illustrated in Fig. 6. Specifically,the cross-section of a B.sp.cf.B. viminea nepioconch(Fig. 6A: shaded area; 6B) is remarkably similar to the crosssectionof a putative adult B. synarmostes (Fig.6C),withvalve heights, thicknesses and inflation virtually identical.Furthermore, both are characterized by strong commarginalridges on the valve exterior and weak reflection of these ridgeson the valve interior. This similarity suggests a peramorphicheterochronic pattern underlying the transition from GF1 toGF3 in Bothrocorbula (Alberch et al. 1979).The third criterion, which permits identification of heterochronicprocess, requires reliable estimates of ontogenetic agein order to identify changes in timing or rate of developmentbetween ancestor and descendent (Alberch et al. 1979; Jonesand Gould 1999; Roopnarine 2001b). We established age estimatesthrough sclerochronologic analysis. Representativesof each of four Bothrocorbula species examined were sampledto determine their ontogenetic age (Table S1). (For the fifthBothrocorbula species, B. viminea, we determined its growthform but did not section valves due to a paucity of availablematerial.) To establish maximum ontogenetic ages, specimensselected for isotopic analysis were drawn from the largest sizeclass of each species. Micromilled profiles were chosen becausethey provide more complete records of d 18 O variability,and accordingly more accurate age estimates, than pointsamplingapproaches (Goodwin et al. 2003). Nevertheless, thepoint-sampled d 18 O profiles show similar patterns of variation(Fig. S2).Isotopic variation in B. synarmostes indicates that the lifespanof this individual was approximately one year, likelyrepresenting shell deposition from late winter/early springthrough late fall/early winter (Fig. 7A). The sinusoidal profileof B. radiatula d 18 O samples indicates that this individual alsolived for approximately 1 year (Fig. 7B). Shell accretionprobably began in the late summer and continued through thenext year, with death occurring late in the second summer.Similarly, analysis of B. wilcoxii suggests these animals livedfor approximately 1 year. Shell accretion likely began in themid- to late summer, continued through the fall, winter and
Page 1: EVOLUTION & DEVELOPMENT 10:5, 642-6
Page 4 and 5: Goodwin et al.Evolution of novel co
Page 12 and 13: Goodwin et al.Canapa, A., Barucca,
Page 14 and 15: Goodwin et al.were step-scanned wit

Evolutionary origins of novel conchologic growth patterns in tropical ...

Create successful ePaper yourself

Delete template?

Save as template?