that among 100 women students at Wellesley College the order of preference was not very different, being blue (38 per cent), red (18 per cent), yellow, green, violet; in a later investigation the order remained the same, there being only some increase in the preference for red; it was considered that association accounted for the preference for blue, while more conscious as well as more emotional elements entered into the preference for red.
By far the most extensive investigation of color preference was that carried on at Chicago by Professor Jastrow on 4,500 persons, mostly adults, of both sexes and various nationalities. Blue was found to be the favorite color, less than half as many persons preferring red; of every thirty men ten voted for blue and three for red, while of every thirty women five voted for red and four for blue. The men also liked violet and on the whole confined their choice to but few colors, the women also liked pink, green (very seldom chosen by men) and yellow, and showed a tendency to choose light and dainty shades. There was on the whole a decided preference for dark shades; the least favorite colors were yellow and orange. It is evident that, as we should expect, within the elementary field of popular esthetics, women show a more trained feeling for color than men.
It is not quite easy to coördinate the various phenomena of color predilection. Careful and extended observations are still required. It seems to me, however, that the facts, as at present ascertained, do suggest a certain order and harmony in the phenomena. It is difficult not to believe that there really is, both among many uncivilized peoples and also many children at an early age, even to a slight extent among civilized adults, a relative inability, by no means usually absolute, to recognize and distinguish the tones of color at the more refrangible end of the spectrum. The earliest writers on the subject were wrong when they supposed that color nomenclature at all accurately corresponded to color perception, and it is well recognized that there are no peoples who are wholly unable to distinguish between green and blue and black. But as Garbini has clearly shown, there really is a parallelism between color nomenclature and color recognition, and Garbini's wide investigation has confirmed the experiments of Preyer on a single child by showing that there is a certain hesitancy and uncertainty in recognizing the colors at the more refrangible end of the spectrum, long after children are familiar with the less refrangible end. In the same way the important investigations of Rivers have confirmed the earlier observations of Magnus and Almquist in showing that savages in many cases exhibit a certain difficulty in recognizing and distinguishing blue and green, such as they never experience with red and yellow. The vague-
- J. Jastrow, "The Popular Æsthetics of Color," Popular Science Monthly, 1897.