420 FRANCIS B. SUMNER 



4. Certain of the bottoms, notably the natural sands and 

 gravels, form very good standards of comparison in judging of 

 two negatives. If, for example, the familiar dark mixed sand of 

 my experiments has the same value in two cases (either in nega- 

 tives or prints) it is likely that the relative shades of the fishes 

 have been preserved sufficiently well. Samples of these sands and 

 gravels were saved by me and were referred to during the prepara- 

 tion of this paper. 



5. While making some of my exposures, I included a strip of 

 white opaque glass, painted so as to be divided into three bands, 

 white, black and gray respectively. This was of considerable 

 vaule in certain cases, but its application was rather limited. 



6. Finally, a plate commonly bears in itself evidence of its 

 having been over or under-exposed or developed, and this may be 

 allowed for in printing. 



In spite of such errors as may have crept in, owing to imperfect 

 technique in photography, I cannot believe that the value of my 

 results has been very materially impaired. Furthermore, I 

 believe that the consequence of such defects of technique has, on 

 the whole, been rather to decrease than to exaggerate the differ- 

 ences of pigmentation which are portrayed in my illustrations, 

 and to minimi/^e many of these instances of adaptation. For the 

 color (as distinguished from the shade) was frequently an im- 

 portant element in the adaptation and this, of course, is not 

 indicated in the figures. Again, the condition of extreme pallor 

 assumed by many of the specimens upon a white background is 

 impossible to reproduce by ordinary photographic processes, ^^ 

 since the skin of the fish always retains a certain amount of ^^ellow, 

 and this, as is well known, looks disproportionately dark in a 

 photograph. Thus figures lA', 10a, etc., give a very imperfect 

 idea of the degree of blanching which these fishes had undergone 

 upon a white bottom, and fig. 4i greatly minimizes the extent of 

 the adaptation of this specimen to its gray background. Accord- 

 ing to my notes the appearance of the two harmonized very well 

 at the time, but the photograph represents that of the former as 



18 A color screen might have been used, but this was not done. 



