Why do morphological phylogenies vary in quality ? An investigation based on the comparative history of lizard clades

Abstract

Phylogenies based on morphology vary considerably in their quality: some are robust and explicit with little conflict in the data set, whereas others are far more tenuous, with much conflict and many possible alternatives. The main primary reasons for untrue or inexplicit morpho­logical phylogenies are: not enough characters developed between branching points, uncertain character polarity, poorly differentiated character states, homoplasy caused by parallelism or reversal, and extinction, which may remove species entirely from consideration and can make originally conflicting data sets misleadingly compatible, increasing congruence at the expense of truth. Extinction differs from other con­founding factors in not being apparent either in the data set or in subsequent analysis. One possibility is that variation in the quality of morphological phylogenies has resulted from exposure to different ecological situations. To investigate this, it is necessary to compare the histories of the clades concerned. In the case of explicit morphological phylogenies, ecological and behavioural data can be integrated with them and it may then be possible to decide whether morphological characters are likely to have been elicited by the environments through which the clade has passed. The credibility of such results depends not only on the phylogeny being robust but also on its detailed topology: a pectinate phylogeny will often allow more certain and more explicit statements to be made about historical events. In the case of poor phylogenies, it is not possible to produce detailed histories, but they can be compared with robust phy­logenies in the range of ecological situations occupied, and whether they occupy novel situations in comparison with their outgroups. LeQuesne testing can give information about niche homoplasy, and it may also be possible to see if morphological features are functionally associated with ecological parameters, even if the direction of change is unknown. Examination of the robust and explicit phylogeny of the semaphore geckoes (Pristurus) suggests that its quality does stem from a variety of environmental factors. The group has progressed along an ecological continuum, passing through a series of increasingly severe niches that appear to have elicited many morphological changes. The fact that niches are progressively filled reduces the likelihood of species reinvading a previous one with related character reversal. Because the niches of advancedPristurusare virtually unique within the Gekkonidae the morphological changes produced are also very rare and therefore easy to polarize. Ecological changes on the main stem of the phylogeny are abrupt and associated character states consequently well differentiated. The small geographical distribution of the group reduces the chance of different forms entering the same niche and developing morphological parallelisms. Many species are represented because there are many refuge areas within the geographical distribution ofPristurusand also because the fact that niches are well differentiated reduces the chance of extinc­tion by species competition. Examination of ten other lizard clades, mainly in the Lacertidae, shows that the historical features noted inPristurusalso occur in other groups with robust and explicit morphological phylogenies, such asMerolesandGastropholis.On the other hand, they are often absent in assemblages where phylogenetic hypotheses based on anatomy are poorly supported or equivocal, for examplePodarcisandPedioplanis. The amount of alteration of genital features is not as closely associated with environmental factors as in many other morphological characters, possibly because ecological shifts do no affect them so directly and because they have their own special causes of change. This partial independence from the environmental factors that appear to control many characters means that genital features may enable some phylogeny reconstruction to be done, even when other systems are unhelpful. There is no evidence among the studied clades that groups with poor morphological phylogenies have inherent general constraints on the development and maintenance of new structural features. It seems that the quality of morphological phylogenies may give some indication of the kinds of history groups have had. Clearly, if the variety of ways in which evolution proceeds is to be understood, groups with both good and bad morphological phylogenies must be investigated.