control will still be needed. Information on a few commonly used species is summarized 
below (see Ator, 1991, for references and additional coverage). 
With rats, it is especially important to consider the age of the rat and the duration of free 
access to food at which the 100 percent weight was determined if reduction to a percentage of 
that weight is to be used. Rats of some strains (e.g., Sprague-Dawley, Long-Evans hooded) 
are semi-continuous feeders and can gain weight almost indefinitely. In such rats, waiting for 
weight to stabilize in order to determine a free-feeding weight is not practical. If rats attain a 
relatively high weight (e.g., 500 grams), 80 percent of that weight may not be a weight at 
which training will occur rapidly. On the other hand, if a free-feeding weight for a young rat 
is quite low (e.g., 200 grams), 80 percent of that weight maintained over the rat's life span 
may be unnecessarily restrictive (Heiderstadt et al., 2000). The best restricted-weight criterion 
is one at which the rats work reliably for food reinforcers, remain healthy, and live as long as 
possible (i.e., two to three years) in studies in which sacrifice is not an experimental endpoint. 
The weights of mice tend to reach an asymptote relatively quickly, but strains differ 
considerably. Weights should be permitted to rise to a reasonably stable maximum under 
free-feeding conditions before they are decreased by restricted feeding. Although stable 
reduced weights can be maintained easily in mice, accidentally missing a day of feeding may 
prove fatal, in contrast to such regimens with other species. 
Free-feeding guinea pigs steadily gain weight for 12 to 15 months before weight asymptotes. 
Use of food or water reinforcers can be problematic. Some investigators found that restriction 
of either had deleterious effects, but success with particular edible reinforcers (e.g., carrot 
juice, sucrose solutions, a milk and cereal mixture, and commercial guinea pig pellets) has 
been described for guinea pigs maintained under restricted feeding. 
Pigeons tend to self-regulate feeding under free-access conditions, and stabilization of the 
body weight of an adult bird occurs in two to four weeks. The 80 percent body weight 
regimen is most easily used in this species. A typical procedure is to weigh the bird after the 
session to determine the amount of supplemental feeding. The bird is fed the difference (in 
grams) between the current weight and the target weight; with experience, investigators often 
are able to determine an additional amount that can be fed such that the bird will be at, rather 
than below, the target weight for the next experimental session. 
With nonhuman primates, the rate of metabolism and the rate of growth can vary 
significantly even within the same species. Food restriction (e.g., one individualized post- 
session feeding per day), rather than reduction to a specific target weight, usually results in 
stable behavioral baselines. Types of reinforcement used with nonhuman primates vary 
greatly. The one chosen is governed by a complex interaction involving the research question. 
45 
