but Peterson expects that the occasional 

 oscillations we see in some larger ungu- 

 lates may eventually turn out to be part of 

 such long-term cycles. 



My colleagues and I have followed the 

 Soay sheep on Hirta through three cycles, 

 but we weren't the first to observe the phe- 

 nomenon. Previous studies of the island 

 sheep by Morton Boyd, of the British Na- 

 ture Conservancy Council, and by zoolo- 

 gists Peter Jewell and Peter Grubb, of the 

 University of London, show that similar 

 die-offs occurred every third or fourth 

 year during the 1960s. Regular oscilla- 

 tions have not been reported in any wild 

 sheep populations in North America or 

 Asia, nor do other ungulates on Scottish 

 islands show similar peaks and crashes. 

 The number of red deer on the island of 



Although sheep graze all over Hirta, in winter they spend most of 

 their time on low ground, especially in the abandoned fields of 

 Village Bay, left. The dry-stone shelters, or cleits, that dot the 

 lower slopes were buih to dry and store seahirds. A yearling 

 ram, below, is already sexually mature. During rut, rams wander 

 widely in search of ewes in estrus. 



Laurie Campbell 



Rum, for instance, where zoologist Fiona 

 Guinness and I have studied them for 

 more than twenty years, remains remark- 

 ably stable, declining slightly after hard 

 winters and increasing after good ones. 

 Then why should Soay sheep behave like 

 voles or lemmings? 



Over the last three years, Steve Albon, 

 Josephine Pemberton, and I, together with 

 other biologists from Cambridge and Ed- 

 inburgh, have begun to glimpse an answer. 

 After a population crash, sheep numbers 

 increase rapidly. Unlike North American 

 wild sheep, Soay ewes first conceive when 

 they are less than a year old, birthing their 

 first lambs in April, soon after their first 

 birthday. Up to 20 percent of the pregnant 

 females bear twins. Since Hirta has no car- 

 nivores, more than 80 percent of the spring 

 newborns usually survive to the beginning 

 of winter, and animals obviously cannot 

 disperse from the island. When the popu- 

 lation is small, winter mortality of lambs is 

 less than 10 percent, so that by the end of 

 the first year following a crash, total num- 

 bers usually have risen by 50 percent or 

 more. Fecundity and lamb survival remain 

 high through the following year, when the 

 sheep increase by 40 to 50 percent again. 

 In the summer of the third season, they in- 

 crease by another 40 percent. At this stage. 



there are more than three times as many 

 sheep on Hirta as there were immediately 

 after the crash, but they still begin the win- 

 ter in good health. 



In late September or October, however, 

 grass growth ceases at this latitude, and the 

 sheep must winter on the remnants of 

 summer's vegetation. When sheep num- 

 bers are high, little food remains by Janu- 

 ary or early February, and the animals 

 begin to lose weight rapidly. Rams, which 

 bum much of their fat in the November 

 rut, are the first to die, followed by lambs, 

 which suffer more heat loss than ewes be- 

 cause of their smaller size. During Febru- 

 ary and March (the last two months of ges- 

 tation), the energy costs of supporting 

 growing fetuses increase sharply, and 

 pregnant ewes (especially those carrying 

 twins) are the final casualties. 



At least two other factors may con- 

 tribute to the crash. First, the sheep suffer 

 from infestations of nematode worms in 

 their gastrointestinal tracts. As the flock 

 increases, more worms are passed out in 

 their dung, so the density of worms in the 

 pasture also rises. Second, as Dawn 

 Bazely and Mark Vicari of Canada's Uni- 

 versity of York have shown, heavy grazing 

 may reduce the production of summer 

 grasses immediately before a crash, fur- 



31 



