Continuous raw trace of water oxygen saturation in a static intermittent-flow respirometer containing an adult coho salmon (Oncorhynchus kisutch) at 8°C. Negative slopes are due to the oxygen consumption of the fish, while positive slopes indicate when the flush pump intermittently switched on to replenish the respirometer with aerated water. The trace illustrates the danger of periodic rather than continuous sampling of oxygen when measuring . If the raw trace was not available and was measured only once at 8 h after entry of the fish into the respirometer, a resting of 3.7 mg min⁻¹ kg⁻¹ would be concluded. However, the continuous nature of the trace clearly shows that a single measurement of at 8 h would yield a significantly elevated estimate of resting (over five times the correct value of 0.7 mg min⁻¹ kg⁻¹ obtained after 16 h) due to spontaneous activity of the fish. Raw data taken from Clark et al. (Clark et al., 2012).

Typical respirometers used for water-breathing organisms including fishes: (A) swim tunnel respirometer (view from above), (B) static respirometer with a closed-circuit recirculation loop, and (C) static respirometer with a recirculation pump within the main chamber. Numbers correspond to: (1) variable speed motor, (2) propeller shaft, (3) propeller, (4) baffles to assist with achieving laminar flow, (5) honeycomb grid to assist with laminar flow (should be as thick as feasible, at front and back of working section, and anywhere throughout the rest of the swim tunnel where space permits), (6) working section where fish is housed, (7) overflow pipe, which extends above water surface, (8) sealable port for oxygen sensor (A), or clear section to allow measurements of oxygen using fibre-optic sensors and spot material (B,C), (9) flush pump (water exits through 7), (10) one-way flow valve (alternatively, the lengths of hosing in B and C can be extended to perform the same function), and (11) recirculation pump. Grey shading depicts the water in A (1 remains dry, 7 extends above water surface), while everything is submerged in B and C except for the top of 7.

Continuous raw traces of water oxygen concentration in three parallel, static intermittent-flow respirometers each being 27 l and containing a quiescent adult blue-spotted rock cod (Cephalopholis cyanostigma) at 28°C. Respirometers used for the traces shown in A and B are each equipped with a recirculation pump with the stream of water passing across the oxygen sensor, while the respirometer used for C does not contain a recirculation pump. The flush (~25 min) and seal (~20 min) cycles of the flush pump are clearly illustrated in A and B as positive and negative slopes, respectively, whereas the trends are not as clear in C due to stratification of oxygen throughout the respirometer in the absence of a mixing mechanism. Consequently, estimates of across the four measurement cycles vary by a maximum of 6.0% for the fish in A, 9.7% for the fish in B, and 45.4% for the fish in C when using sections aligned identically. Panels D–F contain data from the same three fish as A–C with the only difference being that the recirculation pumps were switched off in D and E for the period indicated by the shaded box (respirometer in C and F was never equipped with a recirculation pump). Body masses: (A,D) 343 g; (B,E) 286 g; (C,F) 235 g. Vertical dashed lines are provided for clarity in D–F to indicate the periods when the flush pump automatically switched on. The mean ± s.e.m. estimate of in D when the recirculation pump was on (three cycles) was 1.04±0.01 mg min⁻¹ kg⁻¹, whereas when the recirculation pump was off (four cycles) the mean estimate was higher and more variable at 1.25±0.11 mg min⁻¹ kg⁻¹. Corresponding values for E were 1.21±0.10 mg min⁻¹ kg⁻¹ (recirculation pump on) and 1.80±0.24 mg min⁻¹ kg⁻¹ (recirculation pump off). The mean ± s.e.m. estimate for F across all measurements (seven cycles) was higher and more variable at 3.66±1.06 mg min⁻¹ kg⁻¹. Note that respirometers that lack a sufficient mixing mechanism yield values with significant error and uncertainty, and any data obtained using such methods should be treated with extreme caution.

Frequency histogram of the rates of oxygen consumption of an adult coral trout (Plectropomus leopardus) at 33°C while undisturbed in a static intermittent-flow respirometer ( measured for 6 min in every 10 min for a total of 18 h). Standard was calculated by taking the lowest 10% of values and excluding outliers (outliers considered to be outside of the mean ± 2 s.d. of the lowest 10% of values, illustrated by ‘X’). Shaded box indicates the points used in calculating standard , which had a mean (±s.d.) value of 2.59±0.09 mg min⁻¹ kg⁻¹ (dashed vertical line). Data have been binned into groups of 0.05 mg min⁻¹ kg⁻¹ for illustrative purposes, but in reality the lowest 10% of the raw (non-binned) data points should be used to calculate standard after excluding outliers.

Aerobic scope (solid line) and factorial aerobic scope (dashed line) as a function of temperature in adult female pink salmon (Oncorhynchus gorbuscha), depicting the opposing conclusions that can be drawn from the same data set depending on how the data are analysed and visualised. Modified from Clark et al. (Clark et al., 2011).