166 U.S. NATIONAL MUSEUM BULLETIN 2 59 



invasion. The range of this species has apparently resulted from the 

 distribution of a rather widespread precursor, which was able to 

 maintain genetic contact throughout its range by dispersal through 

 interstitial habitats near the surface. The present, apparent restric- 

 tion of this species to cave waters was almost surely brought about 

 through the operation of factors that were also responsible for the 

 restriction of other amphipods to central Texas caves — namely, ver- 

 tical migration necessitated by a progressively drier climate. 



The four species postulated to have arisen from ancestral stock A2 

 (fig. 36) share a number of close morphological affinities and occur 

 reasonably close together geographically. The origin and further 

 isolation of this line (A2) could have come about as the result of 

 migration of ancestral stock from coastal areas of the old Mississippian 

 embayment shoreline northwestward along flood plains of the Arkansas 

 River. Present patterns of distribution tend to point to the course 

 of this river as the most logical inland route of dispersal. 



Exact relationships between these four species are not yet clear. 

 For example, males of ^S*. bowmani are unknown, even though the 

 heavily spined telson and structure of the gnathopodal propods of 

 the female suggest a close kinship with S. clantoni. S. elatus, on the 

 other hand, bears a certain superficial resemblance to species genet- 

 ically close to S. alabamensis, but its affinity with S. clantoni and 

 S. ozarkensis is attested by the possession of several small, slender 

 spines on the ventral margin of the fourth peduncular segment of the 

 second antenna and in the overall structure of the gnathopodal propods. 

 The lack of completely mature males of this species, unfortunately, 

 does not allow for full interpretation of its exact genetic affinities. 



In some respects, S. ozarkensis might be considered the most 

 primitive of the four species derived from ancestral stock A2. This is 

 possibly evidenced by its elongate first antenna, more heavily spined 

 third uropod (of the female), and laterally spined telson. In addi- 

 tion, when compared with either S. elatus or ^S*. clantoni, S. ozarkensis 

 demonstrates appreciably more sexual dimorphism. 



The ranges of S. clantoni, S. ozarkensis, and S. elatus, are allopatric 

 and well separated geographically. Precursor populations of these 

 three species were probably extrinsically isolated as ancestral stock 

 migrated through the southern Ozarks and Arkansas Valley toward 

 the Osage Plains of eastern Kansas. Along the southwestern edge 

 of the Ozarks (Springfield Plateau) in northeastern Oldahoma, the 

 range of S. ozarkensis overlaps mth that of S. bowmani. Assuming 

 allopatric speciation, S. ozarkensis might have originated from an 

 earlier isolation of precursor populations further east in the Spring- 

 field Plateau, and then, more recently, it dispersed westward to the 

 margin of this plateau where it has come in contact with S. bowmani. 



