736 



SCIENCE. 



[N. S. Vol. IX. No. 230. 



minated the pigments to be studied, and 

 could be moved along the bar. All colors 

 were brought to the same intensity before 

 observations, that is, the photometer, con- 

 taining the colored cai-d, was set at a defi- 

 nite distance from the right-hand lamp, and 

 the other lamp moved backward or forward 

 until a balance was approximately attained. 

 It was left in this position, and the set of 

 observations on any given color made by 

 moving the photometer head in the usual 

 way. Thus the uncertainties were avoided 

 which arise from working with lights of 

 small intensity. 



Six colored papers were selected from the 

 set published by the Milton Bradley Com- 

 pany — red, orange, yellow, green, green- 

 blue, blue. Each of these was examined 

 by direct vision, and at each of the three 

 angles before mentioned. Two concordant 

 series of observations, each involving a 

 large number of readings, were made on 

 diiferent days, and the mean of the two 

 series taken as the iinal result. It soon be- 

 came evident that the pigments at the red 

 end of the spectrum decreased in bright- 

 ness from the center to the periphery of the 

 retina, while those nearer the blue end in- 

 creased in brightness. It seemed probable 

 that some color must exist for which the 

 brightness would be the same for all parts 

 of the retina, and to locate this color more 

 closely the intermediate pigments yellow- 

 green and green-yellow were added to the 

 set originally selected. The results are 

 exhibited in Table I. 



It is shown by this table that the yellow- 



green remains nearly at the same bright- 

 ness for all angles of vision ; that, in fact, 

 the brightness curve for the whole set of 

 pigments might almost be said to rotate 

 about this color as an axis, the red falling 

 and the blue rising as the periphery is ap- 

 proached. The character of the change is, 

 perhaps, more clearlj^ shown in Table II., 

 which is derived from Table I. by multi- 

 plying each series of figures by a factor 

 which brings the yellow-green value to 

 unity, changing all other results in like 

 proportion. The value for any color at any 

 angle of vision may then be directly read 

 as a percentage of the value for yellow- 

 green. 



Table II. 



0° 30° 50° 70° 



E. .514 .275 .194 .165 



O. 1.303 .637 .495 .471 



Y. 2.070 1.620 1.468 1.381 



G.Y. 1.300 1.167 1.096 1.057 



Y.G. 1.000 1.000 1.000 1.000 



G. .631 .720 .819 .818 



G.B. .529 .680 .717 .717 



B. .231 .324 .381 .404 



It is thus seen that while the red falls in 

 peripheral vision to about one-third of its 

 brightness when viewed directly, and blue 

 is nearly doubled in brightness, yellow is 

 reduced in brightness by about one-third, 

 and yellow-green, that portion of the spec- 

 trum where we should expect the greatest 

 brightness if the peripheral color-blindness 

 were of the same character as ' twilight ' 

 color-blindness, remains practically the 

 same at all angles of vision. Yellow is 

 still the brightest of the colors, and the 

 maximum is shifted but little toward the 

 blue. 



It is to be noted, also, that there is com- 

 paratively little change from 50° of ob- 

 liquity outward. At 50° most colors are still 

 distinguishable; at 70°, none of them. At 

 50° the apparatus which gives us the sen- 

 sation of color must still contribute its 

 quota to the sensation of brightness, as in 



