Fig. 6. Heart rate depends on temperature. (A) Scatter plots show the frequency of the heartbeat in isolated hearts (N=10) in vitro as a function of temperature. Each symbol represents a different heart. The number of hearts failing at temperatures of 18, 20 and 22°C, was 2, 3 and 5, respectively. (B) Heart rates from isolated hearts (filled squares) plotted for comparison with heart rates recorded in intact lobsters in vivo (filled circles) as a function of temperature. In both data sets there are significant differences in mean heart rate across temperature (P<0.0001). Assuming a linear trend, the estimated slopes (means ± s.e.m.) of the lines relating heart rate to temperature in intact animals and in isolated hearts are 0.040±0.004 and 0.026±0.006, respectively. *Data are significantly different at P<0.05; **P<0.001; P-value based on contrast in repeated-measures model. The P-values are not adjusted for the multiple comparisons. Even with adjusted differences the data from 12 to 22°C are still significant.) Only data from beating hearts were included for calculation of the mean, therefore, for isolated hearts values of N are as follows: from 2–16°C, N=10; at 18°C, N=8; at 20°C, N=7; at 22°C, N=5. In intact animals only one heart failed as temperature increased (at 20°C). Heart rates (means ± s.e.m.) measured in intact animals in the wild [open symbols; data reproduced from (Mercaldoallen and Thurberg, 1987)]. (C) Heart rates in intact animals (N=4) measured over time as temperature increased from 2°C to a steady state level of 12°C. Horizontal bar indicates time period of temperature change (2 to 12°C) and time period of steady state temperature at 12°C.