Geoscience Reference
In-Depth Information
cause by various natural enemies. As indicated earlier, mortalities may vary in
density spatially between plots, but these need not be the same as those that
vary temporally in the same plots. Application of the techniques described
early is appropriate only where the data represent time series. However, there
is no reason why the time series could not be obtained from populations that
are experimentally manipulated. A well-known example of such a study is the
kilometer-scale exclosure study of the factors causing fluctuations in snowshoe
hares in Yukon Territories (Krebs et al. 1995).
However, inferring the dynamic consequences of experimental results may
not always be straightforward. A common problem is that experiments are
almost inevitably done on a small scale and it may be inaccurate to extrapolate
to the larger scale of natural populations. For example, Gould et al. (1990)
manipulated gypsy moth density on 1-ha plots and showed that tachinid par-
asitoids decimated populations in a strongly density-dependent way. However,
studies from natural populations revealed far lower levels of parasitism caused
by these species and scant evidence for density dependence. Why the differ-
ence? Evidently the parasitoids aggregated to the 1-ha experimental popula-
tions from surrounding areas of low gypsy moth density. This density-depen-
dent aggregation response would be nullified in natural populations, where
densities rise simultaneously over much larger areas. The spatial scale in this
study was thus crucial. Manipulation on a 1-ha spatial scale, although large rel-
ative to most ecological experiments, was not large enough to mimic the
dynamics of natural populations. In other words, experimental manipulations
may introduce a variety of artifacts that may be difficult to detect. Neverthe-
less, experimental manipulation almost always yields more information than
studies of unmanipulated populations, particularly because unmanipulated
control populations would usually be part of the experimental design.
Conclusions
j
This review has identified many limitations in the methods that ecologists
have used to study the dynamics of populations. For all of these reasons the
dynamics of even well-studied systems that have occupied the talents of the
best minds in ecology remain unresolved and hotly debated. The best advice
we can give to those who are embarking on such studies is to maintain a
healthy skepticism of all the techniques and to take a multipronged approach.
Wherever possible, studies of experimentally manipulated populations should
be coupled with unmanipulated ones.
