Ch. 3— Patterns of Animal Use • 57 
For birds, many of those completing the APHIS 
data sheets voluntarily reported bird use under 
the "wild animal" category. According to these data, 
at least 33,910 birds were used in fiscal year 1982 
and 29,781 in fiscal year 1983. Of these, the Univer- 
sity of Maryland used 17,915 birds in 1982, and 
12,305 in 1983 (46). Since this one institution used 
such a large fraction of the reported total, inquiries 
about other large possible users indicated that 
many of the poultry research institutions (mostly 
land-grant universities in the East and South) did 
not report birds on their APHIS forms . The largest 
of these, in terms of poultry research, is North 
Carolina State University, from whom it was learned 
that approximately 41,000 birds were used for 
poultry science and 1,100 in veterinary schools 
(7). Checking the APHIS data sheets for other land- 
grant institutions showed that most had reported 
bird usage. In addition, discussions with research- 
ers at several institutions established that only 80 
to 85 percent of the poultry science usage is in 
laboratories with the remainder mostly in feed- 
ing, management, and breeding studies . Therefore, 
although there is good detail for many institutions 
on bird use, there is uncertainty in the APHIS data 
about nonreporting institutions and about the pro- 
portion of fowl used in actual experimentation. 
Several individuals have estimated bird use in 
the United States. James Will of the Animal Re- 
source Center at the University of Wisconsin in 
Madison, WI, estimated that 25,000 to 100,000 
avian individuals are used for laboratory experi- 
mentation (54). Andrew N. Rowan of Tufts Univer- 
sity School of Veterinary Medicine in Roston, MA, 
estimated that at least 500,000 birds are used in 
biomedical research (40). Both of these figures are 
based on very weak data and so are assigned a 
confidence rating of "poor. "Thus, using these esti- 
mates and the APHIS bird data, an annual use of 
between 100,000 and 500,000 birds is as accurate 
an estimate as can be made. 
Data on Trends in Animal Use 
Several limited data sources exist that suggest 
trends in animal use in the past several years. At 
Wadsworth Center for Laboratories and Research, 
New York State Department of Health (Albany, NY), 
the use of mandated species decreased 40 percent 
from 2,925 in 1980 to 1,754 in 1983. The use of 
rats and mice also decreased substantially (22 per- 
cent), from 72,796 in 1980 to 56,681 in 1983, at 
a time when total research dollars available con- 
tinued to increase ( 1 1 ) . At The John Hopkins School 
of Hygiene and Public Health in Baltimore, MD, 
the daily census of animals decreased from over 
8,000 in 1975 to approximately 2,000 in 1985 while 
animal care personnel dropped from 10 to 4 and 
research expenditures more than doubled (14). 
These data sources are limited in scope, use differ- 
ent counting mechanisms, and can be considered 
anecdotal in nature. They were assigned a confi- 
dence rating of "poor.” 
Calculating Rat and Mouse Usage 
Using these same data sources, estimates for an- 
nual laboratory use of rats and mice in the United 
States were calculated. The criteria and scales de- 
scribed earlier were also applied to assign confi- 
dence ratings to the estimates. To gauge annual 
laboratory -animal use, minimum average costs of 
$4 per rat and $2 per mouse (6,15,24,30,36,55) 
were assumed to represent conservative prices for 
a typical research subject. This permitted extrap- 
olations based on price to represent an expected 
maximum of animals that could be purchased. 
Three different methods were used to estimate 
the use of rats and mice in the United States. The 
first involved using indirect means for the calcu- 
lations, while a second method used 1978 ILAR 
data. The third, and most reliable, method relied 
on corrected USDA/APHIS data and involved cal- 
culations using regression equations. 
Indirect Estimates 
Possible methods for estimating rat and mouse 
usage under this category involve extrapolations 
from data based on NIH funding, NCI usage, NIH 
total use as a function of intramural use, and ani- 
mal breeder information. For example, an estimate 
based on NIH funding involves the following steps 
and assumptions: 
• NIH funds 37 percent of all national biomedi- 
cal research expenditures (52). 
• In 1983, NIH awarded $582,571,000 in direct 
costs to 5,01 1 extramural projects utilizing rats 
and other species (4). If it is assumed that all 
38-750 O 
86 
3 
