140
D AV I D L . G A R S H E L I S
lines. Habitats are managed based on supposed importance in terms of fitness
for members of the target population. High-quality habitats, by definition,
produce animals with high reproduction and survival, and hence a population
with a high growth rate (and high yield, for harvested species). However, the
assumption that we can infer habitat quality or suitability from studies of habitat selection—that selection, even if accurately measured, is directly related to
each habitat’s potential contribution to individual fitness and hence the population’s growth rate—represents what I consider the second fatal flaw in the
process of habitat evaluation.
One of the best examples of a purported relationship between habitat selection and fitness was provided by a study of leaf-galling aphids (Pemphigus
betae) that parasitize narrowleaf cottonwood trees (Populus angustifolia). These
aphids overwinter in tree bark, then over a short period in the spring migrate
up the tree and become entombed in expanding leaf tissue, where they reproduce. Whitham (1978, 1980) found that aphids apparently selected large
leaves over small leaves (they colonized 100 percent of leaves over 15 cm but
only 3 percent of leaves 5 cm or less), even though large leaves were less available (less than 2 percent of leaves were over 15 cm; more than 30 percent were
5 cm or less); moreover, those that colonized large leaves had better survival
and reproduction than those colonizing small leaves. From this evidence it
appeared that aphids had the ability to select leaves that offered the highest fitness; if not, the size distribution of colonized leaves would have matched the
size distribution of leaves available on the tree (i.e., use would have equaled
availability). However, Rhomberg (1984) observed that large leaves tended to
be those near the tips of twigs (where they receive more light), and these
opened a few days later than basal leaves. When the aphids begin their migration, basal leaves are already too mature to enable them to form a gall, so they
must colonize the more distal leaves; thus, the basal leaves are in essence nonhabitat. The migration appears to be timed so that the aphids inevitably colonize leaves that are destined to be favorable to their fitness, but individually,
they do not select these among less favorable habitats. This is not to say that
other animals do not select habitats favorable to their fitness; certainly they do.
But if the comprehensive and seemingly compelling data on a simple system
such as aphids on cottonwood trees can be misinterpreted, then clearly there is
much room for misinterpreting relationships between habitat selection and fitness in more complicated systems.
The presumed link between selection and fitness is rarely tested, partly
because such tests are difficult, but also because this relationship is viewed
D AV I D L . G A R S H E L I S
lines. Habitats are managed based on supposed importance in terms of fitness
for members of the target population. High-quality habitats, by definition,
produce animals with high reproduction and survival, and hence a population
with a high growth rate (and high yield, for harvested species). However, the
assumption that we can infer habitat quality or suitability from studies of habitat selection—that selection, even if accurately measured, is directly related to
each habitat’s potential contribution to individual fitness and hence the population’s growth rate—represents what I consider the second fatal flaw in the
process of habitat evaluation.
One of the best examples of a purported relationship between habitat selection and fitness was provided by a study of leaf-galling aphids (Pemphigus
betae) that parasitize narrowleaf cottonwood trees (Populus angustifolia). These
aphids overwinter in tree bark, then over a short period in the spring migrate
up the tree and become entombed in expanding leaf tissue, where they reproduce. Whitham (1978, 1980) found that aphids apparently selected large
leaves over small leaves (they colonized 100 percent of leaves over 15 cm but
only 3 percent of leaves 5 cm or less), even though large leaves were less available (less than 2 percent of leaves were over 15 cm; more than 30 percent were
5 cm or less); moreover, those that colonized large leaves had better survival
and reproduction than those colonizing small leaves. From this evidence it
appeared that aphids had the ability to select leaves that offered the highest fitness; if not, the size distribution of colonized leaves would have matched the
size distribution of leaves available on the tree (i.e., use would have equaled
availability). However, Rhomberg (1984) observed that large leaves tended to
be those near the tips of twigs (where they receive more light), and these
opened a few days later than basal leaves. When the aphids begin their migration, basal leaves are already too mature to enable them to form a gall, so they
must colonize the more distal leaves; thus, the basal leaves are in essence nonhabitat. The migration appears to be timed so that the aphids inevitably colonize leaves that are destined to be favorable to their fitness, but individually,
they do not select these among less favorable habitats. This is not to say that
other animals do not select habitats favorable to their fitness; certainly they do.
But if the comprehensive and seemingly compelling data on a simple system
such as aphids on cottonwood trees can be misinterpreted, then clearly there is
much room for misinterpreting relationships between habitat selection and fitness in more complicated systems.
The presumed link between selection and fitness is rarely tested, partly
because such tests are difficult, but also because this relationship is viewed