ANALYSIS O F PROCESSES IN CONTROL O F INSECTS
46
do. The study of the functional and numerical responses of mammalian
predators of the pine sawfly by Holling (1959a) gives a most instructive
account of the responses of these predators. The investigation was
carried out in a sand-plain area of S.W. Ontario. The plantations he
used were 15-20 years old, with a closed canopy reducing ground
vegetation to a trace. The sawflies spend the egg and feeding stages on
the foliage, but in June the larvae drop from the trees and penetrate
the pine needles on the forest floor, where they spin cocoons at the
surface of the soil. Here they remain until the latter part of September,
when most emerge as adults. A certain proportion, however, overwinter
in cocoons, and emerge the following autumn. Only the cocooned stage
is susceptible to attack by the small mammals. These are two species
of shrew (a Sorex and a Blarina) and a deer-mouse (a Peromyscus).
Each species makes a characteristically marked opening in the cocoon
which provides a relatively permanent record of the act of predation
and of the taxonomic identity of the predator.
The effects of cocoon density on predation by Peromyscus were
studied in two plots over three years, and the effects on that of Sorex
and Blarina in one plot in one year. Owing to the treatment of parts
of the area with a spray containing virus pathogenic to the sawfly
larvae, a range of different levels of prey densities was available. The
densities of Peromyscus and Blarina were estimated by trapping ; that
of Sorex had to be calculated from the results of laboratory experiments
on the rate of predation per individual.
1. Functional and Numerical Responses of Sawfly Predators
Holling showed that the number of cocoons opened per mammal per
day increased as the prey density rose, up to a certain value of prey
density above which the rising curve of predation rapidly turned over
and levelled off (Fig. 1 8 ~ ) .
The slope and the maximum was different
for each species of predator, but the curves were of the same general
form. Curves for the individual Peromyscus studied in the laboratory
were also of this form.
Turning to the numerical responses of the predators, Rolling’s
observation period of four summers was too short for more than tentative conclusions. His graphs (Fig. 1 8 ~ )
suggest that Peromyscus and
Sorex, but not Blarina, were more abundant at higher prey densities,
up to a certain level of prey abundance that was different for the two.
2. Combined Raponses of Sawfly Predators
When Holling compounded the estimated numerical and functional
responses (Fig. 18c), the calculated combined response of each predator
species was strongly density-dependent, i.e. percent predation greatly
46
do. The study of the functional and numerical responses of mammalian
predators of the pine sawfly by Holling (1959a) gives a most instructive
account of the responses of these predators. The investigation was
carried out in a sand-plain area of S.W. Ontario. The plantations he
used were 15-20 years old, with a closed canopy reducing ground
vegetation to a trace. The sawflies spend the egg and feeding stages on
the foliage, but in June the larvae drop from the trees and penetrate
the pine needles on the forest floor, where they spin cocoons at the
surface of the soil. Here they remain until the latter part of September,
when most emerge as adults. A certain proportion, however, overwinter
in cocoons, and emerge the following autumn. Only the cocooned stage
is susceptible to attack by the small mammals. These are two species
of shrew (a Sorex and a Blarina) and a deer-mouse (a Peromyscus).
Each species makes a characteristically marked opening in the cocoon
which provides a relatively permanent record of the act of predation
and of the taxonomic identity of the predator.
The effects of cocoon density on predation by Peromyscus were
studied in two plots over three years, and the effects on that of Sorex
and Blarina in one plot in one year. Owing to the treatment of parts
of the area with a spray containing virus pathogenic to the sawfly
larvae, a range of different levels of prey densities was available. The
densities of Peromyscus and Blarina were estimated by trapping ; that
of Sorex had to be calculated from the results of laboratory experiments
on the rate of predation per individual.
1. Functional and Numerical Responses of Sawfly Predators
Holling showed that the number of cocoons opened per mammal per
day increased as the prey density rose, up to a certain value of prey
density above which the rising curve of predation rapidly turned over
and levelled off (Fig. 1 8 ~ ) .
The slope and the maximum was different
for each species of predator, but the curves were of the same general
form. Curves for the individual Peromyscus studied in the laboratory
were also of this form.
Turning to the numerical responses of the predators, Rolling’s
observation period of four summers was too short for more than tentative conclusions. His graphs (Fig. 1 8 ~ )
suggest that Peromyscus and
Sorex, but not Blarina, were more abundant at higher prey densities,
up to a certain level of prey abundance that was different for the two.
2. Combined Raponses of Sawfly Predators
When Holling compounded the estimated numerical and functional
responses (Fig. 18c), the calculated combined response of each predator
species was strongly density-dependent, i.e. percent predation greatly