ANALYSIS O F PROCESSES IN CONTROL OF INSECTS
33
these phenomena were active and their effects cancelled. This seems
unlikely, but since the fluctuations of larvae and pupae followed a
somewhat different pattern from those of the eggs (Table I), there is
an element of accidental coincidence in the similarity of the C.V.
values.
While this result suggests there was no major key factor nor regulatory process in action during juvenile development, there was a good
deal of mortality, as is obvious from the values in Table I. Most of this
was due to Mirid bugs and other predators which together caused
mortalities of larvae and eggs estimated at values from 77.8% in 1957
to 98.9% in 1956. Richards and Waloff wrote: “The population of
mirids is probably in no way dependent on Phytodecta and is controlled
by its own complex of parasites and by fratricidal predation.” But:
“If the quantity of broom is reduced, Phytodecta and its predators
become more concentrated and predation becomes more intense.”
Richards (1963) extended this argument and tabulated figures showing
that percent predation increased when the quantity of broom was
reduced (1956, 1958). When percent predation is graphed against the
density of adults in spring of the same years, there appears to be a
well-marked density-dependent relationship. But when the estimates
of egg density are used instead of adults this relationship does not
hold, except for 1956 compared with the other years as a group. It
seems clear that a heavy reduction of Phytodecta, due to contraction
of its host plants and concentration of its predators, occurred a t least
in 1956. Table I shows the reduced numbers of late larvae and pupae
in this year. It also shows that this reduction overshot the mark, taking
the numbers to an appreciably lower level than in the other four years.
This overshooting (resembling the action of a lagging rather than a
prompt density-dependent factor) contributed to the high value of the
C.V. for late larvae and pupae.
The second hypothesis to be drawn from the ranges of variation
shown in Table I arises from the fact that the coefficient of variation
was 71.8 at the time of sampling the larvae and pupae in the soil and
only 53.1 at the sampling of adults in the autumn. This suggests that
a regulatory effect came between these two events. Evidence in support
of this conclusion can be found by taking from the life-tables the
numbers of 4th stage larvae and pupae in the soil, converting them to
density values, and plotting them against the percent mortality occurring between that stage and the autumn adult stage; these mortality
values, as percentages of the larvae and pupae in the soil, are read from
the life-tables. The resultant graph (Fig. 14) suggests that the mortality
between the two stages was strongly density-dependent. The relationship shown in Fig. 14 is such as to compensate to a considerable extent
33
these phenomena were active and their effects cancelled. This seems
unlikely, but since the fluctuations of larvae and pupae followed a
somewhat different pattern from those of the eggs (Table I), there is
an element of accidental coincidence in the similarity of the C.V.
values.
While this result suggests there was no major key factor nor regulatory process in action during juvenile development, there was a good
deal of mortality, as is obvious from the values in Table I. Most of this
was due to Mirid bugs and other predators which together caused
mortalities of larvae and eggs estimated at values from 77.8% in 1957
to 98.9% in 1956. Richards and Waloff wrote: “The population of
mirids is probably in no way dependent on Phytodecta and is controlled
by its own complex of parasites and by fratricidal predation.” But:
“If the quantity of broom is reduced, Phytodecta and its predators
become more concentrated and predation becomes more intense.”
Richards (1963) extended this argument and tabulated figures showing
that percent predation increased when the quantity of broom was
reduced (1956, 1958). When percent predation is graphed against the
density of adults in spring of the same years, there appears to be a
well-marked density-dependent relationship. But when the estimates
of egg density are used instead of adults this relationship does not
hold, except for 1956 compared with the other years as a group. It
seems clear that a heavy reduction of Phytodecta, due to contraction
of its host plants and concentration of its predators, occurred a t least
in 1956. Table I shows the reduced numbers of late larvae and pupae
in this year. It also shows that this reduction overshot the mark, taking
the numbers to an appreciably lower level than in the other four years.
This overshooting (resembling the action of a lagging rather than a
prompt density-dependent factor) contributed to the high value of the
C.V. for late larvae and pupae.
The second hypothesis to be drawn from the ranges of variation
shown in Table I arises from the fact that the coefficient of variation
was 71.8 at the time of sampling the larvae and pupae in the soil and
only 53.1 at the sampling of adults in the autumn. This suggests that
a regulatory effect came between these two events. Evidence in support
of this conclusion can be found by taking from the life-tables the
numbers of 4th stage larvae and pupae in the soil, converting them to
density values, and plotting them against the percent mortality occurring between that stage and the autumn adult stage; these mortality
values, as percentages of the larvae and pupae in the soil, are read from
the life-tables. The resultant graph (Fig. 14) suggests that the mortality
between the two stages was strongly density-dependent. The relationship shown in Fig. 14 is such as to compensate to a considerable extent