Olfactory and vomeronasal systems of caecilians (Amphibia: Gymnophiona)

The morphology of both the main nasal cavity and the vomeronasal organ differs among species representing six families of caecilians. The main nasal cavity is either divided or undivided. The vomeronasal organ differs in position (mediolateral, lateral), size (large vomeronasal organ in the aquatic species), and shape (mediolateral extension, vomeronasal organ with a lateral rostral projection). The great amount of respiratory epithelium of the main nasal cavity, the large vomeronasal organ, and its extensive innervation in typhlonectids may reflect both phylogeny and habitat adaptation, for these taxa are secondarily aquatic or semiaquatic and have several concomitant morphological and physiological modifications. The vomeronasal organ is associated with the caecilian tentacle as the tentacular ducts open into it. This association is further evidence for the involvement of the caecilian tentacle in vomeronasal chemoperception and may represent the mechanism by which these animals smell though the main nasal cavity is closed during burrowing or swimming. Labelings of primary olfactory and vomeronasal projections by means of horseradish peroxidase reaction reveal that the pattern of vomeronasal projections is similar in Ichthyophis kohtaoensis, Dermophis mexicanus, and Typhlonectes natans, even though T. natans possess stronger vomeronasal projections relative to olfactory projections than I. kohtaoensis and D. mexicanus. However, there are differences with respect to the patterns of olfactory projections. The olfactory projection of I. kohtaoensis is characterized by many displaced glomeruli. T. natans has the smallest olfactory projection. The nervus terminalis is associated with the olfactory system as shown by selective labelings of olfactory projections. Six characters potentially useful for phylogenetic analysis emerge from this study of comparative morphology. The characters were subjected to analysis using PAUP to see (1) if any resolution occurred and (2) if any groups were distinguished, whether they corresponded to phylogenetic arrangements based on other morphological characters. The characters are too few to produce nested dichotomous sets for all cases, but they do support the two typhlonectid genera examined and Dermophis and Gymnopis as sister taxa discrete from other groups, and they show that species within genera cluster together.