13s 



THE AMEBIC AN NATUEALIST [Vol. LII 



ties indicated by the reciprocal of the hyperbola to which 

 the data partially conform. 



Factors other than temperature influence the rate of 

 development, The work of Yung showed that in the case 

 of the frog light is one of these. Unfortunately the light 

 was not measured definitely in the work of either Krogh 

 or Yung. Yung kept one lot of developing frogs in the 

 dark and one in a window but where the sun actually 

 never shone on them. Krogh 's work must have been done 

 in similar light. Yung's larvae were reared under the 

 light conditions which he used, for a month or two months, 

 and thus his data are for older stages than those of Krogh, 

 whose results relate to the appearance of the first cleavage 

 furrow. Accordingly, any comparison of the two sets of 

 data is essentially impossible. However, for the purpose 

 of illustrating a principle which is indicated relative to de- 

 velopment under the influence of various intensities of 

 factors other than temperature, I have called the light con- 

 dition under which Krogh 's work was done 10 units and 

 have shown it on a scale at the right-hand side of the 

 graph. It is probable that too strong light will retard 

 development as well as too weak light. Hence the scale 

 is shown double, 12-8, 14-6, etc. ; either increases or de- 

 creases in light intensity are assumed to increase the time 

 required for development. The cross shown on the graph 

 gives the approximate total temperature for darkness in- 

 dicated by Yung's work. This part of the chart is given 

 merely to indicate a method of chart making— of showing 

 the way in which variations of one other factor change the 

 number of time-temperature units required for develop- 

 ment. 



For practical prediction such a curve must be drawn 

 for the shortest time for development at each tempera- 

 ture. This will be under optimum light, chemical, etc., 

 conditions for the temperature concerned. In establish- 

 ing such a least-total temperature curve a few careful 

 determinations within the straight-line limits with other 

 factors optimum will suffice. Outside these limits the de- 



