662 PHOTOPERIODISM IN VERTEBRATES 



by its stipulated attributes, intermediate pliases cannot be tested for 

 directly, but must be inferred from results of contrasting usually dual- 

 stage experimental conditions. This methodology allows the perform- 

 ance and evaluation in a given length of time of a much greater num- 

 ber of experiments than that in which the terminal test is supplemented 

 by initial and periodic histological diagnoses, but if the writer's under- 

 standing of the argument is correct, it robs some of the conclusions of 

 finality by its greater reliance on inductive inference, quite apart from 

 that implicit in the nonparametric statistical analysis required by the 

 small number of fish per experiment. 



Nevertheless, a plausible pattern of four day length-temperature 

 combinations differently conditioning four successive phases of the 

 sexual cycle (Table I) is built by inferences from an impressive num- 

 ber of ingenious experiments. For example, phase was established 

 by the fact that only very young under-yearling fish subjected to long 

 days at high temperature (for over a year) failed to mature, whereas 

 controls of the same age first exposed to one of the alternative condi- 

 tions (Table I) and later to long days and high temperature did 

 mature. On adults, the effects of long days and high temperature are 

 consistent with their effects on Phoxinus, Notropis, Enneacanthus et 

 alia. As none of the four combinations of day length-temperature 

 (Table I) exerted influence on fish in phase 2 noticeably different 

 from those of the other three combinations, phase 2 may represent 

 merely a stage of maturity so advanced that further maturation is sub- 

 stantially independent of such environmental influences. Under con- 

 stant long days at high temperature (16 hr and 20°C), both sexes of 

 Gasterosteus have about a 200-day cycle in which a reproductive and 

 a nonreproductive period alternate. In females the nonreproductive 

 period averaged about 168 days but in males was somewhat shorter. 

 As Baggerman notes, the durations of the cycle in nature and in the 

 laboratory would differ somewhat, because the extrinsic conditions are 

 different, for in either case the inherent rhythm is doubtless modulated 

 by the coaction of external and internal factors. In regard to the re- 

 fractory periods in the cycles of Gasterosteus and Notropis, it is worth 

 mentioning that in the annual cycles of some fishes, periods of refrac- 

 toriness to hypophysis injection have been reported (Gerbilskii, 1950, 

 1951; Kazanskii, 1952; Vivien, 1941). The presumptive autumn 



