Changes in Populations 135 



Oxford. Thrushes removed snails from the colony, cracked their 

 shells on stones on a nearby bank, and ate the soft parts. Thus a 

 sample of the predated portion of the population could be obtained 

 by collecting the broken shells from around the thrush anvils, and a 

 sample from the entire population by collecting individuals from 

 the bog. Of 560 individuals taken from the bog, 296 (52.8 percent) 

 were unhanded, while of 863 shells collected around the rocks only 

 377 (43.7 percent) were unhanded. This significant difference 

 (P < .02) indicates that unhanded individuals in this colony were 

 less likely to be eaten than banded individuals. 



Similar methods had been used by Sheppard in 1950 to study two 

 other colonies. He found that there was a decrease in the percentage 

 of yellow snails killed as the season progressed and that the rate of 

 decrease appeared to be the same in both localities. There was no 

 evidence that this was due to the thrushes hunting in different areas 

 or to a change in the percentage of yellow shells in the population 

 at large. Apparently the selective value of the yellow phenotype was 

 at least partly a function of the background on which it occurred. 

 Early in the spring, when the woodland floor was predominantly 

 brown, the yellow shells were relatively conspicuous and thus at a 

 selective disadvantage. As the season progressed, the background 

 became greener and this disadvantage lessened. By late April or 

 early May the yellow shells were selectively neutral; by mid-May 

 they were at a selective advantage. 



These data indicate a rather strong selective pressure. Because of 

 this, one would expect populations living on uniform backgrounds to 

 be composed only of unhanded individuals, and those living in rough 

 tangled habitats to contain only banded individuals. How then is the 

 polymorphism maintained? Shifts of selective values with the seasons 

 would delay, but not prevent, the removal of the less favored 

 varieties. Interchange of individuals among colonies in different 

 habitats would account for some of the variability, but the range of 

 movement of snails is too small to support this hypothesis for more 

 isolated colonies. 



The answer is that there are physiological factors genetically cor- 

 related with pattern type. Thus experiments with Cepaea nemoralis 

 have shown that unhanded individuals ( especially yellow ones ) are 

 more heat-resistant than banded individuals. Yellow snails are more 

 resistant to cold than pink snails, and unhanded snails are more 

 cold-resistant than banded snails. These and other similar character- 

 istics indicate that color and banding are subject to strong nonvisual 

 selection because they are associated with important physiological 

 advantages. In some cases, heterozygotes may be expected to be 



