Summary.The scores obtained by the three colours of insects “n,”“l” and “i” on background “n” are 163, 271 and 376 respectively, and show that insects differing greatly from the colour of the background are destroyed, while those agreeing with it are preserved. Insects coloured “l” are only partial misfits and occupy an intermediate condition. These conclusions are confirmed by the experiments on background “i,” where the order of destruction is reversed. The scores of 270, 284, 278 and 274 obtained by insects coloured “1” on backgrounds “a” to “g” respectively suggest that these differences have no significance, but are fluctuations about a mean value (probably 270). The scores obtained by insects coloured “n,” 362, 300, 282 and 270 for back‐grounds “g” to “a” respectively, show that on passing to the lighter backgrounds the scores approximate more and more to those of insects coloured “1.” The insects coloured “i” obtained scores of 174, 232, 244 and 270 for backgrounds “g” to “a,” and show a similar tendency. As the difference between the colour of the background and the insects increases, any variation in the colour of the latter is rendered more and more inconspicuous, and in consequence the fish find more difficulty in distinguishing between the three types of prey. One advantage of the Ostwald Colour Chart is that there is a constant apparent difference between two adjacent standards, and, therefore, the predators must experience the same difficulties in distinuishing between insects coloured “1” and “i” on background “a” as between insects coloured “n” and “1” on background “c.” This means that the scores of 270 obtained by insects coloured “1” and “i” on background “a” are a measure of the extent to which the predator is able to distinguish between the two types of prey differing from the background by four and five standards, and these can therefore be added to the scores of 282 and 284 obtained by insects coloured “n” and “1” on background “c.” Combining these two sets of figures, it will be seen that the scores of 552 and 554 were obtained by insects differing in colour from he background by five and four standards respectively. Experiments performed with greater accuracy might show that predators can distinguish between these two types of prey. Even if the results obtained from the original set of experiments on background “a” are combined with the above figures, the scores are increased to 7718 and 784 respectively, still failing to give any evidence of selection. Another advantage of the constant apparent difference between any two adjacent standards is that it is possible to compare the results of any one set with those of the next, and so combine all the sets and deduce the results which might be obtained were seven sets of insects (coloured. “a” to “n”) used on backgrounds (‘i “and’(n.” The calculations are very simple, as th