Chimpanzee Problem Comprehension : Insufficient Evidence

approximately linear rate to 5.5°C. Control and exposure data were obtained between 3.9° and 5.3°C. A blower (100 ft3/min) provided airflow from above. The airflow, measured below the empty chamber with a hot-wire anemometer (Datametrics Airflow Multimeter Model 800 VTP), was approximately 6 m per minute in the region of the lever and increased to approximately 18 m per minute at the opposite side of the chamber. 10. Microwaves were transmitted to the horn via a coaxial cable to a coaxial-waveguide adapter. The feeder horn with a 6.5 by 7.5 cm rectangular aperture directed the microwaves toward the chamber floor 44.5 cm below with the E field parallel to the axis established by the response lever and infrared lamp. The designated power density specifies that value at the lever. A Narda Model 8315 probe calibrated against an NBS XD-I probe was used to map the field. With the back wall removed, the field was measured at the level of the lever at nine locations in a 10 by 10 cm grid with loci 5 cm apart covering the central area of the chamber. The distribution of power densities varied within 11 percent of the mean value. The field was also mapped with a smaller probe fabricated by J. Ali (U.S. Environmental Protection Agency) which enabled measurements to be made at loci close to the chamber walls. There was close correspondence between the two sets of measurements. The energy absorption rate per unit mass was estimated according to the relationship P = 4.186 CAT/t, where C is the tissue specific heat in calories per gram per degree Celsius (in this analysis, C = 0.83), AT is the temperature increase in degrees Celsius, t is the duration of exposure in seconds, and P is watts per kilogram [C. C. Johnson, J. Microwave Power 10, 249 (1975)]. A YSI 423 probe inserted 6 cm into the colon measured the temperature of a pentobarbitalanesthetized rat encased in a Styrofoam block during a brief exposure to microwaves [Lu et al. (11)]. Delta-t, the rate of temperature change, included a correction for the rate of temperature change immediately preceding the exposure. The absorption rate was approximately 8.4 W/kg at a power density of 41 mW/cm2, resulting in a specific absorption rate of 0.20 W/kg per milliwatt per square centimeter. 11. S.-T. Lu, N. Lebda, S. Michaelson, S. Pettit, D. Rivera, Radio Sci. 12(S), 147 (1977). 12. These data are the results of the second exposure session. After studying three rats, we found that the results from the first session were similar to but more variable than those from the second. Since such an outcome might be attributable to the novelty of the microwaves, an effect often seen during initial exposure to drugs and other stimuli, we decided to focus on the data from the second sessions of those three rats plus three others. The following are the means (standard errors in parentheses) of the proportion of time the heat lamp remained on at each power density for all six rats. For the first exposure session: 0 mW/cm2, 0.364 (0.017); 5 mW/ cm2, 0.333 (0.040); 10 mW/cm2, 0.324 (0.032); and 20 mW/cm2, 0.273 (0.029). For the second exposure session: 0 mW/cm2, 0.326 (0.009); 5 mW/cm2 0.298 (0.015); 10 mW/cm2, 0.264 (0.020); and 20 mW/cm2, 0.199 (0.011). Thus the two functions are similar, but the variability from the second session is less, as would be expected, after previous experience. 13. Serial correlation within individual rats was examined by the Durbin-Watson test [J. Neter and M. Wasserman, Applied Linear Statistical Models (Irwin, Homewood, Ill., 1974)], which confirmed the absence of an effect from the previous 15-minute exposure and the suitability of the linear regression model. 14. Using randomization tests [A. R. Feinstein, Clinical Biostatistics (Mosby, St. Louis, 1977)] we compared results for the six rats at each power density with all control periods, with only the preceding control period, and with the adjacent exposure to a different and nonzero power density. Two-sided P values were less than P = .003 except for 5 mW/cm2 against all zero exposures and 5 mW/cm2 against adjacent 10 mW/cm2 (P = .02). These results provide additional confirmation of the sensitivity of the procedure as well as an independent check of the regression analysis. 15. That interpretation is consistent with the conclusion reached in numerous studies of behavioral thermoregulation, including those of Weiss and Laties (7) and Carlisle (8). 16. See N. W. King, D. R. Justesen, and R. L. Clarke [Science 172, 398 (1971)] for an example of a different sensitive procedure. We did not attempt to determine the limits of the sensitivity in the present study; instead, we wanted to deter-