The present study was performed at Potter Peninsula (King George Island, An
tarctica) with the following goals: (1) to assess the size selection of kel
p gulls on Antarctic limpets during the chick-rearing period, and (2) to ev
aluate the degree of accuracy in the use of populational abundance and size
-frequency distribution as an indicator of prey availability. The average s
ize of limpets available to kelp gulls was significantly smaller than that
of the population as a whole and those actually consumed by gulls. The odds
ratio revealed a strong selectivity for large limpets, when available and
consumed sizes were compared. However, only a proportion of the limpets pre
sent was available to gulls. These differences clearly show that using abun
dance information as an indicator of a resource importance can lead to subs
tantial errors in ecological interpretations.