On the Respiration of the Killifish Oryzias Latipes

It was one of the main purposes of the present investigations to consider these aspects of fish respiration. A reduction of external stimuli was achieved by keeping the fish in darkness and in a constant-temperature room. Most measurements were recorded remotely in order to avoid effects due to the presence of an experimenter in the room.

It is known that the gills of marine fishes show wide variations in their gill areas (Gray, 1954). Measurements of gill dimensions in many species of fish, and determination of the relationship between body size and gill area within single species have also been made (e.g. Hughes, 1966,1970, 1972; Muir & Hughes, 1969). It is of interest to determine the respiratory area in relation to the mechanism of gill ventilation. A series of experiments was therefore carried out in which ventilation volume was determined and also measurements were made of the gill dimensions for the same individuals.

In the present paper an attempt is also described to record potential changes in the water surrounding the body of a fish, as recently found by Iwata, Watanabe & Matsuoka (1969).

The fish used was the killifish, Oryzias latipes, which had been collected from ponds and transferred into glass basins supplied with well-aerated tap water in the laboratory. A relatively small number of individuals were kept together with unicellular algae, e.g. Chlorella ellipsoidea, and microscopic freshwater invertebrates, e.g. rotifers throughout the experiment. The killifish is a common freshwater fish in Japan and small in size. The fish used were usually in the size range 120−290 mg.

Most experiments were carried out from May to September 1969, and normally the measurements were made during the afternoon in order to maintain the level of metabolism depending on the time of day, because various authors (cited in Fry, 1957) have reported that fish often show endogenous daily cycles of activity.

Determinations of oxygen consumption were made on killifish kept in cylindrical tubes of about 5 mm diameter through which water was passed continuously. The fish usually remained quiet and breathed regularly for prolonged periods under these conditions. The of the water before and after passing the fish in the experimental chamber was measured with an oxygen electrode (Oxygen Analyser, Model 777, Beckman-Toshiba); the flow rate of water through the respirometer was also determined. In order to determine the ventilatory frequency the potential changes accompanying respiratory movements were also recorded with electrodes of platinum, active and indifferent, which passed through the cylindrical tube and were placed in the vicinity of the head and the tail tip, respectively. The potentials were amplified (Bio-electrographic Amplifier, Sanei, EB-102; time constant: 1·5 sec) and displayed on a pen-recording oscillograph (Sanei, IR-102) as the potentiometric null instrument.

The experimental arrangement used to determine the oxygen consumption and ventilatory frequency is shown in Fig. 1. The tube containing the killifish, the oxygen electrode and the platinum electrodes remained at a constant temperature throughout the experiment.

Direct measurement of the of the expired water was not possible, but the proportion of oxygen removed from the water during its passage over the fish was determined. The rate of flow of water through the respirometer tube was varied by adjusting the hydrostatic pressure of the tank from which water flowed throughout the experiment. Flow rate was determined by collecting water from the outlet of the respirometer tube at the beginning and end of a series of measurements; in most experiments the two determinations were substantially the same.

In 1939 Black, Fry & Scott (cited in Fry, 1957) observed that ‘the handling of fish required to place them in the respiration chamber is sufficient stimulus to cause them to consume oxygen at their maximum rate for some time afterward*. It is therefore necessary to ascertain the duration of the effects of introducing the fish into the apparatus. An example of the relation of activity or excitement (caused by handling) to oxygen consumption is shown in Fig. 2. Here high initial rates, which gradually subside, are found for about 15 min after introduction of the fish. A delay of at least 15 min after the fish have been handled is therefore required before measurements are made.

Fig. 3 indicates the comparison between the active state (initial high level) and the resting state of oxygen consumption, and on the percentage of oxygen removed from the water during its passage over the fish in relation to the rate of flow of water over the fish. From this figure it is clear that the respiration rate of the fish is influenced by activity or excitement; oxygen consumption in the active state is times the resting level.

The effect of water flow on the respiration of a fish in the respirometer tube was examined. The results obtained (e.g. Fig. 4) indicate that with increasing flow rates there is a rise in oxygen consumption but no change in the ventilatory frequency. During these and other experiments which were carried out at the normal level of in the ambient water, a resting level of metabolism at 25°C was obtained of about 300 c.c. O₂/kg/h at a flow rate of 50 c.c./h and about 500 c.c. O₂/kg/h at flows of 150 c.c./h; the regularity of the ventilatory frequency was noticeable, often remaining constant at about 200/min in spite of changes in flow rate. Thus if the ventilation volume is approximately constant, in spite of changes in flow rate, the oxygen tension of the water inspired by the fish in the respirometer tube at low flow rates will tend to be much lower than at high flow rates. As the oxygen consumption of the fish decreases at low levels in the ambient water (Figs. 5, 9), so at low flow rates a decrease in oxygen consumption would also be expected. Such an increase in oxygen consumption with increase in water flow at normal is also found when the ambient of the water decreases or increases (Fig. 5). Moreover, constancy of the ventilatory frequency at different flow rates is observed, at both low and high ⁠, although the particular frequency increases or decreases with the decrease or increase in ambient (Fig. 8).

With increasing flow rates the proportion of oxygen removed from the water during its passage over the fish correspondingly changes (e.g. Fig. 6, at 25°C). At low flow rates it is high, but with average flow rates (70−100 c.c./h) it is about 20%, falling at flow rates of 120−150 c.c./h, and it seems to rise again at high flow rates (more than 150 c.c./h). In the present experiment such an increase in percentage of oxygen removed from the water at high flow rates does seem to be significant, and is pronounced at lower levels, as is shown in Fig. 5.

The expected relationship between increase in standard metabolic rate with increasing water temperature was obtained (Fig. 7). This figure also indicates that the rate is affected by the water flow, and shows a relatively steeper rise in the higher ranges. It therefore seems that the increase in resting with rising temperature becomes more pronounced as the rate of water flow increases.

On the other hand, the proportion of oxygen removed from the water during its passage over the fish is also influenced by changes in water temperature, and with increasing temperature there was a rise in this percentage (Fig. 7).

Experiments were carried out in which the fish was maintained in water of different partial pressures of oxygen for periods of about 30 min under a constant rate of flow of water over the fish, in the range of 50−200 c.c./h. Changes in the oxygen content of the water passed over the fish were made by bubbling nitrogen or oxygen through the exchange column. No observations were made on the effect of varying the rate of the change. The results (Fig. 8) showed that with decreasing oxygen there was a rise in both the ventilatory frequency and the proportion of oxygen removed from water during its passage over the fish, but a fall in oxygen consumption. However, this percentage of oxygen removal fell markedly once the water was below about 40 mmHg. Conversely, with hyperoxia the ventilatory frequency and the percentage of oxygen removal decreased but the oxygen consumption of the fish increased markedly, as has been found by Hughes & Umezawa (1968a).

In some experiments the changes in were sudden and brief. The results of a typical experiment measuring the effect of lowering the are shown in Fig. 9, in which following a rapid fall in of the inspired water the the water after it had passed over the fish also fell, but not proportionately. Consequently there was a rise in proportion of oxygen removed from water, as mentioned already. However, following a return to air-saturated water, there was not a fall, but a rise followed by subsequent fall, in the proportion of oxygen removed from the water. The transient increase in this percentage of oxygen removal following a brief period of hypoxia seems to be significant.

The most obvious respiratory response of the fish to lowering the of the ambient water was an increase in ventilatory frequency. The fish was found to breathe deeply in a preliminary observation when the change in of the water was fairly slow and it reached a level of about 40 mmHg. During the recovery period ventilation rapidly returned to its normal frequency, whereas the percentage of oxygen removal mentioned above was calculated to increase very much, though it was transient Unfortunately, no determination of stroke volume was made in this series of experiments because of technical difficulties. In the absence of any measurements one can only surmise that the rise in the percentage of oxygen removal during a period of hypoxia was achieved by the increase in ventilatory frequency and that its transient rise following hypoxia or during recovery resulted partially from the fall in frequency. The present data showing that there is a relationship between ambient during hypoxia and both percentage of oxygen removal and ventilatory frequency is not the same as that found for the dragonet, Callionymus (Hughes & Umezawa, 1968b). In Callionymus there was a fall both in percentage utilization and in ventilatory frequency ; the calculated stroke volume increased (percentage utilization was measured for Callionymus but not for killifish).

The oxygen consumption of the fish decreases quickly with a rapid fall in ambient as is to be expected, then increases more or less gradually during a period of hypoxia. Following the renewed breathing of air-saturated water there was a marked overshoot which suggested that the fish was making up an oxygen debt incurred during the hypoxic period, as was pointed out previously (Hughes & Umezawa, 1968b). The observation that the fish consume oxygen extremely rapidly during recovery, in spite of the decrease in ventilatory frequency, seems to indicate that there was an increase in percentage utilization.

In the present experiment measurements were made of oxygen consumption, proportion of oxygen removed from water during its passage over the fish, and ven-tilatory frequency in relation to the changes in light intensity which occurred when fish in the respirometer tube were illuminated by electric light after adaptation to daylight. At first, after the introduction of the fish, readings of oxygen consumption show the high initial rate which gradually subsides, as usual. An example relating photo-stimulus to respiration is given in Fig. 10. The response of the gills was a marked bradypnoea in darkness, the respiratory rate falling quite rapidly after an initial momentary increase and then maintaining a steady level, whereas in light it was a tachypnoea, although some fluctuation was observed at the beginning. The changes between light and darkness also became the effective stimulus producing changes in both the oxygen consumption and the percentage of oxygen removal, in which they show a marked rise at the onset of both the light and darkness. So far as the present experiments carried out in the daytime are concerned, it seems that initial responses of light-adapted fish to dark stimulation are more pronounced than those of dark-adapted fish to continue light. However, after these responses the fish remained very quiet and breathed regularly for prolonged periods when kept in darkness, because no fluctuations in oxygen consumption, proportion of oxygen removed from water or ventilatory frequency were found. At the onset of darkness there was a phase of increasing oxygen consumption, although there was a fall in ventilatory frequency. The former seems to be due to the increased proportion of oxygen removed from inspired water by the fish.

From preliminary attempts to determine the effect of an anaesthetic for fish (MS 222) on survival of the killifish it was found that an adequate concentration in water was 1:1000. When the fish was immersed in 0·1 % MS 222 solution, its respiratory movements were usually interrupted for about 1−2 min after treatment at 20°C.

Even if the fish was kept immersed in the solution or suspended in air for at least 3 min after anaesthesia had begun, it could recover from the anaesthetic whenever it was put back into fresh water or its gills were ventilated with fresh water from a syringe.

Fig. 11 shows a typical result obtained with a fish kept in a cylindrical tube filled with running water with or without MS 222 during the experiment. At the beginning of anaesthesia the fish showed signs of a momentary tachypnoea and simultaneously high rates of oxygen consumption. However, the ventilatory frequency decreased quickly with progressive anaesthesia. Consequently there was a fall in oxygen consumption which reached zero levels during anaesthesia. When the fish showed signs of apnoea fresh water was run over it once more, about 3 min after the anaesthesia had begun. The fish renewed its respiratory movements for about 14 min after the supply of fresh water. These movements were intermittent and very poor at the beginning of the recovery period (Fig. 12). During recovery there was a marked increase in oxygen consumption which declined rapidly again, and this sort of pattern tended to be repeated. Such overshoot in the oxygen consumption suggests that an oxygen debt incurred during the period of apnoea is compensated by an active respiratory response. It is to be expected that this raised oxygen consumption during recovery was achieved largely as a result of the increase in ventilation volume and hence in percentage of oxygen removal. When the fish was recovering fairly well from the anaesthesia a return to the original level could often be obtained in the oxygen consumption, but not in the ventilatory frequency.

Two main methods have been used to measure directly the volume pumped by fish over their gills (Hughes & Umezawa, 1968 a, b). In one of these the external openings of the gill slits were covered by funnelled rubber tubes attached by means of a special cement. The second method consisted in separating a chamber containing spiracles and mouth of the dogfish from the external gill slits by means of rubber membranes.

Attempts to determine ventilation volume of the killifish were made using the second method, as shown in Fig. 13. The thin rubber membrane was placed across the opening of the cylindrical tube and fixed tightly with the opening of the other external cylindrical tube which contained a fish covered with a moistened soft sponge except its head region. Then the tip of the mouth was pushed forward through a hole in the membrane so that the mouth was in the inlet of water, but the external gill slits were in the outlet. In some instances this arrangement was quite satisfactory, but sometimes the fish tended to move and struggle during the course of prolonged measurements. The struggling seems to be due to constriction in the vicinity of the mouth, probably owing to the hole in the rubber membrane being too small.

When the apparatus was kept horizontal the difference in hydrostatic pressure between the water entering the mouth and that outside the external gill opening was about +5 mm H₂O and ventilation volumes of about 0·2 c.c./min (range 0·1−0·3 c.c./ min) were measured for fish in the size range 200−300 mg at 25°C ; stroke volumes in the range of 0·001−0·003 cc. were calculated for these specimens (Fig. 14). This experimental method made possible the determination of ventilation volume. No attempts to measure either oxygen consumption or percentage utilization were made.

Although it was possible to alter the imposed pressure difference by inclining the experimental tube to the horizontal, it was found that many specimens could not pump water out of the opercular cavity against a pressure difference.

Measurements of gill dimensions were made with the purpose of estimating the total area of the respiratory surface. The gills were dissected from the fish and measurements were made immediately on the fresh gills of one side under a microscope using a micrometer eyepiece.

Preliminary observations showed that the length of the gill filaments is not so constant, filaments in the middle region of each gill arch being longer, and that the surface area of an average secondary lamella of the long filaments was greater than that of the short filaments. However, it was found that variations in spacing of the secondary lamellae were small on each size of filament in each gill arch in this fish, as already reported for many fishes by Hughes (1966).

The dimensions in the killifish are shown in Fig. 15, which indicates diagrammati-cally a small part of the sieve of two gill filaments. Estimation of the surface area of one side of an individual secondary lamella as equal to bl was made from the dimensions measured, and hence the total surface area of the secondary lamella equals bl.

To determine the length of filaments, measurements were performed separately on two sizes of filament from each gill arch by means of a sampling method. The four arches on one side were dissected out, the filaments of each size (short and long) were counted (remembering that they form a double row splayed out) and their total length (L_s and L_i) in each arch was determined. These values were doubled to take account of the arches of the other side. From the spacing (d ′) of the secondary lamellae and the filament length (L_s and L_i) of each size in each arch, their total number may be obtained as equal to 2 L_s/d ′ or 2 L_i/d ′, because the secondary lamellae are found on both sides of a filament.

Gill areas and the dimensions that were measured are given in Tables 1 and 2. From these results it can be seen that both the length of filaments and the surface area of secondary lamellae of filaments differ somewhat between different gill arches. Each value for the first and second gill arch is larger than that for the third and fourth arch. It is also found that the total area of the secondary lamellae increases with increase in body weight. For example, the value in the specimen of 179 mg was 236 mm² and that in the specimen of 100 mg was 170 mm². Thus, it is estimated that the gill area per gram of body weight in the small specimens (e.g. 1700 mm²/g) is larger than that in the large specimens (e.g. 1318 mm²/g). The average value of the gill area per gram of body weight in the specimen used was about 1400 mm²/g. The trend obtained here coincides with that obtained by Price (1931) for Micropterus. Recently, Muir & Hughes (1969) and Hughes (1970) have found the same relationship between the body weight and total area.

The number of secondary lamellae per millimetre of gill filament was 43 · 5. This value is rather high as compared with values obtained for active fish (e.g. mackerel with 39) by Hughes (1966). Another important factor, total length of all filaments, was about 250 mm; and the surface area of a secondary lamella was about 0-0084 mm² in fishes with an average body weight of 140 mg.

A killifish was kept on its side in the hollow of a sponge sheet nearly equal to the size of the fish, over which water was passed. The sponge was covered with a transparent board of acrylic resin, in which many holes were made to form a line, and electrodes of platinum, active and indifferent, were placed in the medium surrounding the fish through two of the holes (Fig. 16). The indifferent electrode was placed near the tail tip and the active electrode was placed in a desired position.

Slow potentials were recorded which were related to the rhythm of the respiratory movements, but no spike potentials similar to those identified as an ECG by Iwata et al. (1969) were observed in the present experiments. By shifting the active electrode tailwards along the body it was found that slow potentials which were rather large at the mouth region decreased beside the eye, and then reached a maximum value in a region just posterior to the gill opening and decreased again at the tail region. A typical record obtained is shown in Fig. 17. Fig. 18 shows the relationship between the height of the slow potentials and the position of the active electrode along the body. The highest value of the slow potential was about 900 μ V in this case, but reached 1·1 mV in the other determination. Records of the potential in Fig. 17 also show the change in polarity of the slow potentials, being reversed at the position anterior to the gill opening (Fig. 17 f) as has been pointed out by Iwata et al. (1969).

As mentioned earlier, a number of factors determine the level of oxygen consumption of fish. It is relatively easy to maintain the external conditions constant but more difficult to keep the fish themselves in a constant state. The physiological conditions of the fish therefore cannot be controlled and it is preferable for such experiments to use fish domesticated in some way in order to ascertain the role of different external factors influencing their respiration.

Small size and ease of collection and breeding make Oryzias suitable for this type of study. It is not quite ideal, however, because of its very small size and the consequent difficulties in measuring parameters such as ventilation volume.

Some of the results obtained here indicate that the metabolic rate of the fish is affected by handling, as pointed out by Fry (1957). Soon after the introduction of the fish into the apparatus, its oxygen consumption shows surprisingly high rates which only gradually subside, and in general the initial value of oxygen consumption is double the resting value. In the present experiments it was necessary to wait, usually for about 15 min at least, before measuring the resting rate of oxygen consumption.

Oxygen consumption increases with increase of water flow over fish, whereas the ventilatory frequency is nearly constant and independent of the flow rate, as was found for the dogfish by Hughes & Umezawa (1968a).

Why does the oxygen consumption increase with increase in the rate of water flow over the fish? The most probable interpretation is that the volume of water passing over the gills increases, although no information has been obtained in the present work on the ventilation volume in relation to flow rate of water over fish. In Oryzias the ventilatory frequency remained fairly constant and independent of flow rates; thus with increasing flow rates there is probably a rise in stroke volume.

Now, when the fish is at rest in still water, it is supposed that the amount of water expired by the fish depends entirely upon the activity of the respiratory pumps. In flowing water, however, the increase in stroke volume results not only from a normal active effort, but also from a passive effect by which water freely passes through the gills, since the percentage of oxygen removed from the water during its passage over the fish decreases with increase in the flow rate in some ranges. However, one interesting observation is that this percentage of oxygen removed increases with higher flow rates notwithstanding the rise in oxygen consumption. It is therefore thought that the volume of water passed through the gills does not increase proportionately to the rate of water flow round the fish. Such a further increase in percentage of oxygen removal was not found in the dogfish (Hughes & Umezawa, 1968 a).

There is no direct evidence that the augmentation of activities by flowing water is a result of its mechanical effect upon the configuration of the fish. However, it might be expected from the results obtained in Callionymus (Hughes & Umezawa, 1968b), in which changes in minute volume were produced by altering the hydrostatic pressure across the respiratory system, that gill response would be influenced by flow. Consequently, when the fish is subjected to an increased water flow it is supposed that oxygen consumption increases because of an increased activity caused by water flow. Thus it is surmised that the increased percentage of oxygen removed by fish exposed to higher rates of water flow compensates for the higher rates of oxygen consumption resulting from their extra activity. This may be supported by the fact that the trend of increase in the percentage of oxygen removal with higher flow rates is so marked at low of ambient water.

The results described in the present paper give evidence that the normal resting oxygen consumption of Oryzias latipes is in the range 300−500 c.c./kg/h at 25°C and in the range 250−400 c.c./kg/h at 20 °C with average flow rates (50−150 c.c./h).

Experiments in which the of the water which was passed over the fish at a constant rate was changed show a consistent change in oxygen consumption with change in oxygen tension produced by bubbling of pure nitrogen or oxygen. The linear relationship between in ambient water and oxygen consumption is not necessarily similar to that which was found by Hughes & Umezawa (1968a) for dogfish. An increase in ventilatory frequency and percentage of oxygen removal was observed when the fish was subjected to a lowered oxygen tension. A 1·4-fold increase of frequency and a 3-fold increase in percentage of oxygen removal was observed as water fell from 158 to 40 mmHg, but the percentage removal fell once the was below about 40 mmHg. These results, except the fall in percentage removal at below 40 mmHg, agree fairly well with those obtained by Hughes & Umezawa (1968 a) for the dogfish, but contrast with their observations (1968b) on Callionymus, in which there were also marked changes in frequency and percentage utilization at low ⁠, but in the opposite direction. These differences are probably related to differences in habit of each species and also in the experimental method. Callionymus normally rests on the sea bottom for long periods and has a convenient anatomical arrangement for measuring the ventilation volume directly.

A marked rise in oxygen consumption together with a slight fall in both frequency and percentage of oxygen removal was observed when the inspired was raised. No information was obtained in the present work on the relationship between oxygen content of the inspired water and that of blood. Clearly with increasing ambient ⁠, however, there is a possibility that an increase in gradient of dissolved oxygen across the water/blood interface may occur, which leads to an increase in oxygen uptake at higher levels. It is thus surmised that the fish exposed in higher levels of oxygen tension may be in hyperoxia.

According to Fry (1957) ‘fish often show endogenous daily cycles of activity, so that the level of metabolism may depend on the time of day in spite of the maintenance of constant conditions’. It seems advisable, therefore, that any series of measurements should be made at certain fixed times of the day. When measurements are carried out in the daytime, it may be possible to test the effect of darkness, but not of light, without disregarding endogeneous activity cycles, and vice versa at night. Accordingly, it does not seem possible to determine the metabolic rate for the purpose of an extreme comparison between light-adaptation and dark-adaptation in the fish. However, it may be suggested, from the present experiments which were carried out in the afternoon, that ventilatory frequency decreases during darkness, whereas it increases in the light. Both oxygen consumption and percentage of oxygen removal showed temporary high rates which only gradually subsided when a light-stimulus or a dark-stimulus was applied to fish. For more exact determination of standard metabolism it might be necessary for fish to be kept in constant darkness throughout the experiment.

The experiments carried out in this study give evidence that an adequate concentration of the anaesthetic MS 222 is about 0·1 % for the killifish. This value is higher than the concentrations of the drug in water recommended by the Merck Index (1968) for eels (1:5000), for goldfish and brown trout (1:2000) and for Fundulus (1:3500). At the beginning of anaesthesia the fish shows signs of a momentary tachy-pnoea and simultaneously high rates of oxygen consumption. However, respiratory movements decrease rapidly with progressive anaesthesia and are usually interrupted for about 1−2 min after treatment at 20°C. During recovery from anaesthesia there is a marked increase in oxygen consumption which only gradually returns to its original level. Such overshoot may be due to the paying-off of oxygen debt incurred during the period of apnoea. This raised oxygen consumption may be achieved largely as a result of an increased minute volume and not by increase in the ventilatory frequency, since the original frequency was not attained, even when the fish had recovered fairly well from the MS 222.

Determination of ventilation volume of the killifish, by a method which is similar to that used by Hughes & Umezawa (1968a) for the dogfish, gives values in the range of 0·1−0·3 c.c./min for fish in the size range 200−300 mg, when hydrostatic pressure between water entering the mouth and that outside the external gill opening differs by about + 5 mm H₂O. The value of the ventilation volume measured in the fish is 800 c.c./min/kg at 25 ° C, when positive hydrostatic pressure is about 5 mm of water and this is much higher than the average figure given for different species. For example, it is 120 c.c./min/kg for the dogfish and 300 c.c./min/kg for the dragonet (Hughes & Umezawa, 1968a, b) and 443 c.c./min/kg for dogfish (Millen et al. 1966). It is thought that the ventilation volume per gram body weight depends on size of the fish.

According to Iwata, Watanabe & Matsuoka (1969), by whom the potential changes from the water surrounding the body of fish were detected by external electrodes, the slow potentials recorded simultaneously with the mechanogram of opercular movements are related to the respiratory movements and the polarity of the potentials is reversed just anterior to the gill opening. The present experiments on the killifish have confirmed their observations. In the present data, however, the figures obtained for the slow potential at the position just posterior to the gill openings are 900 − 1100 μ V, and these values are twice as large as the values recorded by them at the position just anterior to the gill opening for the eel. Iwata et al. (1969, and personal communication to S.-I.U., 1970) have also found that the spike potentials which are identified as ECG appear in some kinds of fish. But no spike potentials are recorded from the killifish by means of external electrodes.

1. Factors influencing respiration in this small surface-living freshwater fish (100-290 mg body weight) have been investigated using through-flow respirometers.

2. When the rate of flow over the resting fish is increased, there is an increase in oxygen consumption, but no marked change in ventilatory frequency. Percentage of oxygen removed from the water during its passage over the fish decreases with increase in flow rate in some ranges, and seems to increase again with even greater flow rates notwithstanding a rise in oxygen consumption.

3. The standard oxygen consumption is in the range 250 − 400 c.c./kg/h at 20 ° C, and 300 − 500 c.c./kg/h at 25 ° C with average flow rates (50 − 150 c.c./h).

4. Oxygen consumption of the fish is directly related to ambient over the range 40 − 500 mmHg. Increases both in ventilatory frequency and in percentage of oxygen removed from the water during its passage over the fish occur during hypoxia, but decrease with increases in ambient although O₂ uptake increases.

5. Ventilatory frequency decreases during darkness and increases in light. After stimulation with light both oxygen consumption and percentage of oxygen removed showed temporary high rates which only gradually subsided.

6. Ventilation volume, measured directly by collecting the water which had passed over the gills, is about 800 c.c./min/kg at 25 ° C, when the positive hydrostatic pressure is about 5 mm of water.

7. Determinations of gill areas of the fish gave values in the range of 12111700 mm²/g weight. The larger values in the fish are probably related to its greater oxygen consumption.

8. Slow potentials associated with respiratory movements were recorded around the head region of the fish by means of external platinum electrodes. When the active electrode was shifted tailwards along the body, the size of the potentials changed, being maximal just posterior to the opercular opening and rather large beside the mouth. The higher values obtained in the potential were 900 − 1100 μ V. The polarity of the potential was reversed at a position anterior to the gill opening.

The writers wish to express deepest appreciation to Professor G. M. Hughes, University of Bristol, for his invaluable suggestions and encouragement and for the improvement of the manuscript.