1. The anatomy of the blood system of Sabella pavonina is described.

  2. Differences from Sabella found in the structure of the blood system of the closely related Spirographis spallanzanii are pointed out.

  3. Very young Spirographis show no difference in the structure of the blood system from Sabella. Certain of the differences are, however, found in slightly larger specimens of Spirographis which are smaller than fully grown Sabella. The differences are therefore regarded as specifically significant.

  4. A preliminary account is given of the histology of the blood-vessels. The main longitudinal vessels show a distinct muscular layer, which is absent in circular and capillary vessels.

  5. The course of the circulation is described. An explanation of the mechanism of the sorting of oxygenated and deoxygenated blood is put forward.

  6. Peristalsis in the main longitudinal vessels occasionally reverses its direction ; there is thus no physiological polarization of the direction of peristalsis.

  7. The occurrence of blind-ending vascular structures in other polychaetes is reviewed and suggestions put forward concerning their function.

The respiratory blood-pigment chlorocruorin is found only in sabellid, serpulid, and chlorhaemid polyehaete worms. The peculiarities and function of this pigment, together with the physiology of respiration and blood circulation in sabellids, have been studied by Fox and his collaborators (Fox, 1926, 1932, 1933, 1934, 1938 ; Roche and Fox, 1933 ; R. F. Ewer and Fox, 1940). Before further work can be done on these lines it is desirable to have a more detailed knowledge than hitherto of the anatomy of the blood system in the animals concerned. It was for this reason that the present investigation was undertaken. A study has been made of the anatomy of the blood system of Sabella pavonina Savigny, and this has been compared with that of Spirographis spallanzanii Viviani. The latter species is so closely related to the former that Claparède (1868-71, p. 418) remarked: ‘les jeunes Spirographes sont done de vraies Sabelles.’

The anatomy of the blood system of Sabella was studied by Milne Edwards (1838). He described the arrangement of blood-vessels which could be seen by means of a dorsal dissection. In the same year Grube (1838) gave an account of the anatomy of the blood system of Spirographis. His description was based on the observation of living material and the dissection of fixed specimens. He made accurate observations on the movement of the blood in the crown. Both authors described a dorsal vessel running over the surface of the gut, though Grube was very doubtful about the length of this vessel. They both described the ventral vessel with the curious lateral eoilings of the paired ring vessels which lead out of it in each segment. Grube described the vessels of the thorax, and, with particular accuracy, the two longitudinal lateral vessels in the abdomen.

In 1873 Clarapède redescribed the anatomy of the blood system of Spirographis. He appears to have used only fine hand-cut sections. He was able to show that the earlier authors had been wrong in describing a dorsal vessel over the surface of the gut. He showed that the gut is surrounded by a sinus, replaced anteriorly by a plexus over the oesophagus. He stated that not only the vessels to the crown but also the ventral vessel arise from this plexus.

Jaquet (1886) re-examined the blood system of Spirographis. He rediscovered the paired lateral vessels which had been missed by Claparède. In his description of the thorax he followed that of Claparède in every detail.

Meyer (1888) once again described the blood system of Spirographis. He figured and described only the main vessels. He confirmed’ the existence of a gut sinus, and showed that over the oesophagus the sinus concentrated to form a dorsal vessel which carried the greater part of the blood from the gut sinus to the crown. The blood from the crown returned by definite circum-oesophageal vessels to the ventral vessel. Meyer does not draw or describe the plexus found by Claparède around the oesophagus, but it is apparent from his general description that he was aware that it was there.

De Saint-Joseph (1894) published a description of Sabella in which he confirmed the presence of a gut sinus and also of lateral vessels. He did not describe the thorax.

There have been a number of descriptions of the structure of the crown filaments, which have been reviewed by Nicol (1930), whose paper includes an accurate description of the structure of the vessels of the filaments and pinnules of Sabella.

The accounts of the blood systems of the animals described as Sabella alveolata by Williams (1851) and of Sabella arenilega by Cosmovici (1879) have sometimes been compared with descriptions of Sabella pavonina. They refer, however, to Sabellaria alveolata (L.) and Branchi o mm a vesiculosum (Montagu), respectively.

Sabella was sent to Birmingham from Plymouth, and from Millport in Scotland, while Spirographis was collected at Roscoff in Brittany.1

Paraffin sections cut to a thickness of 10 and 20/z were used. Material was fixed in Bouin-Hollande, Duboscq-Brazil, Schaudinn, Champy, Flemming, Hermann, and Camoy. The fixative described by Keilin (1920) was also used. Of these the best results were obtained with Duboscq-Brazil for Sabella and with Keilin’s fixative for Spirographis. Sections were stained in Heidenhain’s iron haematoxylin, Delafield’s haematoxylin and orange G or eosin, Mallory’s triple stain, and Mann’s methyl blue and eosin.

A modification of the method of staining blood by the benzidine reaction described by Pickworth (1934) was used. Whole worms were fixed in a 4 per cent, solution of formaldehyde in sea-water for two days and then washed repeatedly in sea-water to remove the formaldehyde. The material was afterwards stored in sea-water saturated with thymol. Material thus preserved has retained its staining power for at least a year. When required the material was washed in running water and then sectioned on a freezing microtome and stained as described by Pickworth. Thick sections of 100-250 μ. were used.

Small worms were stained by the modification of the benzidine technique described by Slonimsky (1927). Living worms were first washed in tap water before being treated with the reagents, as sea-water produces a yellow incrustation. This method has the disadvantage of being useful only for small and living specimens. The reagents will not penetrate into large specimens.

Small living worms were used to observe the directions of flow of the blood through the transparent tissues.

The blood system of Sabella will be described in detail and those differences which are found in Spirographis will then be given.

The body of Sabella is divided into two regions ; anteriorly the head and thorax, composed of a variable number of segments from six to twelve, and posteriorly the abdomen, with segments up to the number of about three hundred. In describing the anatomy of the blood system it is convenient to start with the abdomen.

1. The Abdomen

The general organization of the blood system in the abdomen may be seen by an examination of Text-figs. 1 and 2.

Text-fig. 1.

Sabella pavonina. Diagram of the vessels in the abdominal segments seen from the posterior aspect. The body-wall and septa of the right side, and the mesenteries, have been cut away. Capillaries are omitted, d.m.b., dorsal muscle block; g.s., gut sinus; l.c., lateral connective vessel; l.v,, lateral vessel; neu.g., neuropodial gland; neu.v., neuropodial vessel ; not.g., notopodial gland ; not.v., notopodial vessel ; r,v., ring vessel ; s.d.v., segmental dorsal vessel ; I.S., trans-septal vessel ; v.g.s., ventral gland shield ; v.g.s.v., ventral gland-shield vessel ; v.m.b., ventral muscle block ; v.v., ventral vessel

Text-fig. 1.

Sabella pavonina. Diagram of the vessels in the abdominal segments seen from the posterior aspect. The body-wall and septa of the right side, and the mesenteries, have been cut away. Capillaries are omitted, d.m.b., dorsal muscle block; g.s., gut sinus; l.c., lateral connective vessel; l.v,, lateral vessel; neu.g., neuropodial gland; neu.v., neuropodial vessel ; not.g., notopodial gland ; not.v., notopodial vessel ; r,v., ring vessel ; s.d.v., segmental dorsal vessel ; I.S., trans-septal vessel ; v.g.s., ventral gland shield ; v.g.s.v., ventral gland-shield vessel ; v.m.b., ventral muscle block ; v.v., ventral vessel

Text-fig. 2.

Sabella pavonina. Diagrammatic section of an abdominal segment seen from the posterior aspect. The vessels in the muscles and glands of the right side are shown, c. capillaries joining the vessels of the neuropodial and ventral glands ; d.m.b., dorsal muscle block; l.v., lateral vessel ; WIJJ., neuropodial gland ; neu.v., neuropodial vessel ; not.c., notopodial capillaries; not.g., notopodial gland; nol.v., notopodial vessel; s.d.v., segmental dorsal vessel; v.g.s., ventral gland shield; v.g.s.v., ventral gland-shield vessel; v.m.b., ventral muscle block

Text-fig. 2.

Sabella pavonina. Diagrammatic section of an abdominal segment seen from the posterior aspect. The vessels in the muscles and glands of the right side are shown, c. capillaries joining the vessels of the neuropodial and ventral glands ; d.m.b., dorsal muscle block; l.v., lateral vessel ; WIJJ., neuropodial gland ; neu.v., neuropodial vessel ; not.c., notopodial capillaries; not.g., notopodial gland; nol.v., notopodial vessel; s.d.v., segmental dorsal vessel; v.g.s., ventral gland shield; v.g.s.v., ventral gland-shield vessel; v.m.b., ventral muscle block

The vascular system in the abdomen is organized into (a) the longitudinal vessels—gut sinus, paired lateral vessels and ventral vessel—and (b) the circular vessels of each segment—ring vessels, lateral connective vessels, segmental dorsal vessels, notopodial vessels, neuropodial vessels, trans-septal vessels, and ventral gland-shield vessels.

The Gut Sinus (or peri-intestinal sinus)

In a freshly dissected specimen or in a small transparent worm the gut sinus is immediately recognizable (Text-fig. 1, g.s.). The blood can be seen to be carried forward by the peristalsis of the sinus wall. The sinus runs the whole length of the worm from the hindmost segment to the anterior septum of segment III. In each segment it receives a pair of ring vessels (Text-fig. 1, r.v.) which come into it ventro-laterally.

The Ventral Vessel

The ventral vessel (Text-fig. 1, v.v.) runs backwards along the whole length of the worm from the posterior portion of segment II. It lies in the ventral mesentery above the giant fibres and nerve-cord (Text-fig. 3, v.v.). The vessel is circular in cross-section and may reach a diameter of 100μ. It is covered with black cells (Text-fig. 3, Ac.).1 From the ventral vessel arise a pair of ring vessels as lateral branches immediately in front of each septum (Text-fig. 1, r.v. ; Text-fig. 3, r.v.).

Text-fig. 3.

Sabella pavonina. Section showing the relations of the ventral vessel. Blood-vessels shaded grey, b.c., black cells; g.f., giant fibres; n.c., nerve-cord; r.v., ring vessel; v.g., ventral groove; v.g.s., ventral gland shield ; v.m., ventral mesentery ; v.m.b., ventral muscle block ; v.v., ventral vessel

Text-fig. 3.

Sabella pavonina. Section showing the relations of the ventral vessel. Blood-vessels shaded grey, b.c., black cells; g.f., giant fibres; n.c., nerve-cord; r.v., ring vessel; v.g., ventral groove; v.g.s., ventral gland shield ; v.m., ventral mesentery ; v.m.b., ventral muscle block ; v.v., ventral vessel

The Lateral Vessels

These vessels (Text-fig. 1, l.v.) can be seen through the skin in small living worms as two zigzag lines running along the whole length of the body in a latero-dorsal position. They run forward as far as the posterior septum of segment II. The lateral vessels are joined to the ring vessels in each segment by two lateral connective vessels (Text-fig. 1, l.c.). On each side of each segment the lateral vessels give off two blind-ending branches—the segmental dorsal and the notopodial vessels (Text-fig. 1, s.d.v. and not.v.).

If one of the lateral vessels is followed forwards along the dorsal muscle block beneath which it runs, its relations may be understood. Immediately after passing through a septum the lateral vessel is joined by the lateral connective vessel (Text-fig. 1, l.c.) and then runs forwards and obliquely upwards. A short distance farther on it gives off the segmental dorsal vessel (Text-fig. 1, s.d.v.) which runs beneath the dorsal muscle block towards the mid-dorsal line, ending there blindly. The lateral vessel then changes its course and runs forwards and obliquely downwards. Shortly before it reaches the anterior septum of the segment it gives off the notopodial vessel (Text-fig. 1, not.v.) which runs downwards beneath the dorsal muscle block to the coelomic pouch of the notopodium (Text-fig. 2, not.c.). After giving off the notopodial vessel the lateral vessel again changes its direction and, running forwards and obliquely upwards, it passes through the septum to join the lateral connective vessel of the next segment.

Along its course the lateral vessel gives off blind-ending capillaries which project into the coelom (Text-fig. 2, l.v.). The vessel may be as much as 80μ in diameter.

The Ring Vessels

The two ring vessels (Text-fig. 1, r.v.) of each segment arise laterally from the ventral vessel immediately in front of each septum and make a characteristic S-shaped bend at their origin. Each of them runs along the ventral muscle block towards the mouth of the segmental organ. Although close to the septum it is not attached to it in this region. The S-shaped bend is covered with black cells continuous with those covering the ventral vessel. On reaching the segmental organ the ring vessel passes between the coelomostome and the septum, turns round, and runs obliquely upwards attached to the anterior face of the septum to join the gut sinus in a ventro-lateral position. About three-quarters of the way from the segmental organ along the upper Emb of each ring vessel there arises the lateral connective vessel (Text-fig. 1, l.c.) which runs upwards and obliquely outwards on the septum to join the lateral vessel of that side (Text-fig. 1, l.v.).

In many segments there are additional anastomoses on each side between the lateral connective vessel and the dorsal limb of the ring vessel, or redupEcations of the latter. The form of these additional vessels is very variable and they are of irregular distribution. They do not necessarily occur on both sides of the same segment. A number of these structures is shown in Text-fig. 4.

Text-fig. 4.

Sabella pavonina. Diagrams of four different types of anomalous organization of the ring vessels, g.s., gut sinus ; l.c., lateral connective vessel; l.v., lateral vessel; r.v., ring vessel

Text-fig. 4.

Sabella pavonina. Diagrams of four different types of anomalous organization of the ring vessels, g.s., gut sinus ; l.c., lateral connective vessel; l.v., lateral vessel; r.v., ring vessel

The ring vessels may reach a diameter of 100 μ.

The Segmental Dorsal Vessels

The two segmental dorsal vessels of each segment (Text-fig. 1, s.d.v.) arise from the lateral vessels as described above. The vessels curve upwards beneath the dorsal muscle blocks. Along their length blindending capillaries arise and project into the coelom (Text-fig. 2, s.d.v.). The segmental dorsal vessels are blind-ending ; each runs as far as the dorsal mesentery, but never joins the corresponding vessels of the opposite side. The vessels may branch towards their end. They reach a diameter of 40 μ.

The Notopodial Vessels

The abdominal parapodia have dorsal uncini and ventral chaetae. Each parapodium has two parapodial glands (Text-figs. 1 and 2, not.g. and neu.g.), the neuropodial gland lying beneath the chaetal bundle while the notopodial gland surrounds the uncinal papilla of the notopodium on its ventral and anterior sides. The papilla itself encloses a pouch of coelom which is filled with blind-ending capillaries (Text-fig. 2, not.c.). These capillaries are supplied by the notopodial vessel (Text-fig. 2, not.v.) which arises from the lateral vessel of that side as described above. The notopodial vessel curves ventrally beneath the dorsal muscle block and then runs upwards into the coelomic pouch of the notopodium. Blind-ending capillaries which project into the coelom are given off along the length of the vessel.

These vessels may reach a diameter of 30 μ.

The Ventral Gland Shield and Trans -sept al Vessels

Along the ventral surface is a double series of segmentally arranged rectangular blocks of glands known as ventral gland shields (Text-figs. 1 and 2, v.g.s.). These are liberally supplied with blind-ending capillaries.

From the ring vessel on each side of each segment, as it passes from the ventral vessel to the segmental organ, is given off a short vessel—the trans-septal vessel—which runs backwards through the septum along the surface of the ventral muscle block (Text-fig. 1, t.s.). Behind the septum the trans-septal vessel forks to form the neuropodial vessel (Text-fig. 1, neu.v.) and the ventral gland-shield vessel (Text-fig. 1, v.g.s.v.).

The ventral gland-shield vessel runs from the fork of the trans-septal vessel transversely across the surface of the ventral muscle block towards the midline. About two-thirds of the way from the parapodium across the muscle block the vessel turns ventrally and runs down through the muscle block (Text-fig. 2, v.g.s.v.). In the ventral portion of the muscle block the vessel divides into a number of branches which ramify in the ventral gland shield. These branches lead to blind-ending capillaries which can be seen through the skin of the worm.

The ventral gland-shield vessel as it runs across the ventral muscle block gives off blind-ending capillaries which project into the coelom ; there is a large clump of them where the vessel turns down to run into the muscle (Text-fig. 2).

The Neuropodial Vessels

From the fork of the trans-septal vessel, the neuropodial vessel (Text-fig. 1, neu.v.) runs outwards for a short distance and then turns posteriorly and runs along the inner side of the neuropodial gland. Along its length it gives off blind-ending capillaries which ramify among the muscles of the chaetal bundle and in the neuropodial gland (Text-fig. 2, neu.v.’).

In each segment vessels run across from the ventral gland shield of each side beneath the ventral surface of the ventral muscle block to the neuropodial gland, joining branches of the neuropodial vessel (Text-fig. 2, c.).

The Blind-ending Capillaries

These capillaries arise from all the vessels in the abdomen except the ring vessels, the lateral connective vessels, and the gut sinus. They may have a diameter up to 30p.

The Terminal Segments

The general form of organization described in the abdominal region is found as far back as the last chaetigerous segment. In this region the individual main vessels become smaller in diameter and the gut sinus is not so well developed. Capillaries are not found except in the ventral gland shields, where they are few in number but no smaller in diameter than any of the other vessels of this region.

The terminal projection on which the anus is situated is devoid of blood supply.

2. The Head and Thorax

The general organization of the main vessels in the head and the thoracic segments is shown in Text-figs. 5 and 6. It is convenient to describe the thorax first and then the head.

Text-fig. 5.

Sabella pavonina. Diagram of the vessels of the thorax seen in lateral view. The anterior end of the worm is to the left. The body wall and septa of the left side, and the mesenteries, have been removed. Capillaries are not shown g except around the oesophagus. The segments are numbered in Roman numerals. l>.v., branchial vessel; c.o.v., circumoesophageal vessel ; c.v., collecting vessel of the peri-oesophageal plexus ; d.v., dorsal vessel; g.s., gut sinus; I.e., lateral connective vessel; l.d.v., latero-dorsal vessel; l.l.v., vessel supplying the lateral lip; l.l.v.1, vessel supplying the lateral lip, ventral sac, and ventral collar fold; l.v., lateral vessel; neu.v., neuropodial vessel; not.v., notopodial vessel; p.o.p., peri-oesophageal plexus; r.v., ring vessel; s.d.v., segmental dorsal vessel; s.m.v., vessel supplying the chaetal muscles of segment II; l.v., transverse vessel; v.g.s.v., ventral gland-shield vessel; v.v., ventral vessel

Text-fig. 5.

Sabella pavonina. Diagram of the vessels of the thorax seen in lateral view. The anterior end of the worm is to the left. The body wall and septa of the left side, and the mesenteries, have been removed. Capillaries are not shown g except around the oesophagus. The segments are numbered in Roman numerals. l>.v., branchial vessel; c.o.v., circumoesophageal vessel ; c.v., collecting vessel of the peri-oesophageal plexus ; d.v., dorsal vessel; g.s., gut sinus; I.e., lateral connective vessel; l.d.v., latero-dorsal vessel; l.l.v., vessel supplying the lateral lip; l.l.v.1, vessel supplying the lateral lip, ventral sac, and ventral collar fold; l.v., lateral vessel; neu.v., neuropodial vessel; not.v., notopodial vessel; p.o.p., peri-oesophageal plexus; r.v., ring vessel; s.d.v., segmental dorsal vessel; s.m.v., vessel supplying the chaetal muscles of segment II; l.v., transverse vessel; v.g.s.v., ventral gland-shield vessel; v.v., ventral vessel

Text-fig. 6.

Sabella pavonina. Diagram of the vascular system seen in a ventral view of the head and thorax, b.v., branchial vessel; b.ves., branchial vesicle; c.f.c., collar-fold capillaries ; f.v., filament vessel; lateral lip vessels; p.v., palp vessel; pn.v., pinnule vessel; v.s., ventral sac

Text-fig. 6.

Sabella pavonina. Diagram of the vascular system seen in a ventral view of the head and thorax, b.v., branchial vessel; b.ves., branchial vesicle; c.f.c., collar-fold capillaries ; f.v., filament vessel; lateral lip vessels; p.v., palp vessel; pn.v., pinnule vessel; v.s., ventral sac

(a) The Thorax

The thoracic chaetigerous segments (the first of which is segment II) are distinguished from the abdominal segments by the inversion of the position of the uncini and chaetae of the parapodia : in the thorax the chaetae are dorsal and the uncini are ventral. There are no uncini on the first chaetigerous segment. In this region the basic plan of the blood system already described for the abdomen is found except in segment II.

The Gut Sinus, Peri-oesophageal Plexus, and Latero-dorsal Vessels

The gut sinus runs anteriorly from the abdominal region as far as the anterior septum of segment III (Text-fig. 5, g.s.). In segment II the gut sinus is replaced by a plexus—the peri-oesophageal plexus—which lies on the dorso-lateral and lateral surfaces of the oesophagus in this segment (Text-fig. 5, p.o.p.). Above this plexus run two latero-dorsal vessels (Text-fig. 5, l.d.v.) which unite at the anterior margin of segment II to form a single dorsal vessel which runs forwards in segment I (peristomium) over the dorsal surface of the oesophagus (Text-fig. 5, d.v.). The peri-oesophageal plexus runs forwards and downwards through segments II and I. The transition of gut sinus to plexus corresponds exactly with the histological transition from stomach to oesophagus described by Nicol (1930).

The Ventral Vessel

The ventral vessel starts in the thorax at the boundary between segments II and III. It is formed at this level by the fusion of the ventral limbs of the two circum-oesophageal vessels (Text-fig. 5, c.o.v.). The black cells on the ventral vessel persist as far forwards as segment III.

The Lateral and Lateral Connective Vessels

The two lateral vessels (Text-fig. 5, l.v.) run forwards in the thorax as far as the anterior surface of the septum between segments II and III. From segment VI forwards the vessels may break up-into a number of smaller parallel vessels. The connexion of each lateral vessel with the ring vessels by the lateral connective vessels is found in all segments except the first.

The Ring Vessels

The two ring vessels of each segment are found in normal form as far forwards as segment III. In the majority of individuals the reduplications of the ring vessels are more elaborate than those which have been described in the abdomen.

In segment II the dorsal limb of each ring vessel is absent, the ventral limb connecting the anterior end of the lateral vessel of that side with the ventral vessel by way of the lateral connective (Text-fig. 5, l.c.). This absence of the dorsal limb of the ring vessels of segment II is to be correlated with the absence of the gut sinus in this segment.

The Segmental Dorsal Vessels.—These vessels arise normally as far forwards as segment III (Text-fig. 5, s.d.v.). The vessels are missing in segment II, the lateral vessels, from which these vessels normally arise, ending at the posterior septum of segment II.

The Notopodial Vessels

These vessels arise in the normal manner as far forwards as segment III (Text-fig. 5, not.v.). They are absent in segment II, in correlation with the ending of the lateral vessels. The muscles of the notopodial chaeta of segment II are supplied by a unique vessel which arises from the ring vessel of each side of that segment more laterally than the trans-septal vessel. This vessel runs forward and upwards and breaks up into blind-ending capillaries around the chaetal muscles on their ventral side (Text-fig. 5, s.m.v.).

The Ventral Gland Shield and Neuropodial Vessels

These vessels show their normal form as far forwards as segment III (Text-fig. 5, v.g.s.v. and neu.v.). The system is absent in segment II, which has no neuropodium, and there are no ring vessels in segment I from which neuropodial vessels could arise. The ventral gland shields of segments II and III are supplied by capillaries which run back from the ventral collar fold.

In the abdomen the ciliated groove which runs along the ventral surface of the worm divides the ventral gland shields into two blocks in each segment. At the junction with the thorax this ciliated groove runs round into a dorsal position and so the ventral gland shields in the thorax are in the form of single segmental blocks. This does not, however, alter the arrangement of the capillary supply to the glands.

It will be noted that the inversion of the parapodia in the thoracic region is not reflected in the organization of the blood system.

There is no blood-supply to the large thoracic kidneys.

(b) The Head1

The Dorsal Vessel

The dorsal vessel (Text-fig. 5, d.v.) arises by the fusion of the two latero-dorsal vessels and runs forwards over the surface of the oesophagus as far as the supraoesophageal ganglia, behind which it forks to form a T whose cross limb is the transverse vessel (Text-fig. 5, t.v). The vessel has a diameter up to 80μ.

The Transverse Vessel

The transverse vessel is situated behind the supra-oesophageal ganglia and parallel to their transverse axis. Above the root of the circum-oesophageal nerve commissure of each side it forks to form the branchial vessel which runs forwards and the circum-oesophageal vessel which runs downwards and backwards behind the nervous commissure. The vessel reaches a diameter of 100/r.

The Circum-oesophageal Vessels

These two vessels run downwards and backwards behind the circum-oesophageal nerve commissures (Text-fig. 5, c.o.v.). The two vessels do not join in segment I as do the nerve commissures, but run backwards as far as the posterior portion of segment II where they join to form the ventral vessel.

Shortly after their origin from the transverse vessel each circum-oesophageal vessel gives off a vessel supplying the lateral lip (Text-fig. 5, l.l.v.). Further back in segment I the circumoesophageal vessel of each side gives off a vessel which immediately divides into two branches supplying the lateral lip, the ventral sac, and ventral collar fold (Text-fig. 5, l.l.v.1). In the posterior portion of segment I the circum-oesophageal vessel of each side also receives a vessel which drains the peri-oesophageal plexus (Text-fig. 5, c.v.). The circum-oesophageal vessels reach a diameter of about 100μ.

The Peri-oesophageal Plexus

This replaces the gut sinus in segment II and is continued forward into segment I. The plexus runs anteriorly and downwards to surround the oesophagus on the dorso-lateral and lateral surfaces in segment II and on the lateral and ventral surface in segment I (Text-fig. 5, p.o.p.’). The blood from the plexus is collected ventrally on each side into a vessel which empties into the circumoesophageal vessel (Text-fig. 5, c.v.). At the sides of the oesophagus the plexus extends outwards between the dorsal and ventral muscle blocks of segment I.

The Branchial, Filament, and Pinnule Vessels

The two branchial vessels arise from the transverse vessel as described above and run forwards into the crown (Text-fig. 6, b.v.). At the base of the crown each vessel swells out to form the branchial vesicle (Text-fig. 6, b.ves.). From the inner side of each branchial vessel just anterior to the branchial vesicle arises a single slender vessel which runs to the palp ;* it is blind-ending and has no branches (Text-fig. 6, p.v.).

From the branchial vesicle the branchial vessel of each side runs forwards. It sends off blind-ending vessels to each filament of the crown in turn (Text-fig. 6, f.v.) and ends in supplying the most distal filament.

The filaments of the crown give rise to numerous small appendages, the pinnules. Into each of these the filament vessels send a short blind-ending vessel, the pinnule vessel (Text-fig. 6, pn.v.).

The branchial vesicle may attain a diameter of 100μ, and the branchial vessel 75μ The filament vessel reaches a diameter of 60μ and the pinnule vessel 20μ.

The Vessels of the Lateral Lips

The lateral bps are each supplied with a rich plexus of capillaries arising from the circum-oesophageal vessels as described above. The plexus is in the form of branching and blind-ending capillaries (Text-fig, 6, l.l.v.).

The Vessels of the Ventral Sacs and Ventral Collar Folds

The vascular supply to these organs arises from the circum-oesophageal vessels as described above. The main trunk to the ventral collar fold of each side runs down the posterior border of the collar fold. It also sends branches forwards to the ventral sac and backwards to the ventral gland shields of segments II and III.

All the main vessels which have been described in Sabella occur in identical form in Spirographis. Such differences as there are, are in points of detail. Only these detailed differences will be described here.

The Lateral and Associated Vessels

The two lateral vessels are similar to those described in Sabella. The segmental dorsal vessels, however, branch frequently. Additional vessels also arise from the lateral vessels and run up beneath the dorsal muscle blocks parallel to the segmental dorsal vessels. The notopodial vessels, too, occasionally branch. In a few segments additional vessels coming from the lateral vessels run down to the notopodia.

The Ring Vessels

The reduplications of the ring vessels described in Sabella (Text-fig. 4) are much more highly developed in Spirographis. The anterior face of each septum is covered with a plexus of fine vessels joining not only the dorsal limb of each ring vessel with the corresponding lateral connective vessel, but also the two limbs of the ring vessel. This plexus is an elaboration of the corresponding reduplications in Sabella.

The Ventral Gland-shield Vessels and their Derivatives

Unlike the state of affairs in Sabella there are in the ventral muscle blocks of Spirographis fine capillaries, which run parallel to the longitudinal axis of the worm. These ventral muscle capillaries (Text-fig. 7, v.m.c.) originate in each segment from that portion of the ventral gland-shield vessel which runs across the top of the ventral muscle block. Most of these capillaries are continuous with the corresponding capillaries arising from the ventral gland-shield vessels of the adjoining segments, but some of them which run out laterally in the muscle block are blind-ending.

Text-fig. 7.

Spirographis spallanzanii. Diagram of the vessels associated. with the ventral muscle block. Cf. Sabella, fig. 2. ex., extension of the ventral gland-shield vessel; s.n.p., sub-neural plexus; v.g.s.v., ventral gland-shield vessel; v.m.c., capillaries of the ventral muscle block; v.v., ventral vessel

Text-fig. 7.

Spirographis spallanzanii. Diagram of the vessels associated. with the ventral muscle block. Cf. Sabella, fig. 2. ex., extension of the ventral gland-shield vessel; s.n.p., sub-neural plexus; v.g.s.v., ventral gland-shield vessel; v.m.c., capillaries of the ventral muscle block; v.v., ventral vessel

The plexus of capillaries in each ventral gland shield runs up on the median side of the ventral muscle block and forms a subneural plexus beneath the nerve-cord (Text-fig. 7, s.n.p.). In some segments a prolongation of the horizontal coelomic portion of the ventral gland-shield vessel runs over the ventral muscle block to the midline and then downwards to join the subneural plexus (Text-fig. 7, ex.).

The Thoracic Chaetigerous Segments

The vessels of the thoracic chaetigerous segments are similar to those in the abdominal segments, but the capillaries in the ventral muscle blocks are missing.

The Branchial and Filament Vessels

The vessels in the filaments of the crown arise in pairs from the branchial vessel, instead of singly as in Sabella. Short reduplications of the branchial vessel occur. There is no asymmetry of the Vessels at the base of the crown corresponding to the asymmetry of the latter.

The Neural Plexus

-Above the sub-oesophageal ganglia and the ganglia of segment II there is a plexus of fine vessels, the neural plexus, arising from the circum-oesophageal vessels. This is not found in Sabella.

Spirographis grows to a larger size than Sabella. There is therefore the possibility that the differences which have been noted above are connected with the greater size of Spirographis and are not specific1 differences. The arrangement of the blood-vessels in the abdomen of a number of small specimens of Spirographis was therefore examined.

Small worms, 4–5 cm. in length, showed no differences from Sabella as far as the blood system was concerned. There were no capillaries in the ventral muscle blocks, no reduplication of the ring vessels or of the segmental dorsal vessels.

In larger worms, 6–7 cm. in length, capillaries were found in the ventral muscle blocks, reduplications of the segmental dorsal vessels were observed, but the ring vessels were no more complicated than the condition described for reduplications in Sabella.

Some of the peculiar features of the blood system of large Spirographis are thus also found in these small individuals of 6-7 cm. in length ; these are much smaller than some of the Sabella studied, which were up to 20-30 cm. in length. The differences between the vessels of Spirographis and Sabella are therefore real and not due to the former animal being the larger.

While no attempt has been made at a detailed examination of the histology of the blood-vessels, it is possible to draw attention to differences between individual vessels which would appear to be significant in the physiology of the circulation. The structure of the vessels is the same in Sabella and Spirographis.

Claparède (1873) described the fine structure of the ventral vessel of Spirographis. He recognized a columnar endo-thelium, a single layer of muscle-fibres and an outer layer of peritoneal cells.

The histology of the ventral vessel was reinvestigated by Hescheler (in Lang, 1903, p. 222). He agreed with Claparède’s description, but found in place of a columnar endothelium a flattened endothelium separated from the layer of muscle-fibres by a layer of connective tissue fibres.

I have found that the blood-vessels fall into two clear histological categories: (1) vessels with a muscular coat, and (2) thin-walled vessels.

1. Vessels with a Muscular Coat

These vessels include the ventral vessel and the lateral vessels, as well as the latero-dorsal vessels, the dorsal vessel, the branchial vesicles, and the circum-oesophageal vessels. The walls of these vessels are covered externally by a layer of peritoneal cells beneath which is a thin layer of circular muscle-fibres lying on a structureless layer, on whose inner side are endothelial cells. Text-fig. 8 shows the structure of the ventral vessel.

Text-fig. 8.

Sabella pavonina. Section showing the structure of the wall of the ventral vessel, b.c., black cells ; b.g., black granules shed into the coelom from the black cells ; m.f., muscle-fibres ; n.e.c., nuclei of endothelial cells ; p.c., peritoneal cells ; s.l., structureless layer

Text-fig. 8.

Sabella pavonina. Section showing the structure of the wall of the ventral vessel, b.c., black cells ; b.g., black granules shed into the coelom from the black cells ; m.f., muscle-fibres ; n.e.c., nuclei of endothelial cells ; p.c., peritoneal cells ; s.l., structureless layer

The gut sinus (Text-fig. 9) is a special case of this arrangement. It is covered with a layer of peritoneal cells. Beneath this is a single layer of circular muscle-fibres which is about 10/z thick. This layer is, however, greatly thickened in the region of the junction between the stomach and the intestine (Nicol, 1930) and may there reach a thickness of 100μ Beneath the musclelayer is a structureless layer, inside which are endothelial cells. On the inner wall of the sinus is another layer of endothelial cells lying on the basement membrane of the gut cells. I have never observed cells free in the blood in this region as described by Bomieu (1923) in Spirographis.

Text-fig. 9.

Sabella pavonina. Section of the gut sinus at the level of the junction of the stomach and intestine. The sinus is shaded grey. &c., ‘bride contractile ‘; b.m., basement membrane of the gut cells ; e.c., endothelial cells; g.c., gut cells; m.f., muscle-fibres; p.c., peritoneal cells ; s.l., structureless layer

Text-fig. 9.

Sabella pavonina. Section of the gut sinus at the level of the junction of the stomach and intestine. The sinus is shaded grey. &c., ‘bride contractile ‘; b.m., basement membrane of the gut cells ; e.c., endothelial cells; g.c., gut cells; m.f., muscle-fibres; p.c., peritoneal cells ; s.l., structureless layer

This condition is similar to that described by Lee (1912) for the intestinal sinus of serpulids. Evenkamp (1931) has, however, described a different arrangement in the gut sinus of the sabelline Laonome kroyeri, where he found a layer of circular muscles lying between the basal membrane of the gut cells and the endothelial cells of the inner wall of the sinus.

Across the lumen of the sinus run fine threads of the structureless layer covered by scattered endothelial cells (Text-fig. 9, b.c.). The threads were first described by Claparède (1873) in Myxi cola infundibulum and called ‘brides contractiles’. Similar structures have been described by Faulkner (1930) in the serpulid Filograna implexa, and by Lee (1912) in a number of other serpulids.

It will be noted that all the longitudinal trunks are built on the same principle.

2. Thin-walled Vessels

All the remaining vessels have thin walls containing scattered nuclei but devoid of muscle-fibres. In Spirographis the larger of these vessels have in addition a thin structureless layer lining the lumen; this has not been detected in corresponding vessels in Sabella.

The main features of the circulation of Sabella and Spirographis have been described by Fox (1933, 1938).

Previously in Spirographis Delle Chiaje (quoted in Grube, 1838) regarded the ventral vessel as a vein and the other vessels as arteries, but gave no reasons. Grube (1838) gave a good description of the reversible flow of the blood in the crown. He considered that the blood flowed forwards in the lateral vessels to the crown and backwards from the crown, by way of the transverse vessel, into the dorsal vessel.

Claparède (1873) was of the opinion that all the blood which was carried forward by the gut sinus flowed into the perioesophageal plexus. From this plexus the blood flowed into the crown and then back again into the plexus which thus contained mixed blood. Some of the blood from the plexus flowed into the ventral vessel and into the vessels supplying the ventral collar folds. Claparède speaks of the possibility of some type of sorting mechanism in the plexus, but, he says, ‘L’enchevêtrement des vaisseaux dans le plexus est trop considérable pour qu’on puisse espérer une solution à cette question’ (p. 82). It will be seen that this mistaken interpretation of the anatomy of the thoracic vessels made the sorting mechanism of the oxygenated and de-oxygenated blood seem very complex.

Fox (1933, 1938) gives the following description. Almost all the blood-vessels, both trunks and blind-ending capillaries, contract rhythmically. Blood is forced out of the capillaries and of the vessels in the crown by centripetal contractions of the walls of these vessels ; then after a short interval blood from the continuous trunks flows back again into the capillaries. In the continuous trunks blood is moved along by peristalsis. Blood is propelled forwards in the gut sinus, from which it flows into the dorsal vessel and circum-oesophageal vessels, thus reaching the ventral vessel, through which it is forced backwards (Text-fig. 10 A). In the abdomen blood moves from the ventral vessel into the ring vessels of each segment, through which part of the blood reaches the gut sinus, and part, by way of the lateral connectives, reaches the lateral vessels (Text-fig. 10 B). It is then moved forwards along the latter as discrete trains of blood starting from the posterior end of the worm, the vessels being closed for some distance behind that section which is full of blood. The notopodial vessels, segmental dorsal vessels, and capillaries of each segment fill as the trains of blood in the lateral vessels reach them, and empty again when the train of blood passes on. There is no such regularity in the emptying and filling of the gland-shield capillaries. The rates of rhythmical contractions of the various vessels differ; for Sabella at 19° C. the following times in seconds between contractions are given ; pinnule, filament, and branchial vessels 10·1, lateral vessels 8·4, ventral vessel 5·5, gut sinus 2·8. The contractions are not synchronous in the two sides of the crown or in the two lateral vessels.

Text-fig. 10.

A. Sabella pavonina. Diagram of the outlines of the vessels seen in fig. 1. The direction of the blood-flow is indicated by arrows.

Vessels in the abdominal segments seen from the posterior aspect (cf. fig. 1). g.s., gut sinus; l.c., lateral connective vessel; l.v., lateral vessel; neu.v., neuropodial vessel; not.v., notopodial vessel; r.v., ring vessel; s.d.v., segmental dorsal vessel; t-s.v., trans-septal vessel; v.g.s.v., ventral gland-shield vessel; v.v., ventral vessel.

Text-fig. 10.

A. Sabella pavonina. Diagram of the outlines of the vessels seen in fig. 1. The direction of the blood-flow is indicated by arrows.

Vessels in the abdominal segments seen from the posterior aspect (cf. fig. 1). g.s., gut sinus; l.c., lateral connective vessel; l.v., lateral vessel; neu.v., neuropodial vessel; not.v., notopodial vessel; r.v., ring vessel; s.d.v., segmental dorsal vessel; t-s.v., trans-septal vessel; v.g.s.v., ventral gland-shield vessel; v.v., ventral vessel.

TEXT-FIG. 10B.

Sabella pavonina. Diagram of the outline of the vessels seen in fig. 5. The direction of the blood-flow is indicated by arrows.

Vessels in the head and thorax, the anterior end of the worm being to the left (cf. fig. 5). b.v., branchial vessel; c.o.v., circumoesophageal vessel; c.v., collecting vessel of the peri-oesophageal plexus; d.v., dorsal vessel; g.s., gut sinus; l.c., lateral connective vessel; l.d.v., latero-dorsal vessel; l.l.v., vessel supplying the lateral lip; l.l.v.1, vessel supplying the lateral lip, ventral sac, and ventral collar fold; l.v., lateral vessel; neu.v., neuropodial vessel; not.v., notopodial vessel; p.o.p., peri-oesophageal plexus; r.v., ring vessel ; s.d.v., segmental dorsal vessel ; s.m.v., vessel supplying the chaetal muscles of segment II ; t.v., transverse vessel ; v.g.s.v., ventral gland-shield vessel; v.v., ventral vessel.

TEXT-FIG. 10B.

Sabella pavonina. Diagram of the outline of the vessels seen in fig. 5. The direction of the blood-flow is indicated by arrows.

Vessels in the head and thorax, the anterior end of the worm being to the left (cf. fig. 5). b.v., branchial vessel; c.o.v., circumoesophageal vessel; c.v., collecting vessel of the peri-oesophageal plexus; d.v., dorsal vessel; g.s., gut sinus; l.c., lateral connective vessel; l.d.v., latero-dorsal vessel; l.l.v., vessel supplying the lateral lip; l.l.v.1, vessel supplying the lateral lip, ventral sac, and ventral collar fold; l.v., lateral vessel; neu.v., neuropodial vessel; not.v., notopodial vessel; p.o.p., peri-oesophageal plexus; r.v., ring vessel ; s.d.v., segmental dorsal vessel ; s.m.v., vessel supplying the chaetal muscles of segment II ; t.v., transverse vessel ; v.g.s.v., ventral gland-shield vessel; v.v., ventral vessel.

Fox (1933, p. 482) described the coelomic capillaries in a given segment as all contracting at the same time, the contractions in one segment being one or more seconds behind those in the segment immediately posterior to it. It appears to me that this is probably not due to a spontaneous rhythm of capillary contraction but to the fact that many of the coelomic capillaries arise from the lateral vessel or its branches and fill periodically as the rhythmic trains of blood pass forwards along the lateral vessel.

I have made the following further observations on the circulation of Sabella. As in the lateral vessels, so too in the ventral vessels separate trains of blood are carried along by peristalsis.

The neuropodial capillaries fill more frequently than the ventral gland-shield capillaries and blood passes across the connecting capillaries to the ventral gland-shield vessels by peristalsis. There is, however, no definite circulation from the neuropodial to the ventral gland-shield vessels, for occasionally blood may be seen flowing in the opposite direction.

In segments I and II the blood-flow is as follows (Textfig. 10 B). At the anterior end of segment III the blood from the gut sinus passes into the two latero-dorsal vessels and into the peri-oesophageal plexus on each side ; from the peri-oesophageal plexus it flows into the circum-oesophageal vessels. At 13-5° C. the following average periods of contractions in seconds were found: gut sinus in thorax 9-1, latero-dorsal vessels and perioesophageal plexus 12-9. The greater frequency of the gut sinus contractions causes a certain excess of blood to flow back from the level of the peri-oesophageal plexus into the gut sinus after each peristaltic wave of the sinus reaches the anterior end.

From the latero-dorsal vessels the blood is carried into the dorsal vessel and thence into the transverse vessel. Some of the blood then flows into the crown along the branchial vessels and some runs directly into the circum-oesophageal vessels. The blood which flows into the crown is held in the blind-ending filament and pinnule vessels for a short time and is then driven back into the transverse vessel. This arrangement might result in the same portion of blood flowing into the crown, back to the transverse vessel, and into the crown again. There is no valvular arrangement in the vessels which could separate the aerated from the unaerated blood. If, however, the contraction frequencies of the vessels are determined the method of separation becomes clear. At 14° C. the following average frequencies of contraction were found: branchial vessel 16·6 seconds, circum-oesophageal vessels 11·4 seconds. Fox (1933, p. 481) has shown that the vessels of the crown remain empty for just over half the period of the cycle of their contraction. As already stated, I have found that the blood expelled from the branchial vessels is returned to the transverse vessel. At the frequency of contraction of the branchial vessels at 14° C. the blood would remain in the transverse vessel for about 10 seconds. If one of the circum-oesophageal vessels contracts during this period the blood will be carried away to the ventral vessel. The rates of contraction of the branchial and circum-oesophageal vessels given above show that this will happen at least for every alternate contraction of the branchial vessel of each side.

While the gut sinus may be regarded as carrying deoxygenated blood, it is clear that the blood in the ventral vessel is ‘mixed’, for some of it is derived from the peri-oesophageal plexus and the anterior end of the lateral vessels. Moreover, the contraction frequency of the circum-oesophageal vessels implies that some of the blood from the dorsal vessel is carried directly to the ventral vessel without flowing into the crown.

Some of the blood in the circum-oesophageal vessels runs into the capillaries in the lateral lips. These capillaries fill at each contraction of the circum-oesophageal vessels and empty again rapidly. There is, however, no such periodicity recognizable in the capillaries of the ventral collar folds and ventral gland shields of segments II and III.

When the blood flowing forwards in the lateral vessels reaches segment II it must run downwards through the lateral connective and ring vessels of that segment and so into the ventral vessel. This could not be seen, however, nor could it be ascertained whether blood runs upwards in all the thoracic ring vessels as it does in those of the abdomen, or whether the direction is reversed in the anterior lateral connective vessels and ring vessels, thus helping to empty the lateral vessels.

In small worms it is possible to see that the normal direction of circulation in the lateral and ventral vessels is reversible. This shows that there is no ‘polarization’ of the blood-vessels of Sabella as Carlson (1908) has postulated to account for the constancy of the direction of flow of blood in the dorsal vessel of Nereis.

Blind-ending vessels are one of the peculiarities of the blood system of sabellids. Similar vascular structures occur in a number of other polychaetes. Excluding blind-ending vessels in gills, palps, and tentacles, they may be roughly divided into four categories.

1. Blind-ending Vessels arising from the Ring Vessels

Allen (1904) has described bunches of contractile blind-ending vessels arising from the ring vessels of the disomid Poecilochaetus serpens Allen. Müller (1858) and McIntosh (1878) have found contractile sac-like structures in a similar position in Magelona papillicornis F. Müll. Huxley (1883) and Faulkner (1930) have shown the same type of structure in the serpulid Filograna implexa Berkeley. A bulb-like structure in similar position was described by Grube (1838) in Eunice harassii Audouin and Milne Edwards.

These structures might act as accessory hearts, serving to drive the blood around the ring vessels. It is to be noted that in Poecilochaetus Allen (1904) described the hearts as blind-ending segmental pouches arising from the dorsal vessel and lying over the intestine.

2. Blind-ending Capillaries Associated with the Segmental Organs

Fuchs (1907) described blind-ending vessels in the eunicid Marphysa sanguinea (Montagu). I have examined this form at Roscoff and find that these vessels are limited to the vessel supplying the segmental organ. In the terebellid Polymnia nebulosa (Montagu), Jaquet (1886) described blind-ending vessels associated with the anterior segmental organs. In La ni ce conchilega (Pallas) a similar condition is described by Meyer (1888). I have observed these vessels to contract.

In the family Arenicolidae Gamble and Ashworth (1900) described short blind-ending vessels around the coelomostome of the segmental organ of Arenicola grubii Clap, and Arenicola ecaudata Johnston. Benham (1891) described similar vessels in Arenicola marina (L.) and Jaquet (1886) in Arenicola claparedii Levinsen. In the closely related Ophelidae I have found in Travesia forbsii Johnston a blind-ending vessel running the length of the segmental organ and giving off short blind-ending capillaries. I have never observed these vessels to contract. A similar vessel going to the segmental organ was described by Schaeppi (1894) in Ophelia radia ta Delle Chiaje. This vessel itself gives off long blindending capillaries. On the ring vessels of this species there are also blind-ending capillaries which are, however, much finer than the vessels described in 1 above. It is of interest to note that Brown (1938) failed to find any blind-ending capillaries in Ophelia cluthensis McGuire.

Among the Amphictenidae similar short capillaries associated with the segmental organs were described by Rathke (1842) in Pectenaria (Amphictene) auricoma (0. F. Mull) and by Cosmovici (1879) in Pectenaria (Lagis) koreni Malm.

In the absence of any definite knowledge even about the contractility of the majority of these vessels it does not seem advisable to speculate on their function.

3. Blind-ending Vessels Supplying the Gonads.1

Blind-ending capillaries arising from the ring vessels and running into the mass of the gonads were described both in Nephthys hombergii Audouin and Milne Edwards and in the chlorhaemid Flabelligera diplochaitos Otto by Jaquet (1886).

Zürcher (1909) described blind-ending ampullae arising laterally in pairs from the ventral vessel of Owenia fusif or mis Delie Chiaje. These ampullae were observed to contract by von Drasche (1885). The gonads are attached to the walls of these ampullae. In Chaetopterus variopedatus (Renier) Probst (1929) has described blind-ending vessels arising from the ventral vessel and running to the gonads. In the anterior region of the worm he also described blind-ending ampullae arising from the ventral vessel. These he regards as homologous with the hearts of the complete anterior ring vessels described by Claparède (1873) in Telepsavus costarum Clap.

While it is probable that these vessels serve to carry oxygen and nutrients to the developing gonads, the reason for their ending blindly will not be understood until the mechanics of the circulation in these forms has been further elucidated.

4. Blind-ending Vessels generally Distributed

Such an arrangement of vessels has been reported in members of two of the three sub-families of the Sabellidae. Among the sabellines they occur in Sabella, Spirographis, Laonome kroyeri Malm (Evenkamp, 1931), and I have observed them in Branchiomma vesiculosum (Montagu). In the Fabriciinae, Evenkamp (1931) reports their presence in Euchone papillosa M. Sars. In the remaining sub-family, the Myxicolinae, they are absent.

Fox (1938) has put forward two suggestions, not mutually exclusive, about the function of the coelomic capillaries in Sabella. Drawing attention to the absence of any capillaries in the muscles of this animal, he points out that part of the oxygen supply to the muscles must come by way of the coelomic fluid, in which a high oxygen concentration would be maintained by diffusion of oxygen from the coelomic capillaries.1 He also suggests that in the breeding season the capillaries may carry oxygen to the genital products which lie free in the coelomic fluid.

In Spirographis, as has been described above, there are capillaries in the ventral muscle blocks, while in Branchiomma I have found both blind-ending coelomic capillaries and capillaries in both dorsal and ventral muscle blocks. This suggests that in these forms the supply of oxygen to the muscles from the coelomic capillaries by way of the coelomic fluid is not the only function of these capillaries.

The hypothesis that the coelomic capillaries may be important in bringing oxygen to the genital products is strengthened by the fact that in Spirographis some of the coelomic blindending capillaries on the ventral gland-shield vessel run between the cells of the ovary. In the coelom of Sabella and Spirographis there are also fat-bearing eleocytes. In Branchi o mm a are similar but smaller cells (Romieu, 1923). It may be that a further function of the blind-ending capillaries is to carry food and oxygen to these cells.

It is not possible to make any estimate of the relative importance of the several proposed functions for the coelomic capillaries of the Sabellinae.

My sincere thanks are due to Professor H. Munro Fox, F.R.S., who suggested the work and advised me during its progress, and to Dr. D. P. Pielou and Mr. E. W. Bentley of this Department for drawing many of the figures.

11.
Allen
,
E. J.
,
1904
. —“
Anatomy of Poecilochaetus Claparède
”,
‘Quart. Joum. Mier. Sei.’
,
48
.
Benham
,
W. B.
,
1891
. —“
Nephridium of Lumbricus and its Blood Supply; with Remarks on the Nephridia in other Chaetopods
”,
ibid
.,
32
.
Brown
,
R. S.
,
1938
. —“
Anatomy of the Polychaete, Ophelia chithensis
”,
‘Proc. Roy. Soc. Edinb.’
,
58
.
Carlson
,
A. J.
,
1908
. —“
Comparative Physiology of the Invertebrate Heart. X. Physiology of pulsating Blood Vessels in Worms
”,
‘Amer. Joum. Physiol.’
,
22
.
Claparède
,
E.
,
1868-71
.
—’Annélides Chétopodes du Golfe de Naples.’ Genève et Basle
.
Claparède
,
E.
1873
.
—’Structure des Annélides Sédentaires.’
Genève
.
Cosmovici
,
L. C.
,
1879
. —“
Glandes génitales et Organes segmentaires des Annélides polychètes
”,
‘Arch. Zool. exp. gén.’
,
8
.
von Drasche
,
R-
,
1885
.
—’ Beitr. z. feineren Anatomie derPolychaten.’ Wien
.
Evenkamp
,
H.
,
1931
. —“
Morphol. Histol. u. Biol, der Sabellidenspecies Laonome kroyeri u. Euchone papillosa
”,
‘Zool. Jb.’
,
53
.
Ewer
,
R. F.
, and Fox,
H.
Munro
,
1940
. —“
On the function of ehlorocruorin
”,
‘Proc. Roy. Soc.’ B
,
129
.
Faulkner
,
G. H.
,
1930
. —“
Anat. and Histol. of Bud Formation in the Serpulid Filograna implexa
”,
‘Joum. Linn. Soc. (Zool.)’
,
37
.
Fauvel
,
P.
,
1927
. —“
Polychètes Sédentaires
”,
‘Faune de France’
,
16. Paris
.
Fox
,
H. Munro
,
1926
. —“
Chlorocruorin ; A pigment allied to Haemoglobin
”,
‘Proc. Roy. Soc.’ B
,
99
.
Fox
,
H. Munro
,
1932
. —“
Oxygen Affinity of Chlorocruorin
”,
ibid., B
,
111
.
Fox
,
H. Munro
,
1933
. —“
Blood Circulation of Animals possessing Chlorocruorin
”,
ibid., B
,
112
.
Fox
,
H. Munro
,
1934
. —“
Oxygen to Iron Ratio of Oxychlorocruorin … in Spirographis
”,
ibid., B
,
115
.
Fox
,
H. Munro
,
1938
. —“
Blood Circulation and Metabolism of Sabellids
”,
ibid., B
,
125
.
Fraipont
,
J.
,
1887
. —“
Genre Polygordius
”,
‘Fauna u. Flora d. Golfes v
.
Neapel’
,
14
.
Berlin
.
Fuchs
,
K.
,
1907
. —“
Topographie des Blutgefássystems der Chaetopoden
”,
‘Jena Z. Naturw.’
,
42
.
Gamble
,
F. W.
, and
Ashworth
,
J. H.
,
1898
. —“
Habits and Structure of Arenicola marina
”,
‘Quart. Joum. Mier. Sci.’
,
41
.
Gamble
,
F. W.
1900
. —“
Anatomy and Classification of the Arenicolidae
”,
ibid
.,
43
.
Grube
,
E.
,
1838
. —
’Anatomie u. Physiol, der Kiemenwünner.’ Kônigsberg
.
Hempelmann
,
F.
,
1906
. —“
Morphologie von Polygordius lacteus und Polygordius triestinus
”,
‘Z. wiss. Zool.’
,
84
.
Huxley
,
T. H.
,
1883
. —“
Hermaphrodite andfissiparous species of tubicular Annelid
”,
‘Edinb. new Philos. Journ’
,
1
.
Jaquet
,
M.
,
1886
. —“
Système vasculaire des Annélides
”,
‘Mitt. zool. Sta. Neapel.’
,
6
.
Johansson
,
K. E.
,
1927
. —“
Polychaeten-Familien Hermellidae, Sabellidae und Serpulidae
”,
‘Zool. Bidr. Uppsala’
,
11
.
Keilin
,
D.
,
1920
. —“
Pharyngeal or Salivary Gland of the Earthworm
”,
‘Quart. Journ. Micr. Sci.’
,
65
.
Lang
,
A.
,
1904
. —“
Beitr. z. e. Trophocôltheorie
”,
‘Jena. Z. Naturw.’
,
38
.
Lee
,
E.
,
1912
. —“
Beitr. z. K. der Serpuliden, speziell ihres Blutgefasssystems
”,
ibid
.,
48
.
McIntosh
,
W. C.
,
1878
. —“
Anatomie von Magelona
”,
‘Z. wiss. Zool.’
,
31
.
Meyer
,
E.
,
1888
. —“
Kôrperbau der Anneliden II
”,
‘Mitt. zool. Sta. Neapel’
,
8
.
Milne
Edwards
, H
.,
1838
. —“
Histoire de la circulation du sang chez les Annélides
”,
‘Ann. Sci. nat.’, Ser. II
,
10
.
Müller
,
F.
,
1858
. —“
Anneliden Fauna der Insel Santa Catharina an der brasilianischen Küste
”,
‘Arch. Naturgesch.’
,
24
.
Nicol
,
E. A.
,
1930
. —“
Feeding Mechanism, Tube and Digestion in Sabella pavonina
”,
‘Trans. Roy. Soc. Edinb.’
,
56
.
Pickworth
,
F. A.
,
1934
. —“
New Method of Study of Brain Capillaries and its Application to Regional Localization of Mental Disorder
”,
‘Journ. Anat.’
,
69
.
Probst
,
G.
,
1929
. —
“Blutgefassystem von Chaetopterus variopedatus “
,
‘Publ. Sta. zool. Napoli’
,
9
.
Pruvot
,
G.
,
1885
. —“
Rech. anat. et morphol. sur le système nerveux des Annélides Polychètes
”,
‘Arch. Zool. exp. gén.’, Ser. 2
,
3
.
Rathke
,
H.
,
1842
. —“
Vergleich. Anat. u. Physiol
.”,
‘Neuste Schr. d. Naturforsch. Gesellschaft in Danzig.’
Vol.
3
, Heft. 4.
Roche
,
J.
, and Fox,
H.
Munro
,
1933
. —“
Crystalline Chlorocruorin
”,
‘Proc. Roy. Soc.’ B
,
114
.
Romieu
,
M.
,
1923
. —“
Rech. histophysiol. sur le sang et sur le corps cardiaque des Annélides Polychètes
”,
‘Arch. Morph, gén. exp.’, Fase. 17
.
Saint-Joseph
,
Baron de
,
1894
. —“
Annélides polychètes des Côtes de Dinard, 3me partie
”,
‘Ann. Sci. nat.’, Ser. VII
,
17
.
Schaeppi
,
T.
,
1894
. —“
Chloragogen von Ophelia radiata
”,
‘Jena. Z. Naturw.’
,
28
.
Slonimski
,
P.
,
1927
. —“
Modification de 1”ultra-micro-méthode’ de Wu-Hsien et son application à la recherche de l’hémoglobine dans les disques germinatifs des Oiseaux
”,
‘C. R. Soc. Biol. Paris’
,
96
.
Williams
,
T.
,
1851
. —
’British Annelida.’ Report of 21st Meeting Brit. Assoc. Adv. Sci
.
Zürcher
,
L.
,
1909
. —“
HistoL der Kôrperund Darmmuskulatur u. d. Hamocols von Owenia
”,
‘Jena. Z. Naturw.’
,
45
.
1

My thanks are due to the Director and Staff of the Laboratoire LacazeDuthiers at Roscoff for their hospitality and help.

1

These cells containing black granules have been called ‘chloragogen cells’ by Claparède (1873) and by de Saint-Joseph (1894). They are discussed by Meyer (1888) and by Evenkamp (1931).

1

The head is composed of the peristomium (segment I), of which the collar fold is a part and of the prostomium and its appendages, the crown, lateral lips, and ventral sac.

1

Morphologically the whole crown represents the palps of other polychaetes (Pruvot, 1885; Johansson, 1927). The two processes referred to here as palps are outgrowths of the dorsal lip; they are Fauvel’s ‘palpes’ (1927).

1

The word ‘specific’ is written wittingly, since the division of Spirographis and Sabella into separate genera seems hardly justified by the few anatomical differences between them.

1

Hemplemann (1906) described blind-ending sacs on the ring vessels of the archiannelid Polygordius lacteus Schneider. These project into the next posterior segment and their walls are covered with gonads. Similar structures are described by Fraipont (1887) in Polygordius neapolitanus Fraipont.

1

A similar idea was put forward by Wirén to explain the condition in Arenicola in which he believed, incorrectly, that there was no capillary supply to the body-wall. His ideas are expressed in the following manner by Gamble and Ashworth (1898), p. 20. Wirén ‘suggests that the assimilation of food and oxygen by the tissues is effected chiefly through the mediation of the coelom, which he points out is parcelled off in the intermuscular spaces, by a channelling out of the subepidermic tissue into “perihaemal canals” ⃜ The extension of the coelom into the intermuscular and subdermal spaces, has, however, all the appearance of acting as the equivalent of lymph spaces in higher forms.’