Rats were stressed by intravenous injection, tail-warming or moderate restraint for 30s, i.e. by stresses imposed by normal handling during experiment. Liver glutamate concentrations were greatly affected. The results were substantially the same in two varieties of rat (Wistar and Sprague-Dawley), in two laboratories, in experiments carried out by two sets of workers, and after all three stresses. The following detailed results refer to Wistar rats. 1. In starved rats at 20 degrees C and 30 degrees C and in post-absorptive rats at 20 degrees C stress by injection raised liver glutamate concentrations from 1.54, 1.57 and 1.88mumol/g wet wt. 30s after injection to 3.4, 2.7 and 3.6mumol/g wet wt. respectively a few minutes later. In starved rats at 20 degrees C the concentration then fell slowly to 2.3mumol/g wet wt., in starved rats at 30 degrees C it remained steady, and in post-absorptive rats at 20 degrees C it rose slowly to about 4.3mumol/g wet wt. The final values seemed fairly steady and corresponded to an ;alert' state. 2. In starved rats at 20 degrees C anaesthesia, with or without injection or cannulation during it, raised glutamate concentrations to the ;alert' values, which were maintained for 2-3h. 3. Liver alanine concentration in post-absorptive rats initially fell from 1.5 to 0.8mumol/g, and then stayed fairly constant. 4. Aspartate and glutamine concentrations altered only in starved rats, and proportionately much less than those of glutamate. 5. The necessity for knowing the time-dependence of glutamate concentrations after experimental handling is emphasized. 6. There is no wholly satisfactory explanation of the observations.