Lobule shape evolution inRadula(Jungermanniopsida): one rate fits all?

Abstract

The quantification of lobule shape for Radula spp. shows that there is overlap in lobule shape space occupied by subgenera, such that lobule shape does not always reflect relationships. Morphological convergence caused by lineages repeatedly traversing shared regions of morphospace appears commonplace in Radula, and means that many pairs of relatively unrelated species have similar lobule shapes. When observed over time, as in comparisons between fossil and extant species, this may give the impression of stasis if fossil species resemble modern species by chance, independent of their relatedness. This poses a challenge to relating fossils of known age to extant lineages, particularly when fossils are sterile. Significant rate variation between lineages was identified by Adams' Q-mode analysis, with the fastest subgenus evolving 23 times more quickly than the slowest. Species of subgenus Volutoradula and subgenus Metaradula are apparently over-dispersed throughout lobule morphospace according to Sidlauskas' method; morphometric branch lengths and hypervolumes in other subgenera can be explained by a stochastic process. In contrast, Bayesian analysis of macroevolutionary mixtures (BAMM) identified a single evolutionary rate as having the highest posterior probability. Consideration of the three independent accessions into auriculate lobule morphospace by Cladoradula and Radula, wherein convergent lobule shapes result from convergent lobule ontogenies and are correlated with bipinnately branched shoot systems and robust primary stems, leads to an ontogenetic hypothesis driven by structural requirements for light interception, under which auriculate lobules are a spandrel. It is speculated that lobules themselves, however, may be a key innovation facilitating radiation into microsites devoid of or depauperate in fungal endophytes.