Effect of sampling regime on estimation of basal metabolic rate and standard evaporative water loss using flow-through respirometry.

Strict criteria have been established for measurement of basal metabolic rate and standard evaporative water loss to ensure that data can be compared intra- and interspecifically. However, data-sampling regimes vary, from essentially continuous sampling to interrupted (switching) systems with data recorded periodically at more widely spaced intervals. Here we compare one continuous and three interrupted sampling regimes to determine whether sampling regime has a significant effect on estimation of basal metabolic rate or standard evaporative water loss. Compared to continuous 20-s sampling averaged over 20 min, sampling every 6 min and averaging over 60 min overestimated basal metabolic rate and evaporative water loss, sampling every 3 min and averaging over 21 min underestimated basal metabolic rate, and sampling every 12 min and averaging over 36 min showed no difference in estimates. Increasing the period over which the minimum mean was calculated significantly increased estimates of physiological variables. Reducing the frequency of sampling from 20 s to a longer interval of 3, 6, or 12 min underestimated basal metabolic rate but not evaporative water loss. This indicates that sampling frequency per se influences estimates of basal metabolic rate and that differences are not just an artifact of differences in the period over which the mean is calculated. Sampling regime can have a highly significant influence on estimation of standard physiological variables, although the actual differences between sampling regimes were generally small (usually <5%). Although continuous sampling is the preferred sampling regime for open-flow respirometry studies, if time and cost are prohibitive, then use of an appropriate switching system will result in smaller errors than measuring individuals continuously for shorter periods.