E. K. Nabocrs 167 



individuals appeared, aud the expectations in regard to the propor- 

 tionate numbers have been fairly realized. 



The Mendelian assumption that hybrids do not produce gametes 

 representing themselves, but give gametes of the species from which 

 they themselves were formed and that these gametes are produced 

 alternatively in about equal proportions, accounts, with the five 

 exceptions, for all the results which have come from my breeding 

 experiments with the grouse locusts. 



In Exps. I (F,-), II (&) and (c). III (a), (6), (c), (d), and V (a), (b). 

 the ancestry of the parents used was known for one or more generations, 

 and their resulting progeny, in the matter of patterns completely (with 

 the five exceptions in thousands), and in the matter of proportionate 

 numbers, fairly, approximated to the expectations of alternative in- 

 heritance. As the results from known parents are closely approximate 

 to expectation, it seems reasonable to expect the equally regular and 

 similar results from parents whose ancestry was not known at first 

 to lead to the identification of the constitution of the parents them- 

 selves. It has been by this method that the constitutions of the 

 parents from the field and group cultures used in Exps. I, II, and IV 

 have been determined. 



(2) The appearance of lone/ and short winged n ess. A glance at the 

 behaviour of the wing lengths character shows that the short winged 

 required the maximum of time to reach maturity after hatching, and 

 that this great length of time is closely correlated with the time of the 

 year — the fall and winter months. The long winged individuals on 

 the other hand required a minimum of time to reach maturity after 

 hatching, and this minimum time is also closely correlated with the 

 time of the year — the spring and early summer. The time of the year 

 during which growth proceeds seems to determine whether it shall 

 extend over a long period or not. If the time for growth be a long 

 one the wing lengths are likely to be short; if the time required for 

 gi'owth be a short one the wing lengths are likely to be long. The 

 length of the wings of the parents does not condition this character in 

 the progeny. The progeny of short winged individuals become long 

 winged if they grow quickly in the spring. The progeny of the long 

 winged individuals become short winged if the growth take place 

 slowly during a long time. Long winged individuals may produce 

 a majority short winged if the growth take jDlace from October to 

 April, while their brother and sister short winged ones may produce 

 all long winged, if the growth progress from March to June. Nor does 



