The Predation Hypothesis

The Scandinavian school of ecologists has developed the predation hypothesis, which postulates that, in a stable community in temperate climates, many generalist predators combined with some vole specialists can keep vole populations from fluctuating widely. Erlinge et al. (1983, 1984) concluded that voles in southern Sweden start their breeding season surrounded by persistent predators poised to snap up the young as soon as they emerge from their nests. Such predation delays and reduces the recovery of the rodent population from the winter nonbreeding period. Predation is also heavy in autumn, because of the rapid functional response of the generalist predators to the high production of rodents through the breeding season. The net result is that rodents can never escape the attention of predators. The generalists are always there, ready and waiting. In addition, the complementary habitat preferences of different predators reduce the abilities of voles to escape predation altogether (Korpimaki et al. 1996).

A single specialist predatory species cannot exert such a comprehensive effect on multiple prey populations. So, the other side of the predation hypothesis states that predation by the northern rodent specialists, the stoats and least weasels, is actually a necessary prerequisite for vole populations to fluctuate. In northern habitats, winter snow cover is prolonged and least weasels and the small northern races of stoats are the main or only predator on voles for much of the year (Wilson et al. 1999; Norrdahl & Korpimaki 2000a). At the very least, the predation hypothesis asserts that weasel predation in the far north, particularly on Microtus species, synchronizes the cycles of all small rodent species across very large geographic regions, including islands (Henttonen et al. 1987; Heikkila et al. 1994) (Figure 7.2).

Extensive evidence from northern Europe and Greenland confirms the expected correlation between heavy predation by weasels and population crashes of voles and lemmings. Extended predation can then keep rodent populations to low levels for several seasons, allowing the vegetation to recover and, thereby, ensuring future increases in rodent numbers (Korpimaki et al. 1991; Norrdahl 1995; Norrdahl & Korpimaki 1995b; Sittler 1995; Klemola et al. 1997; Korpimaki

Figure 7.2 The linked population fluctuations of weasels and small northern rodents are well illustrated by two examples involving different species and different field techniques. (A) Microtus and least weasels sampled by snow tracking. (Redrawn from Norrdahl 1995.) (B) Collared lemmings and stoats sampled by spring censuses of lemming nests. (Redrawn from Gilg et al. 2003.) In both examples, solid circles and lines represent the weasel population and open squares and dashed lines represent the prey population.

Figure 7.2 The linked population fluctuations of weasels and small northern rodents are well illustrated by two examples involving different species and different field techniques. (A) Microtus and least weasels sampled by snow tracking. (Redrawn from Norrdahl 1995.) (B) Collared lemmings and stoats sampled by spring censuses of lemming nests. (Redrawn from Gilg et al. 2003.) In both examples, solid circles and lines represent the weasel population and open squares and dashed lines represent the prey population.

& Norrdahl 1998; Gilg et al. 2003). It may even be able to explain the so-called "Chitty effect" (the observation that voles tend to be larger in body size during the peak of their numbers than in the decline and low phases). This would be easily explained, say Sundell and Norrdahl (2002), if the only voles to survive the population crashes were the smallest ones able to take refuge in burrows too small for the smallest weasels to enter.

Predation has always been among the possible explanations for vole cycles, right back to the early work of Charles Elton (1942), but it was not widely accepted at first. Reservations about the role of predation seemed reasonable at the time, partly because the influence of Paul Errington (p. 142) lingered well into the 1970s, partly because the proponents of the idea assumed that all vole populations are cyclic when many obviously are not, and partly because the crucially different effects of specialist and generalist predators were not distinguished by anybody. Great confusion was created in the literature by field experiments testing different questions with contradictory results (Norrdahl 1995). The confusion could have been expected, however, since the effects of predation, parasites/diseases, and malnutrition on small mammal populations are interactive (Korpimaki et al. 2004).

In recent years, the Scandinavian school has developed the predation hypothesis into a set of clearly defined propositions. They have done a lot of field experiments and population modeling that support the hypothesis as an explanation of how rodent numbers change under given conditions (Oksanen et al. 2000; Hanski et al. 2001; Korpimaki et al. 2002; Ekerholm et al. 2004). In its present form the hypothesis even explains some previously puzzling exceptions. For example, Hanski et al. (2001) developed a model that predicts variation with time in the behavior of rodent populations. Field research has confirmed that rodent numbers have fluctuated less regularly in parts of northern Fennoscandia since the early 1980s (Hanski & Henttonen 1996; Hornfeldt 2004).

The predation hypothesis also predicts that a change in the population dynamics of voles should correlate with a change in the community of predators. To test this prediction, Korpimaki and his colleagues (2002) organized a large-scale, comprehensive experiment comparing the densities of field voles on four pairs of large (2 to 3 km2) study areas. Each pair included predator-removal and nontreatment areas, and was followed for 3 years. The densities of the vole populations free of predation grew twice as fast as the nontreatment populations (subject to predation) in the increase phase, and doubled their autumn densities in the peak phase (Figure 7.3).

A model based on these results, that included seasonal effects and limited density dependence for the prey, predicted the type of population dynamics that is actually observed in the weasel-vole community in northern Fennoscandia. Korpimaki and his team could induce a change from multiyear cycles to seasonal fluctuations in the simulated vole populations in their model simply by reducing the proportion of specialist predators (especially least weasels) relative to

Figure 7.3 Effect of removing weasels on the dynamics of northern voles. Solid squares and lines represent vole populations where predators were removed; open squares and dashed line represent vole populations subject to predation. In this experiment, vole populations lacking predators reached significantly higher densities than those subject to predation. (Redrawn from Korpimaki et al. 2002.)

Apr Jun Aug Oct Apr Jun Aug Oct Apr Jun Aug Oct

1997 1998 1999

Figure 7.3 Effect of removing weasels on the dynamics of northern voles. Solid squares and lines represent vole populations where predators were removed; open squares and dashed line represent vole populations subject to predation. In this experiment, vole populations lacking predators reached significantly higher densities than those subject to predation. (Redrawn from Korpimaki et al. 2002.)

generalist predators. And in turn, the relative importance of predation by weasels was closely correlated with the duration of winter snow cover (see Figure 1.5), which itself is only loosely correlated with latitude (Strann et al. 2002).

Was this article helpful?

0 0

Post a comment