BIOLOGY OF SATURNIID MOTHS—BLEST 459 
experience is of value to the social unit. Despite the plausibility 
of this argument, and the clear evidence for the action of group selec- 
tion which the elaborately differentiated sterile castes of the social 
insects unequivocally present, Williams (1957) has dismissed the 
role of group selection as an important factor in the modulation of 
postreproductive longevities. 
The types of protective coloration that have been described above 
all have one feature in common: they impose patterns of learning 
upon the predators against which they are directed. Consider the 
outcome, first, of an encounter between a naive predator and a palat- 
able procryptic prey. The loss of the individual prey cannot influ- 
ence the reproductive capacity of the residual prey population di- 
rectly ; but the predator has now learned to a greater or lesser degree 
how to find more preys, and the lesson must act to the disadvantage of 
the prey population. The probability that the predator population will 
learn to find preys will be increased the longer the preys live. Thus, if 
group selection is able to act, we must expect it to reduce the post- 
reproductive longevity. 
Conversely, an unpalatable aposematic postreproductive prey 
trains the naive predator to avoid other members of the population, 
and group selection should act to augment their postreproductive 
lifespan. How well do the hemileucines fit in with this evolutionary 
scheme? 
In the first place, group selection can surely act in this case: hemi- 
leucine females lay their total complement of eggs in batches, over a 
period of from one toa few days. The larvae are typically gregarious 
until the final instars, grown synchronously, pupate at about the same 
time, and emerge over a period of a few days in phase with a lunar 
cycle. Thus the siblings derived from each female are available at 
the same time in the same ecological areas for the action of group 
selection. 
So far the data available appear to support the hypothesis well: 
the procryptic Lonomia cynira is short lived—at 26° to 29° C. none 
out of 40 odd individuals hatched from pupae lived for longer than 
4 days, and most had died by the third day from eclosion. The 
distasteful Dirphia (Periphoba) hircia is long lived, certainly for 
10 to 14 days after capture, and similar longevities have been noted 
for D. (Dirphiopsis) eumedide and D. (D.) agis. However, this 
part of the hypothesis may only be reliably tested when more data, 
obtained under controlled conditions, become available from further 
field and laboratory studies. As far as life tables are concerned, at 
the moment we have none, but there seems to be every hope that they 
will bear out the present tentative hypothesis. 
We can, however, use the rapidly accumulating information about 
the significance of the rocking response to provide a speculation about 
