ABSTRACT A method is described for obtaining statistical results on the phototaxis of G. chevreuxi. A long tube is illuminated from one end and the numbers of animals in two arbitrarily delimited end-sections counted at regular intervals. Wild-type specimens in normal conditions show a moderate degree of negative phototaxis. Animals with one eye varnished show circus movements; hence the phototaxis is true tropotaxis (Frankel, 1931). The sense of reaction can be reversed and the animals made to show a moderate positive phototaxis by the addition of acetic acid. Caffein has no effect. Red-eyed mutants, which lack most of the melanin eye pigment, behave similarly to the wild type, though there are indications that they are often rather more sensitive to light, as shown by stronger negative phototaxis in normal conditions, weaker positive phototaxis after addition of acid. The variability of the results, however, is too great to permit of definitive conclusions being drawn. Albino and colourless mutants, which possess neither retinulae nor optic nerves, show no phototaxis.