PERFECT SOCIETIES OF INSECTS. 167 
he mentions one which amounted to more than three 
times that number (40,000). A swarm seldom or never 
takes place except when the sun shines and the air is 
calm. Sometimes, when every thing seems to prognos- 
ticate swarming, a cloud passing over the sun calms the 
agitation; and afterwards, upon his shining forth again, 
the tumult is renewed, keeps augmenting, and the swarm 
departs*. On this account the confinement of the queens, 
before related, is observed to be more protracted in bad 
weather. 
The longest interval between the swarms is from seven 
to nine days, which usually is the space that intervenes 
between the first and the second. The next flies sooner, 
and the last sometimes departs the day after that which 
preceded it. Fifteen or eighteen days, in favourable 
weather, are usually sufficient for throwing the four 
swarms. ‘The old queen, when she takes flight with the 
first swarm, leaves plenty of brood in the cells, which soon 
renew the population ®. 
It is not without example, though it rarely happens, 
that a swarm conducted by the old queen increases so 
much in the space of three weeks as to send forth a new 
colony. Being already impregnated, she is in a condi- 
tion to oviposit as soon as there are cells ready to receive 
her eggs: and an all-wise Providence has so ordered it, 
that at this time she lays only such as produce workers. 
And it is the first employment of her subjects to con- 
4 Bees are generally thought to foresee the state of the weather : 
but they are not always right in their prognostics ; for Reaumur wit- 
nessed a swarm, which after leaving the hive at half-past one o'clock | 
were overtaken by a very heavy shower at three. 
6 Huber,i. 271. 
