The population density of herbivores depends on the spatial scale as well a
s the temporal scale. In a small-scale, short-term experiment, the number o
f individuals entering from the surrounding area will be most influential i
n determining the herbivore density. In large-scale, longterm experiments,
however, the density of herbivores will rather be influenced by the surviva
l rate of individuals inside the field because most of the herbivorous popu
lation derives from the parents that developed inside the held. If we want
to predict the large-scale long-term density of herbivores, therefore, emph
asis should be placed on the estimation of survival rate. To elucidate the
effects of plant density on the large-scale long-term abundance of cabbage
pests, we examined the survival rates of three lepidopterous pests, the sma
ll white butterfly Pieris rapae crucivora Boisduval (Pieridae), the beet se
mi-looper Autographa nigrisigna (Walker) (Noctuidae), and the diamondback m
oth Plutella xylostella (Linnaeus) (Yponomeutidae) under two levels of plan
t spacing (sparse plot, 2 m x 2 m interval; dense plot, 0.5 m x 0.5 m inter
val). The experiment with four blocks was repeated in two seasons. The numb
er of eggs per plant was larger in the sparse plots than in the dense plots
for all species. The survival rate of eggs and larvae, on the contrary, wa
s lower in the sparse plots than in the dense plots. The lower survival rat
e of eggs in the sparse plots was mainly caused by the density dependency,
while the lower survival rate of larvae in the sparse plots was mainly caus
ed by the direct effects of plant density. It was thus suggested that the d
ensity of herbivores may become lower in the sparsely planted field in the
long run because of the higher mortality of larvae.