ADJUSTMENT OF FLATFISHES 419 



shade, lying upon a white or a black background may be caused to 

 vary in appearance from pale gray to nearly black, without in 

 the least affecting the value of the background. In the case of the 

 present subject, not only might the shade of the fish as a whole 

 be caused to vary enormously, but spots which were in reality 

 very conspicuous could be toned doM^n, almost to the point of 

 disappearance. Thus it would be quite possible to produce some 

 (though by no means all) of the differences which appear in my 

 plates by differences of exposure, development or printing. ^^ 

 There is therefore abundant opportunity for self-deception and 

 unintentional exaggeration of one's results, unless this source of 

 error is guarded against. The precautions taken, in my own case, 

 were as follows: 



1. My notes upon the experiments give important clues, in 

 respect to fishes to be compared, as for example, when one is 

 said to be ''much darker " than the other, or to have become "much 

 paler" in the course of 24 hours. After considerable experience, 

 one acquires fairly accurate standards of comparison and is able 

 to pass such judgments with confidence. 



2. The negative was usually developed verj^ soon after the 

 exposure was made, and frequently while the fish still remained 

 upon the bottom which had been employed. Thus it was possible 

 either to compare the negative directly with the object, or at 

 least to pass upon the former while the appearance of the latter 

 was fresh in memory. Likewise, prints were generally made before 

 the lapse of many days. 



3. In some cases, fishes which were to be compared were photo- 

 graphed together upon the same plate (e.g., some of the figures on 

 plates 12 and 13, and many others which I have not included as 

 illustrations). 



'^ For example, as I know by actual experience, such differences as those shown 

 in fig. Ij and 11, plate 3, or between fig. 4d and 4e plate 7 might be brought about very 

 readily by differences of printing; and in such a case, the backgrounds would 

 afford little evidence of this misrepresentation. On the other hand, differences 

 such as those between the skin patterns assumed upon gravel and upon sand or 

 between the latter and the highly contrasted effects assumed upon some of the 

 checker patterns would be utterly impossibe to bring about by any differences of 

 photographic manipulation. 



