738 REPRODUCTION AND MIGRATION IN BIRDS 



to October. On the other hand, Schildmacher (1956) found that 

 Chloris chloris showed a considerable degree of testicular develop- 

 ment on 8-hr photoperiods, although several months behind the 

 attainment of corresponding degrees of development under normal 

 daily photoperiods. Zonotrichia leucophrys nuttalli and Chloris chloris 

 are both early breeders. In our laboratory (Farner, unpublished) we 

 have noted similar responses with a small number of Passer domesti- 

 ciis, another early breeder, held on short daily photoperiods. Miller 

 (1955) has interpreted his data as indication of an "innate" reproduc- 

 tive rhythm. It is possible that this interpretation is correct, although 

 I think that it would be wiser to defer such a desisnation until the 

 rate of testicular development (k of equation 1 ) as a function of the 

 photoperiod (p) is established experimentally, for it is possible that 

 each of the above cited cases may still represent a photoperiodic 

 response. If, for these species, the curves representing k as a function 

 of p are similar to that in Fig. 2 but with lower tails such that A: is a 

 continuous positive function of p until p is equal to, or approaches, 

 zero, then these testicular developments could be interpreted properly 

 as photoperiodic. 



Although the photoperiodic testicular response has been demon- 

 strated extensively in birds (Table I), it is only recently (Farner and 

 Wilson, 1957b) that a quantitative demonstration has been made of 

 its dominant importance in the timing of the testicular cycle. Knowing 

 k empirically as a function of p (equation 2), and knowing the 

 natural duration of the daily photoperiods, we were able to "predict" 

 testicular weight for a particular date by, 



n 



logio W„ = logio TFo + S ki (3) 



1 



where n is the number of days beyond the date of beginning of testicu- 

 lar development. Fluctuations in light intensity because of cloudiness 

 and other factors were ignored because of the low maximal light in- 

 tensity of the response. It was also necessary to ignore possible differ- 

 ences in wavelength composition between the artificial light, which was 

 used in establishing the experimental rates, and natural light; differ- 

 ences in response resulting from such differences are probably small, 

 again because of the low maximal threshold of the response. Data on 



