Hemipenial variation within the six recognized species of the African snake genus Crotaphopeltis is described. The hemipenial ornamentation of the widely distributed species, C. hotamboeia and C. hippocrepis, seems fairly constant whereas the ornamentation of C. degeni and C. tornieri, species with disjunct distributions, displays relatively large intraspecific variation. In spite of the variation, the size and ornamentation of the hemipenes serve to distinguish between sympatric and parapatric species.

The tribe Dispholidini (Bourgeois 1968) is reviewed, paying particular attention to dentition and visceral anatomy. A new genus and species, Xyelodontophis uluguruensis, is described from the Uluguru Mountains in Tanzania. All five genera have enlarged rear maxillary teeth. Thrasops seems to be basal, Rhamnophis shows the development of dagger-like teeth tapering from base to tip, then the other three genera appear to radiate, with Xyelodontophis having more derived dagger teeth broadest in the middle, while Dispholidus and Thelotornis seem to have independently developed enlarged grooved rear fangs.

Thrasops schmidti is recognised as a relict evolutionary species. No subspecies of Rhamnophis aethiopissa or Dispholidus typus are recognised, but the population of Dispholidus on Pemba Island probably represents an undescribed species.

An examination of scalation and dentition of specimens in Brussels (IRSN), Tervuren (MRAC) – mostly Bredo collection, and London (BMNH) from Angola, Congo-Kinshasa and Zambia suggests the existence of a species which is neither P. sibilans leopardinus of which the type is from Namibia, nor P. ‘sibilans’ [mossambicus] of Congo-Kinshasa and Zambia, but a new species previously unnamed.

Historically, the definition of species in the Neotropical snake genus Tropidophis has been difficult because of intraspecific variation in scalation and a paucity of specimens of most taxa. There were 13 species recognized at the time of the last review in 1960, but additional species have since been discovered and a taxonomic review and update is needed. Data on morphological variation are presented here and used to clarify the status of the described taxa. Because many taxa are allopatric with their closest relatives, it is necessary to make decisions as to their status as species or subspecies. As a gauge of species status in the genus, character divergence in ten pairs of closely related sympatric species was examined. Typically, such species are differentiated by two non-overlapping colour pattern differences, often in combination with a diagnostic (non-overlapping) or overlapping difference in scalation. Using this criterion, seven taxa previously considered as subspecies are here elevated to species status, whereas seven other taxa are retained as subspecies, although in some cases they are allocated to different species. As a result, the genus Tropidophis is considered here to comprise 29 species, 26 of which are West Indian and 15 of those are restricted to Cuba.

Dedication 91

History 91

Taxonomy 91

Distribution 92

Behaviour 92

Venom apparatus 93

Venom 93

Evolution 97

Acknowledgements 98

References 98

The family Atractaspididae is a highly modified derivative of a lineage that apparently arose early in the history of ‘colubroid’ snakes, and its taxonomy and relationship with other ophidian groups is still uncertain. Snakes of the genus Atractaspis have a characteristic venom apparatus, including the structure and function of the striking unit and of the venom glands. The composition of their venom is also unique in containing several low-molecular weight components, the sarafotoxins, which affect the cardiovascular system and are similar to the mammalian endothelins.

We use mitochondrial DNA sequences to infer the origin and phylogenetic position of the Lesser Antillean species of the pitviper genus Bothrops, B. caribbaeus and B. lanceolatus. The two species form a monophyletic group, which in turn forms the sister clade to the Bothrops asper-atrox complex. High levels of sequence divergence among the Caribbean species, and between them and the nearest mainland relatives, suggest a relatively ancient origin of these snakes. The hypothesis that the Lesser Antillean Bothrops are the result of a recent colonisation event from within the South American B. atrox complex is rejected, as is the hypothesis that they were introduced to their island habitats by aboriginal humans. The high level of morphological apomorphy displayed by B. lanceolatus suggests a stepping-stone colonisation, St. Lucia being colonised first and then Martinique from St. Lucia. The medical implications of these findings are discussed: a recent case of envenoming from Saint Lucia suggests that Bothrops caribbaeus causes the same thrombotic syndrome of envenoming as B. lanceolatus.

The systematics of two Southeast Asian green pitviper species, Trimeresurus hageni and T. sumatranus, are investigated by canonical variate analysis. Preliminary results reveal two morphological forms corresponding to mainly T. hageni in West Malaysia, Thailand and Singapore and T. sumatranus in Borneo. Allopatric populations of both taxa are examined from Sumatra. Geographic variation is present in both species, which are distinguished mainly by head scalation, but also by colour and pattern.

The publication in 1954 of Underwood's ‘On the classification and evolution of geckos’ was the first comprehensive attempt to understand the systematics, evolution and biogeography of this group of lizards. Combining the use of the exploration of novel characters with a global overview of geckos, Underwood erected hypotheses of relationship and patterns of distribution. In the 48 years since that landmark publication much has changed, but much has stayed the same. Underwood's division of geckos into four major clusters is still recognised today, although the sphaerodactyls are now regarded as a group derived from within the gekkonines, and the diplodactylines have been diminished by the removal of several genera and their placement in the gekkonines. The framework that Underwood established has resulted in generic and/or species level phylogenies being generated for the eublepharids, some sphaerodactyls, the carphodactyline diplodactylines and some clusters within the gekkonines. The latter group, because of its size, has remained intractable to detailed systematic analysis at the generic level, although the recognition of many discrete monophyletic clusters within the Gekkonidae (the Gekkoninae of Underwood) holds out the possibility that greater levels of intergeneric resolution are close to realisation.

Underwood's initial approach to the systematic analysis of geckos was distinguished by its use of novel characters of the visual system that led to new insights. It is possible that the next breakthrough in higher level systematic analysis of geckos may again come from the exploitation of new character sources. Some examples of these possibilities are discussed.

The skull anatomy of uropeltines is reviewed, and new data is presented on the highly derived otico-occipital region. A phylogenetic analysis of uropeltine interrelationships using parsimony is performed using characters derived from skull structure. The basal position of the genus Melanophidium is confirmed; Pseudotyphlops is a relatively derived uropeltine, in spite of its relatively large size. The monophyly of the genera Melanophidium and Rhinophis requires further testing.

The poorly known marine squamate Mesoleptos is reassessed based on two previously known specimens and a newly referred specimen. The three specimens of Mesoleptos zendrinii share unique characters such as long, posteriorly tapering centra and distally straight but non-pachyostotic ribs. Mesoleptos had a narrow neck (and presumably small head), long laterally compressed body, and small fore- and hindlimbs. Phylogenetic analysis suggests that Mesoleptos is the nearest relative of snakes; this phylogenetic position is consistent with its morphology being intermediate between typical marine squamates (e.g. mosasauroids) and primitive marine snakes (pachyophiids). However, this interpretation remains tentative because Mesoleptos is very poorly known, and many of the characters uniting it with mosasauroids and primitive snakes are correlates of marine habits and/or limb reduction.

Introduction 143

Abbreviation used in text 144

Abbreviations used in figures 144

Morphology 144

Disposition of the cloaca 144

Divisions of the cloaca 146

Urodeum 146

Blind sacs 146

Anterior phallodeum 147

Posterior phallodeum 148

Phallodeal ornamentation 148

Composition of phallodeal structures 148

Relationship between the uneverted cloaca and the phallus 149

Systematics 149

Is phallus morphology species specific? 149

Species differentiation and generic identity 152

Discussion 152

Acknowledgements 153

References 153

The cloaca of male caecilian amphibians (Gymnophiona) is a tube that comprises an anterior urodeum and a posterior phallodeum. The phallodeum everts (with the urodeum lying inside it) to form a phallus used for direct sperm transfer in copulation. Phallodeal morphology is rich in detail and variation, and has therefore been considered a potentially useful and much needed tool for caecilian phylogenetics and species-level taxonomy. Despite this, it has been almost entirely ignored in caecilian systematics, there is confusion regarding some aspects of morphology, and variation within and among species is poorly understood. A short review and reconsideration of phallus morphology is presented, and the systematic potential assessed. The anterior part of the phallodeum appears to offer the most obvious systematic potential, and the morphology of longitudinal ridges and their ornamentation here seem to have diagnostic and/or phylogenetic value for some taxa. Although there is evidence of intraspecific variation, at least some of which is associated with ontogeny and reproductive condition, individuals of the same species generally have a common pattern of phallodeal ridges and ornamentation, and congeners often share a similar pattern. However, these patterns are not universally species specific, at least among uraeotyphlids. Although variation needs to be better understood, the male cloaca offers great potential for caecilian systematics.

Holaspis is the most morphologically apomorphic lacertid taxon with 42 or more derived morphological features arising on its exclusive lineage. Nearly all of these confer advantages in three specialised activities, or ameliorate problems resulting from them. The activities are: climbing on the often vertical open surfaces on tree boles and branches, utilising very narrow crevices in wood and beneath bark, and the ability, unique among lacertids, to glide from tree to tree. Although many of the features related to these activities are likely to result from direct adaptation to the situations concerned, exaptation has been critical in the development of gliding. Two behaviours present in the earliest lacertids have been coopted to this activity: rib spreading associated with basking contributes to an effective aerofoil, and balance control associated with running helps maintain appropriate posture in the air. Features originally developed in the context of crevice use also contribute to the aerofoil and a high surface: weight ratio. So, while natural selection has moulded Holaspis for its present activities, multiple accidents of history have also been important, as they also have in the evolution of bird flight.