CROSSES OF JAPANESE TUUTLE-DOVES WITH lil.ONI) \\D \\HITI. Id \(.-li< ,\ I -. 73 



however, with the results found generally in the other crosses, and a consideration 

 of this matter may be undertaken in connection with these data. The effects of 

 "season" upon sex are , after all, not easily separated from other usuries, which 

 also influence the production of sex. Chief among these latter are the matt. 

 the "crowding" of the clutches (overwork), the strength, vigor, health, and a^ 

 parents, and the frequent relative "weakness" of the very first egg, or pair of eggs, 

 of the season. It will be observed in these data, and in (hose of other similar 

 crosses, that "pairs of males" rarely arise in the autumn from crowded clutches i.e., 

 from pairs of eggs produced immediately 8 after other eggs. Likewise "pairs of 

 females" rarely arise from the early part of the season from uncrowded clutch.-. 

 The health and vigor of parents can not always be known in advance, and for this 

 reason the date of death, when known, of all parents is given in the several tabu- 

 lations. The very first egg or pair of eggs in any season are found to produce a 

 higher proportion of females than the next few succeeding clutches; they produce 

 also a higher proportion of infertile eggs and of short-lived offspring than the im- 

 mediately succeeding clutches, and in these several ways warrant their classification 

 as "weak" eggs. 



A condensed and classified tabulation ^f the "shifting of dominance" of 

 of the alba x orientalis crosses is given in tafile 45. Here an arbitrary date July 1 

 in all cases 9 is chosen as the point of separation of early and late -i ,-i-on. It will 

 be noted that before July 1 there were produced 29 males to 10 females; after July 1 , 

 31 males were produced to 33 females. Of course an arbitrary date July 1 in 

 this case can not represent the proper turning-point of predominant maleness 

 to predominant femaleness in all of the 9 series tabulated, since, for example, in 

 some series the maximum production of males continued after July 1, and this 

 tends in the summary to make the number of "late" season males too high. The 

 numbers concerned are not large, but it will be observed that in the "early period," 

 where the method of lumping the data least violates the individuality of the data, 

 8 of the 9 series show a strong predominance of males, and the ninth series is a 

 small one having 1 male and 1 female. In 1908 (see table 42), 3 of the first 4 

 of the series yielded males, while the 10 succeeding eggs gave rise to 10 females. 

 The same pair of parents during the previous year threw 4 males and 1 female before 

 July 1, but 3 females and 1 male after July 1. 



A consideration of the sex data from these crosses is not complete without 

 reference to the evident predominance of males over females in the offspring, 

 notwithstanding the considerable amount of "overwork" effected in several of the 

 series. This predominance is greater in the alba x orientalis cross than in the 

 reciprocal cross; but the average for the two is greater than when orientaKa is mated 

 with orientalis or when the latter is mated with lurtnr, which is only specifically 

 distinct from it. Alba is separated generically from orientalis, though they are 

 closely related genera, having been grouped formerly within a single genus. Again 

 the predominance of males is most pronounced (Chapter IV) when ori< ntuli'* is 

 crossed with Columba, and these two forms are separated by differences of family 



8 About days is the shortest time that an egg may be produced and laid after tin; second egg of the previous 

 clutch was laid. 



9 In the record of 1008, where every egg laid was hatched and the resulting sex UCerUined U 

 division-point is April 1, after which no males were produced, but the 10 further eggs produced females in every case. 



