THE KILLIFISH 35 



believe that their researches on the whole are favorable 

 to Bert's view (Carnot, 1896; Sollaud, 1908; Redfield, 

 1 91 8; Kahn, 1922; Giersberg, 1930; Smith, 1931*2) are 

 about equal to those who take the opposite stand (Spaeth 

 and Barbour, 191 7; Hogben, 1924; Gilson, 1926; Sand, 

 1935). The inconclusiveness of much of the work on 

 this question probably resulted from the fact that it 

 involved experiments with drugs mainly from the stand- 

 point of mammalian physiology, a somewhat uncertain 

 procedure when transferred to the lower vertebrates, 

 and further that no small part of it was done on am- 

 phibians where as we now know, thanks principally to 

 Hogben and his associates, chromatophoral nerves are 

 mostly, if not entirely, absent. 



It is not my purpose here to attempt to deal with 

 this question in a general way, for, as will be shown 

 toward the end of this essay, the truth is that double 

 innervation is not a general question. Melanophores 

 may have double innervation, as I believe to be the case 

 in Fundulus, or single innervation as has been shown to 

 be true of the dogfish, or no innervation at all as appears 

 to be the case in lampreys (Young, 1935) and in the 

 frog. The point that I wish to discuss here is whether 

 there is reason to assume double innervation for the 

 melanophores in Fundulus. 



Some of the evidence leading up to this view has 

 already been given, but there remain certain important 

 aspects of this question still to be considered. In 1932 

 Mills pointed out that a close scrutiny of the melano- 

 phores at the edge of a caudal band in Fundulus showed 

 very important differences depending upon the state of 

 the color-cells. The differences here considered are best 

 seen when the band itself is somewhat blanched. If 

 under such circumstances a given fish is darkened by 

 being kept in an illuminated black-walled vessel, the 



