Changes in Populations j 137 



Island Water Snakes 



Camin and Ehrlich have studied microevolution in populations of 

 water snakes (Natrix sipedoii) on the islands in western Lake Erie. 

 The snakes on these islands have variable banding patterns. These 

 have been divided into four classes: A, B, C, D (type A snakes being 

 unhanded and type D snakes being completely banded; see Fig. 

 7.5). Except in the area of western Lake Erie (and one locality in 

 Tennessee), all known N. sipedon populations are made up of type 

 D individuals. Virtually all the snakes taken from the mainland sur- 

 rounding Lake Erie are type D. On the islands a large proportion of 

 the snakes are of the other types, including numerous individuals 

 that are totally unhanded. The islands have very little inland water, 

 and the snakes are restricted largely to the lake shore. The marginal 

 flat limestone rocks, limestone cliffs, and pebble beaches are the only 

 suitable habitats. 



Large samples of adult snakes were taken from the islands, and 

 pregnant females were kept alive until their young were born. The 

 distribution of pattern types in the litters was compared with that 

 of the adults. The distribution of the adult and litter pattern types 

 from one group of islands is shown in Fig. 7.6. In spite of the diffi- 

 culties of statistical comparison of the cluster-sampled litter popu- 

 lation with the random-sampled adult population, it was possible 

 to show a significant difference in banding pattern between the two 

 populations. The percentage of relatively unhanded individuals (A 

 and B) was higher in the adult population than in the litter popu- 

 lation. 



The observed significant differences between the young and adult 

 populations can be accounted for only by differential elimination of 

 pattern types or by pattern changes in the individual snakes. The 

 evidence is overwhelmingly in favor of the former hypothesis, as 

 snakes kept in the laboratory show no evidence of pattern change in 

 ontogeny. In addition, individuals of all pattern types have been 

 recorded from both adult and litter populations; only the frequencies 

 differ. 



Evolutionary agencies other than selection are easily disqualified. 

 A high proportion of unhanded pattern types might be maintained 

 by migration from other unhanded populations, but the nearest such 

 population is in Tennessee. To maintain the unhanded genotype by 

 mutation alone would require a mutation rate far above that known 

 for any locus ever studied in any organism, even if it is assumed that 

 color pattern is a single factor trait ( which it is not ) . Nor will genetic 

 drift account for the observed differences. There is no sign that the 

 populations on the islands ever approach a size at which drift is 



