Investigations Into the Distribution of Blood in the Heart and Aortic Arches of Xenopus Laevis (Daud.)

Except for typical fishes with their single circulation and post-embryonic birds and mammals with their complete double circulation, vertebrates possess an incomplete double circulation, having various degrees of anatomical division in the heart structure.

Anuran Amphibia are a case in point. They have, as is well known, a complete atrial septum, no septal division in the ventricle and only a partial division of the bulbus cordis by a spiral valve. Up to some twenty years ago it was generally accepted that, in spite of this imperfect separation, some physiological ‘attempt’ was made by the animal to aproach the perfect double circulation. The first reasoned account of the circulation of blood through the anuran heart was given by Brücke (1852). His ideas of the basic mechanism were modified and extended by Sabatier (1873), and have now become known as the ‘classical theory’. This hypothesis may be briefly summarized (see also Foxon, 1955).

In spite of the absence of ventricular septum and only partial division of the bulbus cordis, there is a functional separation between oxygenated blood (from lungs to left atrium), and de-oxygenated blood (from body to right atrium). On atrial contraction these different types of blood would have little opportunity to mix in the ventricle on account of, first, the propulsion gained from the atrial contraction ; secondly, the projection of the atrial septum down through the atrio-ventricular orifice; and thirdly, the simultaneous dilation of the ventricle which draws the blood into its trabecular meshwork. During ventricular contraction the first blood to leave is oxygen-poor blood, and this then passes into the cavum aorticum of the bulbus at pressure sufficient to distend the bulbus wall. The spiral valve does not now touch the opposite wall of the bulbus, and thus the blood has free access to the cavum pulmo-cutaneum. Since it is assumed that the peripheral resistance of the pulmo-cutaneous circuit is less than that of carotid or systemic circuits, an easier flow and a quicker reduction of pressure in the pulmo-cutaneous arches result. Thus, at the beginning of ventricular systole, oxygen-poor blood passes into the pulmo-cutaneous arches. At a later stage of ventricular systole the bulbus contracts, bringing the margin of the spiral valve into contact with the wall, thus completely separating the two cava of the bulbus. The cavum aorticum now receives the more oxygenated blood from the left side of the ventricle. This blood is channelled into the carotico-systemic arches. However, the blood does not enter the carotid arteries until at the last stage of ventricular contraction, for only then can the high resistance in these arteries be overcome. It is assumed in this connexion that the carotid labyrinths are the structures which help to maintain the high resistance.

This ‘classical theory’ was derived originally from anatomical considerations and from direct visual observation on the beating heart. Apparently the only concrete evidence is the observation that in a pithed frog subjected to artificial respiration the blood entering the left atrium is lighter red in colour than that entering the right atrium. This distinction is maintained in the ventricle and the arterial arches, so that the pulmo-cutaneous arches receive the darker blood and the carotid and systemic arches the lighter. Gompertz (1884) found that during artificial respiration in Rana this division in coloration could be seen, whereas it promptly disappeared when respiration was suspended. Further, Ozorio de Almeida (1923) found a distinct line of division in the ventricle when he applied artificial respiration to Leptodactylus ocellatus. Noble (1925) injected indian ink into the pulmonary veins of a series of Anura and Urodela. He reports that in all species which frequently use their lungs and possess a functionally complete atrial septum and a spiral valve (Scaphiopus holbrooki, Hyla crucifer, Acris gryllus and Rana clamitans), none of the ink was passed into the pulmo-cutaneous arches. Acolat (19316), working with Rana sp., replaced the blood by solutions of dyes and observed a clear separation in the ventricle. From further considerations (1931a, b, 1938a, b), however, he concluded that some mixture must take place. Simons & Michaelis (1953), using fluorescein injection, found selective distribution in two out of eight Hyla caerulea. Hazelhoff (1952) also advocated the ‘classical theory’.

Other workers have criticized the ‘classical theory’. As early as 1869 Fritsch claimed that the difference in coloration between the two sides of the heart and the different arches is solely due to difference in the transparency of the walls of the system. Gompertz (1884), although buttressing the ‘classical theory’, said that the spiral valve is at no time pressed against the opposite wall of the bulbus. Vandervael (1933) and Foxon (1947), using techniques similar to those of Noble on Rana and Bufo, maintained that the distribution is random, i.e. the blood from the two atria is thoroughly mixed by the time it leaves the bulbus cordis. Savolin (1949), injecting suspensions of starch or unicellular algae into the hepatic veins of Bufo bufo, also found an almost random distribution. Foxon (1948, 1951, 1953), applying a radiographic technique on Rana temporaria and Bufo bufo, confirmed the correctness of his earlier conclusions concerning random distribution. Finally, it must be mentioned that Simons & Michaelis (1953) found random distribution in several individuals of Hyla caerulea.

As can be seen, a relatively small number of species have been used for experimental work, and considering the disagreement about the distribution even in these, no conclusions may be drawn as to whether there are any specific differences or not. The few pressure recordings in the arches made with inaccurate technique by Gompertz (1884) and Acolat (1938a) do not help us. Neither do the considerations on cutaneous respiration of Vandervael (1933). The ‘classical theory’ is based on uncertain ground, and it is hoped that the following investigations on Xenopus will help to clarify the problem.

Although full descriptions of the blood vascular system of Xenopus and its development are available (Millard, 1941, 1942, 1945, 1949), so far there has been no published account of the heart structure. Nevertheless, dissection and serial sections show a fairly close agreement with the descriptions of Rana escalenta and R. fusca (Gaupp, 1899).

The sinus venosus is formed dorsally by the confluence of the anterior and posterior vena cavae and the large hepatic veins, and empties by way of a transversely oval aperture into the right atrium near to the mid-line (Text-fig. 1). The sinuatrial aperture is guarded by anterior and posterior thickenings of the atrial wall, of which the anterior is the more prominent. These presumably act as valves, preventing reflux of blood into the sinus venosus. The pulmonary veins run forwards dorsal to the sinus and are closely bound to it. They unite to form a short common trunk which empties into the left atrium close to the inter-atrial septum, but slightly anterior to the sinu-atrial aperture (Text-fig. 1). No valvular structures could be made out in the pulmonary aperture.

The two atria are completely separated by a thin, median, vertical septum and their interior surface, in the relaxed state, has a rugose appearance due to the projection of muscular bundles into the lumen. The position of the septum is not noticeable from the exterior, nor is there any external groove to indicate its line of attachment to the atrial wall. The margins of both atria, particularly where they overlap the ventricle ventrally, are partially subdivided into a series of pockets.

These are more prominent than diagrams and descriptions of the hearts of R. escalenta and R. fusca would indicate. Unlike Rana, Xenopus shows no distinct difference in anatomical size between the two atria.

The large atrio-ventricular orifice is guarded by two large, thick valves dorsally and ventrally. These valves are attached to the wall of the orifice anteriorly, but posteriorly have a number of chordae tendinae passing down and inserting into the ventricular musculature. Hardly to be described as semi-lunar, they are quite bulky in appearance, and have the same peculiar histological structure as the spiral valve of the bulbus cordis, which is presumably composed of hardened and modified endocardial tissue. On either side of the atrio-ventricular orifice is a weakly developed, semi-lunar valve (Text-fig. 1).

Posteriorly, the inter-atrial septum is fused with the middle of each of the large dorsal and ventral valves and projects for a short distance into the lumen of the ventricle.

The ventricle has a clear central chamber of relatively small size, into which the atrio-ventricular orifice leads, but the rest of its interior is broken up by the crossing of a large number of muscular trabeculae which, in section, give the appearance of a thick spongy wall. The outer wall of the ventricle, however, is relatively thin. From the right dorsal part of the central chamber an aperture leads into the bulbus cordis (Text-fig. 1). The latter is provided at its base with a series of three semilunar valves, of which the largest is ventral and the other two are smaller and dorso-lateral in position. Attached ventrally to the bulbus wall within the free margin of the ventral valve is the base of the large spiral ‘valve’. This valve then follows a spiral course round the wall of the bulbus in a clockwise fashion when seen from the ventricle, until at the anterior end it is attached to the right side, having turned through about 270°. Although at the posterior end it is only a comparatively small projection from the wall, it rapidly increases in size and, in dissections of fresh material, it would seem almost to fill the lumen of the bulbus. In serial sections its free margin extends about two-thirds across the bulbus for most of its length (Textfig-3A) but it is quite obvious that some shrinkage occurred during preparation.

The spiral valve, then, partially divides the bulbus into two channels, a cavum aorticum lying to the right posteriorly and a cavum pulmo-cutaneum to the left. Anteriorly, however, due to the twisting of the spiral valve, these come to lie ventrally and dorsally respectively. Due to the position of the spiral valve at the posterior end, the bulbo-ventricular orifice appears to lead into the cavum aorticum only, and this junction is guarded by the three valves already mentioned. The cavum pulmo-cutaneum seems to have no direct communication with the ventricle, although there is some doubt about this.

Anteriorly, the bulbus is followed by a thinner-walled tube, the truncus arteriosus, which is completely divided by a horizontal septum (septum principale) into dorsal and ventral chambers, corresponding to cavum pulmo-cutaneum and cavum aorticum of the bulbus. This septum is continuous behind with the spiral valve (Textfig. 3B, C). At the anterior end of the bulbus is a series of semi-lunar valves, guarding its exit to the truncus. Two of these (valves IA and IB of Gaupp, 1899) are attached to the spiral valve and have their free margins projecting into the cavum aorticum and cavum pulmo-cutaneum respectively, while the others (valves II and III) are attached to the bulbus wall and lie opposite the first two (Text-figs. 2, 3B).

Slightly anterior to the bulbus, the dorsal chamber of the truncus is split into right and left pulmo-cutaneous canals by a vertical septum (Text-fig. 3C). Still farther forward oblique and vertical septa almost simultaneously divide the ventral chamber of the truncus into four canals, which continue to form the paired systemic and carotid arches (Text-figs. 2,3 D, E). This is somewhat different from the arrangement in Rana, where both carotid canals arise from the right side of a main vertical septum, but it resembles the condition in urodeles and also in Microhyla and Ramanella, described by Rao & Ramanna (1925).

Projecting posteriorly from the vertical septum into the ventral chamber of the truncus, and attached to the ventral surface of the septum principale, is a block of tissue similar in composition to the spiral valve. This partially divides the chamber at its anterior end into two divisions, leading to the systemico-carotid canals on each side (Text-figs. 2, 3 C). No valvulae paradoxae could be found in the systemic canals of the truncus or in the systemic arches themselves.

A coronary artery arises from the base of one of the carotid canals of the truncus and runs backwards across the surface of the bulbus, but reaches no farther. Venous drainage from the wall of the bulbus passes into a coronary vein, which runs down laterally between the ventricle and right atrium to join the sinus venosus. There is no coronary supply to the rest of the heart.

The basic question as to whether or not there is a selective distribution of blood would seem to be the one most sorely in need of clarification. Most attention has therefore been given to this aspect of the problem. Several methods are available for this study.

Measuring the oxygen content of the blood arriving at, and departing from, the heart by various channels, has not yet been employed. The small quantities of blood which can be removed without risking disturbance in the pressure conditions of the system make an approach along this line very difficult.

Injection and recovery of some identifiable substance (Savolin, 1949) does not permit the tracing of substances through the ventricle and bulbus, although it is possible, in principle, to obtain quantitative results.

Injection of radioactive substances and their tracing is an obvious possibility. Simons (personal communication), however, has attempted the method without success. The main difficulty, apparently, is to obtain a great enough concentration in the arterial arches to give significant measurements.

The radiographic method used by Foxon (1951) is limited by a slow frequency of photographic exposure, due to the necessity of moving large masses of sensitive material, and of using lead shutters. In addition to this there are difficulties which prevented Foxon from obtaining more than about 20% ‘satisfactory’ results.

Cinematographic tracing of injected fluorescein was used by Simons & Michaelis (1953). This technique does not suffer from the disadvantages encountered by Foxon. Both methods, however, allow one to trace the injected material through the heart itself.

All in all, the fluorescein technique seems the most promising one, and this was the reason why we concentrated on this method, adapted from Simons & Michaelis (1953). Certain precautions are necessary in order to avoid the possible introduction of considerable errors.

First of all the animal must not lose much blood during the experiment, since this would upset normal pressure conditions. This does not seem to have bothered other workers. Vandervael (1933), in fact, considers the blood loss and slowing of the heart resulting from pithing as advantageous, in that the action of the heart could more readily be observed.

Secondly, the amount of material injected should not cause an appreciable rise in pressure in the injected vessel; Foxon (1951) states that for a thorotrast solution to be discernible in the systemic arch it had to be in a concentration of at least 25 %. Considering the dilution of the medium in its passage through the heart, it would be necessary to inject undiluted thorotrast into a pulmonary vein in quantities at least equal to the volume of blood normally flowing through this blood vessel. Simons & Michaelis (1953) found it necessary to inject volumes of about 1 ml. fluorescein. Such quantities, apart from changing the viscosity of the blood, increase the blood volume and thus change pressure conditions quite considerably, with unpredictable results.

Thirdly, the injected substance should mix easily with the blood and not form a separate viscous mass. Foxon (1951) points out how this may affect the blood circulation. One of the reasons for the failure of some of his experiments, he states, was that the thorotrast interfered with the action of the atrio-ventricular valves, and on contraction of the ventricle it was regurgitated into the atrium.

Finally, the injected substance should exert no chemical effect on the heart. Fluorescein, as is well known, is physiologically neutral and has been used for a range of physiological experiments, even on man.

In addition, it may be mentioned that the animal should be subjected to a minimum disturbance from normal conditions, or should be subjected only to disturbances which may be determined and taken into consideration. For most injection techniques it is necessary to expose the heart. Insufficient attention has been paid to the abnormalities which may result from this procedure. Unfortunately, it is difficult to determine the possible deviations in question. Some attention has been given, in the present investigations, to the effect of opening the body cavity. This operation causes the removal of pressure exerted on the internal organs by the abdominal musculature and thus might lead to changes in the blood pressure of the pulmonary circuit. However, observations on animals in a pressure chamber, and with distended lungs, showed that no significant changes in blood pressures or pulse curves occurred if the toads were not subjected to completely unphysiological pressure conditions.

The following techniques were adopted for our purposes:

In most cases a preliminary subcutaneous injection of about 0-6 ml. of 20% ethylurethane was given as an anaesthetic. As is well known, this has the effect of dilating the blood vessels and would thus cause a general lowering of the pressure in the system. In other cases, therefore, unanaesthetized toads were used as controls. The experiments showed that urethane had no apparent effect on the type of result obtained.

The toad was fixed to a cork mat, the abdomen opened, and a median incision made through the pectoral girdle. The two sides of the pectoral girdle were then loosely pinned back to expose the heart and arterial arches. For injections into the pulmonary vein, or for observing the circulation through a lung, the latter was cleared of the mesenteries and the lung gently moved to one side. When carefully done, the blood loss resulting from these operations was insignificant. All exposed parts were kept moist during the course of the experiment with drops of Ringer solution. Female Xenopus were found to be unsatisfactory subjects because the pulmo-cutaneous arch was usually hidden, and the carotid arch was also frequently in an unsuitable position for photography. Thus male toads of 40-50 g., having a much more favourable arrangement of the arterial arches, were used in almost all cases. The injections and photographic recording were carried out with the toads lying horizontally. Before the injections were started care was taken to verify that there was no interference with the normal free flow of blood in the vessels leading to and from the heart, and that the heart was not beating in an obviously abnormal fashion.

An Agla micrometer syringe was used as the injection unit. By fixing a length of thin lead tubing on the end of the barrel of the syringe and a very fine cannula (record no. 21) at the other end of the lead tube, the system was sufficiently flexible for adjustment of the cannula position without elastic recoil. This appeared to eliminate difficulties encountered by Foxon (1951), who frequently introduced air into the blood vessel at the same time as his contrast medium. The cannulae used were found to be small enough to allow insertion even into the pulmonary veins of the smaller specimens, without blood loss and without materially obstructing the flow of blood. The syringe itself was fixed in a clamp.

Various concentrations of fluorescein-Ringer solutions were tried, the most satisfactory being a 1 % solution. At concentrations higher than this, the fluorescein was found to diffuse into the blood before the actual injection was started, and this led to a preliminary fluorescence of the heart and aortic arches. The amounts injected varied between 0·15 and 0·004 ml., the average figure being 0-046 ml. It cannot be supposed that such small volumes could have any significantly adverse effect on the blood circulation, either through raising the pressure or by decreasing the viscosity of the blood. The solution was injected slowly and continuously during each experiment.

Ultra-violet illumination was provided by three or four 125 W. Philips mercury-discharge lamps, type 57202 E/70, mounted in polished aluminium reflectors. At a distance of about 1-5 ft. these reflectors focused the light into an intense bluish patch of about 2 in. diameter. Any unevenness of illumination was counterbalanced by the superimposition of the other lamp’s brightest areas. Since the fluorescein solution was strongly fluorescent to the eye, both in the brightest patches of focused light and immediately outside them, it seemed unlikely that the passage of any fluorescein would not be recorded by the camera, due to inadequate ultra-violet irradiation in some areas.

To test for any possible differences of penetration of the ultra-violet light through the walls of the different arterial arches, the following check was made. Small sections of the arches were removed from a toad, washed out with Ringer solution and slipped over the end of a glass tube of bore approximately equal to that of the arch in its normal state. Various concentrations of fluorescein-Ringer solution were then drawn up into the tube and exposed to the camera under ultra-violet light. The fluorescein solution was clearly visible on the negative down to a concentration of 1:10,000. It is true that the sections did not always transmit the fluorescence with equal intensity. This can probably be explained by the difficulty of ensuring that each of the sections was stretched to the same extent. Histological examination of the walls of the arches showed no significant differences in structure or thickness, and it was considered that, since such low concentrations of fluorescein gave a distinct record, the relative penetrability of the blood vessel walls need not be considered in the evaluation of qualitative results.

For photographic recording a Paillard Bolex 16 mm. camera with a parallax adjusting mechanism was employed. Using a 75 mm. lens with a short extension of 7 mm., and a 2-dioptre accessory lens, a field of about 1 × 1$12$ in. could be photographed at an aperture of f 3·5 with sufficient depth of focus for clarity. In order to eliminate reflected ultra-violet light a Kodak Wratten K-2 gelatine filter was used, and in addition a Voigt lander G-2 yellow filter to eliminate the visible bluish light emitted by the lamps. Kodak Super XX panchromatic film was found to be eminently suited to the purpose, since it was possible to photograph at a speed of up to 24 frames per second and more. The films were slightly overdeveloped with Ilford ID-11 M.Q. Borax fine-grain developer. In this way it was possible to obtain negatives of quite satisfactory contrast. The field covered included the heart, main arterial arches and some surrounding tissue. A certain amount of the fine detail was lost, especially movements of the bulbus. Attempts were made to obtain a more detailed idea of the sequence of events, but it was found that the depth of focus was so much restricted by narrowing the field that the results could not be interpreted. As will appear, the main features exposed by the technique were the filling of the atria and ventricle, and the distribution of the fluorescein to the arterial arches.

Injections were made into the left anterior vena cava or into one of the hepatic veins. A typical recording is reproduced in Pl. 3, being one out of fourteen similar recordings made.

The injection site and the proximal portion of the vein was always marked by a dark mass, due to the accumulation of fluorescein solution as a result of the slow movement of blood in these large vessels (Pl. 3, 1). The subsequent course of events was somewhat variable, although not different in principle. After one or two heart beats, the fluorescein might become visible first in either the right atrium or the ventricle, or both simultaneously. In Pl. 3, 1-3, the right atrium does not fill completely with fluorescein at its first diastole, although atrium and ventricle start to fluoresce simultaneously. In all cases, the brilliance of the fluorescence took some time to build up in the atrium. The marginal pockets in particular never obtained a full coloration until a few heart beats had occurred. Passage of the fluorescein into the sinus venosus or between the atrio-ventricular valves was obviously not recorded (due to the thickness of the heart wall?).

In ten cases the appearance of fluorescein in the ventricle was only on the right side, with a more or less sharp line of division between this and the clear left side (Pl. 3, 1-12). This discriminating coloration is interpretable as being due to the penetration of the fluorescein-laden blood in the intertrabecular spaces near the surface of the ventricle. In four cases the line of demarcation between the two sides was not sharp, only a small portion of the ventricle on its extreme right or at its apex becoming coloured. This may have been due to insufficient penetration of the dye into the trabecular meshwork.

During contraction of the ventricle the fluorescence remained, since some fluoresceinated blood had been retained in the muscular meshwork (Pls. 3 and 4). Foxon (1951) has proposed that this remnant, which he found in Rana and Bufo, might be blood concerned with the supply of oxygen to the ventricular musculature.

The visible division of the ventricle was obscured some three or four beats after its first appearance, when the left side also became coloured. Sometimes this occurred quite suddenly in the space of one ventricular diastole (Pl. 3, 12‒14), or sometimes more gradually, and is associated with the recirculation of fluorescein through the pulmonary circuit into the left atrium. This is proved by the fact that the division never breaks down before the pulmonary circuit, as shown by lung and pulmonary vein, has obtained a bright fluorescence. In Pl. 3, for example, the lung starts to fluoresce at frames 6‒8, the pulmonary vein at 10, the left atrium at 11‒12.

On the first or second heart beat, but not later, the fluorescein was quite distinctly detectable in the pulmo-cutaneous arches. No sign of it, however, appeared either in the ventral chamber of the truncus, carotid or systemic arches at that time. This can be clearly seen in Pl. 3, at the first ventricular systole (frames 3‒6), fluorescein appears very darkly in the pulmo-cutaneous arches only. The distinct appearance of fluorescein in the ventral chamber of the truncus was later, usually delayed until the third, fourth, or later heart beat (Pl. 3, 12), i.e. when there was already a coloration of the left side of the ventricle, or there was at least the probability of recirculation of fluorescein through the pulmonary circuit.

In most cases, the pulmo-cutaneous arches remained distinctly more fluorescent throughout the length of the record (e.g. Pl. 3). In two cases the systemic or carotid arches attained a similar density at the fifth or sixth heart beat. In Pl. 3 no appreciable quantity of fluorescein is passed into the systemic and carotid arches until the third beat, when recirculation of fluorescein through the pulmonary circuit has occurred. Only a little fluorescein is visible in the systemic (particularly the right one) and carotid arches during the second beat. It is noticeable that the fluorescein in the right systemic arch only reaches about half-way up the visible part of that arch during the second ventricular systole (9-10), but progresses the rest of the way in the third ventricular systole (12). This should be compared with the complete filling of the visible part of the pulmo-cutaneous arch at the first ventricular systole (3‒6). In all cases the maximum density of fluorescence in the pulmo-cutaneous arches was reached by the second or third heart beat. It must be mentioned, however, that the arches did not always reach the same brightness on both sides. It is likely that some of the arches were shaded from ultra-violet light in these cases, due to the position of the heart, for it was not a constant occurrence. The fluorescein sometimes appeared first in the form of streaks in the arterial arches (e.g. Pl. 3).

The speed of flow of blood through the pulmo-cutaneous arches and the pulmonary artery was obviously high, for the fluorescein appeared in the lungs at the first or second beat. In contrast to this there was never any recordable fluorescence in the tissues supplied by the carotid or systemic arches before the seventh or eighth heart beat.

Attention must be drawn to an aspect of fluorescence of the bulbus in relation to statements by former workers. In these experiments the main part of the bulbus visible from the ventral surface was that part which corresponds to the cavum aorticum. At most, only a small section of the cavum pulmo-cutaneum would be visible at the base and on the left side of the bulbus, from this angle. This would be separated from the cavum aorticum by the free margin of the spiral valve, which follows a diagonal line from the bottom centre to the top left part of the visible region of the bulbus. The rest of the cavum pulmo-cutaneum would be hidden beneath the spiral valve. With these structural arrangements in mind, it must be noted that in most cases either no coloration of the bulbus was visible until the third beat, or the anterior part was clear while a dark basal part was discernible with a diagonal border (which may well be interpreted as the free margin of the spiral valve). This state of affairs is consistent with the supposition that the fluorescein-laden blood from the right side of the ventricle was either passing directly into the cavum pulmo-cutaneum, or was traversing only the lower part of the cavum aorticum before passing over the spiral valve into the cavum pulmo-cutaneum. In four cases, however, a ‘pulse ‘of fluorescein was visible which in one case clearly passed through the bulbus from the right to the left side (Pl. 3, 3 and 8, arrows). During this time no great fluorescence appeared in the ventral chamber of the truncus or in the four anterior arches. It would therefore appear that blood may be admitted to the cavum pulmo-cutaneum by first travelling higher up the cavum aorticum with subsequent leakage over a more anterior part of the spiral valve. These observations are placed on record as an indication that the statements by Ozorio de Almeida (1923), Vandervael (1933) and Foxon (1947), that the blood passes up as two separate streams in the cava, do not always apply in Xenopus, since blood can pass over the spiral valve from the cavum aorticum to the cavum pulmo-cutaneum.

Injections were made into the left or right pulmonary vein. A typical recording is reproduced in Pl. 4, one out of eighteen cases. Certain significant differences emerged from the results of injections into the pulmonary veins. Where no differences are remarked upon in what follows, the picture was the same as with injections via the right atrium.

In the records, the pulmonary vein injection site was not always clearly visible either because this was not included in the area filmed or because it was hidden by the heart. When, however, the site of the injection could be seen, it showed that the injected solution was rapidly washed away by a wave of non-fluoresceinated blood at each atrial diastole, in contradistinction to injection into a vena cava or hepatic vein. This indicates that the flow through the pulmonary veins was considerably faster than that through the anterior vena cava or a hepatic vein.

There was again a gradual filling of the marginal pockets of the atria over a number of heart beats (Pl. 4). The initial filling of the ventricle was somewhat variable. In the majority of experiments (thirteen cases) either the whole ventricle appeared coloured at the first beat (e.g. Pl. 4, 1-3), or spots appeared on both sides at the first beat, denser on the left side. In four other cases only the left side of the ventricle was coloured at the first beat. Two of these cases showed a line of demarcation between right and left sides which lay well to the right of the median line. When a division occurred this tended to disappear rapidly. Unlike the injections via the right side, this was not to be correlated with a recirculation of fluorescein which only occurred much later.

The pulmo-cutaneous arches showed coloration at the first beat in fifteen cases (Pl. 4, 4‒13), and at the second beat in three cases. In twelve cases coloration of the bases of the carotid and systemic arches was simultaneous with that of the pulmo-cutaneous arches (Pl. 4, 5,‒6), in five cases a little after, and in one case slightly earlier than the pulmo-cutaneous arch coloration. In fifteen experiments the pulmo-cutaneous arches were more fluorescent than the others for at least the first four beats (Pl. 4, 4‒18). In four records they even remained darker throughout the length of the film strip (six to ten heart beats). This indicates that the pulmo-cutaneous arches were receiving a considerable proportion of the fluorescein injected into the pulmonary vein. Fluorescence of the lungs (Pl. 4, 8, 9) and pulmonary veins occurred at about the same time as with injections in the hepatic vein, but coloration of other tissues such as the distributional area of the external carotid artery tended to occur sooner than by injections via the right atrium. The cavum aorticum of the bulbus was usually fluorescent from the first beat, or at least from the time of the first appearance of dye in the ventral chamber of the truncus.

The results obtained lead to several conclusions discussed below

1. Upon injection via the right atrium a division in the ventricle appears, so that only the right side is fluorescent until the division is obscured by recirculation of the dye. However, a note of caution is advisable. The line of demarcation as shown in the records represents only a surface picture of the ventricle. In the fluorescein technique a picture of the whole ventricle is not obtained, since both the ultra-violet fight and the resulting fluorescence can only penetrate a short distance through the ventricular trabeculae, and the blood in between them. The line of demarcation between fluoresceinated and non-fluoresceinated blood, through the depth of the ventricle, might either be in a vertical plane or might incline considerably in a diagonal fashion. Yet it is not to be expected that the line of division would differ much from the vertical plane in which the inter-atrial septum projects into the central chamber of the ventricle. The sharp median line of demarcation between light and dark sides of the ventricle appearing in many of the records is therefore interpreted as being due to absorption of fluoresceinated blood between the trabeculae on the right side only. It is clear, therefore, that a form of division between the blood from left and right atria occurs in the ventricle, in so far as right atrial blood is absorbed only into the trabecular meshwork on the right side of the ventricle.

2. Upon injection via the left atrium the fluorescein either tends to spread all over the ventricle at once, or if, exceptionally, a division in the ventricle occurs, the line of demarcation may lie well to the right side of the median line, and anyway disappears before the commencing fluorescence of the right side can be attributed to recirculation. Comparing this with the results obtained with injection of the right side, the following statement can be made : Whereas the output of the right atrium is not, or only in negligible quantities, transferred to the left side of the ventricle, the output of the left atrium shifts, to a considerable extent, to the right half of the ventricle.

3. By comparing injections via left and right atrium it can be seen that the left atrial output occupies more ventricular space than the right output (see, for example, Pls. 3 and 4). The blood from the left atrium either fills the left half and, moreover, some of it mingles with the right-side blood, or, if there is an initial separation, the blood of the left side may extend considerably into the right side. In contradistinction to this the output of the right atrium does not occupy more than the right half of the ventricle. We, therefore, draw the following conclusion : the left atrium has a larger output than the right atrium. No anatomical data conflict with this statement.

4. After injection via the right atrium the pulmo-cutaneous arches fluoresce distinctly within a short time, but no significant coloration of the systemic and carotid arches occurs before recirculation of the die commences. It is true that in some experiments a faint tingeing of the systemic and/or carotid arches did appear before it could be said with certainty that recirculation had occurred. However, one cannot expect that a heart without complete anatomical division can possess a complete selective distribution. It is clear, anyway, that almost all the blood expelled from the right atrium is sent to the pulmo-cutaneous arches.

5. After injections via the left atrium fluorescein tends to appear simultaneously in the bases of all the arches at the first beat, i.e. the appearance of dye in the carotid and systemic arches now tends to coincide with its appearance in the pulmo-cutaneous arches. In addition, it must be stressed that the pulmo-cutaneous arches were more fluorescent than the others for at least the first four beats. In contradistinction to the flow of the output of the right atrium, the blood from the left atrium is distributed to all the arterial arches, and the pulmo-cutaneous arches receive a considerable proportion of this blood.

6. By comparison of the film recordings (see, for example, Pls. 3 and 4) one obtains the strong impression that, at the very least, three-quarters of the right atrial blood plus one-third of the left atrial blood is received by the pulmo-cutaneous arches. Since the left atrium has a larger output, these approximate figures do not necessarily indicate that the pulmo-cutaneous arches receive more than half of the blood from the heart. However, other considerations lead to that conclusion : the left atrium has a larger output, output must be equal to input, thus the pulmonary veins must supply the left atrium with more blood than delivered by the sinus venosus to the right atrium. Consequently the capillary bed of the lungs, and hence the pulmonary arteries, must carry more blood than the body circuit. Since the pulmonary arteries transport the blood delivered by the pulmo-cutaneous arches minus the blood which is diverted to the cutaneous arteries, the pulmo-cutaneous arches must absorb, very definitely, more blood than the other arches together. The drainage by the cutaneous arteries, and the supply through the cutaneous veins to the right atrium (via anterior venae cavae) is not very considerable. The cutaneous arteries are very small in Xenopus, having as measured, in fresh preparations, about one-ninth of the cross-sectional area of the pulmonary arteries. Nevertheless, some drainage of the pulmonary system will occur. All in all, it seems safe to conclude that more blood is flowing through the pulmo-cutaneous arches at each beat than is being sent through the carotid and systemic arches together.

Additional evidence is produced by observations of the velocity on blood flow. In all cases the fluorescein passed up the whole visible part of the pulmo-cutaneous arch during the same beat at which it appeared at the base of the arch. It was visible in the lungs at about the second beat, and in the pulmonary vein at about the third beat. This indicates a very rapid rate of flow through the pulmonary circuit. By comparison, the impressions gained of the speed of flow in the systemic and carotid arches are as follows.

In no case was fluorescein visible at the ends of the exposed portions of the carotid or systemic arches before the second beat after its first appearance at the base of those arches (the length of the exposed portions of all the arches were roughly equal). Usually it took some three or four beats for the dye to reach this distance. In some instances the dye could be seen to progress a short distance up the arch with each beat, although no clear wave front was visible. The appearance of fluorescein in the peripheral distributional area of the carotid or systemic arches occurred later than in the lungs. In no case was any recirculation of fluorescein visible in the right atrium during the course of pulmonary vein injection experiments, even when these lasted for seventeen or eighteen heart beats. In fact, from direct visual observation, return of fluorescein through the anterior venae cavae was not detectable before about the twenty-fifth beat.

The much slower rate of flow in the systemic and carotid vessels could also be seen by direct observation of the blood corpuscles through the arterial walls, under a dissecting microscope, with strong side illumination. In the carotid and systemic arches these movements took place in a jerky fashion, consisting of three phases. First, a sudden rapid movement over a short distance which corresponded with ventricular systole, subsequently a brief complete halt or nearly so, followed by an increasing and moderate speed of flow, which appeared to correspond with contraction of the bulbus. It is significant that by using the same method the movement of corpuscles in the pulmo-cutaneous arches could not be observed, until some obstruction was applied, i.e. the blood was moving too fast all the time for the corpuscles to be visible.

The speed of flow in the pulmo-cutaneous vessels must be at least twice as high as in the other arches, which strengthens the conclusion that they are dealing with more blood per beat than the rest of the system.

Only two references have been found to previous experimental work in which simultaneous pressure measurements have been made in the pulmo-cutaneous and systemic or carotid arches of any anuran (Gompertz, 1884; Acolat, 1938a). This is rather surprising considering the importance of pressure differences in the mechanism proposed by the ‘classical theory’. Both workers used mercury manometers which, however small, have too low a frequency response to give an accurate idea of the shapes of the pulse curves or their amplitude. Moreover, their results do not agree with one another. Repetition of these experiments with Xenopus, using more adequate methods, was therefore an obvious necessity.

What was required for the present experimental work was a form of apparatus which would create as little disturbance in the animal as possible, record pressures accurately, and respond rapidly and faithfully to sudden pressure changes. The mercury manometer does not fulfil these requirements, and thus could not be considered for the present investigations. The optical manometer, operating by transmitting pressure changes to an elastic membrane on which a small mirror is fixed, seemed much more promising. Wiggers (1928) and Hamilton (1946) have published accounts of the mathematical principles of such systems. The Wiggers universal manometer has the disadvantage of requiring the use of a wide cannula which is obviously out of the question with animals as small as frogs. Hamilton, Brewer & Brotman (1934) and Gregg, Eckstein & Fineberg (1937) have described instruments which overcome this difficulty. In these the adverse effect of a narrow hypodermic cannula is counterbalanced by the use of a more rigid elastic membrane of small area.

In pressure transducer units the distortion of a membrane is measured by electrical means through a Wheatstone bridge and amplification system. One form of pressure transducer produced by the Statham Laboratories, California, was tried, but was found to be unsatisfactory, since it was not really designed for use with very small cannulae, or such low-pressure ranges. An optical manometer system was therefore used in the present investigations.

The design adopted was essentially the same as that described by Hamilton et al. (1934), but with a few modifications (Text-fig. 4). Two manometers were constructed, as identical as possible. A few brief comments follow on features differing from the Hamilton manometer.

The most satisfactory membrane was found to be rubber sheeting 0·7 mm. thick. As Gregg et al. (1937) point out this has the advantage that the excursion of the optical lever is linearly related to the pressure applied. The rubber sheeting was tied with wire over the end of the manometer tube, making sure that it was stretched equally in all directions, and to the same extent in the two manometers. On this was mounted a small triangle of the same rubber, glued with two points on the rim of the manometer tube and one point in the centre. The mirror, made from an 0-75 dioptre planoconvex wafer lens silvered on the flat side and cut down to a 5 mm. square, was mounted on the rubber triangle. This allowed the mirror to move on the segment principle (see Wiggers, 1928) in the horizontal axis only. As a precaution against misinterpretation of accidental jolting, base-line mirrors were attached to the outer housing.

The cannulae used were $12$ in., no. 20 hypodermic needles which were of a diameter small enough to be inserted into the arterial arches, without disturbing the normal flow of blood significantly. These were somewhat smaller than those used by Hamilton or Gregg and their associates, and had the effect of overdamping the system, as will be seen.

The manometers were filled with a boiled 5 % sodium citrate solution, great care being taken to avoid air bubbles. To make quite sure that all bubbles were eliminated before each series of experiments, the whole manometer and lead connecting tube was treated in a vacuum chamber until the citrate solution boiled for some time.

Illumination was provided by a microscope projection lamp with a Philips 8 V. 6 amp. globe which could be overloaded to 10 V. An adjustable vertical slit in front of the lens of the lamp allowed a series of beams to be cast out from the vertical filaments of the lamp, and these could be trained on the mirrors of both manometers at the same time.

Recordings were taken with a photo-kymograph, constructed to a design by Prof. R. Goetz of the Surgical Research department of our University.

In order to check the inertia of the system tests were carried out at intervals, using the method described by Hamilton et al. (1934). Although the manometers followed a sudden change in pressure very rapidly, they were aperiodic, due to the damping effect of the fine cannulae. Various ways of overcoming this were tried. Increasing the size of the cannulae to no. 18 had no appreciable effect, and in any case these were too large to use without disturbing the blood flow. Increasing the thickness of the rubber membrane cut down the sensitivity to such a degree that the optical lever had to be extended to an unwieldy length and involved difficulties in obtaining a sharp record of sufficient intensity. Stretching the membrane more tightly had no appreciable effect. However, it was felt that, considering the quick response which the manometers showed to the sudden rise in pressure in these tests, giving an almost square wave form to the recording (ascending slope lasting 0·2 sec., descending slope 0·05 sec. when applying a pressure difference of 40 cm. water), the results obtained would be significant for present purposes. A good idea would at least be obtained of the pulse pressures, the timing relationship between the arches and the general form of the pulse curves. The apparatus showed promise of giving significant comparative readings for the different arches, which was the prime purpose of the experiments.

In order to be quite sure that differences between the pressures recorded were not due to a lack of absolute uniformity in the two manometers used, the manometers were often reversed in successive experiments. Readings in the systemic arch, for example, were not all taken with the same manometer.

Following each experiment the manometers were simultaneously calibrated against a citrate solution, thus cleaning out the cannulae by flushing them with citrate at the same time. The calibrations were entered in the previous experimental recording.

It was found by a series of tests that the accuracy of measurement was ± 0-25 cm. citrate solution pressure (ca. 0·2 mm. Hg), and so all measurements were taken to the nearest 0·5 cm. Readings were subsequently converted to their equivalents in mm. Hg.

The toads were prepared in exactly the same way as had been done for the fluorescein experiments, except that anaesthetized toads were always used, because of the difficulty of preventing struggling during the course of the experiments. The position of the needles was adjusted so that they were in corresponding positions in the blood vessels, lying well to the side to prevent any blockage. Tests with the cannulae of the manometers facing in opposite directions in the same artery showed that at the blood velocities involved there was no measurable difference in the pressure recording due to the cannula position. Any slight differences in the angle of the cannulae in the different arches therefore are entirely negligible. The recordings were taken as quickly as possible, because of the danger of the fine needles becoming blocked with clotted blood in spite of the manometers being filled with citrate solution.

The basic results, and those upon which most attention is to be focused, were those of comparative readings in the arterial arches (summarized in Table 1). After conversion to mm. Hg the readings appear to be smaller than those obtained by other workers (e.g. Schulz, 1906, Rana esculenta, in arterial arches up to 60 mm. Hg). The highest measurements obtained in any of the present experiments were of the order of 45 mm. Hg. It would thus appear that the normal average blood pressure of Xenopus is lower than that of Rana or other Anura. Another immediately striking point is the rather wide range of pressures in different experiments (see lines (d), (e), (f) in Table 1). There appeared to be no relation between the size of animal used and the blood pressure, but a certain percentage of the variations may be due to the extent to which the animals were narcotized.

The main features which are brought out by these results are that the pulmocutaneous arch pressures are consistently lower than either the carotid or systemic pressures (Text-fig. 5), and that the pressures in the carotid and systemic arches are remarkably similar (Table 1).^* The pulmo-cutaneous arches have a pressure at systole about 1 mm. Hg lower and at diastole about 7 mm. Hg lower than the other two arches, and the diastolic pressure differences amount to about one-third of the systemic pressure—quite a considerable amount. These figures are higher than those obtained by Acolat (1938a).

Another interesting result of the pressure measurements is the shape of the pulse curves which showed some variation. In general, the pulse curve showed two waves. In the pulmo-cutaneous arch the first wave was by far the major one, the second wave usually only faintly indicated (Text-fig. 5). In the systemic arch the second wave, although small in the great majority of cases, was somewhat more pronounced (Text-fig. 5). There was no difference between the curves of the carotid and systemic arches.

The first wave of the pulse is due to the main propulsive force of the ventricular contraction, while the secondary curve is caused by the contraction of the bulbus. The cavum aorticum distended to a greater extent on ventricular systole and subsequently contracted more forcibly than the cavum pulmo-cutaneum. Moreover, the more rapid fall of pressure in the pulmo-cutaneous arches tends to reduce the effects of bulbar contraction there. For these reasons the second wave of the pulse is naturally more pronounced in the systemic and carotid arches than in the pulmo-cutaneous arches. It is to be noted particularly that this bulbus contraction does not as a rule play an important part in raising the blood pressure in the arches. The main pressure rise is solely due to the ventricular contraction.

There is no doubt that the features described represent the normal situation. It is true that in a few experiments the second wave of the pulse curve was more marked, in one case it even dominated the picture. However, this phenomenon was always associated with an unusual low blood pressure which could be traced back to rather heavy anaesthesia or some loss of blood. It is, therefore, quite probable that the relatively larger role of the bulbus contraction in those cases can be accounted for by a weaker action of the heart itself. This was confirmed by direct visual observation. In those cases where the ventricular contraction was weak, the bulbus was seen to contract quite markedly, so that the cavum aorticum became almost empty, while in a normally beating heart it did not empty itself completely, and appeared to act more as a pressure relaying mechanism.

The question arises as to whether there is actually a contraction of the muscular wall of the bulbus, or merely an elastic recoil. Histological examination of the bulbus wall reveals no marked elastic fibres, as is the case in the arterial arches. This would seem to indicate that the bulbus is not merely a pressure chamber. Besides, one would not expect an increase in the effect of the bulbar contraction with weakening heart action if the bulbus were merely an elastic chamber. It would appear that the anaesthetic which produced the anomaly has less effect on the bulbar than on the ventricular contraction.

By projecting the records through an epidiascope and tracing out the curves on paper, it was possible to obtain a more detailed idea of the sequence of events. The curves shown in Text-fig. 6 have been chosen because they exhibit typical characteristics, and also represent successive experiments in which the manometers were reversed. It is evident from a comparison of these curves that there was a slight lag in the time of response of one of the manometers, or more probably a parallax error in the optical system. In Text-fig. 6 A the response in the systemic arch and in Text-fig. 6B the response of the pulmo-cutaneous arch are somewhat delayed (same manometer). Nevertheless, the curves show quite clearly what happens. The pulmo-cutaneous pressure rises abruptly and rapidly, whereas there is a slight but true delay in the beginning of the rise in the systemic arch (ca. 0·1 sec.). When the latter pressure does increase, it does so more gradually. Both curves (allowing for the time difference in one manometer) reach their maxima at about the same time, although the pulmo-cutaneous pressure at no time becomes equal to that in the systemic arch. The subsequent fall in pressure in the pulmo-cutaneous arch is invariably more rapid and is more suddenly interrupted by the succeeding ventricular systole.

There cannot be any question of the spiral valve blocking off the entrance to the pulmo-cutaneous arches before the major propulsive force of the ventricular contraction is expended. This is shown by the pulse curves of the pulmo-cutaneous and carotid-systemic arches, which reach their maxima at about the same time. The timing of the bulbus contraction is later, and it can be only then that the spiral valve may come into contact with the opposite wall of the bulbus. Under these circumstances the significance of the spiral valve is not clear. It might well be that its main function is not to bring about a separation of the blood streams in the way which the ‘classical theory ‘implies, but to ensure, as a ramp, that the carotid and systemic arches receive sufficient blood, in spite of their greater peripheral resistance.

A few venous-pressure measurements were taken. These indicated that the pressure in the pulmonary veins was about 3-5 mm. Hg in a lightly anaesthetized toad, while the hepatic vein pressure was only about 1-5 mm. Hg. A small but distinct pulse was visible in the pulmonary vein. By connecting up the manometers simultaneously to the pulmo-cutaneous arch and pulmonary vein it was possible to interpret the main wave of the pulmonary vein pulse as being due to pressure transmitted through the lung capillary bed. In the hepatic vein, on the other hand, a pulse is either absent or only very faintly indicated.

Recordings taken of pressures in the systemic arches of opposite sides showed no significant difference in amplitude, shape or timing of the pulse curves.

The results partly explain why the pulmo-cutaneous arches receive more of the blood which leaves the ventricle than the carotid and systemic arches together. There is a pressure in the pulmo-cutaneous arches some 7 mm. Hg less than in the systemic and carotid arches, at the beginning of systole. There is a more rapid drop in pressure in the pulmo-cutaneous arches. There is a distinct pulse in the pulmonary veins. The pressure in the pulmonary veins is higher than in the hepatic veins, illustrating a more direct transmission of pressure through the pulmonary capillary bed. The pulmonary capillaries have a rather large diameter. These facts all clearly indicate that there must be less peripheral resistance to the flow of blood in the pulmonary circuit than through the general circuit. Consequently there is obviously a tendency for the blood to flow into the pulmo-cutaneous arches. Considering the way in which the systemic and carotid arches branch off from the ventral chamber of the truncus, it is hardly surprising that no significant differences appeared between their pulse curves. The lack of difference between the pulse curves of opposite systemic arches is also not surprising. It is true that the right systemic canal branches off from the truncus at a sharper angle, but since the blood flow is not very rapid (as indicated by the fluorescein experiments) this anatomical arrangement could hardly be expected to cause a difference in the pressures of the two sides.

The present investigations on Xenopus have revealed some unexpected results, as neither the ‘classical theory’ nor the existence of random distribution have been confirmed. There is a kind of selective distribution, but not of the type as required by the ‘classical theory ‘. There is a certain mixing, but not so complete as demanded by random distribution. The main differences from the ‘classical theory ‘concern which side of the heart and which peripheral circuit shows the greater volume flow per unit of time. The difference is partly one of accent, but the accent alters the picture considerably. The situation may be illustrated by the two diagrams of Text-fig. 7. According to the ‘classical theory’ shown in Text-fig. 7B, the output of the left atrium plus some of the right atrial output supplies the body, and the remainder of right atrial blood goes to lungs and skin. The body circuit would transport more blood than the pulmonary circuit.

The postulate that the lungs of Anura are not able to deal with the same quantity of blood in unit time as the systemic circuit has been advocated for a long time and was emphasized by Acolat (1931b). The same idea of a short-circuiting mechanism, i. e. leakage to the left side to relieve the load upon the lungs, has been stressed in relation to the vertebrate series as a whole. The present results sharply contradict such suggestions, because in Xenopus this is obviously not the case.

It is true that the peripheral resistance in the pulmonary circuit is less than in the body circuit, and it is equally true that there is some form of separation in the ventricle. However, although the right atrial blood does not ‘contaminate’ the left side of the ventricle, the output of the left atrium (being the larger of the two), considerably ‘contaminates’ the right ventricular blood. The statement of Acolat (19316), that the output of the right atrium is larger, does not hold for Xenopus. In Xenopus blood from the right side is not given off to the body circulation, but a proportion of left-side blood is conveyed to the pulmonary circulation (Text-fig. 7 A). It is the left atrium which supplies all arterial branches, whereas the right atrium merely supplies the pulmo-cutaneous system.^*

There must be a physiological connexion between the right side of the ventricle and the pulmo-cutaneous arch, and between the left side of the ventricle and the other arches. This conclusion is confirmed by a few unusual cases in which the distribution pattern was changed. In one record of injection via the right atrium there was a division in the ventricle lying well to the left of the mid-line. It was notably the only record of that series in which fluorescein appeared simultaneously in all the arterial arches. Thus the right atrial blood has access to the carotid-systemic arches as soon as, and only if, it occupies a proportion of the left side of the ventricle. In two exceptional cases merely the left half of the ventricle coloured after injection via the left side, and only in these cases were the pulmo-cutaneous arches not darker than the others for at least four beats. Apparently the output of the left atrium has access to the body circuit from the left side of the ventricle, but only when it enters the right side of the ventricle does it obtain full access to the pulmonary circulation.

The above-mentioned exceptional cases also show that the ratio of the outputs of the atria may vary, and the outputs are usually less than their anatomical structure leads one to expect. Our observation that the fluorescence in the marginal pockets builds up gradually shows that the atria hold a certain reserve in these pockets. The main bulk of blood is taken in and expelled from the central parts. This is in good accord with observations by Foxon (1951) and Acolat (1938b). The atria are in fact acting as reservoirs which are not emptied completely at each beat, and whose degree of emptying may possibly vary from time to time. This, in turn, might change the distribution pattern to some extent.

One of the postulates of the ‘classical theory ‘is the assumption that the carotid labyrinths help to maintain a higher pressure in the carotids than in the systemic arches. This idea has to be discarded, at least for Xenopus. Pressures in carotid and systemic arches are equal, hence the carotid labyrinths cannot have such a function.

It is not certain exactly how the bulbus cordis with the spiral valve and its other structural peculiarities contributes to the selective distribution. The experimental records do not show how the blood moves in these regions, apart from the observation that the blood from the right side can pass over the rim of the spiral valve from the cavum aorticum to the cavum pulmo-cutaneum. Other considerations lead to the conclusion that the spiral valve cannot block off the entrance of the pulmo-cutaneous arches before the bulbus has started its contraction. These facts are not sufficient to explain why the blood entering the body circuit from the left ventricular side does not mix to any extent with the crossing stream of right-side blood going to the pulmonary circulation. This problem therefore remains unsolved at the moment.

If the scheme of Text-fig. 7 A applies to other Anura as well, it may explain some of the contractions in the results obtained by other workers. Vandervael (1933) and Foxon (1951) injected only into the pulmonary vein. That they found an apparently random distribution would be as expected, if the distribution were as proposed here. On the other hand, Simons & Michaelis (1953) injected into veins leading to the right atrium and reported a selective distribution in some cases, a result again to be expected according to the present proposals. The scheme may also explain the observations of the authors of the ‘classical theory’, Gompertz (1884), Ozorio de Almeida (1923) and Acolat (1931b), with regard to a physiological division of the ventricle. It is interesting to note that in spite of his disagreement with the ‘classical theory’, Foxon (1951) does report signs of a physiological division of the ventricle. Foxon (1947) and Foxon & Walls (1947), however, report having made injections via the anterior vena cava with a resulting distribution not unlike the pulmonary vein injections. The scheme will also not explain the results of Noble (1925), or of Savolin (1949).

The light in which the ‘classical theory’ was considered to be a satisfactory account was largely in the consideration of the oxygen needs of the tissues. By the same sort of consideration, however, Vandervael (1933) draws the conclusion that a random distribution would be more efficient! A re-examination of Vandervael’s theoretical objections seems worth while. His argument consists of a deduction of the oxygen distribution in an animal with non-functioning lungs (i.e. during hibernation). If the ‘classical theory’ applies under these conditions, he argues, there would be a tendency for the body tissues to receive the de-oxygenated blood, while the skin which is now the functional respiratory organ would have the oxygenated blood sent back to it via the pulmo-cutaneous arch and the arteria cutanea magna. The lungs would also be receiving the more highly oxygenated blood. The argument, however, has not been carried far enough. Under these circumstances it is hardly to be expected that the lung tissues would extract for their own metabolism much of the oxygen from the blood sent to them. The large supply of blood in proportion to the small bulk of the lung tissues denies such a supposition. If the circulation is taken a step further, therefore, it is seen that the left atrium would not be receiving de-oxygenated blood, but blood with a fair amount of oxygen. Apart from this, the ‘classical theory’ proposes that some of the right atrial blood (more highly oxygenated in this case) must pass into the systemic arch in order to balance the differences between the blood capacities of the lungs and the other tissues. This blood, plus the partly oxygenated blood from the left atrium, would then be distributed to the body. Bearing in mind that during hibernation the animal has a much decreased oxygen requirement, Vandervael’s argument can be seen to be incorrect. There would be the seemingly undesirable physiological arrangement whereby the main respiratory organ, the skin under these conditions, would be receiving the more highly oxygenated blood through the cutaneous artery, but this does not affect the argument about the oxygen needs of the body.

Nevertheless, these objections against Vandervael’s argument do not confirm the ‘classical theory’, and a random distribution of blood might suffice equally well for the transportation of oxygen to the tissues. Yet a form of selective distribution is present in Xenopus. What is its significance in the life of the animal?

The body and the head receive the most highly oxygenated blood, only scarcely contaminated with oxygen-poor blood from the right side. This seems to be a better arrangement for the animal than the conditions proposed by the ‘classical theory’, according to which mixed blood is supplied to the body. According to the ‘classical theory ‘the lungs receive oxygen-poor blood while according to the present scheme they receive mixed blood. Does this result in a more gradual utilization of the air in the lungs, and if so, has this any advantage for the animal? Xenopus may stay under water for considerable lengths of time. During such periods, apparently, skin respiration supplies the body with blood oxygenated to a similarly low degree as proposed by the ‘classical theory’.

However, it must be admitted that selective distribution possibly has less meaning in relation to the oxygen needs of the animal than has been advocated. It has been shown (de Graaf, 1957) that the oxygen requirements of Xenopus may at times be extremely low. They are able to survive, for example, with most of the haemoglobin of their blood blocked by carbon monoxide. They are also able to survive by cutaneous respiration alone, in spite of the very small size of their cutaneous arteries. It would appear, therefore, that when their lungs are functional, the blood contains more oxygen than is required in many circumstances. We may thus ask whether selective distribution is possibly related to vital needs other than oxygen requirements.

Another point which escapes explanation at the moment is the significance of a hydrodynamic arrangement whereby there is a slow and volumetrically smaller circulation through the general body tissues, while there is a rapid and volumetrically greater flow in the pulmonary circuit. Moreover, a proportion of the blood circulates two or more times through the lungs before being sent to the rest of the body. Is it purely an accident arising from the design of the blood vascular system? This seems unlikely. Pure accidents in physiological design are seldom if ever encountered in nature.

Further research must take a wider form in order to clarify the position. It should include investigations not only into the actual distributional patterns observed, but also into the mechanism whereby it is achieved, into the volume-flow relationships in the different blood vessels and vascular beds, and into the oxygen requirements of the tissues in different seasons.

If the outcome of the present investigations is not proved by further work to be merely an eccentricity of Xenopus laevis, it will require a revision of the generally accepted principles of the functioning and physiological design of the blood vascular system of the Anura, and of other vertebrates with imperfectly divided hearts.

I must express my gratitude and indebtedness to: Prof. B. J. Krijgsman, under whose supervision the work was carried out, for constant encouragement and advice; Prof. J. H. Day, for his interest in the project and for a number of suggestions; Prof. R. Goetz, of the Surgical Research Department, for the loan of the photokymograph and other assistance; Mr R. L. Liversidge, for the use of the cinematographic equipment and for much help in the construction of experimental apparatus ; the Technical Staff of the Physics Department for the construction of the optical manometers and for other services; Dr N. Millard, for the loan of a microscope projection lamp; and Mr T. Stafford-Smith, for the loan of a parallax adjustor for the ciné camera.

Male Xenopus. No anaesthetic administered. Heart rate 40/min. 0·025 ml 1% fluorescein-Ringer injected into a hepatic vein. Filming: i6/sec. at/3·5. Three subsequent heart beats are shown. About every fifth frame printed. Ventricular systole starts at 3, 9, 12. Ventricular diastole starts at 1, 6, 11, 14. C. carotid arches; CA. bulbus cordis; H.V. hepatic vein; L. lung; LA. left atrium; P.C. pulmo-cutaneous arch; P.V. pulmonary vein; RA. right atrium; S. systemic arch; V. ventricle; V.C. ventral chamber of truncus; Vv. valves IA and III (see Text-fig. 2, p. 146, valves at distal end of bulbus).

Male Xenopus. 0·6 ml. urethane as anaesthetic. Heart rate 45/min. 0-12 ml. 1 % fluorescein-Ringer injected into a pulmonary vein. Filming: 8/sec. at/4·0. Two subsequent heart beats are shown. Every fourth frame printed. Ventricular systole starts at 4 and 16, ventricular diastole starts at 1 and 13.