SUBTERRANEAN AMPHIPOD STYGONECTES 165 



to each other but that differ consistently from those in the caves 

 north and east of the Llano uplift area. A fifth cave, Spanish Wells, 

 is located nearby, but individuals in this population (poorly known, 

 from a sample of only two females) show minor morphological differ- 

 ences from individuals in the other four caves. With the possible 

 exception of Spanish WeUs, regular gene flow is predicted between 

 populations in the caves just northwest of Austm and would account 

 for the homogeneity of these populations. A complicatmg factor, 

 however, is that populations of S. russelli from Cave X and Irelands 

 Cave, both of which lie just south of Austin, and one of which lies 

 only about eight miles south of two of the four caves (mentioned 

 above) northwest of Austui, show closer morphological alliance with 

 the populations from north and east of the Llano uplift area. More- 

 over, in some respects these two popidations differ slightly from popu- 

 lations anywhere else in the range of this species. Two populations 

 from the western part of the range, i.e., Zesch Ranch Cave (Mason 

 County) and Stowers Cave (Kerr County), differ to some extent 

 from each other, but in turn, each population has a number of charac- 

 ters that overlap with those of populations from both the Austin and 

 the Llano uplift areas. Only one small female is known from Cascade 

 Caverns in Kendall County, and this specimen falls morphologically 

 closer to the Irelands Cave population than to any other. 



The pattern of irregular geographic variation in S. russelli implies 

 incomplete gene flow and pockets of partially or completely isolated 

 demes. Additional collecting and the application of more refined 

 biometric techniques will undoubtedly assist in clearly defining geo- 

 graphic patterns and will aid further in unmasking cryptic trends 

 m morphological variation. S. russelli is almost certainly a polytypic 

 form, and when completely studied and fully understood, it may 

 turn out to be a cluster of several sibling species. 



As denoted by the striking similarity in the structure of the telson, 

 S. russelli shows a certain close affinity with both S. balconis and S. 

 hijurcatus, but it differs from these species by a defined difference in 

 the gnathopodal propods. Other, more minor differences, are seen 

 in the proportionately longer fifth pereopod and in the generally less 

 convex posterior margins of the abdominal side plates of this species. 

 Nevertheless, it is apparent that S. russelli was derived from the 

 parental stock represented by line B2 (cf., fig. 36), but its exact 

 relationship to other species of this phyletic assemblage is not abun- 

 dantly clear. The fact that S. russelli occurs sympatrically mth both 

 5. balconis and S. hijurcatus would indicate multiple invasion of the 

 habitats occupied by these species, and because of the wider range 

 and greater degree of morphological variation of S. russelli, I am 

 inchned to view this species as representative of the more recent 



