FLUCTUATIONS IN A PARTRIDGE POPULATION 123 



the death of both adults and juveniles, but the main factor is the movement 

 of young hens (paired) beyond the estate boundary. Loss from this cause is 

 most marked in those coveys whose winter ranges were in the peripheral 

 region of the estate, and has shown relatively little annual variation. 



While some variations in breeding density have occurred, the policy of 

 exploitation has deliberately maintained the breeding stock at a relatively 

 high level, and after an initial period of increase annual variations have largely 

 depended on crop distribution during the late winter and spring. 



Variations in the rate of summer gain (as measured by the September/ 

 March census or by the young/old samples in August) have been very large 

 and are mainly responsible for the fluctuations of the September population 

 densities. If these populations are expressed as percentage gains over the 

 March populations, then the range is from 494 per cent in 1949 to only 

 38 per cent in 1954. In terms of young/old ratios in August the variation has 

 been from 4-17 to 0-55 young per old bird, and the causes of these very- 

 marked differences will now be considered. The possible causes of these 

 variations may be arbitrarily divided into those which are effective prior to 

 the eggs hatching, and those which are mainly effective in the post-hatching 

 period. 



In countries with a more rigorous winter cHmate and with a relatively 

 short period between the melting of the winter snows and the commence- 

 ment of laying, Siivonen (1956) has claimed that the weather during this 

 period of gonad development has a marked effect upon the subsequent 

 production rates of game birds. The effects of weather at this time are said 

 to be apparent in variations in clutch-size and hatchabihty, as well as in 

 the survival rates of the subsequent broods. In England, where prolonged 

 snow covering at this time is rare, weather conditions in early spring do not 

 appear to be critical, and the minor variations in clutch-size and hatcha- 

 bility show no correlation with the chick survival rates that follow. Indeed 

 the partridge usually lays a second clutch if the first is unsuccessful. Repeat 

 clutches are smaller than first clutches (Middleton (1936), Lack (1947), 

 Blank & Ash (in press)), and the average clutch-size in any one year would 

 depend largely on the number of repeat nests included in the sample. On 

 the area studied, where every effortds made to protect nests, nesting success 

 has shown considerable variation from year to year, and has had much more 

 effect on the chick production rate than variations in either clutch-size or 

 hatchability (Table III). And yet from 1949-59 the chick production rate has 

 averaged 9-89 chicks per pair with a range of only ±2 -20. (If 1953 and 1958 

 are excluded, the range is from +o*73 to —0-87.) Nor do these relatively 

 minor variations necessarily show any correlation with the subsequent 

 number of young birds per pair when six weeks old. 



