A Comparative Study of Gill Ventilation in Marine Teleosts

Relatively little work has been done on the gill ventilation of British marine teleost fishes since that of McKendrick (1879) who made one of the earliest comparative studies of fish respiratory movements. He was surprised by the variability, not only in the frequency but also in the shape of the wave forms recorded by tambours placed lightly against the operculum of different species of fresh-water and marine fishes. Apart from noting these variations he did not attempt any classification of the different mechanisms. Baglioni (1907) was the first to do this when he divided marine fishes into four main groups according to their habits of life and drew attention to the accompanying variations in their respiratory mechanisms. These variations are largely in the degree of development of the branchiostegal apparatus which co-operates with the operculum. The first group he distinguished consists of pelagic fishes which never rest on the bottom. In these the operculum itself is well developed and the branchiostegal apparatus relatively small. The second group is intermediate between these forms and the third group which includes the true bottom-living fishes. In the second group he included such families as the Cottidae, Gobiidae and Blenniidae and pointed out that their respiratory movements are characterized by the greater role of the branchiostegal apparatus. The importance of this apparatus becomes even greater in the third group, some of which frequently burrow in the mud or sand of the sea floor, e.g. Trachinidae Pleuronectidae. In the fourth group, however, a true branchiostegal apparatus is absent. This group is not a homogenous one and it includes such families as the Muraenidae and Syn-gnathidae. Willem (1927, 1947) has suggested a similar classification but he recognizes six different groups. The two additional ones are mainly due to his detailed studies of the peculiar mechanisms of the plectognaths. Most recently Bertin (1958) has largely combined these two systems but recognizes a further group to include the interesting adaptations found in forms that live in torrential streams (Hora, 1933).

In a study of the respiratory mechanisms of three fresh-water fishes Hughes & Shelton (1957, 1958) have substantiated the view (Woskoboinikoff & Balabai, 1936, 1937; van Dam, 1938; Henschel, 1939) that the respiratory current is maintained by the action of two pumps, one which forces water through the gills (the buccal force pump) and another (the opercular suction pump) which draws water through this sieve. Diagrammatically such a mechanism can be represented as in Fig. 1 where it can be seen that the possible variables of such a system are at least five in number. The relative sizes of the two pumps can be varied as can the resistance of the gill sieve through which the water must pass. Furthermore, the size and nature of the mouth and the opercular opening and of their valves are also obvious variables. The classification of Baglioni was largely made before such a double pumping mechanism was envisaged and Willem (1940) has categorically denied the importance of the suction pump as described by Woskoboinikoff (1932). The present study was therefore undertaken to see whether all teleost fishes would fit into Baglioni’s scheme and, if so, whether a study of the pressure relations would form a quantitative basis for his classification.

The choice of fish used was largely determined by their availability in good condition for experimentation. In this respect shore-living and bottom-living species are best, as complications due to swim-bladder expansion following capture are absent or very small, but some specimens of pelagic species were also obtained. The species used were: horse mackerel (Trachurus trachurus (L.)), herring (Clupea harengus L.), whiting (Gadus merlangos L.), five-bearded rockling (Onos mustela (L.)), wrasse (Crenilabrus melops (L.)), bullhead (Cottus bubalis Euphrasén), butterfly blenny (Blenmus ocellaris L.), grey gurnard (Trigla gurnardus L.), dragonet (Callionymus lyra L.), plaice (Pleuronectos platessa L.), merry sole (Microstomus kitt (Walbaum)), conger eel (Conger conger (L.)), great pipefish (Syngnathus acus L.). I am indebted to the staff of the Marine Biological Laboratory, Plymouth, for the trouble they took in collecting this material.

The fish usually weighed 40–100 g. and were kept in the sea-water circulation of the laboratory. The methods used were essentially similar to those described by Hughes & Shelton (1958) but were improved in a few details. The anaesthetic M.S. 222 (Sandoz) was used as an alternative to urethane and for some species was preferred, although no significant difference could be found between recordings obtained from animals under the different anaesthetics. Most fish were held in modified versions of the clamp used previously. The fish was first anaesthetized in a more concentrated solution of the anaesthetic and after it had been fixed in the clamp it was placed in a large tank containing 25 l. of constantly aerated sea water having a much lower concentration (one-tenth) of the anaesthetic. In the case of urethane this was about 0·1 % and for M.S. 222 it was usually 1 in 50,000 or 1 in 75,000. Under these conditions the specimens continued breathing regularly while at rest and were left for about $12$ hr. before records were taken of pressure and movement. In some cases they remained in this condition for more than 3 hr. without any apparent distress.

Two Hansen condenser manometers, one utilizing a modified high-frequency circuit (Machin, 1958), were used to record the pressure in the buccal and opercular cavities simultaneously, and long levers attached to RCA 5734 mechano-transducer valves were placed lightly against lower jaw and operculum to record their movements. These four channels were fed into an Ediswan pen recorder and any two of them could be recorded simultaneously on a Cossor double-beam oscilloscope. Photographic records from the latter (Fig. 5) were used for more detailed analysis of the wave forms and also served to check any distortions due to the curvature of the pen records. The frequency response of the pen recorder (flat from 0 to 90 cyc./sec.) was conveniently the same as that of the electromanometers. In several instances ciné films were taken of the unanaesthetized animal and the respiratory current was observed by means of dyes such as methylene blue.

Rather than give a detailed description of the respiration of each species studied the author has preferred to divide them into four main groups according to Baglioni’s classification; the observations made on members of each group are considered together.

Group I. The fishes in this group were the most difficult to study mainly because for them the conditions of the experiments are inevitably abnormal, the animal being stationary relative to the water, which is very rarely the case in life. Furthermore, of all fishes pelagic species are the most difficult to obtain in good condition for experimental purposes. The results described below, however, are based upon specimens most of which were in excellent condition.

Four-channel recordings from herring and whiting (Fig. 2), and from horse mackerel, wrasse and rockling (Fig. 3) are reproduced. The frequency of the movements ranges from 30/min. in the whiting to between 60 and 120/min. for the other four species. In the whiting the movements of the mouth precede those of the operculum by about one-tenth of a cycle. The form and frequency of the movement records are very similar to those recorded by McKendrick (1879) using unanaesthetized animals. The pressure curves (Fig. 5) show that both the buccal and opercular pumps are present and perhaps the latter is slightly more active than the buccal pump in producing the flow of water across the gills. This is in agreement with the fact that the phase of expansion of the two cavities is about three times longer than the closing phase. In the case of the herring these two phases are more or less equal and similarly the excess positive buccal pressure is more or less equal to the excess negative opercular pressure during the opercular expansion, suggesting that both pumps are about equally developed. The particular specimen from which this record was made was in first-class condition.

In the horse mackerel, however, although many more specimens of this species were investigated than of the others, it was less certain that the animals were in good condition. In some cases (Fig. 3 C) the peculiarity was noticed that the operculum tended to remain in the abducted condition with the opercular valve open and was only briefly adducted shortly after the mouth had closed. The frequency of these movements was also low (10–15/min.). In such cases the pressure curves showed large positive pressures in both cavities, the buccal pressure always exceeding the opercular pressure, whereas the negative pressures recorded were relatively small. At first it was thought that these were completely abnormal specimens especially as more normal curves were later obtained—see Fig. 3 D. It can be seen in these recordings that the opercular pressure curve has a very distinct negative phase, although the buccal pressure curve still has only the sharp positive peak which occurs during the closing of the mouth. However, it is possible that the first type of record obtained is not entirely abnormal but is representative of the movements which occur during active swimming. In this case the maintained abduction of the operculum with the mouth open at the same time would ensure a continuous flow of water across the gills as a result of the animal’s forward progression. Attention was drawn to this type of ventilation by Willem (1947) in his first group which also comprised nektonic forms. He noted that each inspiration was followed by a relatively long pause during which the mouth and opercular clefts remained open. Direct observation of another carangid, the California yellowtail (Seriola dorsalis) at the Marineland Aquarium, Los Angeles, has confirmed this description.

Another species investigated which will be included in this group is the wrasse (Crenilabrus melops), in which the pressure curves (Figs. 3 A and 5) are very similar to those obtained from the trout (Hughes & Shelton, 1957, 1958). The frequency of movements is about 60/min., also as in the trout. The mouth begins to open about one-sixth of a cycle before the operculum abducts but closes only one-twelfth of a cycle earlier. In the pressure curves there is a well-defined transition with reversal of the differential pressure.

The differential pressure curves indicate that the two pumps are fairly well balanced in the wrasse (Fig. 4 B) and herring. The buccal pump may predominate in the horse mackerel (Fig. 4 A) and rockling but the opercular pump plays the more important role in the whiting (Fig. 4 C). Some differential curves also showed that there was frequently no reversal of the pressure gradient in the horse mackerel and five-bearded rockling.

Group II. In the families which Baglioni included in this group the following species were studied: Cottus bubalis, Blennius ocellaris, Trigla gurnardus, and Callionymus lyra. These are largely bottom-living fishes in which the branchiostegal apparatus is very much better developed than in the preceding group of pelagic fishes. Many of the species included in this group were found to be very amenable to investigation. Details of their respiratory movements vary but a fairly definite series can be recognized which shows the increasing predominance of the opercular suction pump as the main mechanism in producing the respiratory current (Fig. 4). In some forms such as Trigla and Callionymus the exhalant current is directed dorsally from the operculum, an adaptation which will produce least disturbance of the sand or mud of their benthic habitats. It is difficult to measure the relative volumes of the buccal and opercular cavities, but there can be no doubt that the opercular cavity is relatively much larger in most of the fish of this group. The frequency of the breathing movements is generally low, especially in forms such as Cottus and Callionymus. In the latter the frequency was 12–20/min. Blennius ocellaris has a more rapid rhythm of 70/min. and the positive pressure in the buccal cavity is here relatively greater than in the other species. This animal has well-marked maxillary and mandibular valves which are very effective as is shown by the marked difference in the pressure curve recorded just outside and just inside these valves (Fig. 6B). The mouth movements recorded from this species were smaller and often more complicated than in other forms, but it is possible that this is an artifact due to the position of the transducer lever. Cottus (Fig. 6C, D) also has maxillary and mandibular valves but here the positive pressure in the buccal cavity is accompanied by a relatively large positive pressure in the opercular cavity and hence the differential pressure during this phase of the cycle is not so marked as in Blennius. During the phase of opercular expansion the negative pressure in the opercular cavity is much greater than that in the buccal cavity.

In discussing the breathing of Cottus, Willem (1940) considered that the respiratory current was maintained solely by the action of the branchiostegal apparatus and that neither buccal nor opercular movements played an active role in the pumping mechanism. Such buccal movements as were discernible he considered to be entirely passive and due to the influence of pressure changes in the buccal cavity produced as a result of those in the opercular cavity. The present observations are not entirely in agreement with this description, but there can be no doubt that the branchiostegal apparatus plays a very important role in the production of pressure changes in the opercular cavity. In Cottus then, the opercular suction pump certainly plays a predominant part in gill ventilation, whereas in Blennius the buccal pump is still about equally important Fig. 9 A. The mechanism of Trigla appears to be similar to that of Blennius in that the positive pressure in the buccal cavity exceeds that in the opercular cavity quite markedly. In Callionymus, however, the opercular pump is even more important than in Cottus and for this reason it was studied in greater detail.

Willem (1947) is the only previous author to have noted the extremely interesting mechanism found in Callionymus and he drew attention to the lengthening of the inspiratory phase and to its complex nature. The latter is shown in the movement records of the operculum which, as in many other cases, is rapid to begin with but continues more gradually during the phase of sustained negative pressure. He also noted the reduction in gill area found in this species which has been confirmed quantitatively (Hughes, unpublished). Some films were taken whilst specimens breathed water containing indian ink and it was found that water entered the mouth for about two-thirds of the cycle but was ejected dorsally through the restricted opening of the contracting opercular cavity during only one-fifth of a cycle. This observation fits in with recordings of the movements and pressures as shown in Fig. 7. It can be seen that there is a positive pressure in both cavities for about one-fifth of a cycle, and that the pressure is only slightly greater in the buccal cavity than in the opercular cavity. This suggests that during this phase a complete emptying of the whole system occurs and that subsequently there is first an expansion of the buccal cavity and then a gradual expansion of the large opercular cavities accompanied by a sustained negative pressure which draws water through the gills. The differential pressure curve (Fig. 4E) indicates this very marked predominance of the opercular suction pump as a mechanism for drawing water through the gills. Both the buccal and opercular pumps seem to play an important role in the ejection of water from the whole system. It is possible that during the phase of ejection the gill resistance is very low so that the two cavities function almost as a single one. It was frequently found that fish of this species showed periodic departures from their normal pattern, similar to the’coughing’ found in the Cyprinidae. Thus in Fig. 7B it can be seen that during such departures the operculum and mouth expand more than normally after which the opercular cavity suddenly decreases in volume to produce a very large positive pressure which exceeds that in the buccal cavity, suggesting that there is a reversal in the flow through the gills during the’cough’.

Group III. According to Baglioni (1907) fishes in which the branchiostegal apparatus plays the most prominent role during breathing are found in the families Scorpaenidae, Trachinidae and the order Pediculati (Lophiiformes).These are true benthic forms as also is the family Pleuronectidae. As the latter is the only one of these four families which it has so far been possible to investigate, it will be discussed here ; Baglioni classified the Pleuronectidae in Group II on the basis of their type of breathing but in Group III on the basis of their habitat.

These flatfish were not easy animals to study because of the difficulties of restraining them. The most effective method found was to fasten the fish to a brick with a light bandage so that only the head projected ; by placing the fish in a vertical position it was possible to record from both opercula simultaneously. Alternatively, the fish can be placed in its normal orientation, but then only the buccal and dorsal opercular pressures can be recorded. There are wide discrepancies in the descriptions of the breathing of flatfish under normal conditions, but apart from Henschel’s (1941) work few studies have been made specifically of their respiratory adaptations. Some observers maintain that water only leaves from the lower operculum and they point out the adaptive significance of this arrangement because if it left from the upper operculum disturbances of the sand would destroy the camouflage so characteristic of many species. Other workers assert the opposite. Schmidt (1915), for instance, pointed out that normal respiration would be a difficult matter for a fish lying on its side on the bottom. Schmidt’s paper is in Russian and Norman (1934) gives the following quotation:’Not only would considerable force be required to raise the operculum of the blind side, but the action of the exhalant current of water would tend to lift the body of the fish from the bottom; further, the danger of clogging the delicate gill-lamellae with particles of sand or mud which might enter the lower branchial chamber would be a very real one.’ Schmidt therefore considered that only the upper operculum is functional when the animal is at rest. His work was based entirely on preserved specimens and provides a valuable comparative account of the different types of opercular valves found in flat fish, although some of his conclusions regarding their function have been disputed by Henschel (1941).

Orcutt (1950) has described how when at rest a flounder is supported by the arching of its dorsal and ventral fins in such a way that it is held slightly away from the substratum. Plaice also do this and have often been observed using both opérenla for breathing when at rest in the Plymouth aquarium. When observed in a glass-bottomed tank, dye drawn in at the mouth can readily be seen to be pumped out from both opercular cavities. As far as can be judged by this method, these observations suggest that although a greater volume emerges from the upper opercular cavity the difference is not substantial. The gill area is identical on the two sides (Hughes, unpublished) and the opercula are almost completely symmetrical, but Schmidt noticed that there are differences in the detailed structure of the valves on the two sides.

The animals certainly used both opercula under the conditions of the experiments discussed here. When movement and pressure were recorded from both opercula almost identical records were obtained from the two sides (Fig. 8 B). If anything the pressures were a little less on the blind side. The rhythm is about 30/min. (Pleuronectes) or 60/min. (Microstomus). Adduction occupies one-quarter to one-third of a cycle when the pressure is positive with respect to the outside. During the longer abduction movements the pressure remains negative while water is drawn through the gills. This negative phase is particularly well developed in Microstomus and in contrast the positive phase predominates in the buccal cavity and occurs when the mouth is closed. Closing of the mouth precedes adduction of the opercula by about one-tenth of a cycle and the buccal pressure becomes positive before the opercular pressure. It is significant, however, that the opercular pressure begins to fall before the buccal pressure (Fig. 5). In a typical teleost such as the trout the opposite is true and the buccal pressure also becomes negative before the opercular pressure. This is the transitional phase (4) when the differential pressure shows a briet reversal in most fishes, but in both the species of flatfish investigated no such reversal has been found (Fig. 9B, C). Clearly, the opercular suction pump predominates although the buccal pressure pump plays quite an important role especially in Microstomus.

The apparent absence of a reversal in the pressure gradient may well be associated with the existence of an active mechanism for closing the opercular cavity. Such a system would allow an earlier fall in pressure than is possible in one requiring the development of a negative pressure before the valve can close. The detailed working of this mechanism is beyond the scope of the present work and needs more detailed investigation especially in view of Schmidt’s (1915) comparative account of differences in the valve structure among flat fish. The adaptive value of such a mechanism is obvious, however, for as in the skate (Hughes, 1960) it will restrict the entry of particles into the opercular cavity even if the fish is completely buried.

Group IV. Included here are a variety of families which have lost a true branchiostegal apparatus. Baglioni (1907) recognized this as a heterogeneous group in which there is no common breathing mechanism. This is well illustrated by the two specimens investigated in the present work; in one, Conger conger, the buccal force pump is the more important part of the mechanism whereas in the other, Syngnathus acus, ventilation of the gills is largely achieved by the action of the opercular suction pump.

The recordings from the conger eel clearly show much larger pressure changes in the buccal cavity than in the opercular cavity (Figs. 5, 10B). This was unexpected at first as the opercular cavity is greatly extended posteriorly and appears to play quite an active part during breathing, peristaltic waves being seen to pass along it. However, perhaps it is better to look upon this extension as a greatly enlarged opercular valve and not as a part of the active pumping mechanism of the operculum. Mechanical recordings (Fig. 10 A) showed that closing of the mouth precedes adduction of the operculum by one-sixteenth of a cycle in a fish in which the breathing frequency was 30/min. Adduction is not simultaneous along the whole length of the opercular cavity as both films and recordings show that a wave of contraction passes backwards. The positive pressure recorded in the buccal cavity is much larger than that in the opercular cavity during these movements. During the expansion phase, however, abduction first occurs in a region about two-thirds along the opercular cavity. This region changes shape most rapidly especially during the change-over from adduction to abduction (Fig. 10A). Expansion of the buccal cavity probably takes place before the earliest abduction of the operculum but this is not always obvious from the movement records. The fall in pressure in the buccal cavity is quite large immediately the mouth opens, but soon the opercular cavity becomes more negative than the buccal cavity and water is drawn across the gills.

The differential pressure curve (Fig. 9 E) clearly shows that the buccal force pump is the main mechanism causing water to flow over the gills and that the phase of the opercular suction pump, although longer than that of the buccal force pump, is of relatively little importance. The presence of the backward extension of the opercular cavity serves to damp out any pressure changes in it and also, as was pointed out by Willem (1947), will diminish variations in the initial speed of the respiratory current and so produce little disturbances of the medium. Such a mechanism probably is adaptive in a form with reduced paired fins ; in other fishes the paired fins are often used to back-paddle when the fish is stationary in order to counteract the slight propulsive effect of the exhalant stream from the operculum. The only occasion on which relatively large negative pressures were recorded from the opercular cavity was when the fish was firmly biting the recording needle in its mouth and hence movements of the latter were smaller than normal. It is therefore possible that the opercular suction pump becomes of greater importance during feeding than is indicated by the records of Figs. 5, 10B. Another variation was the occasional’coughing’ behaviour which occurred about once every twenty cycles. As shown during the first cycle of Fig. 10B the operculum is held wide open while the buccal cavity decreases in volume, but it then adducts quite rapidly. During these movements there are relatively small pressure changes, but after a slight expansion the buccal cavity decreases in volume still further and its pressure rises sharply. The operculum also has an extra abduction and then it closes before expanding gradually. The negative pressure in the buccal cavity is slightly greater than normal but that in the opercular cavity remains unchanged. During the succeeding respiratory cycle there is a prolongation of the positive phase in both cavities.

In contrast to the predominance of the buccal pressure pump in the conger eel is the predominance of the opercular suction pump in the pipefish. Even with quite a large specimen it was impossible to hold the animal stationary and to record both movement and pressure without excessively interfering with normal respiration. It was not possible to record movements but the differential pressure curve derived from the pressure records show quite clearly the predominance of the opercular suction pump (Fig. 9 D). This is of interest for, as mentioned above, there is no branchiostegal apparatus in this group. The opercular cavity, however, is relatively enlarged because of the well-developed operculum and it is closed except for a small dorsal aperture through which the water is ejected. Very small pressure changes were recorded from the mouth, as might be expected because of the relatively slight movements of the lower jaw. The narrow mouth together with the reduced opercular pressure produces a rapid flow into the mouth which may be used in feeding.

Leiner (1938) has pointed out the importance of the active opercular suction in the breathing and also in the feeding of the sea-horse, another member of the Syngnathidae.

This comparative survey of gill ventilation in marine teleosts has confirmed the results obtained previously in three fresh-water species (Hughes & Shelton, 1958). The opercular and buccal pressure curves have essentially the same form in all the species so far investigated and demonstrate the general validity of the view that water passes across the gills by the action of a buccal pressure pump in front and of opercular suction pumps behind the gills. The results reported above show that differences in the relative importance of the pumps occur in various fishes although they all utilize this same basic mechanism. It has been found that the differential pressure curve gives a good indication of these differences, despite the possible influence of variations in gill resistance which may occur during a respiratory cycle (Hughes, 1960). The pressure curves show the overall effect of the pumping mechanism, however, and give little information about the relative importance of the constituent parts of the pumps. The action of the suction pump, for instance, depends upon the co-operation of movements of opercular bones and of the branchiostegal apparatus, both of which produce a reduction in the pressure recorded from the opercular cavity. Thus while in general it is found that the suction pump becomes more important in species with an increased development of the branchiostegal rays this is not necessarily the case because the opercular bones may be correspondingly reduced. Nor is the converse true, as is well illustrated by the pipefish in which the suction pump is very well developed but the branchiostegal apparatus is absent.

It is clear, however, that the degree of development of the opercular suction pump is one of the main variables between fish of different habitats. This finding correlates very well with the classifications of fish respiratory movements (Baglioni, 1907 ; Willem, 1947; Bertin, 1958) that have been based almost entirely on morphological features. The correlation between the habitat of the animals and the physiology of their ventilation mechanism has now been demonstrated and indicates further problems to be studied. The adaptive value of an increased development of the suction pump in benthic forms is fairly clear when it is remembered that most of the time they live in water which is almost stationary. Selection has favoured the evolution of a mechanism which is well adapted to drawing a current across the gills during a relatively long part of the respiratory cycle. Such mechanisms are admirably adapted to ensure a steady flow across the gills without creating any disturbance of the muddy or sandy bottom. On the other hand, the buccal pump operates with the opercular pump during a very brief period to produce a rapid evacuation of most of the respiratory system. Thus in this part of the cycle the de-oxygenated water is ejected at high velocity often through a narrow, dorsal and posteriorly directed opening. The adaptations of sponges to produce such a jet and so decrease fouling of their own water have been discussed by Bidder (1923) and his concept of a’diameter of supply’might well be applied here. At the other extreme it is not surprising that fishes which lead an active life and mainly swim upstream in flowing water should not possess such adaptations.

As stated in the introduction, most previous classifications of teleost respiratory movements are fundamentally the same, but there are certain details in which they differ and some of the present work is of interest in this connexion. For instance, both Willem and Bertin classify Callionymus and eels in the same group, whereas we have seen that the former is an excellent example of a fish with a dominant suction pump, whereas at least in the conger eel the opposite is true. Baglioni’s system seems best in this respect although it involves the creation of a heterogeneous group (IV) of species with reduced branchiostegal apparatus. Willem also creates a special group including some nektonic forms like the mackerel and some salmonids and cyprinoids which swim continuously and make few definite respiratory movements but vary the flow by changing the size of the mouth opening. This distinction is a fine one and is of the same order as that between Baglioni’s Groups II and III. What is quite apparent from all attempts at classification is that there is a continuous spectrum of forms between types such as the mackerel through the trout and wrasse to the bullheads, dragonet and flatfish.

There are, however, several interesting features which the pressure curves have revealed and which previously have not been noted. On several occasions, for instance, it has been observed that the differential pressure curve does not include a transitional phase with a reversal in the gradient. This has been found several times in the plaice, merry sole and blenny, but also occasionally in the herring, rockling and horse mackerel. Its existence in the two flatfish is almost certainly correlated with an active branchiostegal valve mechanism and it is hoped that a detailed study of this will be of assistance in interpreting this phenomenon in other species. A further point of interest in the curves is provided by the variations found in the differential pressure in relation to the maximum absolute pressure change in one of the cavities. Thus during the buccal pump phase of the dragonet the maximum buccal positive pressure is as great as the maximum negative opercular pressure during the opercular pump phase. The differential pressure, however, is very much less during the former phase, because the pressure in the opercular cavity is also quite large and the two curves are very close to one another. The converse is true in many specimens of tench (Hughes & Shelton, 1958) where the two pressure curves are almost identical during their negative phases. Such observations suggest that variations in gill resistance occur during the respiratory cycle and that their nature depends on the species of fish. It is clear from this discussion that more detailed information is required about the properties not only of this resistance but also of the other two resistances shown in Fig. 1, i.e. those of the mouth and opercular openings and their valves.

I wish to thank the Director and staff of the Plymouth Laboratory for their assistance and for the facilities they provided during the course of this work.