33 
practically to rule their lives. They go abroad in search of food for 
the family, and to the family they must return, however far they may 
travel and into whatever difficulties and adventures they may fall by 
the way. 
These facts make it necessary that there should be in the economy 
of ant society some self-acting-, self-regulating apparatus for keeping 
the family groups wholly separate and distinct, for making sure that 
every foraging worker shall find its way back to its own companions 
— that it shall neither be liable to lose its way completely nor likely to 
attach itself to an alien family. From this point of view it is greatly 
to the general interest that it should find a welcome and a permanent 
home only among its immediate kindred and habitual companions, and 
nothing could make this condition more secure than a universally 
hostile reception to the wanderer in every other family group. Clan- 
nishness in ants has thus the same justification that it has among 
savage men. It is a means of maintaining the necessary concentration 
of the group for the care of the young, and hence for the preservation 
of the race. 
Area occupied by a Single Colony 
Practical advantage was taken by us of this intolerance of ants- 
towards those of other parentage to ascertain the limits in the corn- 
field of a single colony or family group. Assuming that specimens 
from adjacent hills of corn are members of the same family if they 
affiliate peaceably, but that they belong to diflferent families if they 
fight, I instructed my field assistant to bring the inhabitants of adja- 
cent hills into contact with each other in artificial nests until all those 
afifiliating with the original group used as a test were distinguished 
from those about them who refused peaceable affiliation. By a care- 
ful application of this method at Elliott, May 23, 1906, two family 
areas were thus marked out. 
The group of hills included in the first experiment occupied an 
area 10 hills in length by 7 in width, and 20 of these 70 hills were in- 
fested by ants. Selecting the inhabitants of a hill from the center of 
this tract for the purposes of this test, ants from the other infested 
hills were placed with them successively. By this means it was found 
that the inhabitants of ten of these hills were members of one com- 
munity, harmonizing with each other perfectly when commingled, but 
that those of the other nine were strangers to them, since the repre- 
sentatives of these groups instantly fought when placed together. The 
ten nests thus identified as related were distributed over an irregular 
area about 30 feet long by 14 feet in greatest width, and were proba- 
bly all connected by a network of underground channels. Hills inhab- 
ited by hostile ants were in many cases immediately next to others 
occupied by members of this family group. 
The second experiment covered a plot of 130 hills, 32 of ,which 
were occupied by the corn-field ant. The occupants of 12 of these 
hills affiliated peaceably, while the remaining 20 were rejected by the 
