This proposal describes three studies of play behavior among nursery school children, designed to test a theory of the function of play in young mammals. It grows out of the PI's published work with sable antelope. The basic idea is that play provides opportunities for "self assesment," i.e., testing motor abilities in absolute and socially comparative terms, practicing these to some optimal level, and then selecting new activties that promote growth. Self assesment is thought to be stressful, so stress will be assesed through cortisol secretions after playground periods, in studies 2 and 3.

The theory seems to be well formulated and grounded in relevant literature. The methods of behavior observation and cortisol measurement are well worked out. Careful attention is paid to controlling for circadian variation and individual differences among children. This project would give the PI an opportunity to develop new methodological skills and re-locate her research program to a setting and population that is much more convenient and reliably available.

The only major question I had about the research design stems from the slow time course (about 30 min.) of corticol secretions. This means that measurements only make sense after, rather than during, play periods. So what other variables that distinguish among play periods might affect cortisol levels? Does activity level generally have an effect, apart from activities appropriate for self-assesment? Would social conflict (arguments or fights) have an effect? How will these potential effects be controlled for?