The Effects of Reduced Gill Area and Hyperoxia on the Oxygen Consumption and Swimming Speed of Rainbow Trout

When a fish swims at its maximum sustained speed (U_crit), its oxygen consumption approaches a maximum level (Brett, 1964) reflecting maximal aerobic requirements of those tissues involved in locomotion (Jones & Randall, 1978). If it is assumed that at ss the gills are fully perfused by blood (Randall & Daxboeck, 1982) and that oxygen transfer proceeds at an optimum rate then any reduction in functional gill area will lead to a corresponding reduction in indicating a direct limitation on oxygen uptake at the level of the gills. However, any deviation from a proportional decrease in max as a result of gill area reduction would indicate limitation of oxygen transport at the circulatory or tissue level.

The purpose of the present study was to ascertain whether oxygen uptake is directly limited by gill area by effecting artificial reduction in respiratory area of rainbow trout and observing the effects on and swimming performance. Further, in the absence of any means of increasing gill area, fish were exposed to hyperoxia in an effort to increase the rate of oxygen transfer across the gills.

Rainbow trout (mean±S.D.; length 31 ± 1cm, mass 286 ± 25g, N=48) were obtained from a fish farm near Nailsworth, Gloucester and maintained in stocktanks at 15 ± 0·2°C for 2 weeks prior to surgical manipulations. Fish were fed every second day with ‘Puwoods Silver Cup’ trout pellets, supplemented weekly with chopped liver, and starved for 2 days prior to swimming trials. To avoid interbatch variation in swimming performance, the same group of fish was used in all experiments and individuals were used only once to avoid possible training effects (Beamish, 1978).

After anaesthesia (MS222, 0·1 g I⁻¹), fish were placed on the bench on their sides and the opercular flap held open. A hot wire cauterizer was applied smoothly along the length of the anterior hemibranch of either the first or second gill arch at the point of filamentar attachment, resulting in haemostasis with minimal blood loss. The procedure was repeated on the posterior hemibranch of the same arch, and on the corresponding arch on the other side of the fish, thus accounting for about 30% of gill surface area (Hughes, in Paling, 1968). The operation was completed within 3 min. Mortality rate was 15 % within the first 24h, thereafter survival was 100%.

Subsequent to the operation, fish were transferred to water containing antibiotic (oxytetracycline hydrochloride, 0·05gl⁻¹) for 5 days prior to being returned to stocktanks for 9–14 days before swimming trials were conducted. By this stage damaged filaments were white in appearance and obviously non-functional (see Fig. 1). Feeding resumed within 48 h of the operation. Control fish were subjected to a sham-operation and identical subsequent treatment.

Relationships between swimming speed (U) and oxygen consumption were constructed essentially as described in Webb (1971). In brief, fish were transferred to a Brett (1964) respirometer and left overnight in a water flow of 15cms⁻¹. The following morning, velocity increments of 0·2 bodylengths per second (Ls⁻¹) of 30 min duration were successively imposed until the fish was unable to remove itself from the retaining grid. In trials where hyperoxia was employed oxygen was bubbled through a series of aeration columns supplying the respirometer until a of 300 mmHg was attained. During each velocity increment the respirometer was closed off for 15 min and declining monitored with a Radiometer E5046 oxygen electrode. U_crit was calculated as described in Beamish (1978). at U_crit was assumed to be ⁠. Standard oxygen consumption (⁠⁠) was estimated by extrapolation of the regression to zero U.

An example of necrotic filaments of the second gill arch 2 weeks after the cautery procedure is shown in Figs 1 and 2. Undamaged filaments are present at the isthmus and base of the arch where access by the cauterizer was restricted.

Table 1 gives the regressions, and U_crit values for all treatments. Estimates of the induced reduction of gill area are calculated from day for rainbow trout given in Paling (1968).

In normoxia, there was no significant difference between slopes or elevations of the regressions (P> 0·05; ANCOVA, Snedecor & Cochran, 1972) when fish with reduced gill area were compared to controls, and consequently the extrapolated estimates of did not differ significantly (P>0·05; Student’s t-test). However, reduction in functional gill area of 27 % as a result of cautery of the first arches and 30% due to cautery of the second arches produced significant reductions in of 29 % and 27 %, respectively, in comparison to controls. U_crit values were significantly reduced (P < 0·05) as a consequence of gill area reduction.

Hyperoxia did not significantly increase or U_crit of control fish compared to those values obtained when swimming trials were conducted in normoxia. In addition, no significant increases in or U_crit were observed in hyperoxia in fish with reduced gill area when compared to corresponding values obtained in normoxia.

Artificial reduction of functional gill area induced a directly proportional reduction in indicating that the fish were unable to make major compensatory adjustments in circulation to enhance oxygen transport. Under conditions of maximum aerobic demand, all available gill area would appear to be fully perfused. As and at lower swimming speeds were unaffected by the imposed gill reduction, under resting or routine conditions total gill area is likely to be incompletely utilized, agreeing with suggestions of Booth (1978) and Hughes (1984).

The present experiments corroborate suggestions that the gas exchange surface of fish relates to maximum oxygen uptake capacity rather than to that incurred under resting conditions (Hughes, 1984). Both and gill area scale to increasing body mass by an allometric coefficient of 0·9 (Brett & Glass, 1973; Hughes, 1980). As Btand scales to body mass by approximately 0·75 (Winberg, 1956), no direct link between gill area and should be expected and it is not apparent in the present study.

Attempts to increase the rate of oxygen transfer at the gills by hyperoxic exposure failed to increase in controls or in fish with reduced gill area, suggesting a direct limit on oxygen uptake at the gills independent of environmental when elevated above normoxia.

Damage to the filaments of either the first or second arches had a similar effect on ⁠. Davis (1971) found that trout with first arch ligation had a reduced arterial and incurred pseudobranch necrosis, although no similar effect occurred if other arches were ligated. Such observations have led to suggestions for the localized presence of oxygen receptors within the region of the first gill arch (Randall, 1982). No pseudobranch necrosis was observed in the present study. Therefore such receptors may not be exclusively present in the area of the first arch and/or sufficient oxygen receptors to regulate ventilation and maintain arterial at lower activity levels are present on the remaining filaments of the first arch which escaped cautery (Fig. 1).

These experiments indicate that trout can survive with at least a 30 % reduction in gill area. Gill damage may be of ecological importance only at higher levels of sustained activity. However, other energetic costs such as reproduction and digestion may require a relatively greater proportion of the reduced aerobic scope that results from gill damage, thus altering the relationship between activity levels and mortality (Priede, 1977).

The authors are grateful to NERC for post-doctoral funding (grant no. GR3/4235) and express appreciation to Dr I. G. Priede for advice on revisions of the manuscript.