162 U.S. NATIONAL MUSEUM BULLETIN 2 59 



may be of further significance to point out that two-thirds of all 

 collections of S. a. alabamensis from the northern Ozarks in Alissouri 

 have been from caves, while ten out of eleven collections of this species 

 from the southern Ozarks in Ai'kansas have been from seeps and 

 springs. A reasonable explanation for this might be that caves of 

 Arkansas are still poorly known biologically, but on the contrary, it 

 should be emphasized that during the summer of 1964, a thorough 

 investigation of six strategically located caves in the Arkansas Ozarks 

 did not yield a single amphipod. Furthermore, during the same field 

 work, one seep, out of only a few that were casually visited, did 

 produce a specimen of S. a. alabamensis. 



In summary, the range of S. alabamensis appears to have resulted 

 from a rather recent (perhaps early Quaternary), widespread dispersal 

 of this species or its immediate past ancestor into a number of far- 

 reaching areas of the south-central United States. Despite the 

 extreme vagility still shown by this species, populations are gradually 

 becoming isolated through the effect of extrinsic barriers, which are 

 being created primarily by erosion of the Ozark Plateau region and 

 concomitant downcutting of adjacent areas. The fact that pheno- 

 types display only minor, usually irregular, and often imperceptible 

 changes over wide areas would indicate that the process of isolation 

 is coming about very slowly. 



The isolated range of aS'. balconis to two caves in the Edwards 

 limestone of northern Hays County and southern Travis County in 

 Texas, does not unpose a major problem of interpretation. This 

 species appears to be one of several isolation products of the former 

 range of alabamensis stock, which apparently once occupied parts of 

 central Texas. The close morphological similarity of S. balconis and 

 S. alabamensis was discussed in an earher paper (Holsinger, 1966). 

 Examination of additional material during the present study, however, 

 has led to the recognition of a niunber of minor differences between 

 these two species not noted by me previously, but both species are 

 still regarded as having a rather recent, similar genetic background. 

 S. balconis, Uke 5'. a. occidentalis, has lost all pereonite and pleonite 

 sternal gills and seventh pereopod coxal gills, but differs in having 

 undergone a reduction from bifurcate to simple lateral sternal gills. 

 On the other hand, S. balconis has retained the marked palmar margin 

 concavity and the proportionately long first antenna, both characters 

 of which are typical of /S. alabamensis s. lat. 



S. bifurcatus is considered closely related to both aS*. balconis and 

 S. alabamensis but has undergone a reduction in the length of the 

 first antenna and in the palmar margin concavity of the gnathopodal 

 propods. This species still retains bifurcate lateral sternal gills, 

 although these structures are somewhat reduced. Sternal gills are 



