290
H . KLOMP
rising egg density (k, = 0.124 + 0.07 log E), but the regression coefficient does not differ significantly from zero.
Egg mortality is mainly due to the egg parasite, Trichogramma, and
if this parasite should operate as a delayed factor, a simple relation betveen k, and density cannot be expected.
The k-value of the remaining miscellaneous mortality factors affecting the egg, which can be computed separately using the data of Table
XPII, does not appear to be correlated with egg density either, and from
this we conclude that in the egg stage direct density dependent mortality cannot be shown to occur.
2. Mortality i n the Larval Stage
It is shown in Table XXV that the addition of larval mortality gives
a remarkable improvement of the regression coefficient (cf. Fig. 3 3 4
and C), and this indicates that powerful regulation occurs in this stage.
As already pointed out larvae are subjected to a complex system of
factors operating in succession over a long period of time and, therefore,
the density dependence in this system will be studied for the various
age-intervals mentioned in Table XXVI.
( a ) Juvenile Mortality. The improvement of the regression coefficient
from 0.73 to 0-82 in Table XXVI indicates the possible existence of a
weak density relationship (see also Fig. 35). The k-values of juvenile
mortality (k2-3), when plotted over first instar larval density (LI), do
not tend to increase with a rising density, and a direct density dependency of juvenile mortality can safely be excluded.
( b ) Mortality of Advanced Larvae. Table XXVI reveals an enormous
increase of the regression coefficient when the density of surviving
advanced larvae (nymphs) is plotted over pupal density (Fig. 35-B and
C). This might indicate the occurrence of density dependence in this
age-interval mortality. However, such an effect cannot make the coefficient exceed unity. That this occurs is due to the abnormally high prepupal mortality in 1957, while the density is then at its minimum
(Fig. 27). This makes the relation to be inversely density dependent, resulting in the high value of the regression coefficient (Fig. 35-C). This
inverse effect is also incorporated in the regression diagrams of Fig. 35-A
and B because in all graphs pupal density has been plotted. It is much
less obvious, however, because its influence is counterbalanced by the
density related processes acting on the advanced larvae (see below).
The exclusion of these processes results in a full expression of the inverse
effect in Fig. 35-C.
The mortality of advanced larvae proves to be density dependent.
This is shown by plotting the relevant k-values (k4--6) over larval
H . KLOMP
rising egg density (k, = 0.124 + 0.07 log E), but the regression coefficient does not differ significantly from zero.
Egg mortality is mainly due to the egg parasite, Trichogramma, and
if this parasite should operate as a delayed factor, a simple relation betveen k, and density cannot be expected.
The k-value of the remaining miscellaneous mortality factors affecting the egg, which can be computed separately using the data of Table
XPII, does not appear to be correlated with egg density either, and from
this we conclude that in the egg stage direct density dependent mortality cannot be shown to occur.
2. Mortality i n the Larval Stage
It is shown in Table XXV that the addition of larval mortality gives
a remarkable improvement of the regression coefficient (cf. Fig. 3 3 4
and C), and this indicates that powerful regulation occurs in this stage.
As already pointed out larvae are subjected to a complex system of
factors operating in succession over a long period of time and, therefore,
the density dependence in this system will be studied for the various
age-intervals mentioned in Table XXVI.
( a ) Juvenile Mortality. The improvement of the regression coefficient
from 0.73 to 0-82 in Table XXVI indicates the possible existence of a
weak density relationship (see also Fig. 35). The k-values of juvenile
mortality (k2-3), when plotted over first instar larval density (LI), do
not tend to increase with a rising density, and a direct density dependency of juvenile mortality can safely be excluded.
( b ) Mortality of Advanced Larvae. Table XXVI reveals an enormous
increase of the regression coefficient when the density of surviving
advanced larvae (nymphs) is plotted over pupal density (Fig. 35-B and
C). This might indicate the occurrence of density dependence in this
age-interval mortality. However, such an effect cannot make the coefficient exceed unity. That this occurs is due to the abnormally high prepupal mortality in 1957, while the density is then at its minimum
(Fig. 27). This makes the relation to be inversely density dependent, resulting in the high value of the regression coefficient (Fig. 35-C). This
inverse effect is also incorporated in the regression diagrams of Fig. 35-A
and B because in all graphs pupal density has been plotted. It is much
less obvious, however, because its influence is counterbalanced by the
density related processes acting on the advanced larvae (see below).
The exclusion of these processes results in a full expression of the inverse
effect in Fig. 35-C.
The mortality of advanced larvae proves to be density dependent.
This is shown by plotting the relevant k-values (k4--6) over larval