in numbers of the predator. The cycle would thus 

 continue indefinitely. According to the differential 

 equations, the predator will never be able completely 

 to destroy the prey, nor will the predator species 

 ever completely disappear by reason of starvation. 



There is considerable controversy concerning 

 this theory (Andrewartha and Birch 1954). Cause 

 (1934, 1935) conducted a test in a classic series of 

 experiments with protozoan cultures. An experi- 

 mental food chain was established : boiled oat- 

 meal ^bacteria ^Paraniccium caudatuin *Di- 



diniiim nasntiini. When five Paramecium were in- 

 troduced one day and three Didiniiim two days later, 

 the population of Paramecium was exterminated by 

 the predator. The predator, left without food, disap- 

 peared soon after. In another experiment, cover sedi- 

 ment was introduced into the microcosm, in which 

 the Paramecium could hide, thus to escape the attacks 

 of Didinium. The same number of each species was 

 introduced at the same time. The number of preda- 

 tors increased, and they devoured many of the prey. 

 However, the remaining prey escaped into the cover, 

 and the predators died of starvation. When this hap- 

 pened, the prey, now unchecked, increased in an 

 unlimited manner. When, on the other hand, a micro- 

 cosm was prepared in which there was no refuge, and 

 one Paramecium and one Didinium were introduced 

 on every third day, a series of oscillations resulted. 

 It is suggested that continual cycling of prey and 

 predator populations could be maintained only with 

 immigration of individuals from the outside. In other 

 experiments, interrelated cycles of Paramecium and 

 the yeast on which it fed were established (Cause 

 1935). 



In experimental greenhouse plots of strawberries, 

 populations of an herbivorous mite, Tarsonemus 

 pallidus, and its predator, another mite, Typhlodro- 

 mus. fluctuated regularly in relation to each other. 

 At low populations, the prey species was relatively 

 secure in the cover ofifefed by hairs, spines, and leaf 

 crevices, thus avoiding annihilation. The predator 

 species survived because it utilized honeydew and 

 other nourishment as substitute food until the prey 

 species again increased in numbers (Huffaker and 

 Kennett 1956). 



Reciprocal fluctuations in the density of the azuki 

 bean weevil and its larval parasitoid, a braconid 

 wasp, were sustained experimentally under constant 

 conditions for 112 successive generations. Appar- 

 ently the prey was able to survive in the low of the 

 cycle because of the difficulty the parasitoid experi- 

 enced in finding the surviving individuals ; the para- 

 sitoid, however, never became extinct (Utida 1957). 



These examples indicate that oscillations in the 

 populations of predators and prey can be sustained 

 for relatively long periods of time if such factors as 

 cover, buffer food, or immigration are introduced into 



the experiment. This background of experimental 

 studies is useful in the analysis of possible causes of 

 the more or less regular oscillations in animal num- 

 bers that are observed under natural conditions. 



In an area near Point Barrow in northern Alaska, 

 Siberian lemmings were scarce from 1949 to 1951, 

 increased in 1952, and were near or at a peak in 1953. 

 Associated with this cyclic rise in the lemming popu- 

 lation was a marked increase in the number of preda- 

 tors. There was no breeding in 1951 of pomarine 

 jaegers, snowy owls, and short-eared owls : very few 

 were even seen. In 1953, however, breeding pairs 

 were recorded in densities respectively of about 18, 

 0.3, and 3^ per 250 hectares (per square mile). 

 Least weasels and Arctic and red foxes increased 

 from scarce or no record to common. Because of this 

 heavy predation, the lemming population was re- 

 duced by mid-July of 1953 to Yio or less of what it 

 had been when the snow cover melted in early June 

 (Pitelka ct al. 1955). 



Cyclic changes between 1929 and 1940 in the 

 collared lemming at Churchill, Manitoba, were ac- 

 companied by marked fluctuations in breeding popula- 

 tions of snowy and short-eared owls and of the rough- 

 legged hawk (Shelford 1943). This rapid build-up of 

 predator populations must be attributed to their abil- 

 ity to shift from one region to another according to 

 availability of local prey. The lemming becomes more 

 vulnerable to predation when large populations con- 

 sume the vegetative cover. Influxes of predators suf- 

 ficient to exert a controlling role in outbreaks of 

 mice (Banfield 1947), ruffed grouse, and snow- 

 shoe rabbit (Morse 1939), and bobwhite (Jackson 

 1947) have been reported for regions as far south as 

 Toronto. Minnesota, and northwest Texas. 



The collared lemming breeds in the winter, at 

 least to some extent, as well as during the summer 

 (Sutton and Hamilton 1932), as does, apparently, the 

 Siberian lemming (Pitelka et al. 1955). When lem- 

 mings are exposed to heavy predation during the 

 summer, it is likely that the main population growth 

 comes between August and the following June, dur- 

 ing which time they are protected by a snow cover. 

 When snow is inadequate, heavy predation doubtless 

 continues throughout the year. With the lack of snow 

 insulation, considerable mortality may also result 

 from effects of low temperature (Shelford 1943). 

 There is no evidence, however, that snow cover oc- 

 curs in cyclic harmony with the lemming populations, 

 necessary were snow the critical factor producing the 

 lemming cycle. 



A number of general theories of possible intrinsic 

 cycle causes (Dymond 1947, Crange 1949, Lack 

 1954a) have been found inadequate. Particular 

 cycles have been explained for several species, such 

 as for Daphnia under experimental conditions (Slo- 

 bodkin 1954), sockeye salmon in the Eraser River 



240 Ecological processes and dynamics 



