STUDIES IN CAVE LIFE EVOLUTION - A RATIONALE FOR FUTURE THEORETICAL DEVELOPMENTS USING PHYLOGENETIC INFERENCE

As in every field of comparative biology, phylogeny provides an indepe ndent reference system in studies on cave life evolution to test curre nt theoretical proposals. Using phylogeny, sound hypotheses on the anc estral states of characters and their subsequent changes can be made b y polarizing the characters between related taxa. Hypotheses on evolut ionary processes can also be tested by comparing the patterns they imp ly with independently inferred phylogenetic patterns. The power of the rests relies upon the independence of phylogenetic patterns (built wi th cladistics using Wagner parsimony) and the theoretical proposals un der study. Classical assumptions on the evolution of troglobitic life are analysed with this methodology. The following points are discussed : what is a troglobitic taxon? Are there features characteristic of tr oglobitic taxa? Is troglobitic life an evolutionary dead end? What cir cumstances favour troglobitic evolution? Using phylogenetic analysis, the presence or absence of so-called troglomorphic features were infer red in troglobitic taxa. In fact these taxa can be characterized only by their behavioural ecology. Pre-adaptations (exaptations) call also be precisely defined. Cave living does not appear to be an evolutionar y dead end. Two patterns subsequent to cave life appearance have been documented: speciation of troglobitic taxa in the subterranean environ ment, and reversal to an epigean habitat. Troglobitic life thus turns out to be one step in the diversification of clades. Troglobitic life is usually explained as an evolution under the pressure of unfavourabl e environmental conditions, or the conquest of a new resource, or the result of biological interactions (competition, predation). Phylogenet ic analyses show that none of these hypotheses propose clear alternati ves on cave life evolution. Moreover most of their a priori statements cannot easily be falsified. As such they have only limited explanator y power.