58 • Alternatives to Animal Use in Research, Testing, and Education 
expenditures went to projects that used only 
rats, an upper limit can be extrapolated for 
rats purchasable using NIH funds. 
• Twelve percent of direct costs of NIH- 
sponsored research funds go toward the pur- 
chase of supplies, glassware, chemicals, re- 
search animals, and items listed as expenda- 
ble (55). 
• If it is assumed that half of the supply funds 
went toward the purchase of animals, then 
$34,954,260 would be available for the pur- 
chase of rats. 
• At $4 a rat, 8 . 7 million rats could be purchased . 
• In 1983, NIH awarded $531,519,000 in direct 
costs to 4,080 projects using mice . At an aver- 
age cost of $2 per mouse, 16 million mice could 
be purchased with NIH funds. 
• Assuming that NIH supports 3 7 percent of ani- 
mal use in the country, then the potential num- 
ber of these two species purchasable in the 
United States is estimated at 23.6 million rats 
and 43.1 million mice .This indirect method 
(whether it uses NIH data or NCI data or ani- 
mal breeder information) involves many as- 
sumptions, limited data sources, and cannot 
be considered very reliable. It was assigned 
a “poor” confidence rating. 
Estimate Using Corrected ILAR Data, 1978 
The results of the 197 8 National Survey (22) per- 
mit approximation of animal use for all users with 
techniques that fill in the missing data of non- 
respondents based on a method such as the fol- 
lowing: 
• The NIH -grant -eligible nonprofit biomedical 
research organizations responding to the sur- 
vey reported biomedical research expendi- 
tures of $2.2 billion for 1978. 
• Total national biomedical research expendi- 
tures are estimated at $6.27 billion for 1978 
(52). 
• If it is assumed that animal use (in numbers) 
is proportional to the dollar amount spent on 
research utilizing them and that the usage rate 
of animals by all institutions is proportional 
to that of nonprofit institutions, national usage 
equals (nonprofit ILAR 1980) X 6.27 -5- 2.2. This 
yields an estimate of 16 million mice and 5.6 
million rats used in 1978. Such methods do in- 
volve some assumptions not easily justifiable 
and so the confidence rating is somewhat 
lower than for the ILAR data on which they 
are based. In addition, they are based on in- 
formation already 8 years old. 
Estimate Using Corrected APHIS Data 
About two -thirds of the institutions completing 
APHIS annual reports for 1982 and 1983 volun- 
teered information on the number of rats and mice 
used. Regression equations based on those insti- 
tutions reporting the specific species on the An- 
nual Report of Research Facility forms were used 
to estimate the numbers of rats and mice for those 
institutions not reporting these species (17). The 
estimates obtained using these regression equa- 
tions and then simply applying the mean value for 
reporting institutions to the nonreporters are 
shown in table 3-4 (which summarizes all the esti- 
mates discussed). These regression equations yield 
estimates of 8.5 million mice and between 3.4 mil- 
lion and 3.7 million rats used annually in 1982 and 
1983; applying the mean value for reporting insti- 
tutions to those that did not report yields higher 
estimates. Given the fairly detailed database to 
which the regression equations were applied, these 
estimates received a confidence rating of "good.” 
The estimates generated from these corrected 
APHIS data are likely the most accurate that can 
be obtained with data currently available. 
Summary and Analysis 
of Estimates 
Table 3-5 summarizes the various estimates on 
animal use discussed in this chapter. Several fac- 
tors reduce the usefulness of these data, however: 
APHIS’s definition of animal (which excludes rats, 
mice, and birds) and the exemption from regula- 
tion of research facilities that do in-house breed- 
ing and receive no Federal funds. These limitations 
may cause the numbers generated from the APHIS 
data to be underestimations of the total animal use 
in the United States for research, testing, and edu- 
cation. For example, the Directory of Toxicology 
Testing Laboratories published by the Chemical 
Specialties Manufacturers Association, Inc., lists 
110 facilities in the United States. In checking these 
against the list of APHIS registered research facil- 
