Recent palaeontological evidence (Jarvik, 1967; White, 1966) indicates that the Dipnoi should be regarded as a specialized group of fishes, which have been structurally different from the Rhipidistia (now thought to have included the prototetrapods) for as long as the fossil record can yet show. However, the homology of the dipnoan lung with that of the tetrapod, and the similarities of organization and development between living Dipnoi and Amphibia, might indicate a closer relationship between the two groups than can be revealed by palaeontological studies. The level of organization seen in the living Dipnoi is not thought to be basically different from that of the Devonian forms and a study of the physiology of these animals must expand our knowledge of the development of vertebrate physiological mechanisms. Since the Dipnoi are adapted to a habitat very similar to that in which the first terrestrial verte-brates are thought to have evolved, and since they possess functional aerial and aquatic respiratory systems, a study of respiratory function in these animals may help to elucidate the manner in which the ancestral tetrapods became able to colonize the terrestrial habitat.

Recently our knowledge of the mechanical basis of the respiratory pumps in fishes has been very greatly increased by the introduction of modern techniques which allow the accurate measurement of respiratory parameters with the minimum of disturbance of the fish and the minimum distortion of the parameter under study. (Hughes & Shelton, 1958; Hughes 1960a, b; Saunders, 1961 ; Hughes & Ballintijn, 1965). Ventilation of the gills in teleosts (Hughes & Shelton, 1958; Hughes, 1960 b) and in elasmo-branchs (Hughes, 1960 a) has been shown to be achieved by the interaction of two pumps: a buccal pressure-pump forcing, and an opercular suction-pump drawing, water over the gills. That such a mechanism is basic in the vertebrates is indicated by the presence of an essentially similar mechanism in the cyclostomes (Roberts, 1950). The involvement of a buccal force-pump in both air-breathing and water-breathing in the amphibia has been demonstrated by Willem (1920 a, b, 1929) and the existence of a buccal force-pump component has been described by Grigg (1965 a) and Johansen, Lenfant & Grigg (1967) in the breathing movements of the Australian lungfish Neo-ceratodus and by Jesse, Shub & Fishman (1968) in those of the African lungfish Protopterus sp. Bishop & Foxon (1968) have also demonstrated the importance of the buccal force-pump in both air-breathing and water-breathing in the South American lungfish Lepidosiren. The only other functional analysis of the mechanics of dipnoan respiration was that of Dubois (1892) on air-breathing in the aestivating Protopterus. Szarski (1962) discusses possible mechanisms of ventilation in the Crossopterygii but quotes no experimental work. No complete analysis of the breathing mechanisms has been published for any dipnoan and this paucity of information leaves a serious gap in our knowledge of the development of respiratory mechanisms in the vertebrates.

In the present study the mechanisms of aerial and aquatic respiratory movements in Protopterus aethiopicus are described and compared, and the development of a lung-ventilation mechanism from the branchial irrigation mechanism of the water-breathing vertebrates is discussed. To avoid confusion in describing the movements of air and water through the buccal area of air-breathing fishes, the movement of water across the gills of these fishes will be termed ‘irrigation’ (Saunders, 1961) instead of the commonly used ‘ventilation’ (Hughes & Shelton, 1958). In this account the term ‘ventilation’ will be used only to describe the movement of air into and out of the lungs.

Four main lines of approach were used :

  • (A) Anatomical studies of the head and branchial region.

  • (B) Analysis of X-ray cinematograph studies of the animal, during the respiratory cycles.

  • (C) Analysis of simultaneous recordings of the pressures developed in opercular, buccal and intra-pulmonary cavities.

  • (D) Analysis of the electromyograms, recorded from muscles active in the repiratory cycles.

The animals used in this study were collected from the margins of Lake Victoria in Uganda and transported to England by scheduled air services. They were maintained in well-aerated water at a temperature of 25 ± 1° C. The animals were fed on a diet of ox heart or liver, plus an occasional live fish when these were available. Animals were starved for 1–2 days before any operative or experimental procedures were carried out. The animals were identified as Protopterus aethiopicus from the data presented in Sterba (1962) and in Greenwood (1958).

Two methods were used to determine the gross structure of the head and branchial region. Dissection revealed the position and attachments of the major muscular, ligamentous and skeletal elements, and this information was supplemented by the study of thick sections cut transversely and longitudinally (using a circular saw) from the bodies of deep-frozen animals.

The pressures developed in buccal and opercular cavities were recorded by a method adapted from Saunders (1961). A flexible plastic cannula (Portex PP. 160) was introduced into the buccal cavity by way of a hole drilled through the posterior snout region using a dental drill and handpiece. The size of the burr was chosen to produce a hole that fitted the cannula tubing closely. The hole was drilled from a point mid-way between the snout and eyes and passed downward through the skull entering the buccal cavity immediately posterior to the upper tooth-plate. The cannula was passed through the hole, out through the open mouth and a flange was formed on the proximal end. This end of the cannula was then drawn back into the buccal cavity until the flange fitted tightly against the buccal roof. A short length of close-fitting rubber tubing was slid down the cannula until it fitted closely against the outside of the head, thus anchoring the cannula in position. The opercular cannula was introduced through the roof of the opercular cavity by means of a large-bore hypodermic needle which was passed in through the tissues of the opercular roof and out through the opercular flap. The distal end of the cannula was forced into the bore of the needle and, as the needle was carefully withdrawn, was carried out through the opercular roof. The proximal end of this cannula had already been flanged and fitted with a fine rubber dam-washer. These were drawn up tight against the tissues of the opercular roof and secured in position with a length of rubber tubing, as for the buccal cannula. Cannulae inserted in this manner were well tolerated by the animals and remained patent and leak-free for a number of days at least, and in one case a patent buccal cannula was retained for over 6 months.

It was planned to record the pressures from a number of different sites in each cavity, but in fact the choice of sites was strictly limited. Satisfactory placement of the opercular cannula was only possible via the dorsal cavity wall. Implantation laterally through the operculum or through the opercular valve caused distortion of the opercular wall and impeded respiratory movements and flows. Implantation through the ventral opercular region involved passing the cannula through muscle systems found to be active during the respiratory cycles. Possible placement of the buccal cannula was also found to be limited. Cannulation at the anterior end of the snout region introduced the cannula into the buccal area in front of the upper tooth-row, and thus in front of the buccal pressure-seal mechanism (discussed later), with the result that no pressure changes were recorded. Successful recordings were made from implantations entering the buccal cavity immediately posterior to the tooth-row but implantations a few millimetres further back often caused damage to the brain or associated structures.

In a limited number of experiments a cannula constructed from a No. 18 gauge hypodermic needle attached to length of flexible cannula tubing was inserted through the dorsal segmental muscle and into the anterior median sac of the lung. This method of cannulation was hazardous, however, due to the extremely vascular nature of the lung wall, and it was discontinued after a minimum number of repeatable results had been obtained (3 animals). Attempts were also made to cannulate the lung by means of a cannula passing in through the glottis and pneumatic duct. This procedure failed, however, for the cannula appeared to irritate the walls of the duct and over-production of mucus occurred blocking the duct, and eventually the cannula.

The insertion of all cannulae was performed under deep anaesthesia induced by immersion of the animal in a 0·5–10% solution of tricaine methane-sulphonate (M.S. 222. Sandoz). Anaesthesia and operation times were kept as short as possible, as the animals could not breathe during the anaesthesia and rapidly became anoxic. Changes in pressure were detected using either Sanborn 267 B or Devices pressure-transducer systems, and the outputs of the transducers were displayed and recorded using either a pen recorder or an oscilloscope and camera. During the recording of buccal and opercular pressures, cannulae and transducers were flushed and filled with freshly distilled water in order to minimize the risk of damping due to the formation of air bubbles in the system. During the recording of intra-puhnonary pressure both transducer and cannula were filled with moist air.

Electromyograms (EMG’s) of the muscles of the head and branchial region were recorded by a method modified from Hughes & Ballintijn (1965). Initially insulated pins were used as recording electrodes but as it was feared that these might cause mechanical damage to the muscle fibres moving about them, they were replaced by pairs of insulated stainless steel wires. Pairs of wires, cleaned and hooked at the tip, were threaded through a hypodermic needle. This needle was inserted through the skin to the desired site and was then withdrawn carefully, leaving the electrodes hooked into the muscle. The potentials recorded from the muscles were amplified and then displayed on the separate channels of four-trace or six-trace, chopped-beam oscillo-scopes. Normally one or more pressure recordings would be displayed with the EMG’s from a number of different muscles (Fig. 4). Electrodes implanted as above were well tolerated by the animals and did not appear to cause any mechanical damage or other impairment of the respiratory processes.

X-ray cinematograph studies were carried out using a Watson R 600 Generator and a Watson Windsor tube holder and stand. The X-ray image was received on a 9 in. Phillips image intensifier which was filmed by an Arriflex II BV 35 mm. cine camera. The animal was filmed under water in a chamber constructed from thin Perspex sheeting. The chamber was made just large enough to hold the fish so that the small size and the thinness of the walls helped to reduce the loss of contrast caused by the water and Perspex surrounding the animal. Films were taken of animals displaying both aquatic and aerial respiratory movements from lateral and antero-posterior aspects. Films were usually shot at 16 frames/sec. on Pan F film. Developed films were viewed on an X-ray film projector, and for the purpose of analysis and display selected frames were enlarged and printed on high-contrast paper. Various combinations of KV. and mA. were used to try to increase the contrast of the resulting films, but a high degree of contrast was in fact never really achieved because of the scattering of the X-ray beam by the plastic and water surrounding the animal. A rather higher degree of contrast was obtained in ‘still’ photographs taken on Ilflex film.

During the pressure and electromyographic recordings the animals were contained in a Perspex tank fitted with a movable partition which could be adjusted so as to contain the animal in as short a length of the tank as was desired. The animal was given sufficient room to manoeuvre its way to the surface to breathe, but not enough to enable it to swim forcefully. The tank was narrow but had sufficient room for the animal to turn without difficulty, for if the tank was too narrow the animal struggled violently to turn and removed cannulae and electrodes. The water in the tank was static but well aerated, and was maintained at about 25° C. in a thermostatically controlled room. Usually the animal remained in a resting condition, only moving to come to the surface to take air. Sudden movements on the part of the observer could, however, startle the fish into violent swimming in which cannulae and electrodes were lost. To minimize this danger either sudden movements were avoided and/or the experimental chamber was shielded with polystyrene sheeting, except for a small observation window.

(a) Anatomical studies

X-ray photographic and dissection techniques were used to determine the gross skeletal and muscular anatomy of the head and branchial region (Pl. 1 ; Text-fig. 1). A detailed anatomical description of the cranial muscles in Protopterus has been published elsewhere (Edgeworth, 1935) and will not be repeated here, but a short account of the positions and principal points of attachment of major muscular and skeletal elements is included with results of the EMG experiments below. Certain muscles, found to be active in the respiratory cycles, have not apparently been described previously. As the homologies of the musculature in the Dipnoi are uncertain these muscles will be referred to by terms indicating their position and attachments until they can be named with greater certainty than is possible at present. The musculature of the head region of Protopterus is well developed, perhaps as a consequence of the evolution of the powerful suctorial predation methods shown in this form. The majority of the muscles were found to be active in both feeding and respiratory movements but the level of electromyographic activity recorded during respiratory movements was usually less than during feeding. The nomenclature used in this account is taken from Edgeworth (1935).

Fig. 1.

Diagrammatic representations of head and branchial anatomy of Protopterus to show position and attachments of muscles and skeletal structure. A. lateral aspect. B. Sagittal section. C. ventral aspect.

Abbreviations: a.o.p., ant-orbital process; b.b., branchial bars; c., ceratohyal; c.c., chondrocranium; c.r., cranial rib; d.e., dermal ethmoid; f., pectoral fin ; f.p., fronto-parietal; i.v., inner wall of cranium; l.d.e., lateral dermal ethmoid; l.j., lower jaw; n.c., nasal capsule; o., opercular; p.g., pectoral girdle; p.pt., palato-pterygoid; s.o., sub-opercular; sq., squamosal; v., vertebral column; A.M., anterior muscle of cranial ribs; C.Hy., constrictor hyoideus muscle; G.Th., geniothoracicus muscle; I.M., intermandibularis muscle; L.Ma,, levatores mandibulae muscle, ant. ; L.Mp., levatores mandibulae muscle, post. ; P.M., Post muscle of cranial rib ; R.C., Rectus cervicis; 1–5, principal attachment points of the constrictor hyoideus.

Fig. 1.

Diagrammatic representations of head and branchial anatomy of Protopterus to show position and attachments of muscles and skeletal structure. A. lateral aspect. B. Sagittal section. C. ventral aspect.

Abbreviations: a.o.p., ant-orbital process; b.b., branchial bars; c., ceratohyal; c.c., chondrocranium; c.r., cranial rib; d.e., dermal ethmoid; f., pectoral fin ; f.p., fronto-parietal; i.v., inner wall of cranium; l.d.e., lateral dermal ethmoid; l.j., lower jaw; n.c., nasal capsule; o., opercular; p.g., pectoral girdle; p.pt., palato-pterygoid; s.o., sub-opercular; sq., squamosal; v., vertebral column; A.M., anterior muscle of cranial ribs; C.Hy., constrictor hyoideus muscle; G.Th., geniothoracicus muscle; I.M., intermandibularis muscle; L.Ma,, levatores mandibulae muscle, ant. ; L.Mp., levatores mandibulae muscle, post. ; P.M., Post muscle of cranial rib ; R.C., Rectus cervicis; 1–5, principal attachment points of the constrictor hyoideus.

The skull of Protopterus (Pl. 1 ; Text-fig. 1) was found to differ greatly in structure from that of the teleost and elasmobranch skulls described by Ballintijn & Hughes (1965) and Hughes & Ballintijn (1965). Compared with the primitive bony fishes much reduction and fusion of bone has occurred in Protopterus. This indicates that the skeletal couplings shown to be important in the respiratory mechanisms of the teleost and elasmobranch fishes by the above authors are likely to be of much less importance in Protopterus. This reduction of bone is particularly noticeable in the opercular region where the opercular bones are very small and the opercular wall is mostly composed of muscle and elastic tissues. The jaw suspension is autostylic, hyomandibular, maxillary and premaxillary bones having been lost. The lower jaw is short but well developed and articulates directly with the quadrate. The hyoid arch is reduced to a ceratohyal element only. This is a well developed arch of bone connected both to the cranium and to the dorsal flange of the lower jaw by means of ligaments. Ventrally it supports the muscular tongue.

Two further skeletal elements were found to be concerned in the production of the respiratory movements. The pectoral girdle is a partly ossified arch underlying the posterior buccal cavity floor and the opercular cavity. Muscles from the ceratohyal, the branchial apparatus and the cranial ribs are attached to its surface. The cranial ribs are large ossified structures, peculiar to the Dipnoi, which articulate with the occipital region of the skull and are well anchored in the surrounding muscular tissues.

The ribs of Protopterus are not well developed (Pl. 1) and appear to play no part in the respiratory cycles.

(b) X-ray cinematograph studies

Still radiographs of the living animal show very clearly the outline of the lung and the principal skeletal elements of the body. (Pl. 1). Though the contrast was not high, X-ray cine films showed clearly the movements of air and of the skeletal elements of the anterior region during the respiratory cycles.

During the aquatic respiratory cycles the movements of both the ceratohyal and pectoral girdle arches could clearly be seen. Both arches were displaced upward and then downward through the resting position during the course of one aquatic respiratory cycle. It was hoped that films taken from the antero-postero (dorso-ventral) axis would show movements of the opercular bones in a lateral plane, but sufficient contrast was not developed to show this movement clearly.

The principle events in an air-breathing cycle, as demonstrated by analysis of the X-ray films, are shown in Pl. 2. The animal was first seen to flex the anterior part of the body so that the snout was thrust above the surface (Pl. 2, 1–2). Animals in deeper water do not show this head flexure but rise straight up to the surface. At the surface the mouth was opened and the buccal cavity expanded by a posterior-ventral deflexion of the ceratohyal and pectoral girdle elements which caused a downward displacement of the buccal cavity floor thus drawing air in through the open mouth (Pl. 2, 2–4). Towards the end of this buccal cavity expansion phase the volume of the lung appeared to decrease considerably, indicating that pulmonary gas was passing from the lung and mixing with the atmospheric air in the buccal cavity at this time. (Pl. 2 3–5). At the end of this expiratory phase the animal closed the mouth and lowered the head below the surface. An antero-dorsal movement of the buccal floor skeletal elements was now seen, compressing the air contained in the buccal cavity and forcing it back, through the glottis and pneumatic duct, into the lungs (Pl. 2 5–7). Following this inspiratory phase the animal sank to the bottom of the container.

The movements of the skeletal elements of the buccal floor observed during the air-breathing cycle were similar to those seen in the aquatic cycles but were of greater amplitude and could be measured accurately. Text-fig. 2 shows the displacement of the skeletal elements of the head and branchial region plotted together with the degree of flexion of the anterior part of the body for comparison. The plots commence as the animal reached the surface and opened the mouth into the air. Initially all the elements moved together but the degree of movement of the pectoral girdle was greater and more prolonged, producing considerable expansion of posterior buccal cavity. Movements of the cranial ribs were seen closely following those of the pectoral girdle but their exact function was not clear. The movements of the buccal structures thus expanded the buccal cavity during the expiratory phase of lung ventilation. During inspiration the lower jaw moved upwards, closing the mouth. Following this an antero-dorsal movement of the ceratohyal was seen, raising the anterior part of the buccal cavity floor and forcing the contained air back to the posterior part of the buccal cavity. Antero-dorsal movement of the pectoral girdle followed almost immediately, forcing the contained air back into the lung. Inspiration in this animal was thus seen to involve a buccal force-pump mechanism.

Fig. 2.

Movements of the skeletal elements of the buccal floor and angle of head (movement to and from surface) during the latter part of an air-breathing cycle. As plotted from and X-ray cine film.

Fig. 2.

Movements of the skeletal elements of the buccal floor and angle of head (movement to and from surface) during the latter part of an air-breathing cycle. As plotted from and X-ray cine film.

Expiration was seen to occur with the mouth open and atmospheric air entering the buccal cavity. Some admixture of pulmonary and atmospheric air must then occur in the buccal cavity. However, the expulsion of gas is often forceful and an efficient tidal exchange during lung ventilation is indicated by the results of spirometry experiments (Jesse et al. 1968) and by the monitoring of pulmonary oxygen concentration (B. R. McMahon, 1969). In both cases 60–80% exchange of pulmonary gas is indicated at each ventilation.

(c) Pressure studies

The pressures developed in buccal, opercular and intrapulmonary cavities were monitored during both aquatic and aerial respiratory cycles. All recordings were made from non-anaesthetized, free-swimming animals. Considerable differences were seen between the pressure waveforms recorded during air-breathing and water-breathing (Text-fig. 3).

Fig. 3.

Recording of pressures developed in the buccal and opercular cavities during aerial and aquatic respiratory movements. A. Slow recording to show variation in amplitude of aquatic waveforms between air-breathing movements. B. Fast recording to show details of pressure waveforms from aerial and aquatic respiratory cycles.

B.P., buccal pressure; O.P., opercular pressure; Air, aerial respiratory cycle, I, calibration 10 mm. Hg. ; G., branchial respiratory cycle; L1, L2, L3, stages of air breathing cycle.

Fig. 3.

Recording of pressures developed in the buccal and opercular cavities during aerial and aquatic respiratory movements. A. Slow recording to show variation in amplitude of aquatic waveforms between air-breathing movements. B. Fast recording to show details of pressure waveforms from aerial and aquatic respiratory cycles.

B.P., buccal pressure; O.P., opercular pressure; Air, aerial respiratory cycle, I, calibration 10 mm. Hg. ; G., branchial respiratory cycle; L1, L2, L3, stages of air breathing cycle.

(i) Pressures recorded during aquatic {branchial irrigatory) movements

Little difference was seen in the basic form of the pressure waveform recorded from nine different animals. Considerable variation occurred, however, in the amplitude and frequency of the recorded waveforms even in the same animal during periods of apparent rest. Such fluctuations could be correlated with variation of external or physiological activity of the animal. Animals newly introduced into the experimental chamber or recently disturbed by experimental procedures showed branchial irrigatory movements of high frequency and amplitude. After a period of acclimatization both frequency and amplitude fell to a lower level, but would rise again if the animal were disturbed, or subjected to certain forms of respiratory stress. Variations of both frequency and amplitude of the recorded waveforms were also observed in acclimatized animals apparently at rest, and these could often be associated with stages of the airbreathing cycles (Text-fig. 3 A). Immediately after an air breath both frequency and amplitude of the waveforms were high but decreased steadily until just before the next air breath when they were seen to rise again.

During branchial irrigation, pressure changes were detected from the buccal and opercular cavities but never from the intra-pulmonary cavity. The waveforms recorded were basically similar to those recorded by Hughes & Shelton (1958) and Hughes (1960 a, b) for teleost and elasmobranch fishes in that interacting buccal and opercular pressure changes combined to produce a differential pressure across the gill resistance tending to force water over the gills throughout most of the cycle. While the amplitude of the pressure changes recorded from Protopterus was generally higher than that observed for other fishes, the opercular pressures were generally lower than those recorded from the buccal cavity. This suggests that the buccal force-pump is dominant in branchial irrigation in this animal. A similar condition is seen in the eel Conger conger (Hughes, 1960b).

(ii) Pressure waveforms recorded during aerial respiratory (lung ventilation) movements

The waveforms recorded during the air-breathing cycle were more complex than those of branchial irrigation. Three separate buccal and opercular waveforms coincided with a single depression and subsequent re-establishment of the intrapulmonary pressure, which was seen to remain positive throughout the rest of the cycle. Expiration was seen to occur between two of the waveforms, while inspiration occurred simultaneously with a large pressure increase in both opercular and buccal cavities (Text-fig. 5B).

In both branchial and lung-ventilation cycles the phases of the pressure waveforms recorded from the cavities could be related to specific events in the respiratory cycles, but as much additional information can be provided by the analysis of the results of the EMG experiments; the functional significance of the pressures will be discussed together with the results of the EMG studies later in this work.

(d) The analysis of activity in the respiratory muscles

The evidence from the X-ray and pressure studies outlined above indicates that the basic mechanism of both aerial and aquatic respiratory cycles involves movements of the skeletal elements of the head and branchial region resulting in the development of internal pressures which serve to irrigate the gills or to ventilate the lungs. Recordings have been made of the activity in the muscles of the head and branchial region, simultaneously with recordings of the pressures developed (Text-figs. 4 and 5), to elucidate the nature of the muscle system powering the respiratory pumps. The method of recording used here can give direct evidence of electrical activity only at the electrode site, and thus monitors only a very small fraction of the possible motor units involved in the activity of the whole muscle. For this reason it is imperative to examine the recordings taken from a large number of insertions, taken from different sites in the same muscle, in more than one animal, before conclusions can be drawn as to the activity of the muscle as a unit. The diagrammatic representation of the results seen in Text-figs. 5 A and B is compiled from recordings taken from a large number of insertions in six animals and indicates the period and relative strength of activity of the muscles concerned together with simultaneous recordings of respiratory pressures and events. It must be further noted here that these are electrical signals and are only indirectly related to the degree of contraction produced by the muscle. Other factors such as the size of the muscle involved and the degree to which it is affected by the action of other muscles of the system (and the state of skeletal couplings) may also have a considerable effect on the actual force transmitted by the contracting muscle.

Fig. 4.

Pressures developed in buccal, opercular and intrapulmonary cavities recorded with electromyographic activity of the head and branchial muscles during aerial and aquatic respiratory movements.

Abbreviations: D.P., differential pressure between buccal and opercular cavities; B.P., pressure in buccal cadty ; O.P., pressure in opercular cavity; I.P.P., intra-pulmonary pressure; g., aquatic respiratory movement (branchial irrigation); I. aerial respiratory movement (lung ventilation).

Activity in: (1) rectus cérvida muscle, (a) geniothoradcus muscle, (3) levator mandibulae muscles, (4) constrictor hyoideus muscle, (5) posterior muscle of cranial nb, (6) anterior muscle of cranial rib, (7) inter mandibularis muscle, (8) retractor mandibulae muscle, (9) retractor anguli oris muscle, I calibration = 10 mm. Hg.

Fig. 4.

Pressures developed in buccal, opercular and intrapulmonary cavities recorded with electromyographic activity of the head and branchial muscles during aerial and aquatic respiratory movements.

Abbreviations: D.P., differential pressure between buccal and opercular cavities; B.P., pressure in buccal cadty ; O.P., pressure in opercular cavity; I.P.P., intra-pulmonary pressure; g., aquatic respiratory movement (branchial irrigation); I. aerial respiratory movement (lung ventilation).

Activity in: (1) rectus cérvida muscle, (a) geniothoradcus muscle, (3) levator mandibulae muscles, (4) constrictor hyoideus muscle, (5) posterior muscle of cranial nb, (6) anterior muscle of cranial rib, (7) inter mandibularis muscle, (8) retractor mandibulae muscle, (9) retractor anguli oris muscle, I calibration = 10 mm. Hg.

Fig. 5.

Combined data from pressure, electromyograms and observation experiments to show the interrelation of muscular action and pressure development during the aquatic and aerial respiratory cycles.

A. Aquatic cycle. Pressures and electromyograms recorded during the aquatic cycles. P.P. intra-pulmonary pressure. O.P. opercular pressure. B.P. buccal pressure. D.P. differential pressure between buccal and opercular cavities.

Activity in: L..M. levator mandibulae muscle, R.A.O. retractor anguli oris muscle, G.Th. geniothoracicus muscle, I.M. intermandibularis muscle, A.M. antenor muscle of cranial rib, P.M. posterior muscle of cranial rib, C.Hy. constrictor hyoideus muscle.

B. Pressures and electromyograms recorded during the movements of an air-breathing cycle. Abbreviations as in A.

Fig. 5.

Combined data from pressure, electromyograms and observation experiments to show the interrelation of muscular action and pressure development during the aquatic and aerial respiratory cycles.

A. Aquatic cycle. Pressures and electromyograms recorded during the aquatic cycles. P.P. intra-pulmonary pressure. O.P. opercular pressure. B.P. buccal pressure. D.P. differential pressure between buccal and opercular cavities.

Activity in: L..M. levator mandibulae muscle, R.A.O. retractor anguli oris muscle, G.Th. geniothoracicus muscle, I.M. intermandibularis muscle, A.M. antenor muscle of cranial rib, P.M. posterior muscle of cranial rib, C.Hy. constrictor hyoideus muscle.

B. Pressures and electromyograms recorded during the movements of an air-breathing cycle. Abbreviations as in A.

Fig. 5B.

For legend see Fig. 5A.

Fig. 5B.

For legend see Fig. 5A.

The majority of the muscles of the head and branchial region were found to be active in the respiratory cycles. Some muscles, notably those of the branchial basket, were too small or too difficult of access to be investigated here. Whereas it is probable that they play a minor part in the fine control of the respiratory processes, it is contended here that they play no major role in the production of the respiratory movements. The muscles investigated are listed below and a short account given of their positions, principal points of attachment and the phase of the cycles in which they were active. The results are then incorporated into a functional analysis of the respiratory mechanisms of this animal.

The levatores mandibulae muscles

In development this is a single sheet of muscle which divides to form anterior and posterior sections. The levator mandibulae anterior is a very large muscle attached cranially to the ventral side of the lateral dermal ethmoid and the supraorbital process, and to the dorsal side of the fronto-parietal. Ventrally it is inserted into Meckel’s cartilage and into the posterior flange of the lower jaw. The posterior part is also attached to the posterior flange of the lower jaw and is inserted dorsally into the cranium and the squamosal. The position of these muscles is shown in diagrammatic form in Text-fig. 1A.

No difference in the activity periods of the anterior and posterior parts of the muscle could be detected; both were active in aerial and aquatic respiratory movements but at a fairly low level of activity. (Higher levels were routinely recorded when the animal ‘snapped’ during feeding or attack.) It was concluded that because of the very large size of these muscles a low level of activity was sufficient to meet the needs of the respiratory system. Activity in this muscle occurred prior to and during the development of pressure in the buccal cavity. The action of the muscles would appear to lift the lower jaw, closing the mouth and beginning to raise the ceratohyal and interconnected structures. By clamping the lower jaw tightly shut the action of this muscle facilitates the development of the buccal pressure-seal mechanism, and also profoundly affects the action of the buccal floor musculature. The action of this coupling will be discussed later.

The retractor anguli oris muscles

This muscle is inserted into the upper lip at the angle of the jaw and into the cranium dorsally. The upper lip overlaps the lower lip laterally producing a fold of skin as shown in Pl. 3. The activity of this muscle, which is concurrent with that of the levator mandibulae, is thought to tighten this fold of skin, thus preventing any opening of the mouth and effecting a pressure seal laterally along the buccal cavity.

The retractor mandibulae muscles

Formed in development from the constrictor hyoideus muscle sheet, this muscle in the adult is inserted via a ligament on to the posterior ventral part of the lower jaw. Dorsally it lies between the opercular bones and is connected to them. In this region it overlies the constrictor hyoideus muscle sheet and abuts on to this muscle in the region of the opercular wall (Text-fig. 1A). Activity in the retractor mandibulae muscle was always associated with maximal depression of the lower jaw as is seen in the ‘yawning movements’ discussed below. Such maximal opening of the mouth was rarely seen during the respiratory cycles but was occasionally seen, together with low-level activity in the retractor mandibulae muscle, in the expiratory phase of lung ventilation.

The intermandibularis muscles

These small muscles are inserted on the inner surface of the lower jaw and connect in the mid-line (Text-fig. 1C). They were active during the development of pressure in the buccal cavity and probably served both to tense the rami of the lower jaw and to limit the movement of the overlying ceratohyal and associated structures.

The geniothoracicus muscle

A single muscle connected by a broad ligament to the mid-ventral portion of the lower jaw and by loose ligamentous connexions to the ceratohyal. This muscle runs mid-ventrally to become inserted in the ventral segmental muscle in the pectoral region. (Text-fig. 1A-C). The muscle is also connected by ligaments to the overlying constrictor hyoideus muscle sheet. Activity in this muscle was concurrent with that in the rectus cervicis and intermandibularis muscles, and was associated with an increase or decrease of pressure in the buccal cavity, depending on the action of the levator mandibulae muscles. The action of this coupling mechanism will be discussed later.

The rectus cervicis complex

This is a broad mass of muscular tissue containing several transverse inscriptions which forms the basis of the tissues of the tongue and is connected to the ceratohyal anteriorly and the pectoral girdle posteriorly (Text-fig. 1B). The recorded period of activity was similar to that of the geniothoracicus muscle. This muscle complex was also affected by the action of the levator mandibulae muscles.

The constrictor hyoideus muscles

This is a broad muscle sheet surrounding the branchial region ventrolaterally, with attachments to almost all adjacent structures. Postero-laterally the muscle is attached to the opercular bones and forms the wall of the opercular cavity. Postero-ventrally the muscle is firmly attached to the skin along a line running ventrally from the opercular margin halfway to the ventral mid-line. From this point the muscle runs free over the ventral muscles until it meets and fuses with its partner from the other side. Anteriorly the muscle is loosely connected to the lower jaw by ligaments and is inserted along the ventral margin of the ceratohyal. Ventrally the muscle abuts on to the intermandibularis muscle. Some of the attachments of this muscle sheet are shown in Text-fig. 1 A and C. The action of the muscle is complex and occurs at different periods in aquatic and aerial cycles, and will be discussed more fully later.

Un-named muscles

(1) The anterior muscles of the cranial ribs

This is a block of muscle connecting the pectoral girdle with the cranial ribs on each side (Text-fig. i B), and forming part of the walls of the buccal cavity posteriorly and the opercular cavity ventro-laterally. Activity in this muscle was seen at different times in the aquatic and air-breathing cycles and was associated with both compression and expansion of the buccal cavity.

(2) The posterior muscle of the cranial rib

Though very little modified from the surrounding segmental muscle, this muscle which connects the cranial ribs with this segmental muscle (Text-fig. 1B) is always found to be active in the respiratory cycles. Its period of activity was similar to, but slightly preceding, that of the anterior cranial rib muscles.

Muscles of the glottis

These were described by Edgeworth (1935), but successful electrode implantation proved to be extremely difficult. One record, however, indicated the action of two muscle systems in the glottis region, one continually active and the other active only during ventilation of the lungs. This would suggest that the glottis was closed by muscular activity except at the moment of lung ventilation when it was opened actively by dilator muscles. No confirmation of this was obtained but the existence of dilator and constrictor muscle systems of the glottis were noted by Edgeworth (1935). Active closure of the glottis between ventilations was indicated by the occurrence of a maintained pressure of 10–15 cm. of water in the lung.

Segmental muscles

Though EMG’s were recorded from the flank muscles during swimming activity, no recordings were made which showed activity of dorsal or flank segmental muscle which could be associated with respiratory movements.

In order to simplify subsequent discussion it is possible to divide the muscles discussed into groups of muscles which were active at the same period of the cycle. These groups are :

  • Muscles involved in the closure of the lower jaw, including the levatores mandibulae (L.M.) and the retractor anguli oris (R.A.O.).

  • The muscles of the anterior buccal floor region, including the genio-thoracicus (G.Th), the rectis cervicis (R.C.) and the intermandibularis (I.M.).

  • The muscle of the posterior buccal floor region, including the anterior and posterior muscles of the cranial ribs (A.M. and P.M.).

  • The constrictor hyoideus muscle surrounding the buccal cavity ventrally and forming the muscular outer wall of the opercular cavity.

It is now possible to incorporate the results of the experiments in a functional account of the mechanisms involved in the aerial and aquatic respiratory cycles.

(1) The aquatic respiratory cycle

Simultaneous recordings of the pressures developed in pulmonary, opercular and buccal cavities, and the EMG recordings from the muscles active in the respiratory cycles are displayed in diagram form in Text-fig. 5 A. Differential pressures and the periods of mouth and opercular closure are also indicated.

Activity was first recorded from the jaw-closing muscles (L.M. and RAO). Contraction in these muscles raised the lower jaw, closing the mouth and pressure-sealing the buccal cavity. Pressure experiments indicated that the tip of the tongue fitted into a groove immediately posterior to the upper tooth-row as the mouth was closed, and provided a further pressure seal anteriorly. A slight increase in both buccal and opercular pressures resulted from these movements.

Further increase of pressure was now observed, associated with activity in the muscles of the anterior buccal floor region. From purely anatomical evidence the activity of these muscles should cause a depression of the skeletal elements of the buccal floor and thus expansion of the buccal cavity, but here the muscles were obviously active in the compression phase of the cycle. This apparent anomaly was resolved when the effects of the continued activity in the jaw-closing muscles was considered. These powerful muscles were holding the lower jaw tightly shut. Lower jaw, ceratohyal and pectoral girdle (the principal moving parts concerned with buccal floor movement) were all interconnected by muscular and ligamentous tissues, and therefore were fused in position by the immobility of the lower jaw. This coupling thus ensured that no depression of the buccal floor could occur at this time. Contraction in these muscles must therefore have tensed the buccal floor structures, causing the upward deflexion seen in the X-ray studies, and the observed buccal compression (Text-fig. 5 A, phase 2). Initially the pressure increased concurrently in both cavities but later pressure increase was confined to the buccal cavity (phase 3). This produced a differential pressure between buccal and opercular cavities, which could have caused a respiratory water flow. In fact the opercular wall was seen to be distended and the opercular flap forced open by a stream of exhalent water at this time.

As peak pressure was reached in the buccal cavity, activity ceased in the jaw-closing muscles. Activity could still be recorded from the buccal floor musculature (A.M. and P.M.). As the lower jaw was freed from the restraint of the jaw-closing muscles, activity in the buccal musculature could now draw down the buccal floor and expand the buccal cavity. The buccal pressure falls sharply at this time, and as the buccal floor is depressed the mouth opens slightly, causing water to enter the buccal cavity replacing that formerly passed over the gills. Activity is now recorded from the constrictor hyoideus muscle. Contraction in this muscle closed the opercular flap and may have limited the expansion of the buccal cavity. A slight increase in opercular cavity pressure was often associated with the closure of the operculum. At this time a considerable negative differential existed between the pressure in the buccal and opercular cavities. It is possible that this temporarily induced a reverse flow of water across the gills. This back flow was unlikely to be large, because of the closure of the opercular flap and the small size of the opercular cavity.

At the end of the cycle the pressure in the buccal cavity began to rise. No activity was seen in the musculature at this time and this movement was interpreted as being a passive return of the skeletal elements of the buccal floor to the resting position. As the pressure in the buccal cavity rose, a pressure decrease occurred in the opercular cavity due to the elastic recoil of the operculum at the end of active closure. A second, smaller, positive differential pressure pulse thus occurred at this time which could cause a second respiratory flow to occur.

Irrigation of the gills in Protopterus is thus seen to involve the action of two pumps operating on either side of the gills. The mechanism involved is similar to that described by Hughes and other authors for teleost and elasmobranch fishes. The basic pattern of ventilation in the dogfish, as described by Hughes & Ballintijn (1965), comprises three main phases : an active expiratory movement which causes distension of the parabranchial cavity ; a passive inspiratory movement ; and finally a more active inspiratory phase. A similar sequence is seen in Protopterus but no passive inspiratory stage is seen following expiration. Continued activity in the anterior branchial muscles and the release of the coupling effect are responsible for the initial movements of inspiration.

By reason of the differences in the anatomy of these two forms it is very difficult to compare the activity of specific muscles or muscular systems. However, the action of the adductor mandibulae of the dogfish and the levator mandibulae of the lungfish would appear to be similar, as would the action of the constrictores hyoideus and branchiales of the dogfish and the constrictor hyoideus of the lungfish. The action of the ventral (buccal floor) musculature appears to be less important in the dogfish than in Protopterus, but generally the muscle systems involved are similar in basic pattern. Perhaps the greatest differences were found in the degree of involvement of the pectoral girdle, which is immobile in the dogfish.

The mechanism is also basically similar to that described by Ballintijn & Hughes (1965) for the trout. Here, though the anatomical differences are greater, the pressure waveforms are still similar to those recorded from Protopterus but the method of their production differs in detail. The most important differences are the absence of pectoral girdle movement in the trout and the absence in Protopterus of the complex couplings between the skeletal elements of the jaw suspension and operculum. The divergent morphology of the two forms does in fact make comparison very difficult. It would be very interesting to record the mechanisms shown by anatomically (not phylogenetically) intermediate forms such as the eels (the recorded pressure waves of Conger conger (Hughes, 1960b) are rather similar to those recorded for Protopterus) in which both the branchiostegal apparatus and the opercular suction pump are reduced in functional importance, and such as the Australian lungfish Neoceratodus, where the reduction of the opercular and hyomandibular bones is not so great as in Protopterus. A study of the respiratory mechanisms of the former animal would enable us to decide how much of the pattern seen in Protopterus is typical of the Dipnoi, and how much is due to the rather specialized nature of this lungfish. The basic pattern of the respiratory movements is also very similar to that exhibited by the early larval stages of the toad (Willem, 1920a, b).

A second type of aquatic movement was occasionally seen in Protopterus. The mouth was opened maximally by activity in the buccal floor musculature and often in the retractor mandibulae muscles. The mouth was held open for periods of up to 1 min. and then allowed to close slowly. This movement was invariably followed by a typical branchial irrigatory movement. The movement was very similar to the yawning movements seen in many teleost fishes. In the teleost these movements have been demonstrated as being important in the depth-control mechanisms (McCutcheon, 1966), but since they were often seen in the resting lungfish in the absence of depth change their function in this animal remains obscure. If the response were to be respiratory (pulmonary) it would be of great interest, however, since this may have been the ‘primitive’ response which together with the air bladder in the teleost fishes has lost its respiratory function and become concerned with the regulation of depth.

(2) The mechanism involved in the air-breathing movements

Pressure changes and electromyograms have also been recorded during a large number of air-breathing cycles. These results are expressed in diagram form in Text-fig. 5 B. At first sight the cycle seems very complex, including at least three different stages. It is possible, however, to divide the cycle into three sub-cycles, each of which can be compared with a single aquatic cycle.

The first sub-cycle occurs as the animal approaches the surface prior to lung ventilation. Activity was first recorded in the jaw-closing muscles, closing the mouth and pressure-sealing the buccal cavity. This activity was maintained and modified as was the subsequent action of the anterior buccal floor musculature deflecting the buccal floor upwards, and compressing the water contained in the buccal cavity. Water was observed to be lost from the operculum at this time. Up to this stage no difference was observed between this and a typical aquatic movement, but the animal was now at the surface and the subsequent expansion phase occurred with the mouth opening into the air. As air entered the mouth, pressure decreased very rapidly in the buccal and opercular cavities and activity ceased in the majority of the muscles. A decreased level of activity in the rectus cervicis and posterior buccal floor musculature was sufficient to expand the buccal cavity as the incoming air offered little resistance to such expansion, thus no change in pressure was noted, though the expansion of the buccal cavity was almost maximal. Activity was also recorded from the constrictor hyoideus muscle at this time, closing the operculum and preventing the entry of water from below.

Towards the end of the expansion phase of this movement a decrease in pulmonary pressure was recorded. Pulmonary gas appeared to be held in the lungs under pressure by the action of the glottal sphincter muscles. These relaxed at this time allowing the compressed gas to escape. This expiration is almost certainly aided by the natural elasticity of the lung wall. Much smooth muscle has been described in the lungs of Protopterus sp. (Poll, 1962) and Neoceratodus (Grigg, 1965 a) but it was not possible in this study to show that active contraction of this muscle was associated with expiration. These records showed that expiration occurred during the expansion phase of the first sub-cycle at a time when the mouth was open. In this way the lung was indirectly open to the atmosphere during expiration.

When the buccal cavity was fully expanded and the pulmonary pressure reduced the next sub-cycle started immediately. No passive return of the skeletal elements of the buccal floor occurred between these two stages. Activity was immediately seen in the jaw-closing muscles and a little later in the anterior buccal floor muscles, initiating a buccal compression phase similar to that seen in aquatic cycles. However, concurrent activity was also seen in the posterior buccal floor muscles which were now influenced by the coupling effect of the jaw-closing muscles and this became effective in the compression phase. This modification probably helped to produce the higher pressures (up to 20 mm./Hg.) apparently needed in lung ventilation. The resultant pressure rise occurred in both cavities simultaneously and as the presence of air in the cavities often allowed the gill resistance to break down, a differential pressure between the two cavities was not always recorded.

Increased pulmonary pressure was also recorded at this time, indicating that the gas compressed in the buccal cavity was being forced into the lung. In some cases (as in Text-fig. 5 B) a single inspiratory effort was not sufficient and a second, more powerful, inspiratory stage followed the first. The stimulus which instigated a second inspiratory stage would appear to have been pulmonary pressure, for in animals where the pulmonary cavity was experimentally kept at atmospheric pressure repeated inspiratory efforts were seen and the animals remained at the surface; animals which were obliged to surface into mixtures containing no oxygen showed only a single inspiration and then submerged often for quite long periods. A second expiration in one surface visit was never seen.

Following the final inspiration the expansion phase of that movement occurred after the head was submerged. The water thus drawn into the buccal cavity was passed over the gills and lost from the operculum in a stage very similar to a typical irrigatory movement. The movement is always present and is always more forceful than the usual aquatic respiratory movements. It is thought that its function is to expel the water forcibly over the gills thus removing residual air from the system.

The inter-relation of aquatic and aerial respiratory mechanism

Though the pressure waveforms recorded during the aquatic and the aerial cycles were very different in appearance, because of the different properties of the contained water or gas respectively, examination of the EMG records and comparison with the events as recorded by X-ray photography showed that the aerial cycle consisted of a number of consecutive aquatic-type movements, each modified to serve a specific phase of lung ventilation.

Relatively little modification is needed. The compression phase of the first movement is similar to those of typical aquatic cycles but serves an important function in removing the water contained in the buccal cavity, which might otherwise enter the lungs and cause damage or drowning.

The subsequent expansion phase is prolonged, producing the maximal expansion of the buccal cavity necessary to encompass sufficient air to reinflate the lungs, and allowing expiration to occur while the mouth is open and the operculum closed. This facilitated tidal exchange and prevented water entering through the opercular flap during ventilation.

The compression phase of the second stage is modified to serve the function of lung inspiration. Essentially the air contained in the buccal cavity is compressed by a buccal force-pump mechanism and forced through the pneumatic duct and into the lungs. The principal modification seen here is an alteration of the periods of muscular activity so that muscles normally active only in the expansion phase of the cycle (A.M., P.M., C.Hy.) are now active earlier in the cycle, are affected by the lower jaw coupling mechanism, and are thus active in the compression phase. This modification produces greater and more efficient compression as all the buccal floor muscles act together. The action of the constrictor hyoideus is particularly important at this time, for the contraction of this muscle in the compression phase closes the opercular flap and prevents the loss of the highly compressed air through the operculum.

It is not known how these changes in the periods of muscular activity are effected. It is possible that as there is no passive return of the skeletal elements to a resting position before this inspiratory compression begins (a passive return was described in the aquatic cycles) and therefore the action of the powerful jaw-closing musculature acts on a maximally expanded system. The upward movement of the lower jaw would, under these conditions, cause a stretching and tension in the buccal floor muscles and could cause these to become active earlier than occurs in the aquatic cycle. This hypothesis would explain the difference between the aerial and aquatic compression phases, and would be a simple method of generating the high pressures needed for complete gas transfer and thus efficient lung ventilation.

The expansion phase of the final inspiratory movement occurs as the animal sinks below the surface. The water thus drawn into the mouth is forced over the gills, flushing out any residual air which might interfere with further branchial irrigation or with buoyancy or postural regulation. This movement is very little modified from the normal aquatic cycle but is intimately associated with the air-breathing sequence.

The mechanism of lung ventilation in Protopterus is derived from a series of basically aquatic-type cycles, each of which is modified to serve a specific function in the airbreathing cycle. Basically, inspiration corresponds to the buccal compression phase of the branchial cycle while expiration occurs passively in the expansion phase of the cycle. Essentially the mechanism is based on, and easily derived from, a buccalopercular-branchial pumping mechanism similar to that seen in all fishes (Hughes, 1960a,b; Hughes & Shelton, 1962) and in the amphibia (Willem, 1920a, b).

The evolution of the aerial respiratory mechanisms in the vertebrates

Of the many authors who have published observations on the form of the air breath in Protopterus, only one reports experimental work. Dubois (1892) examined the mechanism of air-breathing in aestivating Protopterus using a tambour fixed to the mouth and levers attached to the body wall at various points. His account has never been confirmed, but has provided the basis for a number of erroneous descriptions of the air-breathing mechanism found in this animal (Smith, 1931) and of schemes for the evolution of aerial respiratory mechanisms in vertebrates (Willem, 1929). By observation through the open mouth and by the use of levers Dubois recognized that three stages occurred in the lung-ventilation cycle. He also observed that the glottis was open during buccal expansion (just following the first compression of the airbreathing cycle in Text-fig. 5B) and assumed that this was inspiration. He concluded that as the mouth was open during inspiration then a buccal force-pump mechanism (‘déglutition’) could not occur and that inspiration must occur by an aspiratory mechanism involving the use of the ribs. Dubois admitted that active dilation of the body wall was an anatomically impossible solution but offered no real alternative. In this account it has been demonstrated that expiration, not inspiration, occurs at this time and thus Dubois’ assumptions are fundamentally incorrect. Willem, however (1929), accepted Dubois’ main points and concluded that inspiration occurred partly by the movements of the truncal ribs but mostly by the movements of the cranial ribs (which he equates with the abdominal movements of Dubois). In fact though movements of the cranial ribs are associated with the respiratory cycles (Text-fig. 2) the direction of their movement is the opposite of that needed to substantiate Dubois’ and Willems’ account. No movements of the truncal ribs have been observed in the X-ray studies carried out on Protopterus and none were observed in a careful radiographic study on the air-breathing movements of Lepidosiren (Bishop & Foxon, 1968). It is in any case doubtful as to whether the small movements possible in the reduced truncal ribs or in the cranial ribs could have more than a minor effect on pressure in the extensive lung of Protopterus or Lepidosiren.

Szarski (1962) proposes another theory of the development of vertebrate lungventilation mechanisms. He quotes Schmalhausen (1957, 1958, not read by author) as deducing that the primitive air-breathing fishes utilized the hydrostatic pressure of the surrounding water as the lung-ventilation mechanism. Here, as the fish approaches the surface at right angles, the differential hydrostatic pressure along its length is utilized to force the air from the lungs and out of the open mouth. The buccal cavity is then filled with air and the animal swims until the head is pointing downward ; the differential hydrostatic pressure acting along the body is now imagined as forcing the buccal air into the lung. While it is likely that the hydrostatic pressure could/have a minor beneficial effect on ventilation in a soft-bodied fish, it seems most unlikely that such a mechanism could have provided efficient ventilation in the rigid-bodied rhipidistian fishes. Schmalhausen, Szarski points out, considered that this mechanism would be useless in air, and virtually useless in shallow water. Szarski maintains that as this method was useless in air it was discontinued by the first amphibian animals which then relied on cutaneous exchange for both oxygen uptake and carbon dioxide excretion. However, the habitat of the primitive amphibians is thought to have been mostly in shallow, periodically hypoxic and hypercarbic water, with rare, essential excursions on to the land. Under these conditions neither cutaneous exchange, nor the lung ventilation method described by Szarski, would have been of use. It is more likely that the cutaneous gas-exchanger was evolved because the primitive lung, though Respiratory movements of an African lungfish extremely efficient as an oxygen exchanger, was inefficient as a carbon dioxide excretion mechanism (Rahn, 1966), and this function, which was performed by the gills in the air-breathing fish, was transferred to the skin as the gills lost function in the terrestrial habitat.

Finally it seems unnecessary to postulate the development of a new lung-ventilation mechanism when the existing buccal force-pump can be so easily modified to serve this purpose. Szarski argues that a mechanism by which the mouth and opercular cavity can be closed simultaneously is a prerequisite for the evolution of an aerial buccal force-pump mechanism such as is seen in the recent amphibia. Indeed this is so, but such a mechanism is demonstrated here in the lungfish and must be present in a great number of air-breathing teleosts. Szarski also considers that a ventilation mechanism where the atmospheric and pulmonary air are mixed in the buccal cavity would be very inefficient, yet it has been shown that in Protopterus lung ventilation is very efficient, as the tidal volume is 60-80% of the lung volume (Jesse et al. 1968; B. R. McMahon, 1969).

It is contended here that the primitive fresh-water fishes, in response to an occasionally hypoxic and hypercarbic environment, came to use the buccal force-pump, previously used only to power the branchial irrigatory flow, to pass oxygen-rich air first over the gills, secondly into respiratory pouches and finally into the developing lungs. The buccal force-pump, with only slight modification, could easily be used for this purpose. This system is similar to that in use by recent amphibia except that the nares are not used. It is suggested that this method is most likely to have been that used by the rhipidistian and prototetrapod forms. The use of the ribs, is shown not to be important in the lungfishes and it is unlikely that it was of use in the proto-amphibian types. The evolution of such an aspiratory mechanism is dependent on the possession of powerful and movable rib structures. These were probably developed at a later stage in evolution when the whole skeleton of the animal, including the ribs, was undergoing the massive evolutionary changes needed to provide a new skeletal structure which could more efficiently support the animal’s mass on land. Such increase in rib area is seen in the early amphibia, and it is possible that these animals had evolved the more efficient aspiratory ventilation mechanisms. We can then imagine that these may have become less important in the recent amphibia, together with adaptation to a more aquatic habitat and the reduction of the size of the ribs, but have progressively evolved in the more terrestrial vertebrates to eliminate the buccal pump completely.

If it is contended that the buccal force-pump mechanism seen in the Dipnoi is essentially similar to that seen in the rhipidistian fishes and Amphibia, it is necessary to comment on the use of the nostrils in respiration. These are used in respiration only in the Amphibia, even though internal nostrils are found in both groups. It is now generally recognized that the nares of the Dipnoi are not homologous with those of the rhipidistian tetrapod line (Thompson, 1964; Bertmar, 1965). Kerr (1932) thought that the evolution of the nostril was an aid to more efficient olfaction and not to respiration, and it is argued here that such olfactory devices were evolved separately in the dipnoan and rhipidistian lines. It is not known at what stage the nares became implicated in respiration, nor is it clear what advantage is gained by their use in an aquatic animal. One of the more plausible theories is that the use of the nostrils was of advantage in filling the buccal cavity with air in an air-breathing but primarily aquatic animal. The filling of the buccal cavity with air must have presented problems to the early crossopterygian fishes for in these forms the jaws were extended and the gape very long. In these animals the head must have had to be raised high out of the water for the buccal cavity to be filled with air. The use of the nostrils would have allowed buccal filling with only the tip of the snout protruding above the surface. In the Dipnoi the gape of the jaws is much reduced, and buccal filling is further facilitated by a fold of skin which acts as a pressure-seal in the compression phases but which considerably reduces the gape when the mouth is opened to take air at the surface. (Pl. 3). It is possible that these two different solutions to the problem developed very early in the evolution of the air-breathing forms and are preserved today in the surviving dipnoan and amphibian forms. The use of the buccal force-pump has similarly been preserved in both lines. Though the presence of a hinge in the skull of the rhipidistian fishes is indicative of major structural differences separating them from the Dipnoi, it is unlikely to have seriously affected the respiratory mechanisms. Thompson (1967) has investigated the kinetics of possible movements between the anterior and posterior parts of the skull but is unable to make firm conclusions as to its function. In the opinion of this author the limited degree of movement possible was unlikely to have been responsible for the production of the respiratory movements, but the increased flexibility of the skull may have added to the efficiency of the buccal/ opercular pumping mechanism.

In conclusion, the lung-ventilation mechanism in Protopterus has been shown to incorporate a buccal force-pump mechanism basically similar to that observed both in other fishes and in amphibians. The mechanism has been shown to consist of a series of aquatic-type movements, each of which is slightly modified to serve a specific function in the air-breathing cycle. Little modification is really needed. Expiration occurs in the expansion phase of one movement and inspiration follows in the compression phase of the next. No movement of the ribs has been observed in this or other lungfishes. It is concluded that the air-breathing mechanism exhibited by the earliest tetrapods is likely to have been basically similar to that exhibited by the modem Dipnoi, and that the aspiratory method of inspiration observed in the fully terrestrial vertebrates was evolved later together with the increase in structural importance of the ribs.

  1. The anatomy of the head and branchial region of Protopterus has been studied by dissection and section techniques to show the relation between skeletal and muscular elements. X-ray cinematographic, pressure and electromyographic techniques have been used to show how the muscular and skeletal systems interact to produce the respiratory movements. The mechanisms involved in aquatic and aerial respiration in Protopterus have thus been elucidated.

  2. The mechanisms of branchial irrigation has been shown to be basically similar to that seen in teleost and elasmobranch fishes, and also similar to that seen in larval amphibia.

  3. The aerial cycle is composed of a series of aquatic-type cycles, each of which is modified slightly to serve a specific function in the aerial cycle. Inspiration occurs by a buccal force-pump mechanism. Expiration occurs by the release of compressed pulmonary gas, aided by the elasticity of the lung wall.

  4. In this animal the air-breathing mechanism is derived from the aquatic mechanism. The modifications are relatively simple and produce an efficient ventilation mechanism.

  5. No movements of the ribs can be seen associated with the respiratory cycles. It is suggested that the aspiratory ventilation mechanisms were not present in the prototetrapods and were not evolved until a later, more fully terrestrial stage was reached.

  6. The evidence suggests that the air-breathing mechanism of the tetrapods was powered by a buccal force-pump mechanism which evolved directly from the aquatic system. The evolution of a new mechanism for lung ventilation in the prototetrapods is considered unnecessary

I am indebted to Professor G. M. Hughes, Department of Zoology, University of Bristol, in whose department and under whose supervision this work was conducted, and also to the Science Research Council, who provided financial support.

Atz
,
J. W.
(
1952
).
Narial breathing in fishes and the evolution of the internal nares
.
Q. Rev. Biol
.
37
,
366
77
.
Ballintijn
,
C. M.
&
Hughes
,
G. M.
(
1965
).
The muscular basis of the respiratory pumps in the trout
.
J. exp. Biol
.
43
,
349
62
.
Basmajan
,
J. Y.
&
Stbcko
,
G.
(
1962
).
A new bi-polar electrode for electromyography
.
J. appl. Physiol
.
17
,
849
.
Bertmar
,
G.
(
1965
).
The olfactory organ and the upper lips in Dipnoi, a comparative embryological study
.
Acta Zool., Stockh
.
46
,
1
40
.
Bishop
,
I. R.
&
Foxon
,
G. E. H.
(
1968
).
The mechanism of breathing in the South American lungfish Lepidosiren paradoxa; radiological study
.
J. Zool., Lond
.
154
,
263
72
.
Dubois
,
R.
(
1892
).
Du mechanisme respiratoire du Dipnoiques
.
Annls Soc. Linn. Lyon
,
36
,
65
.
Edgeworth
,
F. H.
(
1935
).
The Cranial Muscles of Vertebrates
.
Cambridge University Press
.
Greenwood
,
P. A. H.
(
1958
).
The Fishes of Uganda
.
Kampala
:
The Uganda Society
.
Grigg
,
C.
(
1965a
).
Studies of the Queensland lungfish Neoceratodus forsteri (Krefft). I. Anatomy, histology and functioning of the lung
.
Aust. J. Zool
.
13
,
343
53
.
Grigg
,
C.
(
1965b
).
Studies of the Queensland lungfish Neoceratodus forsten (Krefft). III. Aerial respiration in relation to habits
.
Aust. J. Zool
.
13
,
413
21
.
Hughes
,
G. M.
(
1960a
).
The mechanism of gill ventilation in the dogfish & skate
.
J. exp. Biol
,
37
,
11
27
.
Hughes
,
G. M.
(
1960b
).
A comparative study of gill ventilation in marine teleosts
.
J. exp. Biol
,
37
,
28
45
.
Hughes
,
G. M.
(
1961
).
Gill ventilation in fishes
.
Rep. Challenger. Soc
.
3
, No.
13
.
Hughes
,
G. M.
&
Ballintijn
,
C. M.
(
1965
).
The muscular basis of the respiratory pumps in the Dogfish (Scyliorhinus canicula)
.
J. exp. Biol
.
43
,
363
83
.
Hughes
,
G. M.
&
Shelton
,
G.
(
1958
).
The mechanism of gill ventilation in three fresh water teleosts
.
J. exp. Biol
.
35
,
807
23
.
Hughes
,
G. M.
&
Shelton
,
G.
(
1962
).
Respiratory mechanisms and their nervous control in fish
.
Advanc. comp. Physiol. Biochem
.
1
,
275
364
.
Jaquet
,
M.
(
1898-9
).
Quoted in Edgeworth (1935)
.
Jarvik
,
E.
(
1967
).
On the structure of the lower jaw in Dipnoans with a description of early Devonian Dipnoan from Canada (Melanognathus canadensis)
.
J. Linn. Soc. Zool
.
47
(
311
)
155
85
.
Jesse
,
M. J.
,
Shub
,
C.
,
Fishman
,
A. P.
(
1968
).
Lung and gill ventilation of the African lungfish
.
Resp. Physiol
.
3
(
3
)
267
, 287.
Johansen
,
K.
,
Leneant
,
C.
&
Grigg
,
C.
(
1967
).
Respiratory control in the lungfish Neoceratodus forsteri (KREFFT)
.
Comp. Biochem. Physiol
.
20
,
835
54
.
Kerr
,
J. G.
(
1932
).
Archaic fishes. Lepidosiren and Protopterus and their bearing on problems of vertebrate morphology. Jena. 2
.
Naturw
,
67
,
419
433
.
Mccutcheon
,
F. W.
(
1966
).
Pressure sensitivity, reflexes, and buoyancy responses in Teleosts
.
Anim Behav
.
14
,
204
17
.
Mcmahon
,
B. R.
(
1969
).
The relative efficiency of gaseous exchange across the lungs and gills of an African lungfish, Protopterus aethiopicus
.
J. exp. Biol. (In the Press)
.
Poll
,
N.
(
1962
).
XI Étude sur la structure adulte et la formation des sacs pulmonaires des Protopteres
.
Annls Mus. r. Afr. cent. Sér
.
800
.
103
,
131
71
.
Rahn
,
H.
(
1966
).
Aquatic gas exchange theory
.
Resp. physiol
.
1
,
1
11
.
Roberts
,
T. D. M.
(
1950
).
The respiratory movements of the lamprey (Lampetra fluviatilis)
.
Proc. R. Soc. Edin. B
64
,
252
3
.
Saunders
,
R. L.
(
1961
).
The irrigation of the gills in fishes. I. Studies of the mechanism of branchial irrigation
.
Can. J. Zool
.
39
,
637
53
.
Schmalhausen
(
1957
).
Quoted in Szarski (1962)
.
Schmalhausen
(
1958
).
Quoted in Szarski (1962)
.
Smith
,
H. W.
(
1931
).
Observations on the African lungfish Protopterus aethiopicus and on the evolution from water to land environments
.
Ecology
12
,
164
81
.
Sterba
,
G.
(
1962
).
Fresh Water Fishes of the World
. Translated by D. W. Tucker.
London
:
Vista Books
(1963).
Szarski
,
H.
(
1962
).
The origin of the amphibia
.
Q. Rev. Biol
.
37
,
189
241
.
Thompson
,
K. S.
(
1964
).
The nasal apparatus in Dipnoi with special reference to Protopterus
.
Proc, zool. Sac. Lond
.
145
,
207
62
.
Thompson
,
K. S.
(
1967
).
Mechanisms of intercranial kinetics in fossil rhipidistian fishes (Cross-opterygii) and their relatives
.
J. Linn. Soc. Zool
.
46
,
223
58
.
White
,
E.
(
1966
).
A little on lungfishes
.
Proc. Linn. Soc. Zool
.
177
(
1
)
1
10
.
Willem
,
V.
(
1920a
).
Observation sur la respiration des Amphibiens
.
Bull. Acad. r. Belg. Cl. Sci
.
6
,
298
314
.
Willem
,
V.
(
1920b
).
Observations sur la respiration des Amphibiens. Deuxième note
.
Bull. Acad. r. Belg. Cl. Sci
.
6
,
341
47
.
Willem
,
V.
(
1929
).
Les manoeuvres respiratoires chez les poissons et les amphibiens
.
Bull. Acad. r. Belg. Cl. Sci
.
15
,
1
194
.
Willem
,
V.
(
1941
).
Synchronisme des mouvements respiratoires et des pulsations cardiaques chez les poissons
.
Bull. Acad. r. Belg. Cl. Sci
.
27
,
49
64
.

Plate 1

X-Ray photograph of a living Protopterus to show the skeleton of the head and branchial region and the extent of the lung.

Plate 2

Selected stills from an X-Ray cine film of Protopterus during air-breathing to show the principal stages in the air-breathing cycle. Explanation in text.

Plate 3

Mechanism of reduction of gape in Protopterus