2 . The Experimental Pen 
The following criteria were used in arriving at 
more exact details of the design of the pen and the 
establishment of the colony: 
1. The pen should be large enough to accom- 
modate several local colonies and yet at the 
same time be small enough that the observer 
from a tower could observe the rats at all 
points in the pen. 
2. The pen should be enclosed by a barrier 
fence which would prevent the emigration of 
resident rats, or the migration into it by aliens. 
3. Predation by vertebrate predators should be 
eliminated. 
4. Food and water should be restricted to a 
single location accessible to rats living at all 
points in the pen. 
5. Some artificial harborage should be provided, 
although the rats should have the opportunity 
to construct burrows at locations of their own 
selection. 
6. Internal barriers with passages through them 
should be provided. This would enable the 
formation and delimitation of local colonies, 
while at the same time it would facilitate the 
observation of interactions among rats. 
7. The rats to be introduced into the pen should 
be members of a wild population and yet be 
as nearly homozygous genetically as possible. 
This precaution was desirable in order to make 
more certain that such variability of data as 
might become evident would primarily be 
due to environmental conditions. 
The manner in which these objectives were 
satisfied are mainly illustrated in figures 3, 4, and 
5. The area enclosed by the pen was approxi- 
mately one-quarter acre. In the city blocks the 
area available to rats ranges from y 4 to acre. 
However, since portions of yards are not available 
to rats, because of their use by man, the area in 
the experimental pen closely approximated the 
average amount available to rats in city “row 
house” blocks. A tower, whose floor was 20 feet 
above the ground, was placed on the southwest 
side of the pen. Although the behavior of rats on 
the half of the pen nearer the tower could be 
observed in more detail, it was still possible to 
identify individual rats at even the farthest points 
with the aid of 6-power binoculars. 
The problem of preventing rats from escaping 
was solved by constructing a fence of hardware 
Figure 3. — Diagram of the experimental area. The numbered 
rectangles in each of the corner areas represent the positions 
of the sunken wooden harborage boxes. The dashed 
lines enclose paths or areas, which were maintained free of 
vegetation. The eight passages through the barrier fences 
are referred to in the text as here numbered. 
cloth (fig. 6) which extended 4 feet above the 
ground and 2 feet underground. In addition, the 
underground portion of the fence had a 2-foot 
shelf of hardware cloth extending out into the pen 
at a depth of 2 feet. Thus, when the rats burrowed 
down, they would strike this shelf. No rats ever 
surmounted this underground baffle although a few 
dug down that far on very rare occasions. At the 
top of the 4 feet of fence above ground there was an 
18-inch overhang sloping downward, which was 
sheathed with “screen glass”. This is window 
screen with the space between the meshes filled 
with a clear plastic. Actually very few rats ever 
climbed this high since a wire carrying a high 
voltage, as used on cattle fences, was placed a few 
inches above the ground on the inside surface of 
the fence. Rats soon learned not to climb fences. 
The problem of eliminating predators was not 
quite so simple but it was effectively solved. Posts 
supporting the limiting fence were placed at 10-foot 
intervals. On the top of each and at the base of 
each a steel trap was left permanently set. Forty- 
nine predators were removed during the 27 months 
of the study: barred owls ( 3 ), great-horned owls 
5 
