POPULATION CYCLES IN SMALL MAMMAL8
271
For example, amume that the increase phase may occur only when
disease x is not present in the population. The absence of disease x
would thus be t i necessary condition for population increase. However,
if the disease is absent but the females are not in reproductive condition, increase will not ensue and hence the absence of disease x is
not a sufficient condition for population increase. What we are actually
looking for is the set of necessary conditions which together are
sufficient to cause the population to increase. We thus attempt to sort
out necessary factors for further analysis.
We begin by attempting to specify simple conditions but if we are
unable to explain population cycles we will try to specify more complex
conditions. At present no one can specify the necessary conditions for
population growth in any species of vole or lemming. Why do we not
abandon this approach? Can we not be more precise by recognizing
the multiplicity of causes of population cycles? Since there is disagreement about the causes of population cycles, perhaps we should abandon
the search for a universal explanation and be content with more restricted hypotheses for single species in particular communities.
The multiple-factor hypothesis is particularly dangerous as a
methodological argument. If taken at its face value as a vague armchair
theory, the multiple-factor hypothesis is certainly true. The factors
which affect a lemming population in Alaska are certainly different
from those affecting a vole population in Kansas. But if we adopt this
hypothesis as our research strategy, we lose one of the most important
checks on scientific speculation-the testability of hypotheses. Suppose
that we adopt the hypothesis that the absence of disease x is a sufficient
condition for population growth for Microtus ochrogaster in Kansas.
We cannot test this hypothesis on Microtus ochrogaster in Nebraska
because this is a different situation. Nor can we test the hypothesis on
M. pennsylvanicus in New York. If carried to an extreme, we cannot
even test the hypothesis on the next cycle of M. ochrogaster in Kansas
because multiple factors may intervene over time as well as space.
We believe that a universal explanation must be sought for population fluctuations in voles and lemmings until we have evidence that two
or more distinct explanations are required. With this approach we can
test alternative hypotheses on any species in any location.
It must be pointed out that we are not interested simply in a list of
factors which influence rodent populations. Abundant food and cover
will certainly harbor populations capable of reaching higher densities,
and predation can sometimes remove a large portion of the population.
Food, weather, and predation are factors which are acting on all
populations. We accept this generalization, but we wish to ask if these
factors are necessary or sufficient causes of population cycles. Often a
271
For example, amume that the increase phase may occur only when
disease x is not present in the population. The absence of disease x
would thus be t i necessary condition for population increase. However,
if the disease is absent but the females are not in reproductive condition, increase will not ensue and hence the absence of disease x is
not a sufficient condition for population increase. What we are actually
looking for is the set of necessary conditions which together are
sufficient to cause the population to increase. We thus attempt to sort
out necessary factors for further analysis.
We begin by attempting to specify simple conditions but if we are
unable to explain population cycles we will try to specify more complex
conditions. At present no one can specify the necessary conditions for
population growth in any species of vole or lemming. Why do we not
abandon this approach? Can we not be more precise by recognizing
the multiplicity of causes of population cycles? Since there is disagreement about the causes of population cycles, perhaps we should abandon
the search for a universal explanation and be content with more restricted hypotheses for single species in particular communities.
The multiple-factor hypothesis is particularly dangerous as a
methodological argument. If taken at its face value as a vague armchair
theory, the multiple-factor hypothesis is certainly true. The factors
which affect a lemming population in Alaska are certainly different
from those affecting a vole population in Kansas. But if we adopt this
hypothesis as our research strategy, we lose one of the most important
checks on scientific speculation-the testability of hypotheses. Suppose
that we adopt the hypothesis that the absence of disease x is a sufficient
condition for population growth for Microtus ochrogaster in Kansas.
We cannot test this hypothesis on Microtus ochrogaster in Nebraska
because this is a different situation. Nor can we test the hypothesis on
M. pennsylvanicus in New York. If carried to an extreme, we cannot
even test the hypothesis on the next cycle of M. ochrogaster in Kansas
because multiple factors may intervene over time as well as space.
We believe that a universal explanation must be sought for population fluctuations in voles and lemmings until we have evidence that two
or more distinct explanations are required. With this approach we can
test alternative hypotheses on any species in any location.
It must be pointed out that we are not interested simply in a list of
factors which influence rodent populations. Abundant food and cover
will certainly harbor populations capable of reaching higher densities,
and predation can sometimes remove a large portion of the population.
Food, weather, and predation are factors which are acting on all
populations. We accept this generalization, but we wish to ask if these
factors are necessary or sufficient causes of population cycles. Often a