Bathygobius soporator

Tornabene, Luke & Pezold, Frank, 2011, Phylogenetic analysis of Western Atlantic Bathygobius (Teleostei: Gobiidae), Zootaxa 3042, pp. 27-36 : 31

Bathygobius soporator group

This group, which includes B. soporator , B. andrei , and B. lacertus , was recovered as monophyletic on each gene tree and on the tree from the concatenated analysis. The Eastern Pacific B. andrei was recovered as the basal member of the group. Three distinct geographic lineages (West Africa, Gulf of Mexico and Caribbean/South America) are apparent within B. soporator (see “ Bathygobius soporator sublineages” section below). Despite the well-supported hypothesis shown by the concatenated phylogeny, support values throughout the B. soporator group on the RAG1 gene tree (not shown here) are very low, and B. soporator itself appears polyphyletic. We attribute the poor resolution of species in this group to the low levels of polymorphism at this nuclear locus, thus demonstrating the ineffectiveness of using a single slow-evolving nuclear gene when trying to resolve fine scale relationships within this group.

The B. soporator group provides an interesting opportunity to investigate the homology of pigmentation patterns in Bathygobius . Both B. soporator and B. andrei are diagnosed from congeners in their respective regions by a possessing a broad dark vertical or slightly diagonal bar on the first dorsal fin. Bathygobius lacertus , which appears to be the sister species to B. soporator , possesses a longitudinal pigmentation pattern on the first dorsal fin, which is the most common pattern seen in Bathygobius species. Similarly, the general body pigmentation of B. andrei and B. soporator are nearly identical, yet B. lacertus possesses a two-row spotted pattern of body pigmentation that more closely resembles B. antilliensis , B. geminatus , B. cocosensis and B. coalitus than it does B. soporator or B. andrei . There are several potential scenarios explaining the discordance between the phylogenetic relationships inferred from our molecular phylogenies and the shared pigmentation patterns observed in this group. If our hypothesis reflects the true phylogeny of the group, that is, if B. soporator is indeed more closely related to B. lacertus than it is to the phenetically similar B. andrei , then this would suggest that B. lacertus has undergone a relatively recent and rapid reversal in body and dorsal fin pigmentation back to the more common and likely ancestral patterns of the two-row spotted body pigmentation pattern (to which there are several variations within the genus) and the horizontal dorsal fin pigmentation pattern. Coinciding with this reversal in pigmentation patterns would be a shortening in the anterior extent of predorsal squamation, as both B. soporator and B. andrei share predorsal squamation that extends well beyond the posterior margin of the preopercle, and B. lacertus does not. An alternative possibility is that the relationships shown here for the B. soporator group (driven largely by two rapidly mitochondrial gene trees, as RAG1 had low resolution for this group) does not reflect the true phylogeny – perhaps a result of extensive homoplasy in the hypervariable mitochondrial genes (discussed below). If this latter scenario is true, B. andrei may in fact be an allopatric sister species of B. soporato r (a “geminate pair”), as suggested by the nearly identical pigmentation.

Previous studies implied a close relationship between B. andrei and B. soporator . Rubinoff & Rubinoff (1971) performed hybridization studies on Eastern Pacific B. ramosus , B. andrei , and Atlantic B. soporator , all of which came from Panama. In no-choice group matings (multiple females of one species, multiple males of the other species), B. andrei x B. soporator had 37 spawnings, whereas B. ramosus x B. soporator had only nine spawnings, and the sympatric B. ramosus and B. andrei did not spawn. While the B. ramosus x B. soporator eggs were not reared to maturity, B. andrei x B. soporator offspring were reared, and a backcross of this F1 generation with parent B. andrei produced viable offspring ( Rubinoff & Rubinoff 1971). Analysis of genetic distance from protein electrophoresis studies on the same species of Panamanian Bathygobius confirmed that B. soporator was more closely related to B. andrei than to B. ramosus ( Gorman et al.1976) . The four trees provided by the morphological analysis of Miller & Smith (1989) provide additional support for a B. soporator - B. andrei sister relationship.

A confounding issue with the Rubinoff & Rubinoff (1971), Gorman et al. (1976), and Miller & Smith (1989) studies is that B. soporator was previously diagnosed based on a suite morphological characters that are also present in B. lacertus (a species previously considered a synonym of B. soporator ) and B. antilliensis (which had not yet been described at the time of the prior studies). Although B. antilliensis is not currently known from Panama ( Tornabene et al. 2010), B. lacertus is, and thus it is possible that Gorman et al. (1976) and Rubinoff & Rubinoff (1971) had both B. soporator and B. lacertus in their analysis. Although the western Atlantic B. soporator material from Miller & Smith (1989) has not been reexamined, it may contain all three species as well. Therefore, any conclusions made from these previous studies about the B. soporator – B. andrei relationship should be interpreted with caution. The combined molecular, morphological, and behavioral evidence available from the current study and previous studies supports the presence of a monophyletic group containing B. andrei , B. soporator , and B. lacertus .