exceptions to the rule, and also seems to interfere with the likening of children to savages, which was formerly so useful. Observations of the children of such tribes have never been made, to my knowledge. Before going on, however, to consider the case of savages, we must look briefly into the problem suggested by the fact that there is choice of any colour. The child’s susceptibility to the cooler colours, and even its perception of them, especially blue and gray, has been questioned. Preyer says: “The inability of my two-year-old child to recognise blue and gray can be argued not only from his occasional failure to do so, but also from the evident difficulty he encounters in connecting the commonly used and familiar names ‘blue’ and ‘gray’ with any special sensations, while ‘yellow’ and ‘red’ were correctly applied several months ago. Were the sensations of blue and gray as clear as those of red and yellow there would be no failure to recognise the colours. The child does not know what green and blue mean, though he does know red and yellow.... Even at four years blue was oftener called green in the morning twilight, though to me it was clearly blue. The child was greatly astonished to find that his blue stocking had become gray overnight. For years very dark green was called black.”[123] These striking observations seem indeed partially to confirm the hypothesis of Geiger, Gladstone, and Magnus, who came to the conclusion, from the study of ancient picture writing, that primeval man distinguished only the three primary colours (the Young-Helmholtz theory)—red, green, and violet. From these were derived orange and yellow, while blue was the very last to be discovered. Yet, indeed, so far as any philological support is concerned, the hypothesis can hardly be maintained either in regard to the ancients or to modern low-standing tribes. In the remains of buildings and plastic works, which are older than any picture writings, traces are found of all the colours of the spectrum, and the philological test, when applied to civilized peoples, does not yield the confirmation which advocates of the theory desire. While it is true that the Esthonians have no word of their own for blue (their _sini_ is borrowed from the Russian), but the apparent deduction from that fact is rendered doubtful, to say the least, by this passage from Raehlmann: “Some time ago I tested an old Esthonian peasant woman with a gray starling. She was not quite sure of the name of the colour, and changed it often. On closer questioning about her ideas of colour, she seemed to have the spectral series correctly in mind, distinguishing the colours as blood, wax, grass, and sky. She had never needed other terms with which to express her sensations, but she took pains to convince me that she had perfectly clear ideas on the subject of colour.”[124]