26 MARGARET FLOY WASHBURN : time a bit of vivid red drapery hanging on the wall of a room. On one occasion, entering the room hastily and glancing toward the red drapery, I saw a distinct patch of green on the white wall, and it was a second or two before I realised that the red cloth had been removed, and that what was actually seen was the white wall itself. I can think of no other explanation for this phenomenon, which was as unmistakable as it was unexpected, than the hypo- thesis that the green patch was a contrast colour, induced by the expected, centrally excited, or subjective red. Nor is this hypothesis really weakened by my failure to find analo- gous results in the experiments described below. In the case of a phenomenon which can occur only when favoured by very complex conditions, one well-marked positive in- stance should outweigh a number of failures. As a matter of fact, the apparent contrast -effect in these experiments turned out later to be due in all probability to alterations in the brightness of the stimulating light. The image used as a basis for the following experiments was that produced in the retinal field of the closed eyes by prolonged white-light stimulation. The conditions were carefully regulated in accordance with those laid down by Helmholtz in the Physiologische Optik. The subjects sat in a room lighted by the upper half of a window whose lower half was screened. The eyes were fixed upon one point of the window-frame for twenty seconds, and were then closed and covered. Nothing but sky could be seen through the window, and careful note was always taken of the degree of illumination. The precaution was, of course, observed to free the eyes entirely between experiments from all traces of the previous image. The subjects experi- mented on were the writer (W.) and three students of psychology (M., 0. and D.). Of these only W. had general practice in psychological experiments. Fortunately, how- ever, a reliable test could be had both of the degree of special practice attained by the subjects in the course of the research and of the influence of external sources of error. This test lay in the uniformity of the colour changes, observed in the ordinary unmodified image. A wholly un- practised observer watching the course of an after-image for the first time reports chaotic results, and no two such observers agree as to the alterations in colour which occur. No results were taken account of from* an} 7 subject until she was sufficiently practised to find the colour changes ap- proximately uniform or affected only by such fluctuations, as could be accounted for from external causes.
