THE UNKNOWN FACTORS OF EVOLUTION. 91 



In Weismann's language, on the other hand, in explanation of 

 dimorphism in hymenoptera and other types, there are certain 

 sets of biophors corresponding to certain possibilities of adult 

 development. Apply this to the celebrated case of the flat-fishes 

 and the remarkable results recently o.btained by Agassiz, Filhol, 

 and Giard in artificially producing more or less symmetrical 

 flat-fishes by retaining the young near the surface. Weismann's 

 interpretation of the evolution of flat-fishes has always been 

 that it was by the selection of asymmetrical and elimination of 

 symmetrical 'determinants.' In the light of these experiments 

 he must now recast this explanation by saying that the flat- 

 fishes have kept in reserve a set of symmetrical ' determinants ' 

 since the period when our first record of the asymmetrical type 

 appears or about three million years ! 



This attack upon the speculations of one writer is a digres- 

 sion. What I really wish to bring out is the necessity of a 

 far more critical analysis of the various kinds of evidence for 

 Buffon's factor. This necessity may be illustrated by the 

 different interpretations of color change in direct response to 

 changed environment. 



The most significant experiments upon color are those of 

 Cunningham upon the flat-fishes. He has proved that during 

 the early metamorphosis of young flat-fishes, when pigment is 

 still present on both sides, the action of reflected light does 

 not prevent the disappearance of this pigment upon the side 

 which is turned towards the bottom, so that the color passes 

 rapidly through a retrograde development ; but prolonged ex- 

 posure to the light upon the lower side causes the pigment to 

 reappear, and upon its reappearance the pigment spots are in 

 all respects similar to those normally present upon the upper 

 side of the fish. It is very important not to confuse these 

 results, of deep mterest as they are, with those obtained where 

 the environment is new in the historic experience of the organ- 

 ism. Experiments upon color, therefore, afford a marked illus- 

 tration of the necessity of drawing a sharp distinction between 

 cenogenic and palingenic variations. We have, in many cases, 

 been mistaking repetitions of ancient types of structure for 

 newly acquired structures. When the pale Proteus is taken 



