136 LESLIE B. AREY 



by pl()ttiiif>; toinperatures as abscissas and the quantitative 

 amount of distal migration as ordinates would show a gradual 

 slope from 0° to 10°C\, from 10° to 15°C. a rapid drop and from 

 15° to 25°C. a nearly horizontal but slightly sloping line. 



The results of experiments performed in the dark, where the 

 pigment is highly contracted, are usually not as clear cut as 

 those just described. The reason for this is because it is im- 

 possible to see the qualitative distribution of the pigment granules 

 and hence decisions regarding the effect of temperature must 

 depend largely on measurements of the width of the narrow 

 pigmented layer. Such a criterion, as has previously been 

 pointed out, is open to the criticism that individual eyes may 

 vary enough in the absolute amount of contained pigment to 

 disconcert judgments concerning the effect of temperature. 

 After having studied a great number of preparations, I do not 

 believe that such an unequal pigmentation is in truth a factor 

 that warrants serious consideration. However this may be, 

 an obvious precaution consists in the prolonged repetition of 

 each type of experiment. It may be said that in the course of 

 my experimentation on the effect of temperature on normal 

 fishes alone, over 200 retinas have been examined. 



The evidence obtained from Ameiurus, was more conclusive 

 than that from any of the other fishes, with the possible excep- 

 tion of Carassius. At 25°C. (fig. 4) the pigment forms a densely 

 contracted layer, the mean width of which is about 25 /u- In 

 contrast with this is the condition at 5°C. (fig. 3), where the cells 

 have short pigmented processes, the total extent of which is 

 approximately 38 ix. The differences apparent at these two 

 extremes of temperature were so slight in comparison with the 

 much greater variation in the light, that thorough experimenta- 

 tion at the intermediate temperature of 15°C. was not attempted, 

 although it was sufficiently demonstrated that the results at 

 this temperature do not vary to any great extent from those at 

 the two extremes, and probably more closely approximate the 

 highly contracted condition at 25°C. 



The results of some of the temperature determinations were not 

 conclusive. Reference has already been made to the fact that 



