A. GLENN RICHARDS 



149 



discrepancy between expected and obtained value for development at 15° ; 

 see fig. 2, curve B). 



But if the day-degree accumulation idea is correct, why is there a finite 

 threshold well removed from zero developmental rate? This threshold is 

 sharp for constant temperature experiments: about one-third the normal 

 hatching i)ercentage occurs at 15°, but zero ])ercentage at 14°C. The thresh- 

 old is also sharp for alternating temperature experiments when only two 

 or three temperature shifts are made. In early stages of development, 

 Oncopeltus eggs will tolerate 35%-45% of the developmental time at below 

 threshold; in later stages they will tolerate 25%-35% of developmental 

 time at below threshold. The effect is sharp, practically a cut-off (fig. 4), 

 and nearly independent of when the subthreshold temperature treatment is 



Fig. 4. Plot showing relation 

 between hatching and length of 

 time at a subthreshold tempera- 

 ture for Oncopcltus eggs. Solid 

 circles: percentages fully 

 hatched ; open circles : same plus 

 those that ruptured the egg shell 

 without emerging completely 

 ('semi-hatched'). After Lin et 

 al, 8. 



UJ 

 .1 Q 



o 



X 

 CO 

 Ul 



a: 



X 



UJ 



:e 



Q. 

 LU O 



a 



i- 

 z 



UJ 



o 



(T 

 UJ 

 Q. 



m 

 d 



o 



100-- 



50-L 



25-- 



10- 



5 -■ 





10 



— I- 

 25 



-•) 



100 



PER CENT HATCHED 



given (fig. 5). Clearly, there is no 'sensitive period', i.e. no one stage in 

 embryonic development which has a threshold requiring for its com- 

 pletion a temperature higher than that of the hatching threshold. 



A number of other possibilities have been eliminated (8). Thus, the 

 periodicity of frequent temperature alternations is not important as shown 

 by alternations arranged on the basis of a table of random numbers. Time 

 itself is not important since it can be varied widely by the use of various 

 combinations of three or four different temperatures which result in the 

 same developmental time being required although the actual time varies 

 considerably. Another obvious suggestion is that the hatching process it- 

 self is critical but we find that if most of development has been undergone 

 at 20° or 25°, hatching will readily occur at temperatures several degrees 

 below the normal or constant-temperature hatching threshold (this agrees 

 with Johnson's (7) data for the bedbug). In fact, numerous attempts to 



