PART II.

ANATOMY OF THE PERITONEUM IN THE SUPRA-COLIC COMPARTMENT OF THE ABDOMEN.

We have already seen that the transverse colon and mesocolon effect a general division of the adult human abdominal cavity into a cephalic supra-colic compartment, situated between the diaphragm and the level of the transverse colon and mesocolon, comprising in general the hypochondriac and epigastric regions, and a larger caudal infra-colic space which includes the entire rest of the abdominal cavity and is continued caudad into the pelvic cavity. The arrangement of the peritoneum and viscera in this latter space has just been considered. The fact will be recalled that the second or descending portion of the duodenum, passing dorsad of the hepatic colic flexure, forms so to speak the visceral connection between the portions of the alimentary tube situated in the supra-colic compartment and those situated in the infra-colic space. The fixation of this segment of the duodenum and its consequent secondary retroperitoneal position in the adult human subject masks this continuity of the alimentary canal to a certain extent so that it requires more than a superficial examination in order to trace correctly the course of the duodenum from the pylorus to the duodeno-jejunal angle, dorsad of the colon, root of transverse mesocolon and mesentery, and under cover of the secondary parietal peritoneum.

We have now to turn our attention to the viscera contained in the cephalic or supra-colic compartment of the abdomen and to consider the disposition of the serous membrane investing them and connecting them with each other and with the abdominal parietes.

The visceral contents of the supra-colic compartment comprise the liver, pancreas, spleen, stomach and the proximal portion of the duodenum, including the hepatic angle and the supra-colic part of the descending duodenum. Less directly the cephalic portions of the right and left kidney and the corresponding suprarenal capsules belong to this visceral group.

In this region of the abdomen we meet with the most extensive modifications of the primitive dorsal peritoneal membrane, producing conditions which, considered without reference to development and comparative anatomy, are complex and difficult of comprehension. These changes lead to the formation of the so-called "lesser sac," a term which in some respects is unfortunate as it implies a more complete degree of separation from the general peritoneal cavity or "greater sac" than actually exists.

In order to clearly understand the adult arrangement of the peritoneum in this region it is advisable to consider the subject in two distinct subdivisions, dealing successively with the two cardinal facts which contribute to effect the change from the simple primitive to the complicated adult condition.

These two main elements are:

1. Developmental changes in the position of the stomach, alterations in the disposition of the proximal part of the primitive dorsal mesentery attached to the stomach, and the development of pancreas and spleen in connection with this membrane.

2. The development of the liver and the successive stages in the production of the final adult vascular and serous relations of this organ.

=1. Stomach and Dorsal Mesogastrium.=--We have already considered the early stages in the differentiation of the stomach from the primitive intestinal tube of uniform caliber (p. 40). It will be recalled that the stomach at a certain period, while it already presents the main structural features familiar in the adult organ, occupies a vertical position in the abdominal cavity, turning its concave margin (lesser curvature) ventrad, while the convex dorsal border (greater curvature) is directed toward the vertebral column, being attached to the same by the layers of the proximal part of the primitive dorsal mesentery. At this time the stomach presents right and left surfaces, and the oesophageal entrance is at the highest or cephalic point of the organ, while the pyloric transition to the small intestine occupies the distal caudal extremity.

The primitive dorsal mesentery, as already stated, passes as a thin double-layered membrane between the ventral surface of the vertebral column and the dorsal border of the stomach, which, as we will presently see, becomes during the later stages of development the caudal (lower) margin or greater curvature.

It will be seen that the embryonic differentiation of the intestinal tract into successive segments justifies the application of a terminology based on this differentiation to the corresponding portions of the primitive common dorsal mesentery.

Thus the proximal portion extending between the vertebral column and the dorsal border or greater curvature of the stomach becomes the _mesogastrium_; we differentiate this portion still further as the "_dorsal mesogastrium_" to distinguish it from a "_ventral mesogastrium_" which we will presently encounter in considering the development of the liver and the connected peritoneum.

In the same way the section of the primitive common dorsal mesentery attached to the duodenal loop becomes the _mesoduodenum_, that connected with the mobile part of the small intestine (jejuno-ileum) the _mesentery_ proper, while the portion passing to the colon forms the _mesocolon_, to be subsequently still further subdivided, after the different segments of the large intestine have become mapped out, as the _ascending_, _transverse_ and _descending mesocolon_, the _mesosigmoidea_ and the _mesorectum_.

In tracing the development of the adult human peritoneum it is well to consider certain stages, which we will find illustrated by the permanent conditions presented by some of the lower vertebrates:

These stages comprise:

(_a_) Changes in the position of the stomach.

(_b_) Changes in the direction and extent of the dorsal mesogastrium.

(_c_) Development of the pancreas and spleen in connection with the mesogastrium.

A. Changes in the Position of the Stomach.

The primitive position of the organ above outlined (p. 41) is changed during the course of further development by a twofold rotation.

1. The primitive vertical position, in which the oesophageal entrance occupies the highest cephalic extremity, while the pyloric opening is at the opposite caudal end, is exchanged for one directed more transversely, approximating the two gastric orifices to the same horizontal level. In human embryos of 13.9 mm. the fundus has already descended, the pylorus moving cephalad and to the right, while the cardia becomes shifted more to the left. At the same time the greater growth and prominence of the convex border or greater curvature becomes marked in comparison with the relatively short extent of the opposite margin or lesser curvature.

2. Coincident with this change in position is a rotation around the vertical axis, by means of which the original left side of the stomach is turned ventrad, becoming the ventral or "anterior" surface, while the original right surface of the organ now looks dorsad toward the vertebral column, becoming the dorsal or "posterior" surface of human anatomy. The oesophageal or cephalic end is placed to the left of the median line, while the caudal or pyloric end is situated on the right side (Figs. 169 and 170).

The original ventral border, now the "lesser curvature" or "upper border," looks cephalad and to the right, toward the caudal surface of the liver, while the original dorsal border, as the "greater curvature" or "lower border" is directed in the main caudad and to the left.

The prominence of this border is still further increased by the greater development of the stomach to the left of the oesophageal entrance resulting in the formation of the "fundus" or "great cul-de-sac."

This rotation of the stomach explains the asymmetrical position of the vagus nerve in the adult, the left side of the embryonic stomach, innervated by the left vagus, becoming the "anterior" surface of adult descriptive anatomy and _vice versa_.

It will be readily appreciated that a comparatively flat organ like the stomach, will, as long as it occupies a sagittal position, with right and left surfaces, help to divide the upper part of the abdominal cavity to a certain extent into a right and left half, even if the peritoneal connections of the organ are left out of consideration. As soon, however, as the above-described changes in position take place and the surfaces of the stomach are directed ventrad and dorsad, the relative arrangement and extent of this right and left abdominal space becomes altered by the different disposition of the septum, _i. e._, the stomach. The original right side of the organ is now directed dorsad, and the rotation of the organ has created a space between this dorsal or "posterior" surface of the stomach and the background of the abdominal cavity, which is the inception of the "lesser peritoneal cavity" or retrogastric space. We will find that this space becomes well defined and circumscribed by the peritoneal connections of the stomach, but we will realize, even at this stage, that the _dorsal_ surface of the stomach will form a part of the general _ventral_ wall of the lesser peritoneal space.

On the other hand, the partial division of the abdomen into a right and left half, effected by the stomach in its primitive sagittal position, disappears after rotation of the organ. We now pass uninterruptedly from left to right across the ventral (original _left_) surface of the stomach.

B. Changes in the Direction and Extent of the Dorsal Mesogastrium.

The effects of the altered position of the stomach on the disposition of the abdominal space have just been considered in relation to the organ itself, without reference to its natural connections with the parietes and with adjacent viscera. Their true significance and their influence on the adult anatomical arrangement of the abdomen is, however, only appreciated when the changes in the arrangement of the peritoneal membrane which they involve, are taken into account.

The dorsal mesogastrium changes more than any other portion of the peritoneum in the course of development. It not only becomes displaced and altered in direction by the rotation of the stomach, but in addition it grows so extensively that it finally hangs down like an apron over the entire mass of small intestines, forming the great omentum.

If we begin with the primitive disposition of the sagittal stomach and dorsal mesogastrium shown in Fig. 171 it will be observed that both structures together actually divide the dorsal portion of the abdominal cavity into symmetrical right and left halves (Fig. 172).

After rotation of the stomach (Fig. 173) the mesogastrium loses its original sagittal direction. It follows the altered position of the original dorsal border of the stomach, which has now become the caudal margin or "greater curvature," by turning caudad and to the left, being at the same time considerably elongated. This occurs during the second month. Hence the dorsal mesogastrium, after leaving the vertebral column, turns ventrad and to the left to reach its gastric attachment along the greater curvature. This is the first indication of the formation of the great omental or epiploic bursa.

The stomach is here considered as developing in situ and as influencing by its growth and change of position the arrangement and direction of the peritoneal layers with which it is connected. As a matter of fact it is well to note that the stomach at first lies above the primitive diaphragm or septum transversum, migrating, however, at an early period into the subhepatic abdominal position. This migration produces a corresponding increase in the length of the oesophagus (Fig. 34) and the stomach, in consequence of this change in position, acquires its ventral and dorsal mesogastrium. For the purpose of explaining the adult peritoneal relations of the organ it is, however, more convenient to regard the stomach as an abdominal organ from the beginning and to deal with the subsequent changes in position from this standpoint. The inaccuracy is slight and renders the comprehension of the succeeding stages easier.

It will be noticed (Fig. 173) that the rudimentary retro-gastric space or "lesser peritoneal sac" is bounded ventrally by the dorsal (the primitive _right_) surface of the stomach, while its dorsal boundary is furnished by the ventral (originally _right_) layer of the dorsal mesogastrium.

In the primitive condition, therefore, dorsal mesogastrium and stomach form together a straight line sagittal in direction and placed in the median plane of the body. As the result of the developmental changes above outlined this straight line becomes bent at the point where the mesogastrium reaches the stomach (Fig. 173, x). The two component elements of the line (stomach and mesogastrium) hinge on each other here, and the angle which they form opens to the right.

The changes which are to be observed in the later stages depend principally upon a peculiar feature characteristic of the development of the dorsal mesogastrium. This feature consists in the extreme redundancy of the membrane which grows out of proportion to the requirements of its visceral connections, and to a certain extent becomes independent of the direct mechanical purpose of carrying blood vessels to the viscera. Hence in a transverse section at this period (Figs. 174 and 175) the mesogastrium no longer passes in a direct line between its points of attachment, viz. the greater curvature of the stomach and the vertebral column, but extends beyond the stomach to the left. We will appreciate the significance of this extensive growth of the mesogastrium especially in considering the development of the spleen and pancreas. For the present it will suffice to note (Figs. 174 and 175) that the growth has carried the mesogastrium well to the left of the stomach, consequently the retrogastric space is now bounded toward the left by the bend which the original right leaf of the primitive sagittal mesogastrium takes in order to reach its gastric attachment. The retrogastric space therefore terminates toward the left in a blind pocket formed by this reduplication of the mesogastrium.

One more factor is to be taken into consideration, namely the tendency, already noted, of peritoneal surfaces to become adherent to each other. Such adhesion involves the apposed surfaces of the mesogastrium and of the primitive parietal peritoneum to the left of the vertebral column. The dorsal (original _left_) layer of the mesogastrium adheres to the parietal peritoneum covering the left side of the abdominal background and the cephalic portion of the ventral surface of the left kidney up to the end of the blind pouch which forms the extreme left limit of the retrogastric space. Hence, after this process of adhesion is completed, the dorsal wall of the retrogastric space is lined by secondary parietal peritoneum covering the left kidney (original right leaf of primitive mesogastrium) (Fig. 175). We obtain (Fig. 175 at x) an apparent continuity of the parietal peritoneum with that portion of the mesogastrium which, derived from the original left layer of the membrane, appears now to extend, as the ventral one of two layers, between the stomach and the abdominal parietes near the lateral border of the left kidney. (Primitive gastro-splenic omentum.)

It should be remembered that the disposition of the peritoneum just indicated is modified by the development of the pancreas and spleen, both of which organs are intimately associated with the mesogastrium. The foregoing statements and diagrams are therefore merely given for the purpose of affording a general view of the extent, growth and changes of the dorsal mesogastrium before proceeding to consider the development of the pancreas and spleen in and from the membrane itself.

In the view directly from in front the redundancy of the peritoneum forming the mesogastrium is shown in Figs. 176 and 177. Just as the membrane extends further to the left than required by its visceral connection with the stomach, so the downward growth exceeds the demand made by the rotation of the attached border (greater curvature) caudad and to the left. The mesogastrium, forming, as it now does, the great omentum, enlarges in descending toward the transverse colon (Fig. 177). The bag thus formed can be distended with air in a foetus of from 8 to 9 cm. vertex-coccygeal measure, as shown in the figure. Consequently in sagittal section the membrane is seen to extend caudad beyond the level of the greater curvature, and must turn on itself and pass again cephalad in order to reach the stomach (Fig. 178). By reason of this excessive growth the limits of the primitive retrogastric space are enlarged, not only toward the left, but more especially in the caudal direction. The bend made by the mesogastrium in returning to the stomach forms the blind extremity of a pouch which continues the retrogastric space caudad beyond the stomach, and whose dorsal and ventral walls are formed by the reduplicated mesogastrium. This pocket or pouch constitutes the _omental_ or _epiploic bursa_ of the lesser peritoneal cavity, for the great omentum is the direct product of this redundant growth of the mesogastrium caudad. It will be observed that the great omentum is made up of four peritoneal layers, the folding of the double-layered mesogastrium naturally producing this result. The first or ventral and the fourth or dorsal layer are derived from the original left layer of the primitive sagittal mesogastrium; the intermediate second and third layers, separated from each other at this stage by the cavity of the omental bursa, are products of the primitive right leaf of the mesogastrium. Since the entire retrogastric space with its extensions becomes the "lesser cavity" of the human adult peritoneum, it will be seen that its serous membrane is derived from the original right leaf of the mesogastrium (second and third omental layers). After the above-described adhesion of the mesogastrium to the parietal peritoneum overlying the ventral surface of the left kidney, the membrane would be traced in sagittal section (Fig. 179) from the dorsal surface of the stomach caudad, lining the interior of the omental bursa (second layer) to the turn or blind end of the pouch; thence cephalad as the third omental layer, forming the dorsal wall of the epiploic bursa, to invest, as secondary parietal peritoneum, the cephalic segment of the ventral surface of the left kidney.

C. Development of Spleen and Pancreas in the Dorsal Mesogastrium and Changes in the Disposition of the Great Omentum.

In order to obtain a correct conception of the adult human conditions it is finally necessary to consider the development of the spleen and pancreas in their connection with the dorsal mesogastrium and to note the changes which are produced by adhesion of portions of the great omentum to adjacent serous surfaces. It will be advisable to discuss these subjects at first separately, and to subsequently combine all the facts in an attempt to gain a correct impression of their share in determining the disposition of the adult human peritoneum.

=1. Development of Spleen.=--The spleen develops from the mesoderm between the layers of the dorsal mesogastrium, near its point of accession to the greater curvature, in the region of the subsequent fundus. It has therefore, like the stomach, originally free peritoneal surfaces. After rotation of the stomach the organ lies between the two layers of the membrane at the extreme left end of the retrogastric space (Fig. 180).

=Vascular Connections.=--The splenic artery accedes to the mesal surface of the spleen from the vessel which originally passed directly to the dorsal border (subsequent greater curvature) of the stomach, between the layers of the mesogastrium.

With the further growth of the spleen the segment of this vessel situated between its origin from the coeliac axis and the hilum of the spleen becomes relatively larger, forming the adult splenic artery, while the continuation of the original vessel to the greater curvature of the stomach appears now as a branch of the splenic artery, viz., the arteria gastro-epiploica sinistra.

Through the development of the spleen the dorsal mesogastrium has been subdivided into a proximal longer vertebro-splenic, and a distal shorter gastro-splenic segment. The former, as we have seen, loses its identity as a free membrane in the human adult, by fusing with the parietal peritoneum investing the ventral surface of the left kidney. Hence, after this adhesion has taken place, the splenic artery courses from the coeliac axis to the spleen behind peritoneum which functions as part of the general parietal membrane, but which is derived from the original right leaf of the proximal vertebro-splenic segment of the primitive mesogastrium (Fig. 181). On the other hand the distal segment of this membrane, beyond the spleen, remains free, carrying, as the gastro-splenic omentum, the left gastro-epiploic artery between its layers from the splenic artery to the greater curvature of the stomach.

The lateral limit of the area of adhesion between mesogastrium and parietal peritoneum is situated along the lateral border of the left kidney. Hence, in the final condition of the parts, the main splenic vessels at the hilum are situated between two peritoneal layers of which the ventral (Fig. 181) appears as the parietal peritoneum forming the dorsal wall of the retro-gastric space, while the dorsal layer (Fig. 181) forms a reflection from the mesal surface of the spleen, along the dorsal margin of the hilum, to the adjacent lateral border of the left kidney (lieno-renal ligament) and to the diaphragm. At this point of adhesion subsequently firmer strands of connective tissue develop in the serous reduplication forming the _ligamentum phrenico-lienale_ of systematic anatomy. This process of adhesion takes place during the second half of intra-uterine life. A connection with the colon, produced by adhesion of the mesogastrium to the splenic flexure of the large intestine, forms the adult _lig. colico-lienale_, while a similar adhesion between great omentum, transverse mesocolon and phrenic parietal peritoneum just caudad of the spleen, gives rise to the _colico-phrenic_ or _costo-colic "supporting" ligament_ of the spleen.

On the other hand, the ventral one of the two layers constituting the gastro-splenic omentum and including between them the left gastro-epiploic artery, is formed by the distal part of the primitive left layer of the mesogastrium, while the dorsal layer of the same fold is the portion of the primitive right layer beyond the spleen, which has not been converted into secondary parietal peritoneum, but forms now part of the ventral wall of the lesser peritoneal sac between the spleen and the stomach (Fig. 181) (lig. gastro-lienale). Since, therefore, the gastro-splenic omentum is a specialized part of the fully-developed dorsal mesogastrium, and since we have seen that the great omentum is formed directly by the excessive growth of this membrane caudad, it is not difficult to understand why in the adult human subject the ventral layer of the gastro-splenic omentum is directly continuous with the ventral layer of the great omentum along the greater curvature of the stomach to which both are attached. The dorsal layer of the gastro-splenic omentum would, in the same way, be continuous with the second layer of the great omentum, lining the ventral wall of the omental bursa, if it were not for the fact that in the adult adhesions usually obliterate the cavity of the bursa.

Fig. 182 shows the stomach, left kidney, spleen and splenic flexure of the colon hardened in situ and removed from the body of a two-year-old child. The great omentum has been divided along the line of adherence to the transverse colon.

In Fig. 183 the spleen has been removed from the preparation by division of its peritoneal and vascular connections, and is shown in its mesal aspect (gastric and renal surfaces, intermediate margin and hilum). It will be seen that the peritoneal reflections are arranged in the form of two concentric elliptical lines. The two ventral lines form the gastro-splenic omentum and correspond to the reflection of the peritoneum from spleen to left end of stomach carrying the gastric branches derived from the splenic artery. The third line from before backwards results from the division of the secondary parietal peritoneum of the lesser sac, covering splenic artery, and ventral surface of pancreas and derived from the dorsal mesogastrium; while the most dorsal fourth line represents the divided reflection of the peritoneum from the renal surface of spleen to lateral border of left kidney and diaphragm (lig. lieno-renale).

Between the second and third lines of peritoneal reflection appears the portion of the mesal surface of the spleen in contact with and invested by the extreme left end of the lesser peritoneal sac.

Fig. 184, taken from an adult human subject with the viscera hardened in situ, shows the left or splenic extension of the lesser peritoneal cavity.

=2. Development of the Pancreas.=--The pancreatic gland is derived from the hypoblast of the enteric tube. The secreting epithelium and that lining the ducts of the adult gland is formed by budding and proliferation of the intestinal epithelium. The gland develops primarily from two outgrowths which are at first separate and distinct from each other.

1. The proximal and dorsal bud grows directly from the hypoblast lining the duodenum immediately beyond the pyloric junction.

In embryos of 8 mm. (four weeks) (Fig. 185) it appears as a small spherical outgrowth connected by a slightly narrower stalk with the epithelial intestinal tube.

2. The distal and ventral outgrowth is separated from the preceding and is from the beginning closely connected with the similar embryonic outgrowth from the enteric tube which is to form the liver. This portion of the pancreas is, strictly speaking, derived primarily from the epithelium of the primitive hepatic duct and not directly from the duodenum. This primary arrangement of the gland, being formed of two main collections of budding hypoblastic cells, corresponds to the adult system of the pancreatic excretory ducts. The proximal or dorsal outgrowth furnishes that portion of the head of the gland whose excretory system terminates in the _secondary pancreatic duct_ or _duct of Santorini_, while the distal (ventral) outgrowth includes within its area the termination of the principal pancreatic duct or _canal of Wirsung_, which is closely connected with the end of the common bile-duct at the intestinal opening common to both (Figs. 186-187). The method of union of the two pancreatic outgrowths and their respective share in building up the adult gland explains the usual adult arrangement of the excretory system and its variations.

In the embryo of five weeks (Fig. 186) the two portions have grown in length. The dorsal or proximal outgrowth, developing between the layers of the mesoduodenum, is at this time the larger of the two, composed of a number of glandular vesicles clustered around the stalk represented by the parent duct.

The distal or ventral pancreatic growth, connected with the liver duct, is as yet small and presents only a few vesicular appendages. The duct of this portion empties in common with the hepatic duct into the duodenum.

In embryos of the sixth to seventh week (Fig. 187), the two glandular outgrowths have become connected with each other at a point which corresponds exactly to the divergence of the duct of Santorini from the main pancreatic duct (canal of Wirsung) in the adult gland (Fig. 188).

The secondary pancreatic duct (of Santorini) of the adult corresponds to that section of the proximal or larger embryonic outgrowth situated between the intestine and the point where the two glandular diverticula fuse with each other. Hence the canal of Wirsung in the adult is a compound product. It includes the duct system developed, in connection with the bile duct, in the head of the gland, forming the intestinal termination of the main duct. Its distal body portion on the other hand is derived from the duct system of the originally larger proximal outgrowth, including the entire peripheral portion which has become secondarily added to the duct of the ventral outgrowth to form together with it the canal of Wirsung. On the other hand the proximal portion of the duct system of this originally larger part becomes secondarily differentiated as the duct of Santorini.

Fig. 188 shows the normal adult arrangement of the pancreatic and biliary ducts in a corrosion preparation of the canal.

The duct of Santorini in this case opened by a separate orifice into the duodenum above the common opening of the biliary and pancreatic ducts (cf. p. 113).

=Explanation of Adult Arrangement of Human Pancreatic Ducts and Their Variations Dependent Upon the Embryonic Development.=--The smaller distal embryonic outgrowth is, as we have seen, from its inception in close connection with the duodenal end of the common bile-duct (Fig. 185).

The proximal outgrowth, situated nearer to pylorus and derived directly from the duodenal epithelium, is the larger and forms the greater part of the bulk of the adult pancreas (Figs. 186, 187).

If, notwithstanding this primitive arrangement, the distal duct (canal of Wirsung) appears as the main pancreatic duct in the adult, while the proximal (duct of Santorini) is secondary, this depends upon a union of the products of the two outgrowths in such a manner that the greater part of the duct system of the proximal and larger portion is transferred to the distal duct to form the adult canal of Wirsung, while the smaller segment of the proximal duct, between its opening into the duodenum and the point of fusion of the two outgrowths, forms the adult secondary duct of Santorini. This duct opens usually into the duodenum upon a small papilla situated about 2.5 cm. above the common duodenal termination of the bile-duct and canal of Wirsung (papilla Vateri) (Fig. 193). The duct of Santorini usually tapers toward the duodenal opening from its point of departure from the main duct, its caliber gradually diminishing in the direction indicated, so that it is smaller at the duodenal opening than at the point of confluence with the main duct (Fig. 189). Hence the secretion from the proximal head portion of the pancreas, conveyed by this duct and its tributaries, passes usually into the main pancreatic duct and not directly into the intestine through the duodenal opening of the duct of Santorini. The latter is, however, thus enabled to vicariously take upon itself the conduct of the pancreatic secretion in cases of obstruction or obliteration of the main duct (calculi, ulcers, cicatrices, etc.). In these cases of obstruction of the main duct the duct of Santorini enlarges and performs its functions.

Occasionally, without obstruction of the main duct, the duodenal opening of the duct of Santorini is large, and the flow of secretion evidently the reverse of the usual, _i. e._, directly into the intestine.

In other cases, also without pathological conditions, the proximal duct is the larger of the two and serves as the principal channel of pancreatic secretion, the canal of Wirsung being small. This is evidently a persistence and further development of the early embryonic relative condition of the two outgrowths above described (Fig. 190). On the other hand the duct of Santorini may not open at all into the duodenum, terminating in small branches which drain the proximal part of the head of the gland (Fig. 191).

Schirmer has examined the arrangement of the pancreatic ducts in 105 specimens. In 56 of these the duct of Santorini passed from the main duct into the duodenum, opening upon a papilla situated 2.5 cm. above the common opening of the bile duct and canal of Wirsung.

In 19 the duct of Santorini was well developed but did not open into the duodenum.

In but 4 cases the duct of Santorini formed the only pancreatic duct, the lower opening being occupied by the bile duct alone (Fig. 192). We may assume in these cases failure of development of the distal outgrowth connected with the primitive hepatic bud, leaving only the proximal duodenal outgrowth to form the entire adult gland.

Figs. 188 and 189 show the normal arrangement of the duodenal openings of the biliary and pancreatic ducts.

Figs. 190 to 192 show schematically the variations in the relative development and the adult arrangement of the pancreatic ducts.

=Diverticulum and Papilla Vateri.=--From what has been said regarding the embryonic union of the distal pancreatic outgrowth with the hepatic bud it will be easy to recognize the corresponding features in the arrangement of the adult duodenal termination of the common bile-duct and canal of Wirsung. The dilated interior of the duodenal papilla (diverticulum Vateri) corresponds to the embryonic segment between the intestinal opening of the primitive liver duct and the point when this duct gives off the distal larger pancreatic outbud (Figs. 186, 187, 188, 193 and 194).

The union of the pancreatic and biliary ducts to form the recess of the diverticulum Vateri, which then opens by a single common orifice into the duodenum, is better marked in some of the lower vertebrates than in man.

Fig. 195 shows the proximal portion of the duodenum of the cassowary (_Casuarius casuarius_) with the biliary and pancreatic ducts and the diverticulum at their confluence in section.

The development of these two main digestive glands as diverticula from the intestinal canal also explains the direct continuity of the mucous membrane of their ducts with that lining the duodenum, a fact which is of considerable importance in the pathological extension of mucous inflammations from the intestine to the duct system of the glands.

=Development of the Pancreas in Lower Vertebrates.=--In the embryo of the _sheep_ two pancreatic buds are found, but the duct of the dorsal (proximal) outgrowth (duct of Santorini) subsequently fuses entirely with the main duct.

In the _cat_ there are likewise two pancreatic outgrowths.

In the _chick_ three pancreatic buds are visible about the fourth day.

_Amphibia_ likewise present three embryonic pancreas buds.

The ventral (distal) outgrowth is double, the two portions proceeding symmetrically from each side of the hepatic duct. The single dorsal outgrowth is derived directly from the duodenal epithelium. Later on all these outgrowths fuse to form the single adult gland.

_Fish_ also possess several (up to four) embryonic pancreatic outgrowths.

Recently in human embryos of 4.9 mm. cervico-coccygeal measure three pancreatic outgrowths have been observed, all entirely distinct from each other, one dorsal, budding from the epithelium of the primitive duodenum and two ventral, proceeding from the grooved gutter which represents the primitive ductus choledochus at this period. In embryos of from 6 to 10 mm. the two ventral outgrowths have already fused, hence only two buds, a single ventral and a dorsal, are now encountered.[4]

[4] Iankelowitz, Arch. f. Mikr. Anat., Bd. 46, 1895.

These observations place the development of the human pancreas in line with the triple pancreatic outgrowths, two ventral and one dorsal characteristic of the majority of the lower vertebrates, which have been hitherto carefully examined. The ventral or distal bud is probably double in the majority of vertebrates. The two segments fuse, however, so early that the derivation of the pancreas from a double outgrowth, as described above for the human embryo, practically obtains. In forms in which the adult gland presents a number of separate openings into the duodenum (cf. p. 118), the development would probably show multiple embryonic outgrowths from the intestinal hypoblast.

In any case the dorsal pancreatic bud appears to have developed in the vertebrate series before the ventral outgrowth and to be hence phylogenetically the older structure.

COMPARATIVE ANATOMY OF THE PANCREAS.

With the exception of _Amphioxus_ and probably also of the _Cyclostomata_, the gland appears to be present in all vertebrates, varying, however, much in size, shape and relation to the intestinal tube. Usually it appears as an elongated, flattened, more or less distinctly lobulated organ, in close apposition to the duodenum between the layers of the mesoduodenum. In all forms in which the gland is found it is connected with the post-gastric intestine and marks the beginning of the midgut. In structure the gland is usually acinous, resembling the salivary glands. It is well developed in the selachians, forming a triangular body connected with the beginning of the midgut (Fig. 202). In some instances the gland elements do not extend beyond the intestine itself, but remain imbedded in the wall of the midgut, as in _Protopterus_. In certain adult teleosts the pancreas is surrounded by the liver (Fig. 196), in others it does not appear as a compact gland but is distributed in the form of finely scattered lobules throughout the mesentery between the two layers of this membrane. On account of this concealed position of the gland it was formerly believed that the adult teleosts did not possess a pancreas. The pyloric caeca (cf. p. 119) found in these forms were consequently considered to be homologous with the pancreas of the higher vertebrates.

In _Myxinoids_ a peculiar lobulated glandular organ is found imbedded in the peritoneal coat of the intestine near the entrance of the bile-duct, into which its lobules open separately. This organ possibly corresponds to the higher vertebrate pancreas.

An organ which may represent a dorsal pancreas is also developed in _Ammocoetes_ (larva of _Petromyzon_), but its exact homology is still doubtful. It is possible that a true pancreas has not yet developed in the cyclostomata. In _Amphioxus_ no trace of a pancreas is found. In all other vertebrates the gland is present. In certain amphibians, as the frog, the single pancreatic duct opens into the common bile duct (Fig. 197).

In lacertilians and in some chelonians a lateral offshoot of the pancreas is directed transversely and is adherent to the spleen. Fig. 113 shows the gland in _Chelydra serpentaria_. While the gland usually has a single duct, yet two ducts are found in a number of animals (many mammals, birds, chelonians and crocodiles). At times three ducts are encountered, as in the chicken and pigeon.

The arrangement of the pancreatic duct system among mammalia presents the following variations:

1. Mammals with _one_ pancreatic duct, either connected with the bile-duct or entering the intestine independently:

Monkeys, most rodents (except the beaver), marsupials, carnivora (except dog and hyena), many ungulates (pig, peccary, hyrax, etc.), most ruminating artiodactyla.

(_a_) The pancreatic duct joins the common bile-duct before entering the duodenum in the monkeys, marsupials, carnivora, in the sheep, goat and camel.

The point of entrance of the combined duct into the intestine varies. In some forms it is near the pylorus, in others at some distance from the same. The common opening is situated 11/2" to 2" beyond the pylorus in carnivora, and one foot behind the same point in the goat and sheep.

(_b_) The pancreatic duct does not join the bile-duct, but empties separately into the intestine, in most rodents and in the calf and pig.

In the calf the pancreatic duct opens into the duodenum 15' beyond the bile-duct and 3' beyond the pylorus.

In the pig the pancreatic opening is 5"-7" beyond that of the bile-duct and 6"-8" behind the pylorus.

2. Mammals with _two_ pancreatic ducts, of which one usually joins the bile-duct: perissodactyla (except the ass according to Meckel), elephant, beaver, several carnivora, dog, hyena, and according to Bernard the cat. In the perissodactyla the proximal of the two pancreatic ducts empties, either combined with the bile-duct, or separate from it, but very close to it, 3"-4" behind the pylorus. The second distal duct is smaller and opens several inches further down.

In most rodents the pancreatic entrance is placed at some distance from the pylorus. Fig. 199 shows the arrangement of the parts in the rabbit, in which animal the main distal pancreatic duct empties at a distance of 13"-14" from the pylorus into the end of the duodenum, which intestine forms a very long loop, while the biliary duct, receiving the smaller proximal pancreatic duct, opens near the pylorus.

In the _beaver_ the smaller proximal duct joins the bile-duct or even enters the duodenum anterior to the bile-duct, nearer the pylorus, while the distal larger pancreatic duct opens into the intestine 16"-18" behind the biliary duct. Of the two ducts found in the dog (Fig. 200) the smaller proximal either joins the bile-duct or opens into the intestine close to it, 1" to 11/2" beyond the pylorus. The larger distal duct opens into the duodenum 1" to 11/2" behind the biliary duct. Fig. 201 shows the dog's stomach and proximal portion of the duodenum in section. The proximal smaller pancreatic duct here joins the biliary duct, and opens with it by a single orifice into the duodenum. The distal larger pancreatic duct opens independently into the intestine further caudad.

The parts in _Hyaena_ present a similar arrangement.

Bernard always found _two_ pancreatic ducts in the _cat_, one large principal duct and a second smaller accessory duct. Of these, the one situated nearest to the pylorus always united with the bile-duct. The pancreatic duct thus joining the bile-duct was sometimes the main duct, sometimes the accessory smaller duct.

Since the main function of the pancreatic juice is the conversion of starch into sugar, the gland appears better developed in general in herbivora than in carnivora, without, however, disappearing in the latter. In fact it is of considerable size in the carnivora, because the secretion also acts on the albuminous food substances and, though to a lesser degree, on the fats.

PYLORIC CAECA OR APPENDICES.

In the _Cyclostomata_ and _Selachians_ the intestinal canal is in the main free from caecal appendages, while a large portion of the tube is provided with a special fold of the mucous membrane which projects into the lumen of the gut (spiral valve). Fig. 43 shows the straight intestinal tract with the spiral valve of the longer distal segment in a cyclostome, _Petromyzon marinus_ or lamprey. In Figs. 202 and 203 the selachian (shark) intestine is represented in two examples, while the similar spiral valve in a Dipnoean or lung fish, _Ceratodus_, is seen in Fig. 204.

On the other hand in the Ganoids and in many Teleosts longer or shorter finger-shaped diverticula of the midgut are found immediately beyond the pylorus in the region of the bile-duct.

These pouches or diverticula of the intestine form the so-called pyloric caeca or appendices of these fish. They vary very much in length, diameter and number in different forms.

Thus but a single diverticulum appears in _Polypterus_ and _Ammodytes_ (Fig. 205). _Rhombus maximus_ and _Echelus conger_ (Figs. 112 and 206) have two, and the same number appear in _Lophius piscatorius_ (Fig. 207). Perca has three and the _Pleuronectidae_ have three to five.

Fig. 208 shows the stomach and the beginning of the midgut with four pyloric caeca in _Pleuronectes maculatus_, and Fig. 209 the same parts of this animal in section.

Fig. 210 shows the stomach and midgut of _Paralichthys dentatus_, the summer flounder, with three well-developed conical pyloric caeca. On the other hand in some forms the number of pyloric appendices is enormously increased, while their caliber diminishes. Thus 191 caecal appendages are found surrounding the beginning of the midgut in _Scomber scomber_. A well-marked example of prolific development of the pyloric appendages is furnished by the common cod, _Gadus callarias_ (Fig. 211). The appendices are in the natural condition bound together by connective tissue and blood vessels, so as to form a compact organ, resembling a gland (Fig. 211, A), and a similar arrangement is found in _Thynnus vulgaris_ and _alalonga_, _Pelamys_ and _Accipenser_ (Fig. 212).

In some Teleosts (Siluroidea, Labroidea, Cyprinodontia, Plectognathi and Leptobranchiates) the appendices are entirely wanting. If there are not more than 8-10 appendices they usually surround the gut and empty into the same in a circle. In other cases they are arranged in a single line, or in a double row, opposite to each other (Fig. 213). Each appendix may open into the intestine independently, this especially where the number is limited and the individual pouches large (cf. Figs. 206-210), or several may unite to form a common duct.

Fig. 211, _B_, shows the appendices in _Gadus callarias_, the cod, freed by dissection from the investing connective and vascular tissue. It will be noticed that a considerable number of the tubes unite to form ducts of larger caliber which open into the intestine, as seen in the section shown in Fig. 214.

The pyloric appendices apparently have the same _significance_ as the spiral intestinal fold of the Selachians, Cyclostomes and Dipnoeans, _i. e._, the production of an increase in the area of the digestive and absorbing surfaces of the intestinal mucous membrane. Hence, as stated, the appendices and the spiral fold are found to vary in inverse ratio to each other. Thus, for example, _Polypterus_ (Fig. 205) still has a fairly well developed spiral fold and only a single pyloric appendix, while _Lepidosteus_, with but slightly developed spiral fold, has numerous appendices. It was formerly held that the pyloric caeca and the pancreas were mutually incompatible structures, and that where one is found the other will be wanting.

Hence the appendices were regarded as homologous with the pancreas of the higher forms. Recent observations have shown that this view is not strictly and entirely correct, while at the same time it merits consideration in several respects.

It is true that the pancreas in certain teleosts is now known to be present although concealed from observation in the liver or scattered in the form of small lobules between the layers of the mesentery (cf. p. 117), and that in a number of fish, such as _Salmo salar_, _Clupea harengus_, _Accipenser sturio_, both the appendices and the pancreas are encountered. Consequently these structures are not identical or even completely homologous, since they occur side by side in the same form.

On the other hand Krukenberg has demonstrated that the appendices pyloricae may function physiologically as a pancreas by yielding a secretion which corresponds to the pancreatic juice in its digestive action. In the majority of forms, however, they apparently merely increase the intestinal absorbing surface, secreting only mucus.

These structures are nevertheless very interesting and instructive since they furnish a perfect gross morphological illustration of the embryonal stages just considered in connection with the development of the mammalian pancreas. In the adult ganoid or teleost these blind diverticula or pouches, varying greatly in shape, number and size, protrude from the intestine immediately beyond the pylorus, usually in close connection with the duodenal entrance of the bile-duct. Two or more of these pouches may unite to form a common duct or canal opening into the intestine.

These forms, therefore, offer direct and valuable morphological illustration of the manner in which the pancreas of the higher vertebrates develops, _i. e._, as a set of hollow outgrowths or diverticula from the hypoblast of the primitive enteric tube. We can establish a consecutive series, beginning with forms in which only one or two diverticula are found, and extending to types in which the number of the little cylindrical pouches reaches nearly two hundred and in which they are bound together by connective tissue and blood vessels so as to closely resemble the structure of a glandular pancreas. This is one of the most striking instances in which the minute embryological stages of the higher types are directly illustrated by the permanent adult conditions found in the lower vertebrates. [The same statement, as we will see, holds good in reference to the development of the _liver_.]

RELATION OF THE PANCREAS TO THE PERITONEUM.

The gland becomes very intimately connected with the serous layers of the primitive dorsal mesentery. In order to clearly comprehend the adult serous relations it is necessary to make a distinction between two divisions or portions of the gland, based upon the altered relations of the primitive dorsal mesentery which result from the differentiation of the primitive simple intestinal tube into stomach and duodenum.

1. The primary outgrowth of the pancreatic tubules from the duodenum, _i. e._, the part which is to form the "head" of the adult gland, is situated between the two layers of that division of the primitive dorsal mesentery which forms, after differentiation of stomach and small intestine, the _mesoduodenum_. Coincident with the rotation of the stomach, as we have seen, the duodenum and mesoduodenum exchange their original sagittal position in the median plane of the body for one to the right of the median line, balancing, so to speak, the extension of the stomach to the left (Fig. 218).

The original right layer of the mesoduodenum and the right surface of the duodenum now look dorsad and rest in contact with the parietal peritoneum investing the right abdominal background and the ventral surface of the right kidney and inferior vena cava. We have already seen that the descending portion of the duodenum in man becomes anchored in this position by adhesion of these apposed peritoneal surfaces. This fixation includes, of course, the structures situated between the layers of the mesoduodenum, _i. e._, the head of the pancreas. Consequently, after rotation and adhesion, this portion of the gland turns one surface ventrad, invested by secondary parietal peritoneum, originally the left leaf of the free mesoduodenum, while the original right surface of the gland has become the dorsal and has lost its mesoduodenal investment by adhesion to the primary parietal peritoneum.

2. In order to understand the way in which the body and tail of the pancreas obtain their final peritoneal relations it is necessary to consider the development of the dorsal mesogastrium to form the omental bag. If we regard the primitive dorsal mesentery in the profile view from the left side (Fig. 215) it will be seen that, as already stated, the mesoduodenum is the first part of the membrane to be invaded by the pancreatic outgrowth from the intestine. Cephalad of the mesoduodenum the primitive dorsal mesogastrium (Fig. 215) is seen to protrude to the left and caudad to form, as already explained, the cavity of the omental bursa of the retrogastric space ("lesser peritoneal sac"). The further growth of the pancreas carries the developing gland from the district of the mesoduodenum into that portion of the dorsal mesogastrium which now forms the dorsal wall of the omental bursa (Fig. 216).

This double relation of the pancreas to the mesoduodenum and to the mesogastrium forming the omental bursa is well seen in foetal pigs between two and three inches in length (Fig. 217).

The head portion of the pancreas is seen developing between the layers of the mesoduodenum, while the body and tail of the gland, extending to the left, grows between the two dorsal layers of the omentum bursa towards the spleen, which organ is found connected with the left and dorsal extremity of the omental sac derived from the dorsal mesogastrium.

Before the growth of the great omentum is pronounced the continuity of the mesoduodenum and dorsal mesogastrium can be readily appreciated (Fig. 218). But after the redundant growth of the membrane has carried the great omentum further caudad, the stomach and the two omental layers attached to the greater curvature lie in front of the structures included between the two dorsal layers and conceal them from view (Fig. 177).

In sagittal sections to the left of the median line (Figs. 221 and 222) the pancreas now appears included between the layers of the great omentum near their point of departure from the vertebral column. (This point is of course identical with the prevertebral attachment of the primitive dorsal mesogastrium from which the omentum is developed.)

The foregoing considerations will, therefore, lead to the conclusion that the pancreas presents, in regard to its peritoneal relations, two distinct segments:

1. The portion adjacent to duodenum (head and neck of the gland) is developed between the layers of the mesoduodenum.

2. The distal portion of the gland, comprising the body and tail, develops between the layers of the great omentum (dorsal segment), derived from the primitive dorsal mesogastrium.

The transections of the dorsal mesogastrium shown in Figs. 180 and 181 will now have to be amplified by the introduction of the body of the pancreas between the two layers of the vertebro-splenic segment, in addition to the splenic artery (Figs. 219 and 220).

Hence the following facts will be understood:

1. In the adult the splenic artery supplies a series of small branches to the pancreas as it courses along the cephalic border of the gland on its way to the spleen.

2. After the above-described adhesion of the original left leaf of the dorsal mesogastrium (vertebro-splenic segment) to the parietal peritoneum (Fig. 220), the dorsal surface of the body of the pancreas loses its peritoneal investment and becomes attached by connective tissue to the ventral surface of the left kidney.

3. The ventral surface of the body of the pancreas is in the adult lined by peritoneum of the "lesser sac"; in other words the organ has practically assumed a "retro-peritoneal" position, its ventral peritoneal covering appearing now as the dorsal parietal peritoneum of the retro-gastric space.

4. When completely developed the extreme end (tail) of the pancreas extends to the left, following the splenic artery, until it touches the mesal aspect of the spleen at the hilus.

5. If we, therefore, leave out of consideration for the moment the transverse colon and duodenum, which will be taken up presently, and confine ourselves to the arrangement of the stomach, pancreas and great omentum, a sagittal section to the left of the median line would result as shown in Fig. 222, after the adult condition of adhesion has been established.

The same process of fixation, which resulted in the anchoring of duodenum and head of pancreas, extends to the body of the gland and the investing omentum. The peritoneum lining the original left, now the dorsal surface of the gland, fuses with the primitive parietal peritoneum covering the diaphragm and the left kidney. The main body of the pancreas in the adult appears prismatic, giving a triangular sagittal section. The dorsal surface is adherent to the ventral surface of the left kidney; the ventral surface is covered by the secondary parietal peritoneum (original right layer of mesogastrium) which lines the dorsal wall of the retrogastric space and omental bursa (lesser peritoneal sac). The great omentum now appears to take its dorsal point of departure along the sharp margin which separates this ventral surface of the pancreas from a third narrower surface directed caudad. This surface, under the conditions which we are at present examining, would be lined by the peritoneum continued onto it from the dorsal layer of the great omentum. This peritoneum merges along the dorsal margin of this caudal surface of the pancreas with the general parietal peritoneum covering the left lumbar region and the caudal part of ventral surface of the left kidney. We have, therefore, along this line a secondary transition from visceral to parietal peritoneum, obtained by the obliteration of the original visceral peritoneum investing the dorsal surface of the pancreas before adhesion to the parietal peritoneum.

The pancreas assumes, therefore, in the adult a secondary retro-peritoneal position, covered on its ventral surface by peritoneum of the "lesser sac," while the caudal surface is lined by part of the general peritoneal membrane of the "greater sac." The dorsal surface, denuded of serous covering by obliteration, is adherent to the crura of the diaphragm, the aorta and the ventral surface of the left kidney.

It is now proper to compare the conclusions just derived from the study of the development of the human dorsal mesogastrium and connected structures (spleen and pancreas) with the conditions presented by the corresponding parts in one of the lower mammalia, which illustrate some of the human embryonal stages. Here again the abdominal cavity of the cat forms an instructive object of study.

The purpose of the following comparison should be twofold:

I. The mesogastrium, spleen and pancreas in the cat will clearly illustrate the process of human development above outlined.

II. The abdominal viscera of the cat, if properly arranged, will enable us to complete the consideration of this region by including the very important relations which the transverse colon and third portion of the duodenum bear in man to the great omentum and pancreas.

I. SPLEEN, PANCREAS AND GREAT OMENTUM OF CAT.

After opening the abdominal cavity it will be seen that the great omentum can be lifted up, exposing the subjacent coils of the small and large intestine, to which it adheres at no point. In other words the entire dorsal surface of that part of the original mesogastrium which forms the great omentum is free. It will be remembered that this is not the case in the adult human subject, because here the dorsal surface of the great omentum adheres to the transverse colon. Consequently in man only that portion of the dorsal surface of the omentum can be seen which extends between the transverse colon and the caudal free edge of the membrane.

It will be noted that on the left side the spleen is connected by its mesal surface to the omentum and through it with the stomach (gastro-splenic omentum). In other words the cat illustrates the human embryonal stage in which the spleen has appeared between the layers of the dorsal mesogastrium at the extreme left or blind end of the retrogastric pouch formed by the rotation of the stomach and elongation of the mesogastric membrane, but _before_ the adhesion has taken place between the original _left_ (now _dorsal_) layer of the vertebro-splenic segment of the mesogastrium and the primitive parietal peritoneum apposed to it (Fig. 219). Consequently the dorsal wall of the "lesser" sac in the cat is still composed of the two layers of the free vertebro-splenic segment of the mesogastrium, the primitive right (now ventral) layer not having been converted, as is the case in man, into secondary parietal peritoneum by adhesion of the original left (now dorsal) layer to the primitive prerenal parietal peritoneum.

If we now examine the relation of the pancreas to the peritoneum we can establish the following facts:

1. The portion of the gland adjacent to the duodenum, corresponding to the "head" of the human organ, is included between the two layers of the mesoduodenum. This membrane is free, so that the dorsal surface of this portion of the pancreas is seen to be invested by the dorsal layer of the mesoduodenum (Fig. 223). The duodenum and the mesoduodenum, the latter containing the head of the pancreas between its layers, can be turned toward the median line, so as to expose the entire ventral surface of the post-cava and right kidney. To illustrate the arrangement which is found in the adult human subject the descending duodenum and pancreas should be allowed to fall over to the right so as to cover the vena cava and the mesal part of the ventral surface of right kidney. The adult human condition will now be produced if we assume that the structures are fixed in this position by the obliteration of the apposed serous surfaces, viz., the parietal peritoneum over kidney and vena cava on the one hand and the right layer of the mesoduodenum and the dorsal visceral peritoneum of the duodenum on the other.

2. In following out the pancreas of the cat in its entire extent, proceeding to the left of the pylorus, it will be seen that the body of the gland has extended between the two dorsal layers of the great omentum (primitive dorsal mesogastrium) over to the spleen (Fig. 223). Consequently the arrangement in the cat corresponds to the stage in the human development shown in Fig. 219 and Fig. 221 in which adhesion of the dorsal surface of the pancreas to the parietal peritoneum has not yet taken place.

It will be quite easy to reconstruct from the facts as demonstrated by the arrangement of the parts in the cat, the stage in the development of the lesser peritoneal sac in which the dorsal wall of the space is still formed by the proximal portion of the free dorsal mesogastrium (great omentum) and the structures included between its two layers.

It must then become apparent that the entire serous surface which in the adult human subject we regard as "parietal peritoneum of the lesser sac" lining the dorsal wall of the retrogastric space is derived from what originally was the right layer of the primitive sagittal dorsal mesogastrium.

II. RELATION OF GREAT OMENTUM TO TRANSVERSE COLON, TRANSVERSE MESOCOLON AND THIRD PART OF DUODENUM.

The second purpose to be accomplished by the study of the cat's abdominal cavity at this stage is the correct appreciation of the adult human conditions which are produced by areas of adhesion between the transverse colon, transverse mesocolon and third part of the duodenum on the one hand, and the dorsal mesogastrium, as great omentum, with the structures contained between its layers, on the other.

Perform the manipulations of the large and small intestine in the cat (see p. 67) which are required in order that the tract may be arranged so that it will correspond in general to the topographical conditions presented by the adult human subject. Locate the transverse colon and mesocolon and the third portion of the duodenum produced by these manipulations in imitation of the corresponding human structures. Then proceed to plot the different parts out successively as they would appear in a sagittal section (Fig. 224).

The following facts are to be noted and indicated on the plan of the section:

1. The great omentum is free, hanging down from the greater curvature of the stomach over the coils of intestine. Turning the omentum up it will be observed that the body of the pancreas is included between the two dorsal layers of the membrane.

2. The omentum, containing the pancreas, can be lifted up, exposing the next succeeding structure, viz., the transverse colon and mesocolon. In the cat the large intestine has been brought over, by the manipulations above indicated, into a transverse position so as to represent the human transverse colon and its mesocolon. It is therefore necessary to remember that in this mammal the fixation of the transverse mesocolon in the position indicated, by adhesion of ascending and descending mesocola to the parietal peritoneum of the abdominal background, has not yet occurred. Consequently the membrane must be held in the transverse position in order to represent the human arrangement.

It will of course be observed that both surfaces of the transverse mesocolon established in this way are free, not adherent to either omentum or pancreas on the one hand, nor to the transverse duodenum on the other.

3. The third or transverse portion of the duodenum is seen to be attached by the distal part of the mesoduodenum, both of the serous surfaces of the membrane being free. The duodenum having been brought from right to left transversely across vertebral column and aorta, underneath the superior mesenteric artery, the mesoduodenum, in the segment corresponding to the transverse duodenum, exchanges its original sagittal position for one in a horizontal plane, with cephalic (primitive left) and caudal (primitive right) surfaces.

Now compare the above arrangement of the intestines and peritoneum in the cat at once with the conditions presented in the adult human subject, reserving certain intermediate stages, as exhibited by some of the lower monkeys, for subsequent study.

The examination of a similar sagittal section representing schematically the adult human arrangement of the parts (Fig. 225) will reveal the following points of difference as compared with the cat:

1. The peritoneum covering the dorsal surface of the pancreas, derived from the primitive dorsal mesogastrium, has become adherent to the parietal peritoneum, as previously described.

2. The cephalic surfaces of the transverse colon and mesocolon fuse with the corresponding area of the dorsal (4th) layer of the great omentum (dorsal mesogastrium).

In the human foetus in the 4th month the connection is still so slight that the omentum can readily be separated from the transverse colon and mesocolon.

Further dorsad the cephalic layer of the transverse mesocolon adheres to the serous investment of the caudal surface of the pancreas, derived, as we have seen, from the same dorsal layer of the great omentum.

3. The duodenum and mesoduodenum are fixed by adhesion on the one hand to the parietal peritoneum, on the other to the caudal layer of the transverse mesocolon near the root of that membrane.

4. The cavity of the omental bursa is usually obliterated in the adult caudad of the level of the transverse colon, by adhesion of the apposed surfaces of the two intermediate omental layers.

We have therefore three general areas of secondary peritoneal adhesion to deal with (Fig. 225), viz.:

1. Dorsal layer of primitive } { Parietal peritoneum, cephalic mesogastrium (great } { layer of transverse omentum) including the } { mesocolon and cephalic surface serous investment of the } to { of transverse colon. dorsal and caudal surfaces } { of the pancreas (Fig. 225, } { 1). } {

2. Transverse duodenum } { Parietal peritoneum and and mesoduodenum (Fig. } to { caudal layer of transverse 225, 2). } { mesocolon.

3. Between the apposed serous surfaces of the intermediate omental layers (Fig. 225, 3).

Fig. 227 shows the abdominal cavity and disposition of the peritoneum in a macaque monkey (_Macacus rhesus_, male) in the ventral view, with the coils of small intestines removed and the omentum lifted up and reflected upon the ventral body wall. The following important points of difference from the arrangement in the _cat_ on the one hand, and in _man_ on the other, are to be noted:

1. The large intestine presents the typical primate course, with an ascending, transverse and descending colon. The ileo-caecal junction is situated in the right iliac fossa.

2. The ascending and descending mesocola are still _free_, not having become adherent to the parietal peritoneum along the dorsal abdominal wall. Hence the caudal portions of the ventral surfaces of the two kidneys are still covered by the _primitive parietal peritoneum_.

3. The great omentum is not yet adherent to the transverse colon and mesocolon except for a short distance on the extreme right. At this point the dorsal layer of the omentum has begun to contract adhesions to the hepatic flexure of the colon and ascending colon, but the rest of the transverse colon is free. Differing from the human arrangement is a line of adhesion, uniformly present in these monkeys, between the dorsal surface of the omentum along its right edge and the ventral surface and right border of the _caecum_ and _ascending colon_, parts which normally are not adherent to the omentum in man.

4. Hence in tracing the omentum to the left of the limited adhesion to the hepatic flexure and ascending colon, _i. e._, nearly throughout the entire extent of the transverse colon, we find the membrane passing freely without adhesion over the cephalic surface of the transverse mesocolon, which preserves its original free condition, independent of the omentum. This arrangement is shown in the schematic sagittal section in Fig. 230.

5. Tracing the omentum dorsad beyond the transverse colon and mesocolon the pancreas is reached. Here we encounter the first extensive area of omental or mesogastric adhesion. The omental peritoneum continues over the ventral and caudal surfaces of the gland, investing the same, but the dorsal surface has lost its serous covering and is anchored to the ventral surface of the left kidney. Hence a sagittal section would show the arrangement of the monkey's omentum as indicated in the schematic Figs. 229 and 230. Making now a general comparison of the peritoneal membrane of this animal with that of man, and of both with the preceding common embryonal condition, we can draw the following conclusions, indicated schematically in the five figures 228-232.

1. The dorsal layer of the monkey's omentum in its proximal segment behaves in the same way as in man, _i. e._, it becomes adherent to the primitive parietal peritoneum down as far as the caudal margin of the dorsal surface of the pancreas included between the primitive mesogastric layers forming by their further growth the omental apron.

Therefore we find, as in the human subject,

(_a_) The pancreas adherent to the ventral surface of the left kidney.

(_b_) A portion of the ventral surface of the kidney, cephalad of the pancreas, and the dorsal wall of the retrogastric (lesser peritoneal) space lined by secondary parietal peritoneum derived from the third layer of the omentum (original right layer of dorsal mesogastrium).

2. The monkey differs from adult man in the behavior of the dorsal omental layer in relation to the cephalic surface of the transverse mesocolon. The adhesion, which in the human subject fuses this layer with the transverse colon and mesocolon, does not occur in the monkey.

Hence we have in this animal the following conditions:

(_a_) The omentum is non-adherent to the transverse colon and transverse mesocolon.

(_b_) The caudal surface of the pancreas is lined by its original mesogastric peritoneum.

(_c_) The transverse mesocolon is formed by the original two layers of the primitive dorsal mesentery; hence its cephalic layer is not "peritoneum of the lesser sac" as is the case in man.

(_d_) The caudal part of the ventral surface of the left kidney below the pancreas, is covered by the original parietal peritoneum.

(_e_) Only one point or line of _secondary peritoneal transition_ exists, where the dorsal layer of the omentum in the adult becomes continuous with the parietal peritoneum covering the caudal surface of the pancreas and the ventral surface of the left kidney.

_Note_: In the schematic sections shown in Figs. 228 to 232 the transverse colon is represented as far removed from the ventral surface of the left kidney, in order to make the peritoneal lines of the mesocolon more clear. Actually a sagittal section which would divide the kidney would cut the transverse colon at its extreme left end, where it turns close to the ventral surface of the left kidney and then follows its lateral border to form the splenic flexure (Fig. 235). The caudal part of the ventral surface of the left kidney in the adult human subject is covered by the peritoneum which, as secondary parietal peritoneum, is derived from the upper part of the right leaf (later ventral leaf) of the descending mesocolon. Hence it should be remembered that these diagrams present _combinations_ of sections. A section which will show the full development of the transverse mesocolon is mesad of the kidney; while a section through the kidney would be too far laterad to show the transverse mesocolon.

Figs. 233, 234 and 235 show sagittal sections through the left kidney with the adult arrangement of the peritoneum and colon and the embryonic and adhesion stages leading to the same.

It will be observed that in all the schematic sections of the early embryonic stages the two layers of the transverse mesocolon are shown without dorsal attachment, as turning with the formation of a fold (Fig. 228 at x) into two layers descending ventrad of the parietal peritoneum. This is because the dorsal attachment of the mesocolon is at this stage still in the median line and would hence not be encountered by a sagittal section through the kidney, and because the two layers of the transverse mesocolon, immediately after rotation of the large intestine, are still directly continuous with the two layers of the descending mesocolon. That is to say, the cephalic layer of the transverse mesocolon is continuous with the dorsal (originally the left) layer of the descending mesocolon, and the caudal layer of the transverse mesocolon with the ventral (originally the right) layer of the descending mesocolon, which is, in the human subject, to assume subsequently the character of parietal peritoneum after the dorsal layer and the primitive parietal peritoneum have become obliterated by adhesion (Fig. 235).

Fig. 236 shows this continuity of the descending and transverse mesocolon as a permanent adult condition in the macaque. The fold of transition between the two is seen at x in Fig. 228. It will be noticed that the ventral surface of the left kidney, caudad of the adherent pancreas, is covered by the primitive parietal peritoneum, corresponding to section in Fig. 230.

RELATIONS OF SPLEEN AND OMENTUM IN _MACACUS RHESUS_.

The spleen in this animal has not contracted any extensive adhesions to the parietal peritoneum (the phrenico-lienal lig. of anthropotomy is not developed). It can be turned mesad so as to expose the lateral border and an adjacent segment of the ventral surface of the left kidney, as well as the dorsal surface of the tail of the pancreas at its tip, still covered by mesogastric peritoneum. Hence in the monkey the adhesion of the original vertebro-splenic segment of the mesogastrium, including the pancreas, to the primitive parietal peritoneum is less complete than in man.

MEDIAN ATTACHMENT OF DESCENDING MESOCOLON AND ITS RELATION TO THE MESOCOLON OF THE SIGMOID FLEXURE IN THE _MACAQUE_.

Fig. 236 shows the abdominal viscera, hardened in situ, of _Macacus cynomolgus_, the Kra monkey, in the ventral view and from the left side.

The great omentum is lifted up, the pancreas is adherent to the ventral surface of the left kidney, the caudal portion of which is covered by the primary parietal peritoneum, which can be exposed by turning the still free descending mesocolon mesad. The mesocolon retains its primitive attachment to the median line ventrad of the large prevertebral blood vessels. It is readily seen that adhesion between the left leaf of this free descending mesocolon and the parietal peritoneum down to the level of the iliac crest would produce the conditions found in the human adult, with an attached descending colon and a free sigmoid flexure; also that limited adhesion of the mesocolon of the sigmoid flexure to the parietal peritoneum would produce, as previously explained (cf. p. 97), the intersigmoid peritoneal fossa.

=2. Ventral Mesogastrium and Liver.=--The peritoneal reflections from the stomach to the liver, and the arrangement of the membrane in connection with the latter organ, remain for consideration.

Certain complicated adult conditions, encountered in this part of the abdominal cavity, make it desirable to arrange the subject for purposes of study under the following subdivisions:

I. The development of the liver and of its vascular system, and the significance of the adult circulation of the liver and of the foetal remnants connected with the organ.

II. The anatomy of the ventral mesogastrium and the changes produced in the arrangement of the membrane by the development of the liver.

=I. A. Development of the Liver.=--The liver, like the pancreas, is developed from the duodenum as an outgrowth from the hypoblast lining the enteric tube. As we have previously noted, the first outgrowth of the hepatic diverticulum is closely associated with the distal pancreatic outbud; in fact the latter arises as a derivative from the hepatic duct rather than as a distinct outbud from the intestinal tube. (This close association of the hepatic duct with the pancreas is well seen in the arrangement of the concealed pancreas of some teleosts (cf. p. 117, Fig. 196).)

In point of time the liver is the first accessory structure to develop by budding from the primitive alimentary canal, the pancreas and lung following.

In the primitive type of development, as seen in _Petromyzon_ and in the Amphibia, the liver appears very early, as a diverticulum of the embryonic intestinal tube, near its cephalic extremity, projecting on the ventral aspect down into the mass of yolk-cells (Fig. 237). The short stretch of the primitive alimentary canal cephalad of the hepatic diverticulum corresponds to the foregut. With the development of the heart the primitive foregut becomes divided into pharynx and post-pharyngeal segment (oesophagus and stomach). The hepatic diverticulum then lies immediately dorsad of the caudal or venous extremity of the heart. Hence it is probable that the liver is an older organ in the ancestral history of the vertebrates than the pharynx or even the heart. The liver diverticulum lies in close connection with the omphalo-mesenteric veins which return the blood from the yolk-sac to the heart. In the course of further development, as will be seen below, the liver comes into very intimate relations with the venous circulation.

In human embryos of 3.2 mm. the primitive hepatic duct appears as a wide hollow pouch composed of hypoblast cells, growing between the two layers of the ventral mesogastrium, which membrane, extending between the ventral border of the primitive stomach and the ventral abdominal wall, will be subsequently considered in detail. The liver, in developing between the layers of the ventral mesogastrium, approaches very early the _septum transversum_ or rudimentary diaphragm and becomes connected with the same. A mass of mesodermal cells, derived from the mesogastrium and from the primitive mesodermal intestinal wall surrounding the hypoblastic lining of the tube, covers the caecal termination of the primitive hepatic duct, forming the so-called embryonic _hepatic ridge_. This mesodermal tissue accompanies the duct in its further growth and branching, forming the connective tissue envelope, known in the adult as the capsule of Glison. The primitive hepatic duct is directed cephalad in the mesogastrium between the vitelline duct and the stomach (Fig. 101).

In embryos measuring 4.25 mm. the duct is 0.24 mm. long. Later (in embryos of 8 mm.) the primitive single duct divides into two secondary branches, indicating, even at an early stage, the adult arrangement of the duct, as formed by the union of the right and left hepatic ducts (Fig. 185).

The gall-bladder in embryos of this size (8 mm.) is a well-defined caecal diverticulum, branching caudad from the main hepatic duct.

The vesicular mucous surface is thus derived from the enteric hypoblast in the same way as the epithelial lining of the bile-ducts and capillaries. The external muscular and fibrous coats of the gall-bladder are developed from the mesoderm of the mesogastrium.

It is to be noted that at an early stage the gall-bladder is derived from the main duct close to the intestine, the latter duct being very short. Later on the common duct grows in length, making the liver more and more a gross anatomical organ distinct from the intestine. The cystic duct develops as the result of a similar increase in length of the cystic diverticulum. The two principal secondary branches of the hepatic duct give origin to sprouts or buds. These are derivatives of the hypoblastic cells of the larger ducts and may from the beginning be hollow, possessing a lumen continuous with that of the parent duct (Selachians, Amphibians). In warm-blooded animals these sprouts are at first solid, forming the s. c. _hepatic cylinders_, and only subsequently become hollowed out with the further development of the biliary duct system of the liver. The rapid growth of the organ leads to a great increase in the number of the hepatic cylinders. They spread out on all sides, finally coalescing with adjacent buds so as to form an interlacing network whose meshes are filled by blood vessels. After the hepatic cylinders have become canalized they preserve the same arrangement, hence the resulting biliary capillaries of the adult form an anastomosing network. Amphioxus and the amphibians have a single hepatic outgrowth (Fig. 49).

In the Selachians the liver arises as a ventral outgrowth at the hinder end of the foregut immediately in front of the vitelline duct, thus bringing the liver from the beginning into close proximity with the vitelline veins entering the heart. Almost as soon as formed the outgrowth develops two lateral diverticula, opening into a median canal. The two diverticula are the rudimentary lobes of the liver and the median canal uniting them is the rudiment of the common bile-duct and gall-bladder.

In the Teleosts the liver arises quite late (in the trout about the 25th day) as a solid outgrowth from the intestinal canal close to the heart. In the Amniota the liver arises in the same position as in the Anamnia, but, at least in birds and mammals, shows its bifurcation almost, if not quite, from the start. The two forks embrace between them the omphalo-mesenteric or vitelline veins just before they empty into the sinus venosus of the heart.

In the chick the liver appears between the 56th and 60th hour, the right fork being always of greater length but less diameter than the left. The hepatic outbud in the rabbit appears during the 10th day, and during the 11th day begins to send out branches.

In man, as above stated, the bud appears well marked in embryos of 3 mm.

[Certain adult variations make it appear possible that there are two human embryonic hepatic buds, a cranial and a caudal, as is the case in birds.]

=I. B. Comparative Anatomy of the Liver.=--The liver, phylogenetically a very old organ, occurs in all vertebrates, for the caecal diverticulum of the intestine of amphioxus (Fig. 49) has probably the significance of a hepatic outbud.

The primitive form of the liver is symmetrically bilobed, a type which is seen well in the chelonian organ (Fig. 238).

In size the liver is subject to great variations. It is usually larger in animals whose food contains much fat. Hence carnivora in general have a larger liver than herbivorous animals.

Its shape also varies considerably, depending on the form of the body cavity and on the amount and disposition of the available space. Hence in the snakes the organ appears long drawn out, flattened, almost ribbon-like (Fig. 239), while the relatively very large coronal diameter of the body cavity in the turtles permits the liver to expand transversely (Fig. 238).

In general, when the liver is large and the available space for its reception limited, it is usually split into several (two to seven) lobes, which permit, by mutual displacement, the accommodation of the organ to varying space-conditions of the body cavity (Fig. 240). Under the opposite circumstances, on the other hand, even the primitive bilobed character may disappear and the liver is then unlobed (Fig. 241).

The presence or absence of a gall-bladder depends apparently largely on the character of the food and on the habitual type of digestion. In many vertebrates digestion is carried on nearly continuously, without marked interruption, especially in many ungulates, ruminants and rodents. In such animals the gall-bladder is absent. It is also absent in several birds (most Parrots, Doves, Ostrich, Rhea americana, the Cuculidae, Rhamphastos, etc.). This variability emphasizes the morphological fact that the biliary bladder is only a modified portion of the hepatic duct system, as shown by the development above outlined.

A great variety is observed in the arrangement of the biliary ducts, through which, at the period of intestinal digestion, bile passes from the liver and gall-bladder into the intestine, while in the intervals of digestion the secretion is only carried from the liver to the bladder. The following main types of the biliary duct system may be recognized:

1. The hepatic duct joins the cystic to form the common bile-duct, entering the duodenum by passing obliquely through the intestinal wall (Fig. 242). This form is encountered in man and in most mammals. It is also found in some birds (_Buceros_), many amphibians, and in some fish (_Lophius_). Instead of one hepatic duct two may join the cystic duct separately to form the common bile duct (_Phoca litorea_), or the number of hepatic ducts may be further increased. The separate hepatic ducts then unite successively with the cystic duct. This occurs in many mammals (as _Tarsius_, _Galeopithecus_, monotremes) and in some fishes (_Xiphias_, _Trigla_, _Accipenser_) (Fig. 243).

2. Of two hepatic ducts only one helps to form with the cystic duct the common duct, while the other leads from the liver transversely into the bladder, especially into the neck, forming the hepatico-cystic duct (Fig. 244). This arrangement is found in several mammals (calf, sheep, dog).

3. No common bile-duct is formed. The hepatic and cystic ducts each empty separately into the intestine (hepato-enteric and cysto-enteric ducts), while a hepato-cystic duct carries the bile directly from the liver to the gall-bladder (Fig. 245).

_Lutra vulgaris_ among mammalia, the majority of the birds and several reptilia present this type.

When the gall-bladder is absent a single large hepato-enteric duct is found, or instead a number of smaller ducts which enter the intestine successively.

=I. C. Development of Vascular System of Liver.=--In order to comprehend the peritoneal relations of the adult liver it is absolutely necessary to have a clear understanding of the development of the vascular system in connection with the gland.

For our purpose, in the first place, a serial consideration of the successive stages, illustrated by schematic diagrams, will prove most practicable. These diagrams represent the structures in the dorsal view, _i. e._, in the position which they would occupy in the adult liver with the gland resting on its upper or convex surface and with the ventral sharp margin turned toward the beholder (see Fig. 259).

The development of the venous system, especially in connection with the liver, presents a somewhat complicated series of successive conditions. After having become familiar with the principal typical embryonal stages, as shown in the following diagrams, the student is strongly recommended to cement this knowledge by the comparative examination of the venous system. The permanent veins of the lower vertebrates, while in many cases not strictly homologous to those of the higher forms, yet are excellent objects for study, since they serve to illustrate temporary stages in the development of the mammalian venous system, and to that extent are of aid in comprehending one of the most difficult and important chapters in human anatomy. At the conclusion of the diagrammatic consideration of the mammalian development a number of comparative facts will be put together for this purpose.

=1. Early Stage.=--In the earlier developmental stages in mammalian embryos the primitive dorsal aorta extends caudad along the ventral aspect of the vertebral axis, giving off paired vitelline or omphalo-mesenteric arteries to the yolk-sac and allantoic arteries to the embryonic urinary bladder or allantois (Figs. 246 and 247).

The blood is returned from the vascular area of the yolk-sac by two vitelline or omphalo-mesenteric veins, which unite near the heart to form a common trunk, continued as the _sinus venosus_ into the caudal or auricular extremity (venous end) of the primitive tubular heart (Figs. 246, 247 and 248).

=2. Development of Allantois. Stage of Placental Circulation.=--The placental circulation, replacing the temporary vitelline circulation of the earliest stages, is inaugurated by the appearance of two umbilical veins, which pass cephalad, imbedded in the tissue of the ventral mesogastrium, to empty into the sinus venosus near the vitelline veins (Fig. 249). The umbilical veins return the oxygenated blood from the placenta to the embryo. At first the right umbilical vein is the larger of the two.

The sinus venosus at this time also receives two large veins, transversely directed, called the ducts of Cuvier, which are formed near the heart by the union of the anterior cardinal (primitive jugular) and posterior cardinal veins, draining respectively the head end of the embryo, and the body walls and Wolffian bodies.

The vitelline veins are placed on each side of the primitive small intestine, and become connected with each other by a broad anastomotic branch (Fig. 249). When the hepatic outgrowth buds from the duodenum the vitelline veins send out branches which break up into a wide-meshed capillary network in the mesodermic tissue enveloping the hepatic cylinders. Hence at this period the circulation in the vitelline veins is made up of three districts:

(_a_) Distal segment of veins, coursing along duodenum, and joined by a transverse anastomosis, before reaching the liver bud (subintestinal veins).

(_b_) Middle segment, from which capillary vessels are derived, ramifying upon and between the developing hepatic cylinders.

(_c_) Proximal segment, formed by the continuation of the proximal part of the vitelline veins into the sinus venosus of the heart.

=3. Formation of Portal Circulation. A.=--With the further development of the liver the direct connection of the distal segment of the vitelline veins with the sinus venosus becomes lost, the intermediate segment being entirely broken up into an intrahepatic network (Fig. 250). Hence all the blood brought to the liver by the vitelline veins (venae hepaticae advehentes) passes through the hepatic capillary circulation, before it is carried by the proximal segment of the vitelline veins (venae hepaticae revehentes) into the sinus venosus. The amount of this blood increases with new connections which the vitelline veins make with the venous radicles developing in the intestinal tract and its appendages. In proportion as, with the development of the placenta and reduction of the yolk-sac, the original significance of the vitelline veins as nutritive and respiratory vessels disappears, this secondary connection of the vitelline veins with the veins of the alimentary tract becomes more and more important, until finally the original vitelline veins, now properly called omphalo-mesenteric veins, return the blood from the intestinal tube, pancreas and spleen to the liver.

The venae hepaticae advehentes, becoming connected in this way with the developing intestine, pancreas and spleen, form the rudiments of the future portal system, while the venae hepaticae revehentes are prototypes of the hepatic veins of the adult circulation.

=B. Development of the Portal Vein.=--The distal subintestinal segments of the vitelline veins are early united by a transverse anastomotic branch. The section of the veins above this anastomosis is seen already in Fig. 250 to have assumed an annular shape, while the veins below the primary anastomosis are approaching each other to form a second ring-like junction.

In Fig. 251 the subintestinal segments of the two vitelline veins are seen to have communicated with each other by transverse anastomotic branches around the duodenum, two of these branches being situated ventrad and one dorsad of the intestinal tube. These branches, and the portions of the primitive vitelline veins between their points of derivation, form two vascular loops or rings, encircling the primitive duodenum (Fig. 251).

The distal portions of the vitelline veins, before reaching the caudal annular duodenal anastomosis, next fuse into a single longitudinal vessel which also receives the veins from the stomach, intestine, spleen, and pancreas, and forms the beginning of the portal vein.

By atrophy of the right half of the lower, and of the left half of the upper duodenal venous ring (Figs. 252 and 253), the proximal portion of the portal vein is formed as a single vessel, taking a spiral course around the duodenum (Fig. 256). Hence in the adult the portal vein and its principal branch (the superior mesenteric vein) crosses over the ventral surface of the duodenum (third portion), turns along the mesal side of the second portion, and then continues to the liver along the dorsal aspect of the first portion (Fig. 254). _Note_--In comparing Fig. 254 with the schematic figures it should be noted that the same presents the parts in the _ventral_ view, while the schemata offer the _dorsal_ aspect.

=4. Changes Leading to the Final Arrangement of the Umbilical Veins.=--A very important rearrangement of the umbilical veins takes place. These veins originally course in the lateral abdominal wall, close to the fold of the amnion (Fig. 255), and then turn cephalad of the developing liver along the septum transversum to empty into the sinus venosus at each end (Figs. 249 and 250). The right umbilical vein is at first the larger.

This symmetrical arrangement, and the direct connection of the umbilical veins with the sinus venosus, now becomes lost by the occurrence of the following changes:

1. At first (Fig. 249) all the blood carried to the liver by the omphalo-mesenteric veins passes through the hepatic capillary network before being conducted by the venae revehentes to the sinus venosus. Very early, however, a new intrahepatic channel develops, the ductus venosus (Figs. 250-253), which passes obliquely between the entrance of the left omphalo-mesenteric vein into the capillary system (l. v. advehens) and the termination of the right omphalo-mesenteric vein (r. vena revehens) in the sinus venosus.

In human embryos of 4 mm. the ductus venosus can already be distinguished, and in embryos of 5 mm. the vessel has assumed considerable proportions.

2. A communication is next established on both sides between the capillary hepatic network in the portion of the liver nearest to the abdominal wall and the umbilical veins as they ascend imbedded in the abdominal wall (Fig. 251).

This connection is usually from the start larger on the left side and connects with the left omphalo-mesenteric vein just at the point where the same is about to be continued into the ductus venosus. This connection becomes rapidly larger, so that the ductus venosus, which at first appeared merely as an anastomotic channel between the left omphalo-mesenteric vein and the terminal portion of the right omphalo-mesenteric vein, now forms the main continuation of the left umbilical vein. This vessel grows very rapidly up to its connection with the ductus venosus and soon exceeds the right umbilical vein in size (Fig. 252). Beyond the ductus venosus on the other hand the proximal segment of the left umbilical vein diminishes in size, and loses its independent character by incorporation in the hepatic circulation. Only its terminal portion, emptying into the sinus venosus, is preserved. This is surrounded by the growing masses of hepatic cylinders and is converted into a vena revehens.

The connection of the right umbilical vein with the liver vessels is at first symmetrical to that on the left side, but less strongly developed. The effect of this connection is to reduce in the same way the proximal segment of the right umbilical vein and to convert its termination into a vena revehens. With the great development of the left vein, however, the vein on the right side gradually diminishes and finally loses its connection with the intrahepatic circulation altogether. The right umbilical vein is now reduced to a vessel of the ventral abdominal wall, which carries blood in the reverse of the original direction, _i. e._, from the abdominal wall caudad _into_ the left umbilical vein (Figs. 253 and 255).

The connection thus established between the umbilical vein and the portal circulation results in the formation of a single large (the original left) umbilical vein which, throughout the remainder of foetal life, returns all of the placental blood (Fig. 253).

The newly developed hepatic portion of the left umbilical vein becomes, however, not only connected with the ductus venosus, but also with the right part of the upper venous ring, derived from the right omphalo-mesenteric vein (Fig. 253). This connection forms the left portal vein of the adult, and enlarges rapidly.

The terminations of the ductus venosus and of the venae hepaticae revehentes undergo a number of secondary changes in relative position. The left hepatic vein loses its direct connection with the sinus venosus, and now opens into the termination of the ductus venosus, into which the right hepatic vein also empties. This common vessel (v. hepatica communis) subsequently forms the proximal segment of the postcava when this vessel develops (Fig. 256).

The blood, therefore, returned to the liver by the left umbilical vein divides at the transverse fissure into three streams. Two of these pass through the connection with the portal vein and through branches developed from the hepatic part of the umbilical vein into the capillary system of the right and left lobe. The third continues through the ductus venosus to the common hepatic vein and sinus venosus (Fig. 256). The ductus venosus thus becomes the chief vessel returning arterialized placental blood to the heart. When the postcava develops fully the hepatic segment of this vessel also joins the terminal part of the ductus venosus (Fig. 256) and gradually replaces the same as the main returning venous channel, the proximal part of the ductus venosus being incorporated in the vena cava (Fig. 257). The postcava then receives the right hepatic veins separately, while the left hepatic veins and ductus venosus open together into the main vein. This condition obtains up to the time of birth and the consequent interruption of the placental circulation.

While at first the ductus venosus communicates throughout its entire length with the meshwork of the hepatic capillary system, a separation into two segments, _i. e._, ductus venosus proper and intrahepatic segment of umbilical vein, is established after the free communication with the left umbilical vein takes place. This condition is exhibited in Fig. 258, which represents the corroded venous system of the foetal liver, and in Fig. 259, showing an injected liver in the foetus at term.

It will be observed that the umbilical vein on entering the liver gives off a large branch to the left lobe, and a smaller branch on the right side to the quadrate lobe, which act as the main venae advehentes of these portions of the liver. Arrived at the transverse fissure the umbilical vein divides into three branches, at right angles to each other. The left branch enters the left lobe, the right branch becomes directly continuous with the left main division of the portal vein, while the central branch, continuing the direction of the umbilical vein, passes dorsad, as the ductus venosus proper, to join the left hepatic vein close to its entrance into the postcava.

=5. Changes Consequent upon the Establishment of Pulmonary Respiration.=--After birth the umbilical vein and its continuation, the ductus venosus, become obliterated, the former constituting the round ligament of the liver, the latter the ligament of the ductus venosus, both structures imbedded in corresponding portions of the sagittal fissure on the caudal and dorsal surfaces of the adult liver (Figs. 284 and 286). The lateral branches of the umbilical vein, however, in its course from the ventral margin of the liver to the transverse fissure (Fig. 258), remain pervious and are transferred to the portal circulation.

It will be noticed, in reference to the _direction_ of the blood current, that at birth a sudden reversal takes place in the right terminal branch of the umbilical vein at the transverse fissure (Figs. 260 and 261). Before birth the blood current of the umbilical vein divides into three streams, right, left and central. The latter enters the ductus venosus. The left enters the liver directly, the right traverses, from left to right, the segment between the termination of the umbilical and the bifurcation of the portal vein. This segment in the adult carries blood from right to left, as left branch of the portal vein. In the foetus, however, the blood traverses this segment from left to right, in passing from the umbilical to the right branch of the portal vein. The blood entering the liver through the portal vein passes chiefly into the right division of that vessel (Fig. 260).

After birth all the venous blood entering the liver passes through the portal vein. In the right division the direction of the current is the same as in the foetus.

On the left side, however, the current is now from right to left, from the bifurcation of the portal into the channels of the left lobe formerly connected with the umbilical vein (Fig. 261).

Hence the direction of the current in this segment is reversed at birth.

SUMMARY OF HEPATIC CIRCULATION.

The foregoing consideration of the development shows us that the hepatic circulation presents successively three main stages:

=1. Omphalo-mesenteric or Vitelline Stage=, which results in the laying down of the primary capillary circulation of the liver and in the establishment of its connection with the developing veins of the alimentary tract (primitive portal channels).

=2. Umbilical or Placental Stage=, in which the greater part of the blood circulating through the liver is oxygenated blood returned from the placenta by the umbilical vein, accounting for the rapid growth and relatively large size of the organ during foetal life.

The placental blood uses the preformed capillary channels of the vitelline or primitive portal system in the liver, and the same rapidly extend and enlarge with the accelerated growth of the gland. During this stage venous blood is also returned from the alimentary tract to the liver by the portal vein, produced by fusion of the distal segments of the primitive vitelline veins and their secondary connection with the mesenteric, splenic and pancreatic veins (omphalo-mesenteric development of primitive vitelline veins).

=3. Adult or Portal Stage.=--With the interruption of the placental circulation the portal vein assumes again its original position as the only vein carrying blood to the liver. With the establishment of intestinal digestion and absorption this vessel grows rapidly in size.

COMPARATIVE ANATOMY OF THE HEPATIC VENOUS CIRCULATION.

For the purpose of fixing the main facts in connection with the development of the higher mammalian hepatic circulation, and in order to obtain a demonstration of the cycle through which the different veins pass, the student is recommended to examine, preferably by personal dissection, a limited series of lower vertebrates which can be readily procured and easily injected. The following series has been selected, but it will be understood that other forms can be substituted, according to the local conditions which govern the supply of the material.

1. _Fish._ _A Selachian_, the common skate (_Raja ocellata_) or dog-fish (_Acanthias vulgaris_).

2. _Amphibian._ (_a_) Urodele. _Necturus maculatus._ (_b_) Anura. The common _frog_.

3. _Reptile._

Preferably, on account of the ease of injection, one of the larger lizards, as _Iguana tuberculata_.

The turtles, although somewhat more difficult objects to prepare, can be substituted.

4. _Bird._ The common fowl.

5. Human foetus at term.

=1. Fish.=--The venous system can be injected by tying a canula in the lateral vein, and injecting both cephalad and caudad, or by injecting cephalad through the caudal vein. The injection of the systemic veins can also be made caudad through one of the ducts of Cuvier, combined with an injection cephalad of the caudal vein.

The following main facts are to be noted in the venous system of the Selachian (Fig. 262):

=1. There are Two Portal Systems.= (_a_) _Renal Portal System._--The caudal vein divides near the vent into two branches which course along the lateral border of the kidneys, sending _afferent_ or _advehent_ veins into the organ. The blood traverses the renal capillaries and is gathered together by the _efferent_ or _revehent_ veins, which empty into median paired vessels, the posterior cardinals.

(_b_) _Hepatic Portal System._--The veins of the digestive tract and appendages unite to form a hepatic portal vein. The blood after traversing the capillary system of the liver is collected by hepatic veins, which form a dilated hepatic sinus emptying into the sinus venosus of the heart.

2. The middle segment of the intestine, presenting a spiral valve in the interior, gives rise to a vein emptying into the portal vein which corresponds to the subintestinal vitelline vein of the mammalian embryo (Fig. 202).

3. The posterior cardinal veins, also greatly dilated and forming the posterior cardinal sinus, join, near the heart, the veins returning blood from the head, the anterior cardinal or jugular, to form a transversely directed trunk, the duct of Cuvier, which empties into the sinus venosus at the auricular extremity of the heart. Into the duct of Cuvier empties on each side a _lateral vein_ returning the blood from the body walls. This vein can be considered, for our present purpose, as representing in general the abdominal vein of amphibians and reptiles, and the umbilical vein of the mammalian embryo.

The adult selachian venous system is therefore to be considered as illustrating the following conditions above encountered in our study of the embryology of the mammalian venous system.

1. The heart illustrates excellently the stage in the mammalian development, in which auricular and ventricular segments have differentiated, but before the division of the cavities into a pulmonary and systemic portion by the development of the auricular and ventricular septa and the division of the arterial trunk into pulmonary artery and aorta.

The sinus venosus still exists, as an ante-chamber to the auricular cavity proper, receiving on each side the ducts of Cuvier, which represent the fusion product of the systemic veins, anterior and posterior cardinal.

2. The hepatic portal circulation corresponds to the mammalian stage in which the vitelline veins have become omphalo-mesenteric by joining the intestinal veins.

The spiral vein remains as a portion of the original vitelline vein corresponding to the subintestinal segment of the mammalian embryo (cf. Figs. 248 and 249).

The selachian portal vein represents the united vitelline veins, into which the veins of the digestive tract open.

In the liver we find a simple system of venae advehentes, derived from the branching of the portal vein, a hepatic capillary network, and venae revehentes, the proximal remnants of the original vitelline veins which carry the liver blood to the sinus venosus. The condition of the hepatic circulation corresponds therefore to the stage shown in Fig. 250 of the mammalian development. There is as yet no association of the hepatic venous system with the representative of the umbilical vein (the lateral vein of the selachian).

3. The lateral veins, which we can, as stated, regard for purposes of illustration, without prejudging their genetic significance, as representing the mammalian embryonic umbilical veins, still present the condition corresponding to the early mammalian embryonal stage shown in Fig. 250. They are veins of the body walls, emptying cephalad of the liver, directly into the ducts of Cuvier, and through them into the sinus venosus of the heart.

Fig. 262 shows the arrangement of the venous system in a typical selachian diagrammatically.

=2. Amphibian.= (_a_) =Urodele.=--The following points are to be noted in comparison with the preceding form:

1. The two ducts of Cuvier entering into the sinus venosus are formed by the anterior cardinal and subclavian veins, which latter, having appeared with the full development of an anterior extremity, receives the posterior cardinal veins, representing the mammalian azygos system.

2. The renal portal circulation persists. The caudal vein is, however, no longer the only afferent vein of this system. With the full development of a posterior extremity an iliac vein returns the blood from the same and gives a large branch (afferent to the portal renal system), while the trunk continues cephalad as an anterior abdominal vein, corresponding to the lateral selachian vein, emptying in the hepatic portal vein.

3. The efferent veins of the renal portal system no longer unite to form the posterior cardinal, as in the Selachian, but empty into a new median vessel, the inferior vena cava, or postcava, which has replaced the distal segments of the posterior cardinal veins.

The postcava now carries the blood from the kidneys directly to the heart. The original posterior cardinal veins still persist in their proximal segments, as smaller trunks connecting the distal part of the postcava with the ducts of Cuvier through the subclavian veins. The ducts of Cuvier represent the precavae (venae cavae superiores) of mammalia and the postcardinals the mammalian azygos veins.

4. The hepatic portal system differs in two respects from the Selachian type.

(_a_) The blood returned to the liver from the digestive tract by the portal vein becomes mixed before entering the gland with the blood returned from the posterior extremities and abdominal walls by the abdominal vein.

This vein, paired below and continuous with the lateral of the two branches into which the iliac vein divides, becomes united into a single trunk above and empties into the portal vein.

The abdominal vein represents the lateral vein of the Selachian and corresponds to the umbilical vein of the higher vertebrates.

(_b_) The venae hepaticae revehentes do not empty directly into the sinus venosus, but into the proximal portion of the postcava.

Hence the adult urodele venous system illustrates, in reference to the mammalian development, these stages:

1. The umbilical (abdominal) vein has lost its direct connection with the sinus venosus. The proximal segment, cephalad of the liver, has disappeared, and its blood now passes directly into the hepatic circulation by its union with the portal vein.

(Cf. stage schema Figs. 251 and 252.)

2. The postcaval vein has made its appearance, largely replacing the posterior cardinal veins, whose proximal segments became converted into secondary vessels (azygos) uniting the system of the postcava with that of the duct of Cuvier (mammalian praecava), while their distal segments are transformed into the distal portion of the postcava.

The postcava, therefore, is made up of two districts:

(_a_) The proximal portion is a new vessel, developed in connection with the hepatic venous system.

(_b_) The distal portion is derived from the distal segments of the original posterior cardinal veins.

The termination of the hepatic veins in the postcava corresponds to the stage shown in schema Fig. 256.

Fig. 263 gives a schematic representation of the arrangement of the venous system in a typical urodele amphibian (_Salamandra maculosa_).

In Fig. 264 the dissected venous system of _Necturus maculatus_, the mud puppy, is shown in an injected preparation.

(_b_) =Anure.=--The venous system of _Rana esculenta_ is shown in Fig. 265. Comparison with venous system of _urodele_:

1. The abdominal vein, corresponding to the mammalian umbilical vein, has assumed a greater importance in reference to the hepatic circulation. It is a large trunk, continuous below with the pelvic vein, terminating above in two branches, which enter the liver as afferent veins, being joined just prior to the division by the hepatic portal vein.

2. A small cardiac vein, coming from the heart, empties into the angle of bifurcation of the abdominal vein.

3. The postcava is well developed, formed by large efferent renal veins. It entirely replaces the posterior cardinal veins which are absent in the adult animal.

4. A right and left praecaval vein is formed by the union of two jugular trunks with the vein of the anterior extremity and a large musculo-cutaneous vein.

Comparison with the mammalian development: the venous system of this amphibian can be used to illustrate the mammalian embryonal stage shown in schema Fig. 252, in which the abdominal or umbilical vein has become the most important vessel in the afferent hepatic venous system.

The communication existing by means of the cardiac vein between the heart and the hepatic afferent system may suggest, but _purely for illustrative purposes_, the direct connection of the umbilical vein with the heart by the ductus venosus in the mammalian embryo (cf. schema Figs. 250-256).

=3. Reptile.=--In _Iguana_ the renal portal system is well developed. The caudal vein, returning the blood from the tail and the cavernous tissue of the genital organs, continues for a short distance upon the fused caudal end of the two kidneys (Fig. 269) and then divides into two afferent renal veins which ascend on the ventral surface of the glands, giving branches to the renal capillary system. About the middle of the kidney each afferent vein is joined by a large transverse branch from the abdominal vein (Fig. 266).

The renal efferent system begins by a number of inter-renal anastomoses which unite along the mesal border of the right kidney into a large ascending trunk, while the corresponding vessel of the left side, starting from the same anastomosis, is considerably smaller (Figs. 266 and 269). Each of these vessels also receives blood from the testis, epididymis, vas deferens and adrenal body in the male, and from the ovary and oviduct in the female. They represent, in fact, the distal functional part of the right and left embryonic postcardinal vein. Just caudad of the left testis the vein of the left side crosses obliquely ventrad of the aorta and joins the right vessel to form the trunk of the postcava, which enters, immediately beyond the cephalic pole of the right testis, the prolonged caval lobe of the liver (Figs. 266 and 269). Ascending in the substance of this gland and receiving the afferent hepatic veins (Fig. 268), the vena cava emerges from the cephalic surface of the liver greatly enlarged and proceeds to the right auricle.

The abdominal vein divides below into two branches which pass caudad on each side of the bladder, receiving tributaries from the same, to the lateral border of the kidneys (Figs. 266 and 269). Here the vessel is connected by the transverse branch above described with the afferent renal portal system derived from the caudal vein. At the same point it receives the sciatic vein, the principal venous vessel of the posterior extremity. Above, the main abdominal vein, resulting from the union of the two branches referred to, ascends on the dorsal surface of the ventral abdominal wall, receiving a few twigs from the ventral mesogastrium within whose free caudal edge the vessel runs. Just before reaching the liver the abdominal vein turns dorsad on the caudal surface of the gland and joins the hepatic portal vein (Figs. 268 and 275). Several accessory veins, two or three in number, belonging to the system of the abdominal vein, pass above this point from the ventral body wall between the layers of the ventral mesogastrium, to enter the liver separately on its convex ventral surface, above the fusion of the main abdominal vein with the portal vein. These additional branches on entering the liver join the portal system, forming a set of ventral accessory portal veins.

The hepatic portal vein derives its principal tributaries from the splenic, gastric, pancreatic and intestinal veins. One or two additional branches (accessory vertebral portal veins), as above stated, connect the system of the segmental and vertebral veins with the portal circulation, entering the liver separately. In like manner one or two gastric veins (accessory gastric portal veins) enter the dorsal aspect of the liver separately, passing from the stomach to the gland between the layers of the gastro-hepatic omentum (Fig. 275).

Compared with the development of the mammalian type, the venous system of Iguana serves to illustrate the stage in the history of the umbilical vein (represented by the abdominal vein of the reptile) in which the connection of the vessel with the portal vein has been formed and transmits the greater part of the blood returned by the umbilical vein to the liver, while the proximal segment above this point, originally continued into the sinus venosus, has begun to disappear, being, however, still represented by the vessels which, as accessory ventral portal veins, pass in the ventral mesogastrium, from the body wall to the liver.

It will be noted that all the hepatic portal blood, whether conducted by the main portal and abdominal vein, or by the accessory portal branches, traverses the capillary circulation of the liver before entering the postcava.

The vertebral and segmental venous system, representing the azygos veins of the mammalia, is very rudimentary (Figs. 266 and 267). The distal portions of the postcardinal veins form the efferent renal branches and the ascending trunks of the postcava.

The next segment of the vertebral veins appears as a trunk on the right side which enters the portal circulation. A second vein higher up is connected with both the gastric portal system and with the longitudinal chain of the vertebral veins. Finally a proximal venous branch on each side of the vertebral column, representing the upper portion of the postcardinal veins, receives the proximal segmental veins and empties into the subclavian vein (Fig. 267).

=4. Bird.=--The characteristic change in the venous system of the bird, as compared with that of the amphibian and reptile, is found in the nearly complete abolition of the renal portal system. The caudal vein bifurcates, sending on each side a large trunk, which receives the pelvic (int. iliac) veins, to the kidney (renal afferent portal vein), but only a few small branches enter the substance of the gland (Fig. 270, afferent renal V). The main vessel continues cephalad through the kidney and, after receiving the vein from the posterior extremity (femoral), unites as common iliac vein with the vessel of the opposite side to form the postcava. This vessel traverses the liver, receiving the hepatic afferent veins of the portal system. The portal vein is formed by tributaries from the intestinal canal, pancreas and spleen, and is also joined by a large coccygeo-mesenteric vein, which is given off at the point of bifurcation of the caudal vein and receives tributaries from the lower part of the alimentary canal. The abdominal vein of amphibians and reptiles is represented probably by the epigastric vein, which returns the blood from the omental mass of fat to the hepatic veins.

Compared with the mammal on the one hand, and with the lower types on the other, the venous circulation of the bird illustrates the following points:

1. Extensive reduction of the renal portal system and direct formation of postcava by the iliac veins, foreshadowing the condition found in the mammal.

2. Complete separation of the portal and systemic venous circulation in the adult. Disappearance of the ventral abdominal vein as a vessel of the body wall.

=5. Human Foetus at Term.=--The student is recommended to examine, by dissection and injection, the venous system of a foetus at term, noting the following facts:

1. Course of umbilical vein in ventral abdominal wall and along free edge of falciform ligament to liver (Fig. 241), corresponding to the position of the amphibian and reptilian abdominal vein (Figs. 264 and 275).

2. Connection of umbilical vein in liver: (_a_) With portal system (Figs. 258 and 271). ({~GREEK SMALL LETTER ALPHA~}) With portal vein. ({~GREEK SMALL LETTER BETA~}) With portal system of left and quadrate lobes by branches derived directly from umbilical vein while situated in the umbilical fissure (Fig. 258). (_b_) With hepatic veins and postcava by the ductus venosus (Figs. 258 and 271).

3. Connection of the postcaval and precaval systems by the azygos veins representing the proximal segments of the embryonic postcardinal veins (Fig. 272).

If possible the dissection of an injected foetus should be combined with the examination of corrosion preparation of the foetal circulation and especially of the venous system of the foetal liver (Figs. 258 and 271).

3. The remnants of foetal structures in the adult liver (round ligament and ligament of the ductus venosus) should be compared with the structures from which they are derived in the foetus at term (umbilical vein and ductus venosus).

II. THE VENTRAL MESOGASTRIUM.

This membrane has been heretofore mentioned on several occasions. It now remains for us to carefully consider its arrangement in detail, both as regards the peritoneal relations of the liver and in reference to its influence on the abdominal space as a whole. We can best accomplish this purpose by considering the membrane in the first place in a purely schematic manner. In contradistinction to the primitive common dorsal mesentery, which extends the entire length of the alimentary tube, the ventral mesentery, or properly the ventral mesogastrium, is confined to the stomach and proximal portion of the duodenum. We can represent the membrane as extending between the ventral abdominal wall and the ventral border (later the lesser curvature) of the stomach and of the hepatic angle of the duodenum. Cephalad it is connected with the embryonic septum transversum (future diaphragm). Caudad its two layers pass into each other in a free concave edge, including between them the umbilical vein (free edge of falciform ligament of adult). Consequently a schematic profile or lateral view of the membrane and its attachments in the earlier stages would appear as represented in Fig. 273, while the arrangement in transection would be as shown in Fig. 274. It will be observed that the separation of the cephalic portion of the abdominal cavity into symmetrical right and left halves, previously indicated in discussing the primitive stomach and the dorsal mesogastrium, is actually completed by the ventral mesogastrium. This complete separation of the lateral halves of the coelom cavity ceases at the point where the ventral mesogastrium terminates in the free concave edge carrying the umbilical vein. Hence caudad of this falciform edge the two halves of the cavity communicate freely with each other ventrad of the intestine and dorsal mesentery.

This difference in the extent of the mesogastria is perhaps best understood by reference to their relation to the first portion of the duodenum. We have seen that the duodenum in the early stages is attached dorsally by a portion of the common dorsal mesentery, which, after differentiation of the intestinal tract, immediately follows the dorsal mesogastrium proper, forming the mesoduodenum (Fig. 172). The proximal portion of the duodenum (hepatic angle) is still included within the fold of the ventral mesogastrium which membrane terminates immediately beyond this point in the free edge surrounding the umbilical vein (subsequent round ligament) (Fig. 172). The remainder of the duodenum is devoid of any ventral attachment, being only connected to the dorsal body wall by the mesoduodenum (Fig. 197).

Subsequently, after the fourth month, while the right surface of the mesoduodenum and descending duodenum adhere to the parietal peritoneum, the peritoneal investment of the first portion or hepatic angle remains free. This peritoneal covering of the proximal duodenal segment is situated at the point where the caudal end of the ventral mesogastrium, after surrounding the first portion of the duodenum, becomes continuous with the dorsal mesentery forming the mesoduodenum. Obliteration of the latter membrane by adhesion to the parietal peritoneum leaves the first portion of the duodenum invested on both surfaces by the _lesser omentum_, derived from the ventral mesogastrium. The ventral surface of the gut is covered by the ventral layer, the dorsal surface by the dorsal layer of the lesser omentum. These two layers become continuous around the right free edge of the lesser omentum (hepato-duodenal ligament) forming the ventral boundary of the foramen of Winslow (cf. infra, p. 177).

Returning to the schematic consideration of the ventral mesogastrium above outlined (Figs. 273 and 274) we have to note the first important change in the arrangement depending upon the development of the liver. This organ, growing, as we have seen, from the duodenum, extends between the two layers of the ventral mesogastrium, receiving a serous investment from the same. At an early period the liver, developing thus between the mesogastric layers, reaches the septum transversum and becomes closely connected with it, laying the foundation for the subsequent extensive attachment of the gland to the diaphragm.

Extending caudad the liver grows beyond the caudal free edge of the ventral mesogastrium on each side, carrying the serosa with it. Consequently the ventral margin of the liver becomes indented at this point; the umbilical vein and subsequently its fibrous remnant, the round ligament, are imbedded in a notch and fissure (umbilical notch and fissure) continued from the ventral margin dorsad along the caudal surface of the liver (Fig. 259).

This growth of the liver has now effected a division of the primitive ventral mesogastrium into two segments:

1. Ventral portion, between diaphragm and liver, forms the broad falciform or suspensory ligament of the liver.

2. The dorsal portion, between liver and stomach, forms the lesser or gastro-hepatic omentum.

The caudal free edge of the ventral mesogastrium extends between the umbilicus and the caudal surface of the liver, carrying the umbilical vein between its layers. The growth of the liver serves to bury this free edge and the contained vein in a fissure on the caudal surface of the liver. The same obtains in the case of the ductus venosus continued from the umbilical vein (umbilical fissure and fissure of ductus venosus of adult liver). Consequently the original continuity of the broad ligament and lesser omentum, as parts of the primitive ventral mesogastrium, is not readily seen in the adult.

The broad ligament extends across the convex cephalic surface of the liver uniting it to the ventral abdominal wall and diaphragm, while its free falciform edge apparently stops at the umbilical notch in the ventral border of the organ. Actually, however, the obliterated vein is surrounded in the bottom of the fissure, by a peritoneal fold which effects the junction between broad ligament and lesser omentum.

We will see later in what way the permanent adult arrangement of the lesser omentum is brought about. For the present we can state, on the hand of the schematic Fig. 273, that the free caudal edge of the falciform ligament containing the umbilical vein, and the free edge of the gastro-hepatic omentum form together originally the caudal free edge of the ventral mesogastrium, which membrane becomes separated, by the growth of the liver, into suspensory or broad ligament and lesser or gastro-hepatic omentum.

This primitive disposition of the ventral mesogastrium and the viscera connected with the same, is well shown in some of the lower vertebrates in whom the development never proceeds beyond the early mammalian stages. Fig. 275 shows in profile view from the right side the situs viscerum and peritoneum in _Iguana tuberculata_.[5] The two dorsal aortic roots are seen to unite to form the main aorta, which descends between the layers of the dorsal mesentery, sending branches to the dorsal margin of oesophagus and stomach. From the opposite border of the stomach the ventral mesogastrium is derived. Its dorsal segment (gastro-hepatic omentum) connects liver and stomach, carrying between its layers the portal vessels, hepatic artery and biliary duct. The ventral segment of the membrane, forming the suspensory or broad ligament, extends between abdominal wall and ventral surface of the liver. Caudad, the lesser omentum and the suspensory ligament are seen to have a common concave falciform edge.

[5] _Iguana tuberculata_, one of the large lizards native of South America. This animal forms an excellent object for the comparative study of the visceral and vascular anatomy of the abdomen. It possesses a well-differentiated intestinal tract, several coils of small intestine, a well-marked caecum and large intestine. The examination of this or a similar reptilian form is to be highly recommended. Iguana is easily obtained in any of our large cities, as a considerable number of these animals are annually imported from Mexico and the South American states.

The ventral abdominal vein ascends between the layers of the suspensory ligament and near the liver becomes connected by a large branch with the portal vein. A few smaller branches are seen passing from the abdominal wall beyond this point. In this reptile, therefore, the permanent vascular arrangement corresponds to an early human embryonic stage.

The reptilian ventral abdominal vein is the homologue of the umbilical vein of the placentalia. The large branch passing to the portal vein represents the connection established in the human embryo between the umbilical and portal veins. The small branches, continuing cephalad between the mesogastric layers, represent the temporary proximal remnants which in the human embryo the umbilical veins form in connection with abdominal walls. The permanent adult arrangement of this part of the vascular system in this animal corresponds therefore to one of the stages of development in the human embryo, as previously indicated (cf. p. 149; Figs. 251 and 252).

PERITONEAL RELATIONS OF LIVER.

It is well to begin the study of the peritoneal connections of the liver with the consideration of the embryonic stage shown in Fig. 273 schematically.

If we imagine this embryonic liver detached from its connections in such a manner as to leave the divided peritoneal layers of the ventral mesogastrium as long as possible, and if we regard the preparation from behind, the appearance of the parts could be represented in Fig. 276.[6]

[6] I am indebted to Dr. J. A. Blake, former Assistant Demonstrator of Anatomy at Columbia University, for the valuable suggestion which led to the preparation of Figs. 276, 277 and 278 together with the correlated text.

It will of course be seen that the area of direct adhesion to the diaphragm, extending transversely, would separate the lesser omentum from the suspensory ligament.

As is seen in the transection (Fig. 274), the right and left layers of the suspensory ligament, at its attachment to the liver, turn into the visceral peritoneum investing the organ on its ventral and cephalic surfaces. Continuing around the borders of the liver this visceral peritoneum then invests in like manner the dorsal or caudal surface directed toward the stomach, until, at the region of the future portal or transverse fissure, this visceral peritoneum becomes in turn continuous with the two layers of the lesser or gastro-hepatic omentum. Consequently in the embryonic detached liver the lines of peritoneal reflection would be nearly cruciform, the vertical limb of the cross being formed on the cephalic surface by the two layers of the suspensory ligament, while on the caudal surface it is formed by the layers of the lesser omentum. The horizontal arm of the cross is formed by the upper and lower limits of the area of diaphragmatic attachment, along which the parietal diaphragmatic peritoneum turns into the visceral hepatic investment (forming the two layers of the primitive coronary ligament). In the liver shown thus schematically from behind we would overlook the dorsal and adjoining portions of the cephalic and caudal surfaces of the adult human liver.

The primitive biliary duct, portal vein and hepatic artery reach the liver between the layers of the lesser omentum. The venae revehentes (hepatic veins) reach the sinus venosus at the attachment of the liver to the septum transversum (primitive diaphragm).

The first important change, resulting in a rearrangement of these peritoneal layers, is produced by the connection of the umbilical with the rudimentary portal vein.

This junction occupies a relatively wide area on the caudal surface of the liver, and the layers of the lesser omentum are separated somewhat at this point to accommodate the enlarging vascular structures between them. More especially is this the case with the right leaf of the primitive gastro-hepatic omentum. A species of lateral diverticulum is formed by this leaf so as to include the umbilical vein at its junction with the portal. The membrane in the region of this diverticulum turns its surfaces dorsad and ventrad, and its free edge toward the right (Fig. 277). With the gradual increase in the size of the vessels, and with the transverse position which the rotation of the stomach imparts to the opposite border of the lesser omentum attached to the lesser curvature, this transversely disposed portion gradually exceeds in length and size the part of the original omentum enclosing the umbilical vein. This vessel and the investing peritoneum become lodged in a sagittal depression on the caudal surface of the liver (rudimentary umbilical fissure), while the transverse portion, developed as indicated, surrounds the structures connected with the liver at the future transverse or portal fissure.

Schematically this rearrangement of the hepatic peritoneal lines of reflection can be shown in Fig. 278.

It will be observed that in this way a small part of the caudal surface of the right lobe has become partially marked off from the remainder as a rudimentary Spigelian lobe, bounded ventrally by the transverse fissure and lesser omentum attached to the same; to the left by the two layers of the lesser omentum containing the ductus venosus; while the limit cephalad is afforded by the reflection of peritoneum from liver to diaphragm, forming part of caudal layer of right coronary ligament. To the right this rudimentary Spigelian surface is directly continuous with the rest of the dorsal and caudal surface of the right lobe (Fig. 277). Finally a definite right limit is given to the Spigelian lobe by the increasing size of the postcava and its closer connection with the liver. This vessel now assumes the position of the main venous trunk entering the heart from below.

This inclusion of the vena cava in the fissure or fossa of that name on the dorsal surface of the liver affords, so to speak, the vertical measure of the non-peritoneal area of the liver attached directly to the diaphragm. As the vein develops the interval between the two layers of the right coronary ligament increases, producing the well-known large non-peritoneal area on the dorsal surface of the adult liver, which is directly attached to the diaphragm.

Immediately to the left of the vena cava, however, the original condition persists. The area of direct diaphragmatic attachment is narrow and consequently the two layers of the coronary ligament are close together at this point.[7]

[7] It should be remembered that in the final adult arrangement of the abdominal viscera the liver shifts relatively backwards, so that the diaphragmatic attachment, originally directed cephalad, now looks dorsad and forms part of the dorsal or "posterior" surface of the adult organ. The original ventral surface looks cephalad, as well as ventrad, forming the convex surface which in the adult rests in contact with the abdominal wall and diaphragmatic vault, while the surface originally directed dorsad toward the stomach finally in large part has an inclination caudad forming the "inferior" surface of human anatomy.

In this way a species of recess (Spigelian recess or hepatic antrum of lesser sac) is formed. A portion of the dorsal liver surface lying just to the left of the vena cava, between it and the ductus venosus, remains invested by peritoneum which is reflected from the boundaries of this space to the diaphragm. This forms the Spigelian lobe (Fig. 278).

The lobe is bounded to the right by the postcava, to the left by the reflection of the lesser omentum to the stomach along the fissure for the ductus venosus; cephalad the boundary is formed by the reflection of the caudal layer of the coronary ligament to the diaphragm.

The caudal boundary is afforded by the transverse position which the lesser omentum has assumed in the region of the transverse or portal fissure.

It will be seen that the original continuity of the Spigelian lobe with the caudal surface of the right lobe is maintained by the narrow bridge of liver tissue connecting the caudal right angle of the rectangular Spigelian lobe with the right lobe. This narrow isthmus, situated between vena cava dorsad and the free right edge of lesser omentum ventrad, forms the so-called _caudate lobe_.

Fig. 279 shows a human foetal liver at the end of the eighth month in the view from below and behind. The original continuity of the layers of the lesser omentum, attached along the fissure for the ductus venosus, with the fold of the falciform ligament occupying the umbilical fissure can still be made out for a short distance beyond the left extremity of the transverse fissure. The section of the lesser omentum which occupies the transverse fissure and, including the portal vein, hepatic artery and duct between its layers, terminates in the free right margin, is evidently derived by a lateral extension from the right layer of the primitive sagittal lesser omentum, whose original direction is preserved along the fissure of the ductus venosus.

In Fig. 280 the lines of peritoneal reflection on the cephalic, dorsal and caudal surfaces of a human foetal liver at term are shown.

We can now proceed to trace the reflection of the peritoneum from the liver to adjacent structures.

Begin with the caudal layer of the coronary ligament on the extreme right, where fusion with the corresponding cephalic layer produces the right triangular ligament. The caudal layer of the coronary ligament proceeds from right to left along the caudal margin of the non-peritoneal dorsal diaphragmatic surface of right lobe, being reflected along this line from the liver to the adjacent portions of the diaphragm and ventral surface of right kidney and suprarenal capsule (hepato-renal ligament). A small cephalic part of ventral surface of right suprarenal capsule lies above this line of reflection, is hence non-peritoneal and firmly connected with the liver just to the left of entrance of vena cava into the caval fissure. Continuing, the caudal layer of the coronary ligament crosses the ventral surface of the vena cava and turns, immediately to the left of the vein, at a right angle, ascending to form the left boundary of the Spigelian recess, being reflected along this line from the left margin of the caval fissure to the pillars of the diaphragm. Arrived at the opening of the central tendon permitting passage of vena cava into pericardium, and at the level of the entrance of the left hepatic vein into the cava, the peritoneum turns again at a right angle and runs from right to left, forming the cephalic limit of the Spigelian recess. Turning caudad along the fissure for the ductus venosus, as right leaf of that portion of the lesser omentum which is attached to this fissure and has preserved its sagittal position, the peritoneal line of reflection reaches the left extremity of the portal or transverse fissure. It now turns to the right following the fissure as the dorsal layer of the transverse segment of the lesser omentum, and becomes continuous, with the formation of a free right edge, with the ventral layer of the same membrane, passing from right to left, the two layers including between them the structures entering and leaving the liver at the transverse fissure (portal vein, hepatic artery, duct). Arriving at the left extremity of the transverse fissure the ventral layer of the transverse segment of the lesser omentum--as we practically trace it in the adult as a free membrane--turns directly into the left leaf of the sagittal segment attached along the fissure for the ductus venosus, and becomes continuous along the dorsal border of the left lobe with the caudal layer of the left coronary ligament. This direct continuity, as just stated, exists practically in the adult. From the development of the membrane, however, it will be seen that the ventral layer of the transverse lesser omentum, at the left extremity of the portal fissure, becomes continuous with the right layer of the primitive mesogastrium enclosing the umbilical vein. After surrounding this vein it is continued into the left leaf of the same membrane, which in turn passes into the left layer of the portion attached along the fissure for the ductus venosus.

This original connection can at times be traced very clearly in young specimens (Fig. 279), and occasionally is also still evident in the adult liver.

Usually, however, the round ligament of the adult and its investing peritoneum is buried so deeply in the umbilical fissure, or even bridged over in part by liver tissue, that the connection is not evident. The ventral layer of the transverse omentum then appears directly continuous with the left layer of the sagittal omentum attached along the fissure for the ductus venosus.

We can sum up the facts just considered as follows:

1. The rotation of the stomach from the sagittal into the transverse position, and the development of the umbilical and portal veins, rearrange the original sagittal plane of the lesser omentum, dividing it into two districts:

(_a_) Cephalic portion, remaining in the original sagittal plane, follows the fissure for the ductus venosus. With the incorporation of the Spigelian lobe in the adult dorsal or "posterior" surface of the liver, this segment of the omentum assumes a vertical direction, forming the left boundary of the Spigelian recess, being reflected from the fissure for the ductus venosus to the abdominal portion of the oesophagus and the part of the lesser curvature of stomach adjacent to the cardia.

(_b_) Distal caudal portion of the lesser omentum is twisted laterally and turned to the right by the change in the position of the stomach and the development of the structures connected with the liver at the transverse fissure. It is reflected from this fissure to the distal part of the lesser curvature and to the first portion of the duodenum. This transverse segment of the lesser omentum is a secondary derivative from the right leaf of the primitive membrane, produced by the enlarged area for entrance of umbilical and portal veins at the transverse fissure. It lies ventrad of caudal border of Spigelian lobe.

2. The distal segment of the original omentum containing the umbilical vein (round ligament), continues imbedded in the umbilical fissure, to the ventral margin of the liver, where it joins the layers of the suspensory ligament passing over the cephalic surface.

3. The adult lesser omentum at the transverse fissure may be regarded as a diverticulum of the right leaf of the primitive embryonal sagittal omentum.

With the reduction of the umbilical vein after birth to form the round ligament this structure becomes deeply buried in the umbilical fissure. The ventral and dorsal layers of the lesser omentum at the transverse fissure thus become continuous with respectively the left and right layers of the second segment of the omentum which ascends vertically along the fissure for the ductus venosus.

4. The cephalic layer of the coronary ligament (Fig. 280) remains practically in the embryonic condition. The adult convex cephalic surface of the liver is traversed in the sagittal direction by the suspensory ligament which connects it with the abdominal surface of the diaphragm, and thus effects the division into right and left lobes on the convex surface. Arrived at the dorsal border of this surface (junction of "superior" and "posterior" surfaces) the right and left leaves of the falciform ligament turn at right angles into the cephalic layer of the right and left coronary ligament, which at each extremity meet the right and left caudal layers to form the triangular ligaments. It will thus be seen that the apparent irregularity in the relative arrangement of the s. c. "upper" and "lower" layers of the coronary ligaments, produced by the Spigelian recess, is only a difference in the interval between the two layers, caused by the vertical extent of the non-peritoneal direct diaphragmatic attachment of the right lobe to the right of the vena cava.

=Comparative Anatomy of Spigelian Lobe and Vena Cava in the Cat.=--The lines of peritoneal reflection in the _cat's_ liver and the arrangement of the Spigelian lobe and recess are seen in Fig. 281, taken from a preparation hardened in situ.

Compared with the human liver it will be noted that the area of diaphragmatic adhesion is much less developed. The dorsal surface of the right lobe to the right of the postcava is peritoneal, there being no extension laterad of the right coronary and triangular ligaments. The postcava enters the liver in a special prolongation of the liver substance (caval lobe).

The boundaries of the Spigelian recess and the lines of attachment of the gastro-hepatic omentum correspond to the human arrangement.

RELATION OF THE HEPATIC PERITONEUM TO THE "LESSER SAC."

_Foramen of Winslow._--We have previously seen that the rotation of the stomach and the further growth of the dorsal mesogastrium lead, in the first instance, to the formation of the "lesser peritoneal cavity." This cavity is in fact primarily the retrogastric space created by the transverse position of the stomach, augmented by the cavity of the omental bursa developed from the dorsal mesogastrium.

We have now to consider the additional boundaries of this space contributed by the peritoneal connection of the lesser curvature with the liver.

The lesser omentum follows, of course, along its gastric attachment to the lesser curvature the general direction of the stomach, passing from the cardia transversely downwards and to the right. We distinguish the two layers of the adult membrane as ventral and dorsal, which meet in the free right edge and include between them the main structures entering and leaving the liver at the transverse fissure, viz.: the portal vein, hepatic artery and bile-duct.

The lesser omentum therefore prolongs the plane of the stomach cephalad towards the liver and thus forms the continuation of the ventral boundary of the lesser peritoneal sac. We can now consider the line of its hepatic attachment in the light of the facts previously adduced, and combine the same with the line of gastric attachment to the lesser curvature. Fig. 282 shows the foetal liver and stomach in their relative position in the dorsal view, and Fig. 283 gives the lines of the peritoneal reflections. The vertical segment of the omentum, occupying the fissure for the ductus venosus, passes to the cardiac part of the lesser curvature, its ventral layer covering the ventral and left side of the oesophagus, while its dorsal layer passes to the dorsal and right side of the oesophagus at its entrance into the stomach. The transverse segment of the omentum, attached on the liver to the portal or transverse fissure, accedes to the pyloric part of the lesser curvature. Of course the ventral and dorsal layers of the omentum are continuous with the serous visceral investment of the ventral and dorsal surfaces of the stomach.

Fig. 284 shows this right-angled course of the lesser omentum at the hepatic line of attachment in a preparation of the abdominal viscera hardened in situ, with the segment of the stomach between the cardiac and pyloric orifices removed. The arrow is passed behind the right free edge of the lesser omentum. This portion of the membrane is still intact, not having been disturbed by the removal of the body of the stomach, and includes between its layers the structures connected with the liver at the transverse fissure (duct, hepatic artery and portal vein). The lesser omentum is seen to be attached to the liver along the transverse fissure (Fig. 284, _A_) and along the fissure for the ductus venosus (Fig. 284, _B_), constituting the transverse and vertical segments above referred to, which pass into each other at the angle of junction between the transverse fissure (left end) and the fissure for the ductus venosus (Fig. 284, _C_). The caudal and left border of the Spigelian lobe is exposed by the division of the omentum, and the extent of the Spigelian or hepatic recess of the lesser peritoneal sac is shown. Fig. 285 shows the liver, stomach and lesser omentum of a Macaque monkey hardened in situ, and demonstrates still more conclusively that the uniform curve of the omentum along the lesser curvature of the stomach becomes a broken line at the hepatic attachment, the angle being placed at the left end of the transverse fissure at the point where the same encounters the fissure for the ductus venosus.

In Fig. 286 finally the hardened abdominal viscera of an adult human subject are shown in the ventral view with the lesser omentum incised. The cut through the lesser omentum exposes the hepatic recess of the lesser peritoneal cavity immediately to the left of the foramen of Winslow. Toward the right free margin of the omentum the divided portal vein, hepatic artery and duct are seen between the layers of the omentum imbedded in the pancreas and coursing behind the first portion of the duodenum on their way to the transverse fissure.

To the left of these structures the omental tuberosity of the pancreas projects above the level of the lesser curvature under cover of the secondary parietal peritoneum forming the dorsal wall of the lesser sac, while the lower edge of the Spigelian lobe appears in the upper angle of the incision.

If we remember that the liver is itself welded to the diaphragm between the layers of the coronary ligament (Fig. 280), it will become apparent that the serous surface of the Spigelian lobe forms part of the ventral wall of a peritoneal recess situated behind the lesser omentum, between this membrane and the diaphragm. Access to this recess, without the division of peritoneal layers, can only be obtained by passing from right to left, along the caudate lobe, between the vena cava behind, covered by parietal peritoneum, and the free right edge of the lesser omentum in front. (In the reverse direction of the arrow shown in Fig. 284.) This hepatic or Spigelian recess of the lesser peritoneal cavity has categorically the following boundaries (Figs. 282 and 283):

Dorsal: Parietal peritoneum, reflected along the line CD, from the caudal layer of the coronary ligament to the diaphragm.

Ventral: Visceral peritoneum investing the Spigelian lobe and the gastro-hepatic omentum.

Right: Reflection of peritoneum along the line DE (caval fissure) to become the parietal peritoneum covering the diaphragm.

Left: Right layer of lesser omentum, reflected along the fissure for the ductus venosus (CB) to the cardiac portion of the lesser curvature, continuous with the dorsal layer of the lesser omentum reflected from the transverse fissure to the pyloric segment of the lesser curvature (AB).

We will presently see that certain relations of the vessels connected with the liver at the transverse fissure and of the duodenum prevent the finger, when passed from right to left behind the free right edge of the lesser omentum and along the caudate lobe of the liver, from proceeding downward at this point. A narrow channel of communication is thus formed between the Spigelian recess and rest of the lesser sac on the one hand, and the general greater peritoneal cavity on the other. This channel is the so-called foramen of Winslow.

Having once passed this narrow space the finger will be in the Spigelian recess and can palpate its boundaries. Further progress cephalad and to the right is barred by the diaphragmatic adhesions of the liver just detailed. But in the direction downward behind the lesser omentum and along the dorsal surface of the stomach, as well as to the left toward the spleen the excursion is limited only by the length of the examining finger.

After opening the abdominal cavity of the human adult, elevating the liver and depressing the stomach, the hepatic attachment of the lesser omentum can be traced as already described. It will then be observed that the gastric attachment of the membrane lies in one plane following the lesser curvature while the hepatic attachment forms a broken line, with the angle situated at the left extremity of the transverse fissure. The vertical segment of the hepatic attachment, occupying the fissure for the ductus venosus, turns at this angle into the transverse segment which follows the transverse fissure to its right extremity where the two layers pass into each other around the right free omental margin (hepato-duodenal ligament). Consequently we overlook, in an abdominal cavity thus exposed, the entire caudal surface of the liver, including the caudal surfaces of right, left, and quadrate lobes. The junction of right and caudate lobes can be seen between vena cava and right edge of the omentum, or rather, it can be felt at this point. But the Spigelian lobe, turning its surface dorsad against the parietal peritoneum covering the diaphragm, forms part of the "posterior" liver surface and is not visible, although--as just stated, it can be palpated by passing the finger through the foramen of Winslow. The Spigelian lobe cannot be overlooked in its entire extent until the liver is removed from the body and regarded from behind. The caudal edge (continuation of its right angle into the caudate lobe and papillary tubercle) can be seen by tearing through the layers of the lesser omentum and lifting the liver up forcibly (Fig. 286).

=Caudal Boundary of Foramen of Winslow.=--We have above referred to the fact that the finger introduced through the foramen of Winslow meets in this canal with resistance if an attempt is made to pass downwards. After passing this constricting point the free excursion into the Spigelian recess and behind the omentum and stomach and toward the spleen can be performed.

In considering the elements which produce this narrowing of the communication between the two peritoneal sacs at the foramen of Winslow we have to deal with two factors, one primary and constant, the other secondary and inconstant.

1. The first of these is afforded by the arrangement of the arterial vessel supplying the liver. The hepatic artery is a branch of the coeliac axis, furnishing arterial blood to the liver tissues and supplying, in addition, branches to the stomach, duodenum and pancreas.

This vessel is, of course, placed primarily, like all other arterial branches supplying the alimentary tract, between the layers of the primitive dorsal mesentery. Originally the vessel supplies the distal (pyloric) portion of the stomach along its dorsal attached border (subsequently the greater curvature) corresponding to the adult gastro-epiploica dextra of the hepatic (gastro-duodenalis).

It likewise gives branches to the adjacent pyloric portion of the duodenum and the pancreas, as that gland develops from the intestine, corresponding to the adult superior pancreatico-duodenal branch, and to the ventral border (lesser curvature) of stomach, corresponding to the adult pyloric branch of the hepatic.

With the development of the liver from the duodenum arterial branches derived from this primitive gastro-duodenal vessel pass to the sprouting hepatic cylinders by continuing around the duodenum, beneath its serous investment, to reach the interval between the two layers of the ventral mesogastrium, in which the liver develops, near the free margin of this membrane.

After the rotation, which turns the right side of the stomach, duodenum and mesoduodenum dorsad, the branch which passes over the dorsal surface of the duodenum to reach the liver becomes more favorably situated and develops into the main hepatic artery which reaches the liver at the transverse fissure between the folds of the lesser omentum. The original right side of the duodenum, now turned dorsad, adheres to the parietal peritoneum. The hepatic artery which reached the liver by passing over this surface of the duodenum, beneath its visceral serous covering, becomes imbedded in connective tissue by the adhesion of the visceral duodenal and the primitive parietal peritoneum. Hence in the adult the hepatic artery courses imbedded in the connective tissue which binds the duodenum to the abdominal background to reach the interval between the two omental layers which carry it to the transverse fissure.

The hepatic artery, therefore, derived from one of the primitive intestinal branches (gastro-duodenal) is, notwithstanding its hidden position in the adult, originally situated between the layers of the free primitive dorsal mesogastrium.

It now becomes necessary to regard the development of the great omentum from the primitive dorsal mesogastrium in relation to this course of the hepatic artery. We have seen that the great omentum and the cavity of the omental bursa is produced by the extension of the dorsal mesogastrium to the left and caudad, subsequent to the rotation of the stomach. The splenic artery and the left gastro-epiploic branch pass from the coeliac axis to the left between the layers of the mesogastrium, as previously seen (Figs. 291 and 292).

The hepatic artery, however, is so to speak placed on the border line between the portion of the primitive mesentery which, as dorsal mesogastrium, is to turn to the left and caudad to form the great omentum, and the portion which, as mesoduodenum, turns to the right and passes to the duodenal loop (Fig. 287).

In the further course of development the dorsal mesogastrium grows more and more, forming the omental bag, while the mesoduodenum on the other hand becomes anchored early and obliterated as a free membrane by adhesion of its original right layer to the primitive parietal peritoneum. The hepatic artery runs on the line dividing these two different mesenteric segments. We can imagine, so to speak, that the redundant growth of the omentum to the left and caudad, takes place over the hepatic artery as a resistant support (Figs. 288 and 289). Cephalad of the hepatic artery is the developing omentum, caudad of the vessel the mesoduodenum. The artery follows the cephalic limit of the mesoduodenum and becomes, as stated, adherent to the abdominal background in the segment between its origin from the coeliac axis and the point where, after having crossed the dorsal surface of the duodenum, it enters the right edge of the lesser omentum on its way to the liver.

=Pancreatico-gastric Folds.=--If we open the lesser peritoneal cavity by dividing the gastro-hepatic omentum and look into the background of the retro-omental space, we will see a fold of the secondary lining parietal peritoneum (derived from the mesogastrium), which can be traced from the cephalic border of the pancreas to the pyloric extremity of the stomach. This fold carries the hepatic artery to the lesser omentum behind the first portion of the duodenum, and is called the right or main pancreatico-gastric fold. A similar fold, further to the left, carries in a like manner the coronary artery of the stomach to the cardiac end of the lesser curvature. This fold forms the left or secondary pancreatico-gastric fold. Between the two folds the caudal margin of the Spigelian lobe projects into the lesser cavity.

The appearance of the two pancreatico-gastric folds in the adult human subject is well seen in Fig. 284.

Fig. 290 shows the abdominal cavity of _Nasua rufa_, with great omentum divided to bring into view the vessels passing from coeliac axis to liver and stomach and elevating the retrogastric parietal peritoneum to produce the pancreatico-gastric folds.

(The course of the hepatic artery from coeliac axis to liver in the dorsal view in the cat is seen in Fig. 223.)

Figs. 291 and 292 represent schematically cross-sections directly through the foramen of Winslow, showing the method by means of which the hepatic artery reaches the upper border of the duodenum and the effect of the adhesion of duodenum and mesoduodenum upon the disposition of the vessel.

The coronary artery, like the splenic, is at first situated between the layers of the dorsal mesogastrium (vertebro-splenic segment). Like the splenic the coronary artery becomes anchored to the abdominal background and placed secondarily behind the parietal peritoneum of the lesser sac by the adhesion of this mesogastric segment to the primitive parietal peritoneum. To reach the lesser curvature at the cardia and to run thence from left to right along the lesser curvature between the layers of the gastro-hepatic omentum, the vessel raises the investing parietal peritoneum (originally the right leaf of the dorsal mesogastrium) into a crescentic fold, extending between its origin from the coeliac axis at cephalic margin of pancreas and the beginning of the lesser curvature of the stomach. Hence this fold is called the left pancreatico-gastric fold. (Seen well in Fig. 284.)

In the next place it must be borne in mind that the relation of the primitive hepatic artery to the vascular supply of the stomach, pancreas and duodenum produces a permanent shortening of the primitive mesentery at this point. This result is indicated in the schematic figures 287, 288 and 289.

In the original condition the dorsal mesentery, passing to a practically straight intestinal tube, is of uniform sagittal measure (Fig. 287).

As development proceeds, and as the liver grows from the duodenum, the hepatic artery develops from the primitive pyloric vessel as above indicated. This vessel, assuming greater importance with the rapid growth of the liver, is not lengthened out as happens with the remaining purely intestinal branches which follow the increase in the length of the intestinal canal. The hepatic artery, therefore, will mark the point where the original short sagittal extent of the primitive mesentery will tend to be preserved. Cephalad of this point the dorsal mesogastrium grows out into the great omentum (Figs. 288 and 289); caudad of the same point the membrane, in following the development of the intestine, becomes drawn out into the permanent mesentery and mesocolon.

The hepatic artery, in addition, marks the cephalic limit of the adhesion which anchors the duodenum and mesoduodenum to the parietal peritoneum. Consequently in the adult the vessel courses in as direct a manner as possible, taking the shortest course from the coeliac axis to the liver, passing dorsad of the duodenum and giving what now appear as secondary branches to supply the intestine, the stomach and pancreas (pyloric and gastro-duodenal arteries (pancreatico-duod. superior and gastro-epiploica dextra)).

Even if no fixation of the duodenum and mesoduodenum takes place this course of the hepatic artery will produce a constricted passage between the liver (caudate lobe) cephalad, abdominal parietes and aorta dorsad, lesser omentum and pyloric duodenum ventrad, and hepatic artery caudad. This passage leading from the general peritoneal cavity into the retrogastric space is the _primitive foramen of Winslow_. This condition is well represented in the abdominal cavity of some of the lower mammalia, in which duodenum and mesoduodenum remain permanently free.

Fig. 293 shows a view of the abdominal cavity from the right side in a specimen of the ant-eater, _Tamandua bivittata_.

The right kidney is seen in the background, covered by the parietal peritoneum. The duodenum and mesoduodenum are free and can be turned toward the median line. The opening of the foramen of Winslow leading into the retrogastric space is seen between the liver cephalad, kidney and vena cava dorsad, lesser omentum and pyloric extremity of the stomach ventrad, and a fold of peritoneum carrying the hepatic artery caudad. Exactly similar conditions prevail in the cat and in many other mammals.

It will be seen in all these instances that neither portal vein nor bile-ducts limit the foramen caudad. These structures can be lifted up and turned toward the median line with the free duodenum and mesoduodenum. But the hepatic artery must pass to the liver from the _retroperitoneal coeliac axis_. In doing this the vessel traverses the cephalic border of the pancreas, and the pyloric extremity of the stomach and duodenum, to reach the lesser omentum which conveys it to the liver.

Consequently there must always be a narrow peritoneal neck between the liver cephalad, aorta dorsad, hepatic artery caudad, and pyloric extremity of stomach and duodenum together with the lesser omentum ventrad. It should be remembered that the vessel which extends after the development of the liver into the lesser omentum as the _hepatic_ artery, was originally destined for the supply of these latter structures. In the adult these primary embryonic terminal branches to the intestine appear as secondary branches derived from the hepatic as the main vessel. Their origin, however, serves to keep the beginning of the small intestine in comparatively close connection with the hepatic artery which courses over the dorsal surface of the duodenum to reach the liver. The narrow space thus left between aorta, hepatic artery, duodenum, lesser omentum and liver forms the framework of the foramen of Winslow and appears always as a confined and narrow channel. This relation is shown in the accompanying schematic Figs. 294 and 295 which represent a sagittal section through the foramen. This primitive foramen is thus bounded cephalad by the liver (caudate lobe, connecting Spigelian and right lobes), ventrad by the first portion of the duodenum and the lesser omentum, with hepatic artery behind the intestine and between the omental layers; dorsad by the abdominal background and large retroperitoneal vessels, and caudad by the coeliac axis and beginning of the hepatic artery.

2. In the forms which possess in the adult an adherent duodenum and mesoduodenum, as in man, the foramen of Winslow obtains a secondary caudal limit by the agglutination of the descending duodenum and the parietal prerenal peritoneum. This is the secondary and inconstant factor referred to above in the caudal boundary of the foramen. The result of this anchoring of duodenum and mesoduodenum is to bring the margin of the foramen further to the right and to bury the hepatic artery still further from view. Thus in the adult human subject the structures bounding the foramen at the margin of the entrance into the narrow channel would be above caudate lobe of liver, behind postcava, below duodenum adherent to ventral surface of right kidney, in front first portion of duodenum and lesser omentum. The hepatic artery will be felt on introducing the finger through the foramen in its original position, but it will be seen that the actual boundaries of the foramen have been moved so to speak a little further to the right by the duodenal adhesion.

Fig. 296 shows a complete dissection of the adult human viscera and vessels concerned in the formation of the foramen, hardened in situ.

The stomach is removed, dividing of course the coronary artery and vein and the left gastro-epiploic artery. The portal vein, hepatic artery and bile-duct are seen entering and leaving the liver at the transverse fissure. Behind them and to the right the vena cava enters the liver. The hepatic artery distributes its pancreatico-duodenal branches to the duodenum and pancreas. The left angle of the Spigelian lobe and the fissure for the ductus venosus appear to the left of the portal vein and hepatic artery. The right angle of the Spigelian lobe and its continuation into the right lobe by means of the caudate lobe is hidden by the structures occupying the transverse fissure. We would enter the beginning of the foramen of Winslow by passing between the vena cava behind, the structures in the transverse fissure (portal vein, hepatic artery and duct) in front, caudate lobe of liver above and duodenum below, the latter in the undisturbed condition of the parts adherent to the right kidney. Continuing to the left the finger would pass between aorta behind, coeliac axis and hepatic artery below and in front, and liver above. These structures bound the permanent and primary narrow channel of communication between the retrogastric or lesser peritoneal space and the general peritoneal cavity, which exists even if a free duodenum and mesoduodenum allow us to lift the intestine away from vena cava and right kidney.

The main facts pertaining to the structure of the lesser peritoneal sac and its connection with the greater peritoneal cavity by means of the foramen of Winslow may be summed up as follows:

The mesogastrium as a whole, expanding originally in the sagittal plane in a fan-shaped manner between the vertebral column and the ventral abdominal wall, from the level of the umbilicus to the septum transversum (diaphragm), divides the cephalic part of the abdominal cavity into a symmetrical right and left half.

Figs. 172 and 273 represent the membrane as seen in a profile view from the left side. We distinguish the segment dorsad of the stomach as the dorsal mesogastrium, directly continuous with the remaining segments of the common primitive dorsal mesentery, while the portion ventrad of the stomach forms the ventral mesogastrium in which the liver develops. The segment of the ventral mesogastrium between liver and stomach becomes the lesser or gastro-hepatic omentum, while that between liver and ventral abdominal wall forms the falciform or suspensory ligament.

A transection, showing the dorsal and ventral mesogastrium at the level of the fundus of the stomach, is given in Fig. 298. The mesogastria are here seen to be short, while in the schematic Figs. 291 and 292 the membrane is, for the sake of distinctness, represented as being of considerable extent.

The ventral mesogastrium surrounding the liver and stomach extends caudad to include the first portion of the duodenum. Beyond this point it terminates in a thickened free edge which includes the umbilical vein. This vein extends from the umbilicus to the transverse fissure of the liver (Fig. 297), lying within the umbilical fissure on the caudal surface of the gland.

At the point where the vein enters the liver the thickened margin of the ventral mesogastrium is continued, as ligamentum hepato-duodenale, to the upper part of the duodenum and forms the ventral boundary of the foramen of Winslow. Between the layers of the mesogastrium which meet in this margin are situated the portal vein, biliary duct and hepatic artery, together with the nerves and lymphatics of the liver.

The mesogastrium originally divided the abdominal cavity between umbilicus and diaphragm into symmetrical right and left halves of equal size and extent. This early symmetrical arrangement becomes disturbed about the seventh week by the rotation of the stomach and the resulting altered course of the mesogastrium, which render the two original equal halves of the abdominal cavity unequal and asymmetrical. The original right half becomes placed behind the stomach and is converted into a blind sac with its opening directed to the right.

The communication of the general abdominal cavity with the retrogastric space by means of this channel is still wide in the embryo, but gradually becomes narrowed in the course of further development to form the foramen of Winslow. This opening is situated between the hepato-duodenal ligament and the parietal peritoneum covering the vena cava. It is constricted from below by the curve of the hepatic artery as this vessel passes from the coeliac axis to reach the liver at the transverse fissure between the layers of the lesser omentum.

The earlier developmental stages of the higher mammalian embryos are in general well illustrated by the permanent adult conditions found in some of the lower vertebrates, in which development does not proceed beyond the primitive condition.

In reptiles, birds and mammals the epiploic bursa is generally formed, while in amphibia the dorsal mesogastrium is very short and connects the stomach directly to the dorsal midline of the abdominal cavity without forming the sac-like extension of the great omentum.

The dorsal mesogastrium with the stomach, and the ventral mesogastrium including the liver between its layers, divides in these animals the cephalic part of the body cavity into two halves, corresponding to the earlier embryonic stages in man and in the higher mammalia.

The foramen of Winslow of the higher forms appears in the lower vertebrates as the wide-open space leading from below into the right half of the coelom cavity. The dorsal mesogastrium remains short, not forming the pouch-like extension of the great omentum. The stomach retains more or less its primitive vertical position without rotation or elevation of the pyloric extremity, and the intestinal canal is simple, short and comparatively straight.