lute). They state that the 8- to 14-micron band 
seems to be the optimum wavelength region 
for infrared detection of warm-blooded ani- 
mals. They say, “To be detected as different 
from its background, an animal must emit (or 
mask) enough energy to produce an instanta- 
neous response, averaged over the entire field- 
of-view, which is greater (or smaller) than the 
response produced by the background alone.” 
In their study the field-of-view was approxi- 
mately 3 feet in diamter; thus, few game ani- 
mals would completely fill it. Ground measure- 
ments were used to test whether or not rodents 
or birds would be confused with larger ani- 
mals. These measurements indicated that at 
the altitude flown, it would be difficult to detect 
a fox and that it would be impossible to detect 
smaller animals. 
The test on the George Reserve was made at 
noon on January 4, 1967, when the Reserve 
had a snow cover of 6 to 8 inches and the air 
temperature was approximately 25° F. Day- 
time sensing with a snow cover and overcast 
sky (warm air) appeared most appropriate. 
Imagery was obtained from an altitude of 
1,000 feet, and seven flight lines were used for 
complete coverage. Imagery was interpreted by 
the infrared physics laboratory of the Univer- 
sity of Michigan, and the hotspots were identi- 
fied as deer by the interpreters. These were sub- 
sequently field checked to be certain that a per- 
manently hot target did not exist at that point. 
Ninety-eight animals were detected. 
In this test, the 7° C. differential in infrared 
emission between the deer and background was 
well within the detection range. The tempera- 
ture differential was nearly ideal, but detection 
would be possible with a lower differential. 
Different vegetation types gave approximately 
the same apparent temperature. The evergreen 
canopy on the Reserve is limited, and no deer 
were in the area during the flight. The noon 
timing of the flight probably reduced any pos- 
sibility of fox or racoon being detected since 
they would tend to be in their dens. The au- 
thors concluded that infrared scanning... . 
“may give good census figures on big game 
populations in relatively open canopies; for ex- 
ample, counting of deer in hardwood stands in 
winter, and in low brush and censusing open 
range animals such as caribou, antelope, bison, 
etc. Since the technique involves heat rather 
than visible light, scanning could be done at 
night when animal activity is frequently high- 
est.” 
Results suggest that under a conifer canopy 
and probably under a hardwood canopy, many 
animals would escape detection. At present the 
method is extremely expensive and is probably 
feasible only when a scanner could be used in 
conjunction with other applications, such as 
detecting forest insect damage or water pol- 
lution. More details are available in a recent 
paper (McCullough, Olson, and Queal 1968). 
The two studies discussed above suggest that 
for some time in the future, sophisticated re- 
mote sensing techniques will not have wide use 
in game census procedures. In special situa- 
tions, they may be valuable—for instance, in 
large-scale photography for censusing concen- 
trations of animals or waterfowl. And cer- 
tainly photography, with or without infrared 
scanning, is valuable for sampling concentra- 
tions of big-game animals in open range like 
that on the African Veldt. 
It appears doubtful that radar scanning will 
have much application in mammal studies in 
the near future. However, radar has an inter- 
esting potential because it can penetrate rain, 
clouds, tree canopies, and even soil; if suitable 
wavelengths can be used, the returns from 
trees, soil, and animals can be separated and 
the animals thus counted. Despite its present 
limitations for mammals, radar scanning has 
been shown to have utility in following bird 
movements, especially where the bird is sil- 
houetted against the sky. Work in this area 
has been in progress since before 1958 (Tedd 
and Lack 1958). A major problem has been ac- 
curately estimating numbers of birds from the 
radar returns. By calculating size of the echoing 
areas of individual birds, Eastwood, Isted, and 
Rider (1962) were able to estimate the number 
of birds in a roost from the radar returns on de- 
parture flights from the roost. The use of 
radar in ornithology has been summarized by 
Eastwood (1957). 
In a recent paper, Dyer (1967) describes a 
scheme for analyzing scan by motion picture 
filming of bird movement on the radar screen. 
He used the technique to determine relative 
densities of redwing blackbird flocks moving in 
a concentration area in North Dakota. Reflec- 
tions from buildings and shelterbelt plantings 
caused problems in the study of the local low- 
flying flocks. 
REMOTE SENSING IN WILDLIFE 
RESEARCH 
What other opportunities are there for aer- 
ial remote sensing in wildlife work? The au- 
thor’s initial interest concerned the potential 
of infrared scanning for investigating the use 
of forest openings by white-tailed deer. We 
were hopeful that by scanning at night we 
might determine animal use for many open- 
ings, as well as the size of openings favored by 
the animals. This is still a possibility, although 
the opportunity for such study is now remote. 
Another application that might be investigated 
197 
