Energy limits to body size in a grazing reptile, the Galapagos marine iguana

Ecology, Oct, 1997 by Martin Wikelski, Victor Carrillo, Fritz Trillmich

Food intake

We caught 111 iguanas (71 on Santa Fe, 40 on Genovesa) that had been observed focally during their foraging time. They were caught immediately at the end of the daily foraging, and their stomachs were flushed (methodological details in Wikelski et al. 1993). Stomach contents were sun-dried in the field. Samples were exported under CITES permissions Numbers 006 IC and 009 IC of the Ecuadorian government. Dry mass of the stomach contents was determined in the laboratory after drying the samples at 60 [degrees] C to constant mass. The food intake per bite was determined as the dry mass of algae divided by the number of bites per day. To calculate the intake for different body length classes, we assumed that the focal animals were representative of their respective body length classes. We then used mean bite rates of focal animals for each body length class, mean body masses of iguanas, and mean foraging times to calculate the mean feeding performance of iguanas of a given body length class (Table 2). In this calculation, we lumped data from the two years for each island and calculated the mean intake per bite.

To convert dry mass of stomach contents into energetic units, we prepared the samples as previously described (Wikelski et al. 1993) and analyzed them by bomb calorimetry. Conversion factors were 13.8 [ or -] 2.1 kJ/g of algae dry mass (n = 24) for Santa Fe, and 10.0 [ or -] 2.7 kJ/g (n = 17) for Genovesa. This difference in energetic content is explained by the differences in algae species composition (see also Nagy and Shoemaker 1984, Wikelski et al. 1993).

Statistical analysis

Data were processed with SPSS (1991) for Windows. Two-tailed test statistics were used. Residuals of regressions were inspected for normality (see Fig. 3 legend for an exception). Data are given as mean [ or -] 1 SD if not indicated otherwise, except for regression equations (mean [ or -] 1 SE). A General Linear Model was used to perform the analyses of variance whenever sample sizes were not equal. Type III sums of squares were used in tests for significance. F value subscripts indicate degrees of freedom of the model and error. Significance for all tests was accepted at the [Alpha] = 0.05% level.

RESULTS

Sea surface temperatures (SST)

Genovesa and Santa Fe differed significantly in temperature of the surrounding surface water (SST), with water temperatures at Santa Fe being about two degrees lower than at Genovesa during both years ([ILLUSTRATION FOR FIGURE 1 OMITTED]); Santa Fe: 1991/1992, 25.3 [degrees] [ or -] 1.8 [degrees] C (n = 119), 1992/1993, 23.4 [degrees] [ or -] 2.1 [degrees] C (n = 154); Genovesa: 1991/1992, 27.3 [degrees] [ or -] 1.7 [degrees] C (n = 109), 1992/1993, 25.6 [degrees] [ or -] 1.5 [degrees] C (n = 92); ANCOVA factorial model: SST = island (P [less than] 0.001) year (P [less than] 0.001) island x year (P [less than] 0.001) date (P [less than] 0.001) error; [F.sub.4,470] = 500, (P [less than] 0.001; higher level interactions were not significant and were therefore dropped). Water temperatures were also markedly higher in the 1991/1992 El Nino year than in the near "normal" year 1992/1993. Water temperatures steadily increased during both years, from November until February (all linear regressions with P [less than] 0.001). Such a temperature increase is expected, because this time period marks the transition from the cold season (approximately June-November) to the warm season (December-May) in the Galapagos.