Glyn Satterley Book; National Tajst for Scotland 



Houses on a village street, below, abandoned by the islanders after 

 1930, have been restored and are used to accommodate work 

 parties and visiting scientists. Right: Sheep graze among the dry- 

 stone cleits. When their numbers peak, the animals closely crop 

 abandoned fields and lower slopes, even devouring rushes. 



Tim Clutton Brock 



ther depressing the autumn food supply. 

 These factors help to answer the immedi- 

 ate question of why the population shows 

 periodic, dramatic die-offs. They do not 

 tell us, however, why Soay sheep popula- 

 tions should oscillate while those of other 

 ungulates are stabihzed by the effects of 

 increasing density on reproduction or 

 mortality. Do similar processes not occur 

 in Soays — and, if not, why not? 



We have found that rising population 

 density has little effect either on the fecun- 

 dity of the ewes or on neonatal mortality in 

 the sheep through the first two years of the 

 cycle. Even in the third year, 90 percent of 

 the flock's adult ewes become pregnant. 

 Why increasing numbers have so little ef- 

 fect on neonatal survival is easy to see: 

 food is plentiful on Hirta, even in the third 

 year of the cycle. On Hirta, which is about 

 as far north as southern Alaska, days are 

 long and nights are short in early summer, 

 and there is a burst of plant production. 

 Growing lambs have plenty of food during 

 their first months of life, even when sum- 

 mer population is highest, so that popula- 

 tion density itself has little or no effect on 

 lamb survival. The relatively high lamb 

 mortality during the summer following a 

 crash — when population size is low but 

 food is plentiful — occurs because light. 



weak lambs have been produced by ewes 

 that have barely survived the winter. 



The same burst of plant growth in early 

 summer helps to explain why the sheep 

 can remain fecund as their population den- 

 sity increases. After the middle of June, 

 lambs suckle infrequently, and their moth- 

 ers then have several months to recover 

 the condition lost during lactation. As a re- 

 sult, they can reach the necessary weight 

 to conceive by the time of the late October 

 rut, and summer numbers have little effect 

 on the proportion that conceive. 



This situation differs from the breeding 

 cycle of most other ungulates, which wean 

 their offspring much later in the year. For 

 example, red deer on Rum bear their 

 calves in June and continue to suckle them 

 until November or December, after the an- 

 nual rut in October. During lactation, espe- 

 cially in the weeks when their milk pro- 

 duction is highest, the daily energy 

 requirements of females increase as much 

 as fourfold, and mothers typically lose a 

 substantial proportion of their body 

 weight. 



Unlike Soay sheep, female red deer 

 cannot begin to regain this lost weight 

 until the latter months of lactation in late 

 summer, when the demands of suckling 

 calves have dropped. By this time, days 



are shortening, plant growth has dropped 

 back, and food is no longer superabun- 

 dant. High numbers of red deer deplete the 

 food supply in late summer. Consequently, 

 many mothers cannot regain body weight 

 before October and fail to conceive during 

 the rut. As a result, when deer density is 

 high, the majority of mothers breed every 

 other year, substantially lowering the 

 growth rate of the deer population. 



To explain the apparent lack of relation- 

 ship between population density and fe- 

 cundity in Saint Kilda's sheep, we needed 

 to compare the weights of mothers that 

 had raised lambs during late summer with 

 the weights of those that had not. But to 

 weigh a sample of ewes, one must first 

 catch them — and these animals are unac- 

 customed to humans. Unfortunately, 

 sheepdogs are of no use, for the sheep 

 scatter, rather than bunch, when they are 



32 Natural History 3/94 



