Population cycles in small rodents
C J Krebs,M S Gaines,B L Keller,J H Myers,R H Tamarin
DOI: https://doi.org/10.1126/science.179.4068.35
IF: 56.9
1973-01-05
Science
Abstract:We conclude that population fluctuations in Microtus in southern Indiana are produced by a syndrome of changes in birth and death rates similar to that found in other species of voles and lemmings. The mechanisms which cause the changes in birth and death rates are demolished by fencing the population so that no dispersal can occur. Dispersal thus seems critical for population regulation in Microtus. Because most dispersal occurs during the increase phase of the population cycle and there is little dispersal during the decline phase, dispersal is not directly related to population density. Hence the quality of dispersing animals must be important, and we have found one case of increased dispersal tendency by one genotype. The failure of population regulation of Microtus in enclosed areas requires an explanation by any hypothesis attempting to explain population cycles in small rodents. It might be suggested that the fence changed the predation pressure on the enclosed populations. However, the fence was only 2 feet (0.6 meter) high and did not stop the entrance of foxes, weasels, shrews, or avian predators. A striking feature was that the habitat in the enclosures quickly recovered from complete devastation by the start of the spring growing season. Obviously the habitat and food quality were sufficient to support Microtus populations of abnormally high densities, and recovery of the habitat was sufficiently quick that the introduction of new animals to these enclosed areas resulted in another population explosion. Finally, hypotheses of population regulation by social stress must account for the finding that Microtus can exist at densities several times greater than normal without "stress" taking an obvious toll. We hypothesize that the prevention of dispersal changes the quality of the populations in the enclosures in comparison to those outside the fence. Voles forced to remain in an overcrowded fenced population do not suffer high mortality rates and continue to reproduce at abnormally high densities until starvation overtakes them. The initial behavioral interactions associated with crowding do not seem sufficient to cause voles to die in situ. What happens to animals during the population decline? Our studies have not answered this question. The animals did not appear to disperse, but it is possible that the method we used to measure dispersal (movement into a vacant habitat) missed a large segment of dispersing voles which did not remain in the vacant area but kept on moving. Perhaps the dispersal during the increase phase of the population cycle is a colonization type of dispersal, and the animals taking part in it are likely to stay in a new habitat, while during the population decline dispersal is a pathological response to high density, and the animals are not attracted to settling even in a vacant habitat. The alternative to this suggestion is that animals are dying in situ during the decline because of physiological or genetically determined behavioral stress. Thus the fencing of a population prevents the change in rates of survival and reproduction, from high rates in the increase phase to low rates in the decline phase, and the fenced populations resemble "mouse plagues." A possible explanation is that the differential dispersal of animals during the phase of increase causes the quality of the voles remaining at peak densities in wild populations to be different from the quality of voles at much higher densities in enclosures. Increased sensitivity to density in Microtus could cause the decline of wild populations at densities lower than those reached by fenced populations in which selection through dispersal has been prevented. Fencing might also alter the social interactions among Microtus in other ways that are not understood. The analysis of colonizing species by MacArthur and Wilson (27) can be applied to our studies of dispersal in populations of Microtus. Groups of organisms with good dispersal and colonizing ability are called r strategists because they have high reproductive potential and are able to exploit a new environment rapidly. Dispersing voles seem to be r strategists. Young females in breeding condition were over-represented in dispersing female Microtus (17). The Tf(C)/Tf(E) females, which were more common among dispersers during the phase of population increase (Fig. 6), also have a slight reproductive advantage over the other Tf genotypes (19). Thus in Microtus populations the animals with the highest reproductive potential, the r strategists, are dispersing. The segment of the population which remains behind after the selection-via-dispersal are those individuals which are less influenced by increasing population densities. These are the individuals which maximize use of the habitat, the K strategists in MacArthur and Wilson's terminology, or voles selected for spacing behavior. Thus we can describe population cycles in Microtus in the same theoretical framework as colonizing species on islands. Our work on Microtus is consistent with the hypothesis of genetic and behavioral effects proposed by Chitty (6) (Fig. 7) in that it shows both behavioral differences in males during the phases of population fluctuation and periods of strong genetic selection. The greatest gaps in our knowledge are in the area of genetic-behavioral interactions which are most difficult to measure. We have no information on the heritability of aggressive behavior in voles. The pathways by which behavioral events are translated into physiological changes which affect reproduction and growth have been carefully analyzed by Christian and his associates (28) for rodents in laboratory situations, but the application of these findings to the complex field events described above remains to be done. Several experiments are suggested by our work. First, other populations of other rodent species should increase to abnormal densities if enclosed in a large fenced area (29). We need to find situations in which this prediction is not fulfilled. Island populations may be an important source of material for such an experiment (30). Second, if one-way exit doors were provided from a fenced area, normal population regulation through dispersal should occur. This experiment would provide another method by which dispersers could be identified. Third, if dispersal were prevented after a population reached peak densities, a normal decline phase should occur. This prediction is based on the assumption that dispersal during the increase phase is sufficient to ensure the decline phase 1 or 2 years later. All these experiments are concerned with the dispersal factor, and our work on Microtus can be summarized by the admonition: study dispersal.