Especially in view of the recent work on the significance of clutch-size (Lack, I947), it is desirable to ascertain to what extent parent birds adjust the frequency of their feeding visits to the number of young in the nest; whether in the amount of food they receive the members of larger broods are at any disadvantage compared with those of smaller broods; and whether a lower feeding-rate is correlated with slower development of the young in the nest. This study is based primarily on about 5000 hr. of observations made by supervised African observers between 1937 and I946. They are selected from the later stages of the nestling periods, when the feeding activities of the parents were not confused by brooding. All the work was done at Amani, in Tanganyika Territory at 50 S., except for one species, the Colletoptera, which was studied about 50 miles away on the shores of the Indian Ocean. Amani is an evergreen locality in which forest of an especially luxuriant type predominates, as described elsewhere (Moreau, I935); but most of the data on which the present analysis is based are derived from hirundines and swifts nesting in clearings. The length of day in these latitudes is always within 20 min. of 12 hr., and the active period of all diurnal birds, including swifts, varies throughout the year only between about i21 and 13 hr. The observations were made in continuous spells of 7 or 8 hr. as a rule, but occasionally from dawn to dusk; and for each nest from 4 to 13 days of observations are available for the present purpose. As I have had occasion to remark before, I am satisfied of the general reliability of the Africans' work. In the detailed accounts I have already published of the breeding biology of some of the birds here dealt with, feeding-rates were calculated on a time-unit of 200 min., since the hour is an inconveniently short unit for recording activity of swifts. However, the use of the 200-min. unit has given rise to some difficulty and misconception. In the present study all feeding-rates have therefore been calculated per hour, although for the swifts the result is a series of fractions. In other species, where the feeding rates are big enough, I have worked to the nearest whole number. The possibility cannot in most species be excluded that on particular days or in particular nests lower feeding-rate may be correlated with, and so off-set by, bigger or more numerous morsels per visit. However, in each species such cases presumably occur irrespective of number in brood, so that they would not invalidate a comparison between the means, or the ranges, of feeding-rates for the respective numbersin-brood.
Read full abstract