Laboratory studies have shown that Antarctic krill Euphausia superba shrink if maintained in conditions of low food availability. Recent studies have also demonstrated that E. superba individuals may be shrinking in the field during winter. If krill shrink during the winter, conclusions reached by length-frequency analysis may be unreliable. In this study, the correlation between the body-length and the crystalline cone number of the compound eye was examined. Samples collected in the late summer show an apparent linear relationship between crystalline cone number and body-length. From a laboratory population, it appears that when krill shrink, the crystalline cone number remains relatively unchanged. If crystalline cone number is little affected by shrinking, then the crystalline cone number may be a more reliable indicator of age than body-length alone. The ratio of crystalline cone number to body-length offers a method for detecting the effect of shrinking in natural populations of krill. On the basis of the crystalline cone number count, it appears that E. superba shrink during winter.