2 



Peter Lesica did a preliminary electrophoretic survey 

 of Arabis fecunda (pers. comm. 1989) . Our work expanded on 

 number and kind of gel and electrode buffers and number of 

 plants screened. We ran a total of 180 plants on 12% starch 

 gels, using eight gel and electrode buffers from Soltis et 

 al. (1983), Rieseberg and Soltis (1987), Clayton and Tretiak 

 (1972) and Ridgeway et al. (1970). Twenty enzyme systems 

 were screened. Enzyme abbreviations and locus scoring 

 follow Gottlieb (1977) and Soltis et al. (1983). Data were 

 analyzed using BIOSYS-1 (Swofford and Selander 1981) . 



RESULTS 



Phosphate-PVP or tris-HCL grinding buffers and grinding 

 method 1 (plant ground before freezing) consistently gave 

 the best banding patterns. Tris-maleate grinding buffer 

 caused bands to migrate slower, to separate less, and to be 

 more diffuse. Distilled water gave no activity for any 

 enzymes. Grinding method 2 (plant ground after freezing) 

 showed no activity for most enzymes. 



Given the low level of enzyme polymorphism in these 

 samples (see below) , we decided to screen a smaller number 

 of individuals per population for as many enzyme systems as 

 possible. Twelve enzyme systems coded by 18 putative loci 

 were resolved (Tables 1 and 2). Eight additional enzyme 

 systems were not used due to poor resolution and/or 



