ANALYSIS O F PROCESSES IN CONTROL O F INSECTS
25
But generally, when parasite-host oscillations and elliptical or
polygonal patterns are involved, the method should be regarded as
only a first step in analysis. Some later steps are illustrated by Morris
The main requirement of the method is a series of estimates over an
adequate run of generations; also, as Morris (Zoc. cit.) pointed out, it is
desirable to have figures for a wide range of densities to work upon, so
that density-effects may show up clearly. Where adequate data are
available, the method can be expected to provide a simple and useful
index of the degree of density-dependence, or regulation. The slope of
the line shows whether the density-dependent relationship comes into
play strongly or only feebly as density rises and falls ; and the scatter
of the points about the fitted line may show the degree of influence of
minor disturbing or modifying factors ; but if consistent densitydependence continues to operate during the course of population
changes imposed by density-independent influences, no such scatter
arises (Fig. 10). Marked changes representing breaks in the densityrelationships tend to give points widely separated from the densitydependent slopes (Fig. 11). A marked change in the “target” equilibrium level gives rise to a new set of points to which a separate line
must be fitted (Fig. 11).
4. By Correlation between Density and Weather Indices
This method can be illustrated by reference to a much discussed
example. Entomologists at the Waite Institute, Adelaide, S. Australia,
collected twenty rose blooms from the garden of the Institute daily
over a number of years and counted the numbers of the apple blossom
thrips, Thrips imaginis Bagnall, in each sample (Davidson and
Andrewartha, 1948a,b). The insect bred in various flowers, but not in
these roses, which served merely as traps. The numbers of thrips in the
samples depended on the weather at the time and on the abundance of
roses and of other flowers, as well as on the abundance of thrips. After
declining from a minor autumn peak, their abundance greatly increased
during the main flowering season in spring (September onwaras),
reaching a peak in November or December, then rapidly fell as the hot
dry season set in.
The relationships between the density of the population in the roses
and the weather factors were set out by Davidson and Andrewartha
(194813). The degree of association between the density of thrips and
the weather during the preceding months was measured by partial
regression. Four components of the physical environment were selected
and 78% of the variance of the population was attributed to their
influence. The maximum density attained in the spring was considered
(ZOC. cit.).
25
But generally, when parasite-host oscillations and elliptical or
polygonal patterns are involved, the method should be regarded as
only a first step in analysis. Some later steps are illustrated by Morris
The main requirement of the method is a series of estimates over an
adequate run of generations; also, as Morris (Zoc. cit.) pointed out, it is
desirable to have figures for a wide range of densities to work upon, so
that density-effects may show up clearly. Where adequate data are
available, the method can be expected to provide a simple and useful
index of the degree of density-dependence, or regulation. The slope of
the line shows whether the density-dependent relationship comes into
play strongly or only feebly as density rises and falls ; and the scatter
of the points about the fitted line may show the degree of influence of
minor disturbing or modifying factors ; but if consistent densitydependence continues to operate during the course of population
changes imposed by density-independent influences, no such scatter
arises (Fig. 10). Marked changes representing breaks in the densityrelationships tend to give points widely separated from the densitydependent slopes (Fig. 11). A marked change in the “target” equilibrium level gives rise to a new set of points to which a separate line
must be fitted (Fig. 11).
4. By Correlation between Density and Weather Indices
This method can be illustrated by reference to a much discussed
example. Entomologists at the Waite Institute, Adelaide, S. Australia,
collected twenty rose blooms from the garden of the Institute daily
over a number of years and counted the numbers of the apple blossom
thrips, Thrips imaginis Bagnall, in each sample (Davidson and
Andrewartha, 1948a,b). The insect bred in various flowers, but not in
these roses, which served merely as traps. The numbers of thrips in the
samples depended on the weather at the time and on the abundance of
roses and of other flowers, as well as on the abundance of thrips. After
declining from a minor autumn peak, their abundance greatly increased
during the main flowering season in spring (September onwaras),
reaching a peak in November or December, then rapidly fell as the hot
dry season set in.
The relationships between the density of the population in the roses
and the weather factors were set out by Davidson and Andrewartha
(194813). The degree of association between the density of thrips and
the weather during the preceding months was measured by partial
regression. Four components of the physical environment were selected
and 78% of the variance of the population was attributed to their
influence. The maximum density attained in the spring was considered
(ZOC. cit.).