1987] 
Windsor — Tortoise beetle, Acromis sparsa 
137 
which predators seem to have an advantage. When A. sparsa larvae 
move between feeding sites they normally move 2-6 cm ahead of 
their mother. In so doing they are unguarded and on several occa- 
sions I observed chalcid wasps (probably Spilochalcis sp.) flying in 
tight circles around and briefly landing on these moving groups of 
larvae. This may be especially common as groups composed of older 
larvae move the 1-2 m down the vine to a pupation site. Mothers 
did not shield or otherwise defend their offspring under those cir- 
cumstances and may not have sensed the danger. Larvae also 
became vulnerable to predators if their aggregation became divided 
into two groups, only one of which could be competently guarded. 
Predacious bugs, especially Stiretrus spp. (Pentatomidae, Asopi- 
nae), were adept at creating and exploiting this situation. Females 
could charge and “bulldoze” a Stiretrus adult from an area with 
larvae. However, repeated approaches by the bug were eventually 
rewarded by the temporary isolation of some larvae, some remain- 
ing on the under surface of the leaf while the others had moved to 
the top, etc. These were quickly speared and sucked dry. Predacious 
bugs tended to remain with a family group for several days, often 
until offspring were gone. Orphaned larval groups were only rarely 
encountered. Individuals within these groups were usually tightly 
aggregated suggesting that the presence of the mother is not crucial 
to maintenance of the aggregation. Poorly aggregated, orphan 
groups were also occasionally found. Individuals in these groups 
were usually few in number implicating predators or parasitoids as 
the disruptive factor. Aggregated larvae did not disperse nor did 
mothers react when I presented injured or crushed members of their 
group indicating the absence of an effective volatile alarm substance. 
Larval mortality in defended groups 
Natural levels of larval mortality were examined at the “field” and 
“forest” sites during August and September 1978. Due to the differ- 
ences in egg mortality described above only 8 of 19 females and 
brood reached the larval stage at the “field” site; all of 22 at the 
“forest” site. Additionally, the initial number of larvae per mother 
differed significantly between sites; 29 larvae per group at the forested 
habitat versus 14 in the old field site (Mann Whitney U Test, 
z = 4.14, p < 0.001). The rate of larval disappearance, roughly 1.2 
larvae per day (Fig. 3), did not differ between sites (Mann Whitney 
U Test, z = 0.133, p = 0.45). Larvae from only one of eight groups 
