54 BULLETIN 647, U. S. DEPARTMENT OF AGRICULTURE. 
The ants soon began tunneling into the ground at short intervals 
along the entire course of the trail, and by October 15 these shafts 
were numerous. The foragers still followed the original trail along 
the surface of the ground, but could no longer be traced for its 
entire length, as the}' were continually disappearing into the tunnels. 
Queens gradually separated from the original colony and took up 
their abode in the tunnels, until finally there was a string of small 
colonies all along the trail from mother colony to sirup. The origi- 
nal purpose of the tunnels doubtless was to protect the workers from 
light and heat while they rested from their labors, but the queens 
found them well adapted for nesting purposes. 
SHELTER STRUCTURES, OR " COW SHEDS." 
In the trees the ants invariably utilize such natural shelters as 
cracks and depressions in the bark, abandoned tunnels of borers, the 
space between touching leaves and fruits, etc., often further exclud- 
ing light by piling particles of trash along the edges of cracks and 
walling in the space between nearly touching leaves, fruits, and 
branches. A portion of the ants foraging in trees almost invariably 
may be seen retracing their steps up the tree, carrying either liquid 
forage or prey with them. If traced, these ants usually will be 
found seeking a rest in the nearest shelter of the sort mentioned. 
Sometimes, while resting, their forage is deposited nearby and occa- 
sionally thereafter forgotten; at other times it is held indefinitely in 
the jaws. 
The erection of the so-called " cow sheds " over scale insects and 
aphids is a further extension of this habit of building shelters in 
which the worker ants can rest. The number of ants attending 
aphid and coccid groups is almost alwa} r s greater than can secure 
honey dew continuously. Some of them, therefore, always must be 
waiting until their hosts have a fresh supply ready. During this 
period of waiting and unrequited solicitation the " cow sheds " serve 
the usual ant-protective purpose. These structures, of course, may 
protect from enemies the particular insects covered by them, but, 
even if this protection were absolute, no great number would benefit 
by it, because comparatively so few are covered. The occasional 
occurrence of parasitized remains of scales under these " cow sheds " 
indicates, furthermore, that the protection afforded even those com- 
paratively few scales is often faulty. 
On orange trees badly infested by the black scale shelter structures 
sometimes are found over groups of mealybugs, and in this case their 
most important function happens to be protection of the ants and 
mealybugs against the honeydew of the scale and its accompanying 
