Delayed trait development and the convergent evolution of shell kinesis in turtles

Abstract

Understanding developmental processes is foundational to clarifying the mechanisms by which convergent evolution occurs. Here, we show how a key convergently evolving trait is slowly 'acquired' in growing turtles. Many functionally relevant traits emerge late in turtle ontogeny, owing to design constraints imposed by the shell. We investigated this trend by examining derived patterns of shell formation associated with the multiple (at least 8) origins of shell kinesis in small-bodied turtles. Using box turtles as a model, we demonstrate that the flexible hinge joint required for shell kinesis differentiates gradually and via extensive repatterning of shell tissue. Disproportionate changes in shell shape and size substantiate that this transformation is a delayed ontogenetic response (3-5 years post-hatching) to structural alterations that arise in embryogenesis. These findings exemplify that the translation of genotype to phenotype may reach far beyond embryonic life stages. Thus, the temporal scope for developmental origins of adaptive morphological change might be broader than generally understood. We propose that delayed trait differentiation via tissue repatterning might facilitate phenotypic diversification and innovation that otherwise would not arise due to developmental constraints.

Keywords: convergent evolution; joint development; ontogeny and phylogeny; shell kinesis; turtle shell evolution.

Figure 1.

Shell kinesis via hinge rotation in box turtles (illustration credit: Jessica Gassman; (a)). In akinetic-shelled turtles (i), the junction or seams of pectoral (Pec) and abdominal (Ab) ectodermal scutes does not overlap (see arrows) with the suture of hyoplastron (Hyo) and hypoplastron (Hyp) bones ((b); modified from Pough et al. [49]), unlike in kinetic-shelled species (ii) such as box turtles ((b); modified from [50]). Shell kinesis (Akinetic = blue; Kinetic = red) in small-bodied turtles (CL = carapace length in mm; see (c)) has evolved at least eight times, probably in conjunction with evolutionary transitions from aquatic (cyan) to terrestrial (yellow) habitats (d); time-calibrated phylogeny from Pereira et al. [51]. (Online version in colour.)

Figure 2.

The immature seam (inlet: haematoxylin and eosin sagittal section) of the pectoral (Pec) and abdominal (Ab) scutes in hatchling Terrapene ornata does not feature a plastral hinge (a). Ossification centres of the hyoplastral (Hyo) and hypoplastral (Hyp) bones (in alizarin red) expand and begin to fuse near the Pec-Ab scute junction a year after hatching in T. ornata (b). These bones are sutured 3 years post-hatching (c). By year 5, the suture is repatterned and features fibrous connective tissue, particularly on the external surface of the plastron (d). This connective tissue proliferates externally and the serrated edges on the interior plastral surface are further reduced in fully grown adults (e); silhouettes are modified from Pritchard PCH [77]. (Online version in colour.)

Figure 3.

Adult plastral buttresses (axillary = Ax; inguinal = Ing) are fully developed in Chrysemys picta (a), partially reduced in Emys blandingii (b), and fully reduced in Terrapene ornata (c). This skeletal difference is pre-patterned in hatchlings, as the Ax buttress in E. blandingii and T. ornata does not achieve contact with the carapace (the distance to carapacial ossification centres is represented by black arrows in (b,c)). This pattern corresponds to akinesis in C. picta, as well as varying degrees of kinesis (arrows) in adult E. blandingii versus T. ornata (c(ii,iii)). Progressive breakdown of bone and scute tissue and replacement with fibrous connective tissue occurs near the Ax buttress in growing T. ornata (d). The hyoplastral–hypoplastral suture (Verhoeff–van Gieson-stained sagittal section) displays an interdigitating pattern in akinetic turtles ((i): Glyptemys insculpta; scale bar = 10 µm), in contrast to the cornified (yellow) and collagen-rich (red) fibrous connective tissue comprising the hinge of kinetic-shelled species ((ii): T. ornata; scale bar = 50 µm) (e). (Online version in colour.)

Figure 4.

Scaling of the plastron during embryonic and post-hatching stages varied among species (a), as T. ornata (TO) displayed a higher growth rate and thus attained a larger adult plastron-to-carapace length ratio compared to C. picta (CP) and E. blandingii (EB) (b). Plastron shape, represented by the first axis of a principal component analysis, displayed greater deformation during post-hatching growth in T. ornata (c). A PCA plot demonstrates that shape divergence was driven by differences in plastron length (points are proportionate to length, i.e. PL). Note that larger individuals plot further to the right along the x-axis, which accounted for most shape variation (PC: 50.6%) (d). A comparison of thin-plate spline visualizations of species-specific means for plastron shape against the consensus plastron shape indicate that the anterior plastral lobe is broader in T. ornata (e). (Online version in colour.)