Two visual pigments, rhodopsin (or visual purple) and porphyropsin (or visual violet) appear to have a widespread distribution. Thus Wald (1949), referring to rod pigments, writes: 'To summarize our present information, true land and sea vertebrates have the rhodopsin system; true fresh-water vertebrates have porphyropsin; while those equivocal forms which can distribute their lives between fresh water and one of the other environments frequently possess both photopigments, mixed or in temporal succession'.These visual pigments are carotenoid-protein complexes. The above conclusion was reached mainly as the result of studies of the decomposition products ofthe carotenoid moieties (Wald, 1935(Wald, a, b, 1936(Wald, a, b, 1942. These products were found to be either vitamin A1 or vitamin A2, the former associated with rhodopsin and the latter with porphyropsin. The absorption spectra of the retinal extracts were found to fall into one of three groups; they had absorption maxima either around 500 mu (rhodopsin group), or 520 m,u (porphyropsin group) or between the two (mixtures of rhodopsin and porphyropsin) (Wald, 1938(Wald, , 1939(Wald, , 1941 However, the presence of differing amounts of yellow impurities in the extracts complicated the inter-species comparison of absorption spectra and wide limits were tolerated by Wald (1938Wald ( , 1939 in assigning a visual pigment to the rhodopsin or porphyropsin groups. Furthermore, no tests ofhomogeneity were applied.It has been shown (Dartnall, 1952 a) that difference spectra-the changes in absorption spectra caused by bleaching-are highly reproducible under certain conditions and hence provide a means of precise comparison between extracts from different species. Moreover, the homogeneity of the extracts can be put to test by comparing the difference spectra obtained by bleaching with lights from different parts of the visible spectrum.