Table 1.
Effects of explanatory variables on VCO2 from rested geckos (n = 48) and recently active geckos (n = 70) analyzed by linear models.
The same explanatory variables were used to assess time to recovery in recently active geckos. The P-values associated with significant predictors are in bold. F-values for these tests are reported in Table B in S2 File.
Table 2.
Effect of explanatory variables on VCO2 found in minimal models obtained by an automated stepwise backward procedure based on AIC values for rested geckos (n = 48) and recently active geckos (n = 70).
The same explanatory variables were used to assess time to recovery in recently active geckos. The excluded explanatory variables by the automated procedure are denoted with a dash symbol. The P-values associated with significant predictors are in bold. F-values for these tests are reported in Table C in S2 File.
Fig 1.
Relationship between mass and number of pentastomes versus log (VCO2) production for rested geckos.
Partial regression plots (y axis: residuals of regressing the response variable on the explanatory variables, but omitting the explanatory variable of interest, and x axis: residuals of each explanatory variable of interest regressed on all remaining explanatory variables) of (a) the relationship of log (gecko mass) to log (VCO2), given the effect of number of pentastomes, (n = 48); and (b) the relationship of number of pentastomes to log (VCO2) given the effect of log (gecko mass). Minimal regression model for the rested gecko experiment: log (VCO2) = (1.023 × log (gecko mass))–(0.005 × number of pentastomes)– 2.465; model adjusted R2 = 0.2093.
Fig 2.
Relationship between mass versus VCO2 production in stationary phase (p0) for recently active geckos.
Partial regression plot (y axis: residuals of regressing the response variable on the explanatory variables, but omitting the explanatory variable of interest, and x axis: residuals of each explanatory variable of interest regressed on all remaining explanatory variables) of the relationship of gecko mass to VCO2 produced in stationary phase (p0) for recently active geckos (n = 70) given the effect of population. Minimal regression model for recently active geckos: VCO2 (p0) = (0.390 × gecko mass)– 0.101; model adjusted R2 = 0.8239.
Fig 3.
Relationship between mass and number of pentastomes versus VCO2 production in secondary phase (p3) for recently active geckos.
Partial regression plots (y axis: residuals of regressing the response variable on the explanatory variables, but omitting the explanatory variable of interest, and x axis: residuals of each explanatory variable of interest regressed on all remaining explanatory variables) of (a) the relationship of gecko mass to VCO2 produced in secondary phase (p3) for recently active geckos (n = 70) given the effect of number of pentastomes and mites, and (b) the relationship of number of pentastomes to CO2 produced given the effect of gecko mass (g) and mites. Minimal regression model for recently active geckos: VCO2 (p3) = (1.411 × gecko mass)–(0.033 × number of pentastomes)–(0.063 × number of mites) + 0.141; model adjusted R2 = 0.7649.
Fig 4.
Relationship between time in secondary phase (p4) versus number of pentastomes for recently active geckos.
Minimal regression model for recently active geckos (n = 70): log (time in stationary phase) (p4) = (0.013 × number of pentastomes) + 6.164; model adjusted R2 = 0.1424. Note log scale used on y-axis.