COLOR SENSITIVITY OF THE PERIPHERAL RETINA. 47 



the physiological effects of these edges were sub-liminal, and that the 

 stimuli which were actually operative physiologically in our experi- 

 ments represent narrower bands than we have given in the table. 



B. METHOD. 



In these experiments only the horizontal meridian, nasal and tem- 

 poral, was explored. Each sitting was preceded by fifteen minutes' 

 adaptation in absolute darkness. The exploration began at the peri- 

 phery and was continued to the paracentral region, within 10 of the 

 center of the fovea. The method employed was a procedure without 

 knowledge. In no case did the observer know what stimulus was to be 

 employed ; in no case did he see it in direct vision until after the series 

 had been completed. 



After the adaptation period had elapsed, the lights were turned on 

 in the two lanterns, and the observer was asked to fixate. The stimulus 

 color was still invisible, since it was hidden by a shutter. When accurate 

 fixation had been secured, signals were exchanged, and the stimulus 

 color was exposed for a period of three seconds. Then the lights were 

 turned off, the observer's record was taken, and the stimulus-lantern 

 was moved to a position ten degrees nearer the fixation-point. After a 

 rest of six or seven minutes the experiment was resumed, as before. 

 The three seconds' period of exposure was at first marked off by a 

 metronome, but it later became possible to estimate the period from 

 memory with a high degree of accuracy. The subject wrapped his head 

 in a heavy cloth while the readings were being taken and the stimulus- 

 lantern was being moved to its new position. The readings and settings 

 were made by the light of a portable electric lamp (the pocket flash- 

 light). This lamp was found to be exceedingly convenient, and admir- 

 ably suited to dark-room experiments. 



The earlier records are not included in our published results. Some 

 time was lost in practice before the observers acquired the power to 

 maintain an accurate fixation during the exposure of the color. It was 

 found, too, that in our earlier experiments the interval of rest which 

 elapsed between consecutive exposures was too short. After a good 

 deal of experimentation it was decided that six minutes was about the 

 optimal interval of rest, under our experimental conditions. A shorter 

 interval could not be employed for reasons which will be discussed later. 



For the sake of control each meridian was explored twice, at differ- 

 ent sittings. During the first exploration the stimulus-lantern was set 

 successively at 90, 80, 70, etc.; during the second, it was set at in- 

 tervening positions, 85, 75, 65, etc. In our tabulated records we have 

 combined the results of each pair of explorations into a single series. 



