View source for Paper - The morphology and permeability of the roof of the fourth ventricle in some mammalian embryos (1938)

{{Ref-CohenDavies1938}}
=The morphology and permeability of the roof of the fourth ventricle in some mammalian embryos=

By Henry Cohen and Sarah Davies

From the Department of Medicine, University of Liverpool

==I. Introduction==

Ina previous communication (Cohen & Davies, 1937) it was reported that in the chick embryo the first escape of cerebrospinal fluid from the central nervous system was coincident with the development of the choroid plexuses of the brain. Two membranous areas, one anterior and the other posterior, in the roof of the fourth ventricle functioned as the sites of fluid escape. The objects of the present investigation are: (1) To discover whether or not a similar correlation exists in mammalian embryos. (2) To identify specialized membranous areas, if such exist, in the roof of the fourth ventricle of these types. (8) To ascertain whether such areas allow of any fluid escape during embryonic life and, if so, whether their permeability is in any way influenced by injections of a hypertonic solution into the blood stream.

==II. Survey of Literature==

Very few workers have investigated the morphology of the roof of the fourth ventricle in vertebrates during development.

In 1917 Weed recorded the presence of two membranous areas, one anterior and the other posterior, in the roof of the fourth ventricle of pig and of human embryos at different stages of development. The anterior membranous area is ’ a transitory structure disappearing early in embryonic life. The posterior area, on the other hand, is persistent and forms the greater part of the inferior choroid velum. He also identified an anterior membranous area in the embryos of chick, rabbit, sheep and cat.


Further, Weed replaced the fluid within the central nervous system of living pig embryos at various stages of development by a 1 % double solution of potassium ferrocyanide and iron ammonium citrate. Subsequent fixation in an acid medium secured precipitation of Prussian blue (ferric ferro-cyanide) in situ and thus revealed the path taken by the injected solution. Weed found that the injected fluid escaped into the surrounding mesenchyme through the membranous areas in the roof of the fourth ventricle, the first escape being coincident with the development of the choroid plexuses of the brain.


Keegan (1917), investigating rabbit and chick embryos, recorded the presence of.an anterior and posterior membranous area in the roof of the fourth ventricle. A detailed description of the complete history of these areas is not given. He injected a 1% double solution of potassium ferrocyanide and iron ammonium citrate into the central nervous system of living embryos of various ages, but found that no extra-ventricular spread occurred, even at stages when the choroid plexuses were well developed.

According to Flexner (1929), a membranous area is present in the roof of the fourth ventricle of amblystoma embryos. In the earlier stages of development it forms the major portion of this roof and is continuous throughout, showing no differentiation into anterior and posterior portions. The subsequent disappearance of this area during later stages is brought about by metamorphosis of its constituent cells into typical ependymal cells and by the development of the choroid plexus.

Owing to the size of amblystoma embryos, injections were not made into the central nervous system. Flexner assumes, however, that the membranous area in the roof of the fourth ventricle functions as the site of extra-ventricular passage of fluid. The basis of his assumption is that a corresponding area in mammalian embryos has definitely been shown (Weed, 1917) to perform this particular function.

Cohen & Davies (1937), investigating chick embryos, identified and described two membranous areas in the roof of the fourth ventricle, one anterior and the other posterior to the developing choroid plexus. The anterior area is a transitory structure, disappearing at the 9th day of incubation, while the posterior area persists beyond this stage. Injections of a 0-8 % double solution of potassium ferrocyanide and iron ammonium citrate were made into the central nervous system of living embryos of different ages. It was found that an extra-ventricular spread of the injected solution occurred, first through the anterior area and later through the posterior area. The first escape of fluid was correlated with the development of the choroid plexuses of the brain.

Much work has been done to determine the influence of varying salt concentrations in the blood on the cerebrospinal fluid circulation in adult Mammalia. Of particular interest are the observations recorded by Najiagas (1921), Weed (1922), Foley (1923), and Forbes e¢ al. (1928).

Najiagas (1921) replaced the intra-ventricular fluid of both normal and hydrocephalic kittens by a 1 % double solution of. potassium ferrocyanide and iron ammonium citrate. This procedure was followed by intravenous injections of a hypertonic salt solution. Subsequent fixation in an acid medium secured precipitation of Prussian blue (ferric ferrocyanide) in situ and thus indicated the channels along which the replacement solution had passed. In both normal and hydrocephalic animals it was found that absorption occurred through the ventricular ependyma into the capillary bed of the nervous system. Intravenous administration of a hypertonic salt solution served to hasten the process _ of absorption. In all animals, normal and hydrocephalic, the epithelial cells of the choroid villi showed no invasion by the Prussian blue granules.

Weed (1922), investigating cats and dogs, increased the salt content of the blood by intravenous injections of a hypertonic saline solution. He then injected a 1% double solufion of potassium ferrocyanide and iron ammonium citrate into the subarachnoid spaces of these animals. His findings are in close ‘agreement with those of Nafiagas. The Prussian blue precipitate had passed through the ependymal lining of the cerebral ventricles into the underlying capillary network. No absorption was observed through the epithelium of the choroid plexuses.

Foley (1928) made an investigation on the lines essentially similar to those adopted by Weed. He found that increased osmotic value of the blood caused a retrograde flow of fluid from the subarachnoid spaces into the ventricles of the brain. The distribution of Prussian blue granules indicated that the foreign solution had passed not only through the ependyma but also through the epithelium of the choroid villi.

Forbes e¢ al. (1928), investigating cats, injected a 1% double solution of potassium ferrocyanide and iron ammonium citrate into the subarachnoid spaces. After a short interval, a hypertonic salt solution was administered either intravenously or into the peritoneal cavity. In all cases examined, precipitated Prussian blue granules were observed in the lumen of the vessels in the choroid plexuses.

Thus, according to Nafiagas (1921), and Weed (1922), the choroid plexuses take no part in the intra-ventricular absorption of the cerebrospinal fluid even when the osmotic pressure of the blood is increased. On the other hand, Foley (1928) and Forbes et al. (1928) state that under the influence of intravenous injections of a hypertonic solution the plexual epithelium absorbs fluid from the ventricles.

A search of the literature has failed to reveal any data concerning the influence of hypertonic solutions in the blood stream on the extra-ventricular passage of cerebrospinal fluid during embryonic life. The Roof of the Fourth Ventricle 433

==III. Material==

The findings herein recorded are based on observations of rabbit, rat and guinea-pig embryos.

Only in the case of rabbits was it possible to obtain embryos of accurately determined ages. In the case of rats and guinea-pigs the dates of mating, and therefore the exact ages of the embryos, were not known. A rough indication of their stage of development was obtained, however, by measurement of the embryos along the mid-dorsal line from the crown of the head to the root of the tail. Since there is considerable variation in size, not only among embryos of the same age from different mothers, but even among embryos from the same uterus, such lengths are clearly no satisfactory criterion of age; nevertheless, they do serve as an approximate indication of the stage of development.

==IV. Methods of Investigation==

In all cases the mother was anaesthetized either with chloroform or ether. The entire uterus with the embryos in situ was quickly removed and placed in normal saline solution warmed to a temperature of 37—38° C.

Immediately prior to injection an incision was made in the uterine wall in such a way that the placenta remained intact. The embryo, still attached by its umbilical stalk to the placenta, was exposed by careful removal of the chorion and amnion and kept in the warm saline solution throughout the injection.

The embryos from each animal were divided into two groups, each group containing at least two embryos. In each embryo of one group injections were made firstly into the blood stream, and then after a period of 4-5 min. into the central nervous system. In each embryo of the other group injections were made into the central nervous system alone.

For the purpose of injection into the vascular system, a 2 % saline solution warmed to a temperature of 37—38° C. was used. The embryonic heart was exposed through an incision in the mid-ventral line of the thoracic wall. The injection was made by means of a short glass tube of 8 mm. bore drawn out into a fine capillary point at one end and provided with a small rubber bulb at the other. A small quantity of the saline solution was injected under very slight pressure into the left ventricle of the heart and this procedure repeated until the blood vessels of the head region became so pale as to be hardly distinguishable.

The fluid used for injection into the central nervous system was a double solution of potassium ferrocyanide and iron ammonium citrate. Other workers in this field (Weed, 1917; Keegan, 1917), who have adopted a similar experimental method, used a solution of 1 % concentration, as this is claimed to be isotonic with mammalian blood. Therefore, in order to make our results valid for comparison with theirs, a solution of similar concentration! was used throughout the present investigation. Since a detailed account of the method employed for injecting the embryonic central nervous system has already been communicated (Cohen & Davies, 1937) a brief statement will suffice here. The solution was injected into the caudal end of the central canal of the spinal cord. Any fluid contained in the central nervous system was simultaneously withdrawn from the lateral ventricle of either cerebral hemisphere, thereby eliminating undue increase in intra-cerebral pressure.


* 1 The solution was prepared by dissolving 0-5 g. of each salt in 100 c.c. of solution.

As soon as the injection was completed the embryo, still in the warm saline solution and attached to its portion of the uterine wall, was placed in an incubator at 37-38° C. Here it was kept alive for about 3-1} hr., according to the size of the embryo, to allow of any possible spread of the injected solution. Owing to the transparency of the integument, particularly in the earlier stages of development, the embryonic heart-beat was easily observed and served as the criterion of prolonged circulation.

The embryos were then fixed overnight in a solution containing 10% formaldehyde and 1% hydrochloric acid. At the same time the acid in the fixing agent secured precipitation in situ of Prussian blue (ferric ferrocyanide) and thus revealed the path taken by the injected solution. After fixation, the embryos were dehydrated in a series of suitably graded alcohol solutions, cleared in xylol and embedded in paraffin wax of 52 or 56° C. m.P. according to season. They were then serially sectioned in the longitudinal plane at a thickness of 4 and stained in Ehrlich’s or Weigert’s haematoxylin and eosin.


==V. The Morphology and Permeability of the Roof of the Fourth Ventricle in Certain Mammalian Embryos at Different Stages of Development==

(a) Results of injections of a true solution into the central nervous system

Histological observations are reported below on rabbit, rat and guinea-pig embryos in which the central nervous system was injected with a 1% double solution of potassium ferrocyanide and iron ammonium citrate.

===Rabbits===

The earliest embryos which could be injected with any measure of success were those of 11 days. Thereafter specimens were injected at intervals of approximately 24 hr. until the 18 days’ stage was reached.

====11 days’ embryos====
The wall of the brain stem was wholly ependymal in nature and was of variable thickness. The roof of both third and fourth ventricles was continuous and unfolded throughout and showed no indication of a choroid formation. The roof of the third ventricle consisted of an ependymal layer 3—4 cells deep. The individual cells were columnar in shape and possessed pale granular cytoplasm and round or slightly elongated nuclei. The nucleoli and nuclear membrane were intensely stained in contradistinction to the pale nucleoplasm. The roof of the fourth ventricle was triangular in shape, the wide base being directed anteriorly and the narrow apex posteriorly. The major portion of this roof consisted of a single layer of cuboidal cells with granular cytoplasm and round clearly defined nuclei. These nuclei were so densely stained that their nuclear membrane and chromatin mesh could only be distinguished with great difficulty. In the anterior region of the roof (see Text-fig. 1) a small median oval area was visible where the cells were extremely flat, consisting of a narrow, elongated nucleus with a thin cytoplasmic investment. This area was abruptly demarcated from the remainder of the roof, its flat cells contrasting sharply with the cuboidal cells of the surrounding epithelium.

The injected fluid was seen wholly within the central nervous system, no extra-ventricular spread of dye being evident. The Prussian blue granules, intermingled with coagulum, extended in uniform distribution along the central canal of the spinal cord and in the cerebral ventricles. They showed no condensation associated with any particular region in the brain stem.




Text-fig. 1. Rabbit, 11 days.


====12 days’ embryos====
The wall of the brain stem was thicker than previously, the greater part consisting of ependyma associated with a narrow stratum of overlying nerve tissue. The thalamencephalic roof, however, was thinner than elsewhere and consisted of ependyma only. It was smooth and unfolded and still showed no indication of villus formation. The myelencephalic roof presented the same general appearance as in the preceding specimens, but the membranous area in its anterior region was relatively larger.

The injected solution still lay wholly within the central nervous system but was no longer uniformly distributed in the cerebral ventricles. There was a slight accumulation of blue granules intermingled with coagulum adhering to, and coextensive with, the ventricular surface of the membranous area in the roof of the fourth ventricle. The individual cells of this area, however, showed no impregnation with the dye.

====18 days’ embryos====
The roof of the thalamencephalon was thinner than in preceding stages but still showed no indication of folding. The roof of the fourth ventricle (see Pl. I, fig. 1) was now divided into anterior and posterior regions by a very shallow depression in the transverse median plane. The cells participating in this invagination tended to be more columnar than cuboidal. Apart from a slight difference in shape they resembled very closely the cells of the surrounding epithelium, showing a similar degree of granulation and wealth of nuclear chromatin content. The membranous area in the anterior portion of the roof was relatively larger than previously, occupying practically the entire roof region anterior to the median depression. In the posterior portion a small, median, oval area was distinguishable where the cells were very low cuboidal in stature. Their nuclei, however, were still rounded or only a little elongated. This slightly differentiated posterior area was not sharply delimited from the surrounding epithelium but showed gradual mergence with it.

The injected fluid was still retained within the spaces of the central nervous system, and was again slightly condensed below the membranous area in the anterior region of the myelencephalic roof. No such condensation was evident below the slightly flattened area in the posterior region, but merely a sparse distribution of granules adhering to its ventricular surface.

====14 days’ embryos====
At this stage the median transverse invagination in the roof of the fourth ventricle was more emphasized than previously and its cells more markedly columnar. The majority of these cells retained on the whole their granular appearance. A few, however, were seen to contain a small spherical vacuole lying in the cytoplasm and closely associated with the.nucleus. The ventro-lateral walls of the median fold now exhibited a slight degree of undulation, but no true choroid villi were evident. The anterior membranous area in the myelencephalic roof was somewhat smaller than in the preceding stage, but was still abruptly demarcated from the surrounding epithelium. The slightly differentiated posterior area, on the other hand, was relatively larger although it still merged imperceptibly into the general roof layer.

As in earlier stages, the injected solution was still confined within the central canal of the spinal cord and ventricular system of the brain. The blue granules showed a similar degree of concentration below the membranous areas of the myelencephalic roof as in the previous specimens.

====15 days’ embryos====
The roof of the thalamencephalon was narrower than hitherto, consisting of a thin ependymal layer 1-2 cells deep. This roof was now slightly undulating and formed several shallow, irregular folds invaded by the overlying mesenchyme. The median depression in the myelencephalic roof (see Pl. I, fig. 2) was deeper than previously and the columnar shape of its constituent cells more pronounced. The undulations in the ventro-lateral walls of this fold, first observed in embryos of 14 days, were here more marked, forming short, unbranched, incipient villi. The epithelial cells of the villi possessed densely granular cytoplasm and round or slightly elongated, welldefined nuclei. In a large number of these, however, the nucleus with its cytoplasmic investment was relegated to the lower (ventricular) portion of the cell, the upper or aventricular portion being occupied by a large spherical vacuole. In a few cells the reverse condition obtained. The membranous area in the anterior region of the fourth ventricle roof was smaller in extent than previously, but its border still showed an abrupt transition from the surrounding epithelium. The posterior area, on the other hand, showed considerable increase in size, exceeding the anterior area both in length and in width. This area was now sharply delimited from the general roof epithelium and its constituent cells were very flat, consisting of a narrow, elongated nucleus with a scanty investment of cytoplasm.

The distribution of Prussian blue granules was similar to that recorded in preceding stages, no extra-ventricular spread of the dye being evident.

====16 days’ embryos====
The invaginations of the thalamencephalic roof were more pronounced than previously. Moreover, they now extended anteriorly for a short distance into the lateral ventricles via the foramina of Munro, but did not form true branching choroid villi. At this stage gross morphological changes were evident in the myelencephalic roof. The median transverse fold was extremely well developed reaching almost to the floor of the ventricle. The incipient villi in the ventro-lateral walls of the fold were more marked than hitherto and exhibited a slight degree of branching, forming an embryonic choroid plexus. In section each villus consisted of an axial core of highly vascular mesenchyme surrounded by an epithelium one cell deep. The cells were columnar in shape and possessed pale granular cytoplasm and round, well-defined hyperchromatic nuclei. In the majority of the cells, the nucleus with its investing cytoplasm was relegated to the lower ventricular region of the cell, while the upper or aventricular portion was entirely occupied by a large vacuole. In a few instances the vacuole was so extensive that it practically filled the entire cell, causing the nucleus to be greatly compressed. The anterior region of the fourth ventricle roof was considerably shorter in longitudinal section than previously, owing apparently to encroachment by the developing cerebellar anlage and choroid plexus. Consequently the membranous area of this region was now partly incorporated into the anterior wall of the median transverse fold. This area was still sharply delimited from the surrounding epithelium, but was somewhat smaller in extent than in the preceding stage. The posterior membranous area, on the other hand, was relatively larger, and occupied the greater part of the roof region posterior to the choroid plexus.

The injected dye was still confined within the spaces of the central nervous system. A dense accumulation of blue granules adhered to the ventricular surface of both membranous areas in the medullary roof, but in each case their constituent cells showed no invasion by the dye.

====17 days’ embryos====
The degree of undulation in the thalamencephalic roof was much more pronounced than hitherto, The resultant villi extended for a considerable distance into the lateral ventricles and exhibited a slight degree of branching, forming an incipient choroid plexus (see Pl. I, fig. 8). Each villus in section consisted of a central core of vascular mesenchyme surrounded by an epithelium one cell deep. The individual cells were columnar in shape and possessed pale, granular cytoplasm and round or slightly elongated, hyperchromatic nuclei. The majority of the cells were uniformly granular, but a few in the extreme anterior portion of the plexus showed a slight degree of vacuolation. In these particular cells a small spherical vacuole was seen in the upper aventricular region, whilst the nucleus and its cytoplasmic investment lay in the lower ventricular portion. The transverse median fold in the roof of the fourth ventricle (see Text-fig. 2) was deeper than previously and the branching of the villi in its ventro-lateral walls was more pronounced. Further, the epithelial cells of the villi showed an even higher degree of vacuolation than in the preceding specimen. The anterior region of the medullary roof was even smaller in extent with the result that the median fold lay immediately behind the posterior lip of the cerebellum. The membranous area in the anterior portion of the roof showed even further reduction in size and was now relegated entirely to the upper part of the anterior wall of the median fold. Moreover, this area no longer showed abrupt demarcation from the surrounding epithelium but gradually merged into it. Its constituent cells had lost their flat appearance and were now low cuboidal in stature. The posterior membranous area was very extensive and at this stage occupied practically the entire posterior region of the roof.


Text-fig. 2. Rabbit, 17 days.



The distribution of dye was similar to that in preceding stages, no extra cerebral passage being evident. 18 days’ embryos. The choroid plexuses of the lateral ventricles were extremely well developed and their villi exhibited a higher degree of branching than previously. Moreover, the epithelial cells of these plexuses now showed a marked degree of vacuolation. In the majority of the cells the nucleus with its investing cytoplasm was relegated to the extreme ventricular periphery, whilst the aventricular portion of the cell was occupied entirely by a large vacuole. The morphology of the medullary roof was essentially similar to that of the preceding stage. The choroid villi of the fourth ventricle, however, were even more numerous, extending anteriorly below the cerebellum, and they showed a higher degree of branching. The epithelial cells of the villi were again markedly columnar in shape and their cytoplasm was extremely vacuolate. The membranous area in the anterior wall of the median fold showed even further regression than previously. The posterior membranous area, on the other hand, was even more extensive and now occupied the entire roof region posterior to the choroid plexus.

As in all earlier stages, the injected dye was wholly retained within the spaces of the central nervous system, no extra-ventricular spread being evident. A particularly dense accumulation of blue granules adhered to, and was coextensive with, the ventricular surface of the membranous area in the posterior region of the medullary roof, but its individual cells were entirely free of the dye. No such condensation was evident below the vestigeal anterior membranous area.

===Rats===

The series of {{rat}} embryos in which the central nervous system was injected ranged from 1-3 to 2-5 cm. in length. Two embryos were obtained measuring 0-6 cm., but owing to their size no injection was made into their central nervous system. These specimens were examined histologically, however, to determine the morphology of the brain at this stage.

====Embryos measuring 0.6 cm====
The wall of the brain stem was wholly ependymal in nature and was practically of uniform thickness. The roof of the fourth ventricle was typically triangular in shape, the ‘wide base being directed anteriorly and the narrow apex posteriorly. This roof was continuous and unfolded throughout and consisted of a cuboidal epithelium one cell deep. The individual cells possessed pale, granular cytoplasm and round, clearly defined nuclei touching both the inner and outer cell walls. The nucleoplasm was pale in contradistinction to the intensely stained nucleoli and nuclear membrane.

====Embryos measuring 1.3 cm====
The wall of the brain stem was thicker than in the preceding stage and consisted of ependyma associated with a very narrow stratum of overlying nerve tissue. The roof of the third and fourth ventricles was in each case smooth and unfolded throughout and showed no indication of a choroid formation. The roof of the thalamencephalon consisted of an ependymal layer 3—4 cells deep. The individual cells were columnar in shape and possessed densely granular cytoplasm. Their nuclei were round or slightly elongated in a plane at right angles to the longitudinal axis of the body, and possessed a deeply staining nuclear membrane and 2-3 nucleoli in their chromatin mesh. The roof of the fourth ventricle again consisted of a single layer of granular cuboidal cells. In the anterior region of this roof (see Textfig. 3) a large, median oval area was distinguishable where the cells were extremely flat, consisting of a narrow elongated nucleus surrounded by a thin cytoplasmic investment. This area was abruptly delimited from the surrounding epithelium, its flat cells contrasting vividly with the cuboidal cells of the general roof layer.

The injected solution was wholly retained within the central canal of the spinal cord and ventricular system of the brain. There was a dense accumulation of Prussian blue granules, intermingled with coagulum, adhering to, and coextensive with, the ventricular surface of the membranous area in the myelencephalic roof. The cells of this area, however, showed no impregnation with the dye. Neither was there any spread of the injected solution from any other region in the brain stem.




Text-fig. 3. Rat, 1-3 cm.

====Embryos measuring 1.5 cm====
The roof of the thalamencephalon was somewhat thinner than previously but was still smooth and unfolded. At this stage the roof of the fourth ventricle was clearly divided into anterior and posterior portions by a shallow depression in the median transverse plane. The cells participating in the fold differed from the typical epithelial roof cells in that they tended to be more columnar than cuboidal and their nuclei oval rather than rounded. The membranous area in the anterior portion of the myelencephalic roof was larger than in the preceding specimens, extending from the cerebellar anlage to the median depression.

The injected solution was strictly confined within the spaces of the central nervous system, no extra-ventricular spread being evident.

====Embryos measuring 1.7 cm====
The thalamencephalic roof was much thinner than hitherto, consisting of a columnar epithelium one cell deep, and was markedly undulating. The simple irregular folds so formed were invaded by the overlying mesenchyme and extended anteriorly for a short distance into the lateral ventricles via the foramina of Munro. The median transverse invagination in the roof of the fourth ventricle was deeper than in the previous stage (see Text-fig. 4) and its cells more columnar in shape. Moreover, its ventro-lateral walls showed a slight degree of folding but no true branching choroid villi were evident. The anterior portion of the myelencephalic roof was shorter in longitudinal section than formerly, owing apparently to encroachment by the developing cerebellum. The membranous area in this region now showed a slight reduction in extent and was partly incorporated into the anterior wall of the median fold. This area was still sharply delimited from the surrounding epithelium and its constituent cells retained their flat appearance. In the posterior region of the roof a small median, oval area was visible where the cells showed a slight tendency to flattening, but their nuclei were still rounded or only a little elongated. This area was not sharply delimited from the surrounding epithelium but showed gradual mergence with it.



Text-fig. 4. Rat, 1-7 cm.

The injected dye was still retained within the central nervous system and was particularly condensed below the anterior membranous area in the roof of the fourth ventricle. No such condensation was seen below the slightly flattened posterior area, but merely a fine distribution of granules adhering to its ventricular surface.

====Embryos measuring 1.8 cm====
At this stage the undulations in the thalamencephalic roof were more pronounced than previously and the incipient choroid villi so formed extended forward for a considerable distance into the lateral ventricles. The median transverse fold in the roof of the fourth ventricle was more emphasized than in the preceding specimen and its ventro-lateral walls more markedly undulating, forming short unbranched incipient villi. The anterior portion of the myelencephalic roof was even shorter in extent, and the membranous area in this region of the roof was incorporated yet further into the anterior wall of the median fold. The area, however, was still sharply 442 _ Henry Cohen and Sarah Davies

delimited from the surrounding epithelium. The slightly flattened posterior membranous area was relatively much larger, but still showed gradual mergence with the remainder of the roof.

The distribution of the injected solution was similar to that in the preceding specimens, no'extra-cerebral passage of dye being evident.

====Embryos measuring 1.9 cm====
These embryos presented the same general histological features as in the previous stage, but the median depression in the myelencephalic roof was deeper and its ventro-lateral walls showed a higher degree of folding. The anterior membranous area of this same roof showed even further reduction in size and was now wholly incorporated into the anterior wall of the fold. The posterior area on the other hand was relatively very large. At this stage it was abruptly demarcated from the surrounding epithelium and its constituent cells were extremely flat.

The dye was still confined within the spaces of the central nervous system and was now particularly condensed below the membranous area in the posterior region of the roof of the fourth ventricle. No such condensation was evident below the anterior area of this roof.

====Embryos measuring 2.1 cm====
In these specimens the villi in the thalamencephalic roof were very pronounced and, as they extended forward into the lateral ventricles, showed a slight degree of branching, forming an embryonic choroid plexus. In section each villus consisted of an axial core of loose-meshed vascular mesenchyme surrounded by an epithelium one cell deep. The cells were columnar in shape and possessed granular cytoplasm and round, welldefined nuclei. The majority of these cells were uniformly granular, but a few in the extreme anterior portion of the plexus showed a marked degree of vacuolation. The nucleus with its investing cytoplasm was relegated to the lower (ventricular) portion of the cell, the upper or aventricular region being occupied by a large spherical vacuole. The median transverse fold in the myelencephalic roof was very deep, extending almost to the floor of the ventricle (see Pl. I, fig. 4), and its constituent cells were markedly columnar. The incipient villi in its ventro-lateral walls were more pronounced than previously and were slightly branched. Many of the cells in the epithelium of the villi contained a large spherical -vacuole in their aventricular portion, the nucleus and its cytoplasmic investment lying in the lower ventricular half. The membranous area in the anterior region of the fourth ventricle roof was even smaller than in the preceding stage and its border was no longer sharply defined but merged gradually into the surrounding epithelium. The constituent cells of this area had lost their flat appearance, being low cuboidal in stature, and their nuclei were now round or only slightly elongated. The posterior membranous area was even more extensive than previously, occupying almost the entire roof region behind the choroid plexus.

The injected dye was still confined within the central canal of the spinal cord and ventricles of the brain. No extra-cerebral spread was observed from any area in the brain stem. The Roof of the Fourth Ventricle 443

====Embryos measuring 2:3cm====
In these specimens the villi of the choroid plexuses of the brain showed an even higher degree of branching than previously and now the majority of their epithelial cells were vacuolate. The anterior membranous area in the medullary roof showed even further regression and was vestigeal in appearance.

The distribution of Prussian blue granules was similar to that of the preceding stage, no extra-ventricular spread being evident.

====Embryos measuring 2:5 cm====
At this stage the choroid villi of the plexuses were very numerous and highly branching, and the cells of their epithelium extremely vacuolate. The anterior portion of the medullary roof was very short in longitudinal section, consequently the median transverse fold lay immediately behind the cerebellar anlage. The vestigeal membranous area in this region of the roof had now vanished entirely. The posterior area was very extensive and occupied the entire roof region posterior to the choroid plexus.

As in all previous specimens, the injected solution was confined within the spaces of the central nervous system. At this stage the blue granules were heavily condensed below the membranous area in the posterior portion of the medullary roof, but the individual cells of this area were entirely free of the dye.

===Guinea-pigs===

The present series of embryos in which injections were made into the central nervous system ranged from 1-6 to 3-2 cm. in length. One embryo measuring 0-4 cm. was available to us, but owing to its size the central nervous system was not injected. Nevertheless, this specimen was examined microscopically to determine the morphology of the brain at this stage.

====Embryo measuring 0.4 cm====
The wall of the brain stem consisted of a layer of ependyma only. The roof of both third and fourth ventricles was continuous and unfolded throughout and showed no indication of villus formation. The roof of the fourth ventricle was triangular in shape, the wide base being directed anteriorly and the narrow apex tapering into the spinal cord. This roof consisted of a single layer of cuboidal cells with granular cytoplasm and round, clearly defined nuclei. In its anterior region a small, median, oval area was distinguishable where the cells showed a slight tendency to flattening but their nuclei were still rounded or only a little elongated. This area was not clearly delimited from the surrounding epithelium but showed gradual mergence with it.

====Embryos measuring 1.2 cm====
The wall of the brain stem was thicker than in the preceding specimen and consisted of ependyma and a narrow stratum of overlying nerve tissue. The roof of the thalamencephalon consisted of a layer of ependyma only and still showed no indication of folding. The myelencephalic roof was also smooth and continuous throughout. The slightly differentiated median area in the anterior region of this roof was now sharply delimited from the surrounding epithelium and its constituent cells were very flat, the cytoplasm forming a thin investment around the narrow, elongated nucleus. 444 . Henry Cohen and Sarah Davies

The injected dye was wholly retained within the central nervous system, no extra-ventricular spread being evident. The blue granules, intermingled with coagulum, extended in uniform distribution along the central canal of the spinal cord and in the cerebral ventricles. They showed no condensation associated with any particular area in the brain stem.

====Embryos measuring 1.6 cm====
The roof of the thalamencephalon was still unfolded but was thinner than in the previous stage, consisting of an ependymal layer 2-8 cells deep. The individual cells possessed pale granular cytoplasm and round or slightly elongated nuclei. The nucleoli and nuclear membrane were deeply stained in contrast to the pale nucleoplasm. The roof of the fourth ventricle (see Pl. I, fig. 5) was now divided into anterior and posterior regions by a slight invagination in the median transverse plane. This depression was invaded by the overlying mesenchyme and its constituent cells tended to be columnar rather than cuboidal. The membranous area in this roof was larger than in the preceding specimen and occupied the greater part of the roof region anterior to the median invagination. In the posterior portion of the roof a small, median, oval area was observed where the cells were low cuboidal in stature but their nuclei were still rounded or only a little elongated. This slightly differentiated area was not sharply defined but merged imperceptibly into the surrounding epithelium.

The injected solution was still confined within the central nervous system, but no longer exhibited the same uniformity of distribution as previously. A dense accumulation of blue granules, intermingled with coagulum, was seen adhering to the ventricular surface of the membranous area in the anterior portion of the myelencephalic roof. The individual cells of this area showed no impregnation with the dye.

====Embryos measuring 1.8 cm====
At this stage, the roof of the thalamencephalon presented a slightly undulating appearance, forming several irregular, shallow folds invaded by the overlying mesenchyme. The median transverse depression in the myelencephalic roof was deeper than previously and its constituent cells more markedly columnar. Its ventro-lateral walls exhibited a slight degree of folding but nd true branching villi were apparent. The anterior portion of the roof was shorter in longitudinal section than in the preceding stage, owing to encroachment by the developing cerebellum. The median membranous area in this anterior region showed a slight reduction in extent and was now partly incorporated into the anterior wall of the median fold. It was still sharply delimited from the surrounding epithelium and its constituent cells retained their flat appearance. The slightly differentiated posterior area, on the other hand, showed considerable increase in size, exceeding the anterior area both in length and width. It still merged into the general roof layer and its cells were again low cuboidal in stature.

The injected dye was still retained within the central nervous system and was particularly condensed below the membranous area in the anterior portion of the fourth ventricle roof. No such condensation was evident below the slightly differentiated posterior area, but merely a fine distribution of granules adhering to its ventricular surface.

====Embryos measuring 1.9 cm====
The invaginations in the thalamencephalic roof were more pronounced than in the preceding stage and extended anteriorly for a short distance into the lateral ventricles via the foramina of Munro. The median transverse fold in the roof of the fourth ventricle was more emphasized and its ventro-lateral walls exhibited a higher degree of undulation. The membranous area in the anterior region of this roof was smaller in extent than previously but its border was still clearly defined. The slightly flattened posterior area was relatively larger and again merged gradually into the surrounding roof epithelium.

The distribution of dye was similar to that of the preceding stage and the blue granules were still wholly confined within the central nervous system.

====Embryos measuring 2 cm====
These specimens presented the same histological picture as in the previous stage, but the median transverse fold in the medullary roof was much deeper. The invaginations in its ventro-lateral walls were now very pronounced, forming short, unbranched incipient villi. At this stage, the membranous area in the posterior region of this same roof showed slight increase in extent and was now abruptly demarcated from the surrounding epithelium. Its constituent cells were extremely flat, the cytoplasm forming a thin investment around the very elongated nucleus.

The distribution of injected dye resembled that of preceding stages, no extra-ventricular passage being observed.

====Embryos measuring 2.3 cm====
The folds of the thalamencephalic roof were more marked than hitherto and, as they extended forward into the lateral ventricles, showed a slight degree of branching, forming slender choroid villi. Each villus in section consisted of a central axis of vascular mesenchyme surrounded by an epithelium one cell deep. The individual cells were low columnar in shape and possessed pale granular cytoplasm and round or slightly elongated nuclei with a rich chromatin content. The median transverse fold in the medullary roof (see Text-fig. 5) was very pronounced, extending almost to the floor of the ventricle. The villi in its ventro-lateral walls were more numerous than previously and were now slightly branched, forming an embryonic choroid plexus. In section, each villus consisted of an axial core of mesenchyme containing minute capillaries in its intercellular meshes and surrounded by an epithelium one cell deep. The cells were markedly columnar and possessed pale, granular cytoplasm and slightly elongated, clearly defined nuclei. In a large number, the nucleus and its cytoplasmic investment were relegated to the lower or ventricular region of the cell, the upper aventricular portion containing a large spherical vacuole. The developing cerebellum had encroached on almost the whole of the anterior portion of the medullary roof with the result that the membranous area of this region was now wholly incorporated into the anterior wall of the median fold. This area showed even further reduction in size and was no longer sharply delimited from the surrounding epithelium but merged gradually into it. Its constituent cells had lost their flat appearance and were now low cuboidal in stature. The posterior area, on the other hand, was very extensive and occupied practically the entire roof region behind the choroid plexus.

The injected dye was still strictly confined within the central nervous system, no extra-ventricular spread being evident. There was a dense accumulation of blue granules below the posterior area in the medullary roof, but the individual cells showed no invasion by the dye. No such condensation was evident below the vestigeal anterior area.

====Embryos measuring 2.8 cm====
In these specimens, the histological appearance of the brain was essentially similar to that of the previous stage, but the choroid villi of the lateral ventricles showed a slight degree of vacuolation.


Text-fig, 5. Guinea-pig, 2-3 cm.

====Embryos measuring 3.2 cm====
The choroid villi of the lateral ventricles were more numerous than previously and exhibited a higher degree of branching. The median transverse fold in the medullary roof was very pronounced, and the branching villi in its ventro-lateral walls constituted a well-developed choroid formation extending anteriorly below the cerebellum. The epithelial cells of the villi were markedly columnar and the majority of them were now extremely vacuolate. The anterior portion of the fourth ventricle roof was very short in section, consequently the posterior cerebellar lip lay in close approximation to the anterior wall of the median fold. The membranous area in this region of the roof was no longer visible, even as a vestigeal structure. The posterior area had developed to such an extent that it now occupied the whole of the posterior portion of the roof.

As in all preceding stages, the injected dye was wholly retained within the central nervous system, no extra-cerebral extension being evident.

To summarize: the sequence of events in the morphological development of the roof of the fourth ventricle is strikingly similar in rabbit, rat and guinea-pig embryos. In all cases two membranous areas, one anterior and the ofher posterior, are present in this roof at various stages of intra-uterine life. Both areas are sharply demarcated from the surrounding cuboidal epithelium and consist of extremely flat cells with narrow, elongated nuclei. The anterior area is rapidly differentiated at an early period of development. It then undergoes a gradual regression and is finally obliterated, owing to encroachment by the enlarging cerebellum and choroid plexus. The posterior area, on the other hand, appears at a relatively later stage of development. It increases in size and eventually occupies the major portion of the medullary roof.

The first choroid villi to develop are those of the fourth ventricle. They originate in the ventro-lateral walls of a median transverse fold in the medullary roof, and rapidly assume a branching appearance. In earlier stages the epithelial cells of the villi are uniformly granular but as development proceeds they become vacuolate. The nucleus and its surrounding cytoplasm is relegated ‘to the ventricular periphery of the cell, the aventricular portion containing a large spherical vacuole. The plexuses of the lateral ventricles originate as undulations in the thalamencephalic roof. The folds become increasingly marked and, as they extend forward into the lateral ventricles, form slender branching choroid villi. The epithelial cells of the villi are, at first, granular throughout, but later assume a vacuolate appearance.

In all embryos examined the injected solution remained entirely within the central canal of the spinal cord and ventricular system of the brain. An extra-cerebral passage of dye was never observed at any stage of development, either through the membranous areas in the medullary roof or through the epithelium of the choroid villi.

The stages at which the salient morphological changes occur in the different embryonic types are tabulated below. Naturally such stages are not com‘parable with each other. Moreover, the lengths given of rat and guinea-pig embryos are, for reasons already stated, no reliable criterion of age. They are quoted, however, since they serve as a very rough indication of the stage of development.

Disappearance Differentiation Formation of Formation of of anterior of posterior choroid plexus choroid plexus membranous membranous of fourth of lateral Embryo area area ventricle ventricles Rabbit 17 days 15 days 16 days 17 days Rat 2-1 cm. 1-9 cm. 2-1 cm. 2-Lem. Guinea-pig 2-3 cm. 2 cm. 2-3 cm. 2-3 om.

(b) Effect of intra-vascular injections of a hypertonic solution on the distribution of a true solution injected into the central nervous system

The present section is based on observations of rabbit, rat and guinea-pig embryos in which a 2 % saline solution was injected into the left ventricle of the heart, After a period of 4-5 min, had elapsed a 1% double solution of potassium ferrocyanide and iron ammonium citrate was injected into the central nervous system. Since a detailed account of the development of the medullary roof has been given in the previous section, observations relevant to the distribution of the injected double solution alone are recorded below.

===Rabbits===

In embryos of 18, 14 and 15 days, the choroid plexuses of the brain were not fully differentiated. At these stages, the injected fluid was wholly retained within the spaces of the central nervous system. At 16 days, when the choroid plexus of the fourth ventricle had developed, the foreign solution had invaded the dorsal medullary mesenchyme (see Text-fig. 6). Both membranous areas in the myelencephalic roof functioned as the sites of fluid escape. Not only was there a dense accumulation of blue granules below these areas but in each case their constituent cells were impregnated with the dye. Blue precipitate was also seen adhering to the ventricular surface of the epithelial cells in the choroid villi, but in this and in all succeeding stages the individual cells were entirely free of the dye. In embryos of 17 and 18 days, the posterior membranous area alone permitted of any fluid escape. There was a particularly heavy condensation of the injected solution adhering to its ventricular surface. Its cells were deeply impregnated with the dye and the mesenchyme immediately above this area contained in its intercellular spaces masses of coagulum intermingled with blue granules.



Text-fig. 6. Rabbit, 16 days.


===Rats===

In embryos measuring 1-6, 1-7 and 1-9 cm. no true branching choroid villi were apparent in the brain. At these stages the injected fluid was strictly confined within the central canal of the spinal cord and in the cerebral ventricles. In embryos measuring 2-1 cm. the choroid plexus of the fourth ventricle was fully differentiated (see Text-fig. 7). Accumulations of blue granules were seen adhering to the epithelium of the villi, but in these and in all later embryos the cells showed no invasion by the dye. At this stage the foreign solution had extended into the overlying mesenchyme through the posterior membranous area in the medullary roof. A slight condensation of blue granules adhered to the ventricular surface of the vestigeal anterior membranous area, but its cells were entirely free of the dye. In embryos 2-2 and 2-4cm. long the choroid plexuses of the lateral ventricles were well developed and the extra-ventricular passage of the injected solution was more marked than hitherto. Anteriorly the dye had spread over the cerebellar anlage and posteriorly for a short distance along the dorsal surface of the spinal cord.



Text-fig. 7. Rat, 2-1 cm.

===Guinea-pigs===

In embryos measuring 1-2, 1-5, 1-8 and 2-1 cm. no extra-cerebral passage of dye was evident; neither had the choroid plexuses of the brain developed. In embryos 8 cm. in length branching choroid villi were present in both lateral ventricles and in the fourth ventricle. At this stage a marked periaxial spread had occurred through the posterior membranous area in the medullary roof. There was a dense accumulation of blue granules adhering to the ventricular surface of this area. Its cells showed heavy invasion by the dye and the meshes of the overlying mesenchyme contained blue precipitate intermingled with coagulum. A marked condensation of blue granules was seen below the epithelial cells of the plexuses, but in these and in all subsequent specimens the individual cells were free of the dye. In embryos measuring 3-6 cm. the choroid villi of the cerebral ventricles showed a high degree of branching and the extra-ventricular spread of the injected solution was very pronounced (see Text-fig. 8). Anteriorly the dye had extended over the optic lobes and posteriorly along both dorsal and ventral surfaces of the spinal cord.

Thus, in embryos subjected to injections of a hypertonic saline solution: an extra-ventricular passage of dye was observed, but only at stages when the choroid plewuses of the brain were fully differentiated, The injected fluid escaped 450 Henry Cohen and Sarah Davies

into the surrounding mesenchyme through the membranous areas in the roof of the fourth ventricle. Never, at any stage, was a spread observed through the epithelial cells of the choroid villi.

==VI. Appendix==

During the course of the present investigation, five mouse embryos came to hand incidentally. Their individual lengths (crown-rump measurement) were 1, 1-1, 1-3, 1-4 and 1-5 cm. respectively. No injections were made into the central nervous system of these embryos. They were examined histologically, however, to discover whether or not specialized membranous areas were present in the myelencephalic roof at any stage of development. Our observations are briefly recorded below.



Text-fig. 8. Guinea-pig, 3-6 cm.

Embryo measuring 1 cm. The major portion of the roof of the fourth ventricle consisted of a cuboidal epithelium one cell deep. The individual cells possessed pale, granular cytoplasm and round, well-defined nuclei situated in the centre of the cell. This roof was divided into anterior and posterior regions by a shallow depression in the median transverse plane. The depression was invaded by the overlying mesenchyme and its constituent cells tended to be columnar rather than cuboidal. In the anterior region of this roof a large median, oval area was distinguishable where the cells were extremely flat, consisting of a narrow, elongated nucleus with a thin, granular, cytoplasmic investment. This area was abruptly demarcated from the remainder of the roof, its flat cells contrasting sharply with the cuboidal cells of the surrounding epithelium, The Roof of the Fourth Ventricle 451

Embryo measuring 1-1 cm. This specimen presented very much the same appearance as the previous one, but the depression in the roof of the fourth ventricle was more emphasized.

Embryo measuring 1-3 cm. In this specimen (see Text-fig. 9) the invagination in the myelencephalic roof was deeper than previously, forming a definite fold, and its constituent cells were more markedly columnar. The membranous area in the anterior region of the roof was relatively larger, extending from the cerebellar anlage to the median transverse depression. In the posterior region a small, median, oval area was visible where the cells were low cuboidal in stature. Their nuclei, however, retained on the whole their rounded shape. This slightly differentiated area was not sharply delimited from the surrounding epithelium but showed gradual mergence with it.



Text-fig. 9. Mouse, 1-3 cm.

Embryo measuring 1-4cm. At this stage (see Text-fig. 10) the median transverse fold in the myelencephalic roof was very pronounced, extending almost to the floor of the ventricle, and its constituent cells were very clearly columnar. The ventro-lateral walls of this fold were strongly undulating and formed short, unbranched incipient villi supported by overlying mesenchyme. The majority of the epithelial cells in each villus possessed pale, uniformly granular cytoplasm, but in some cells one or two small vacuoles were seen lying close to the nucleus. The anterior portion of the fourth ventricle roof was considerably shorter in longitudinal section than in the preceding stage, consequently the median transverse fold lay immediately behind the cerebellar anlage. The membranous area in the anterior region of the roof was smaller in extent than previously and was now incorporated into the anterior wall of the median fold. This area, however, was still sharply delimited from the surrounding epithelium and its constituent cells retained their flat appearance. Practically the whole of the posterior region of the roof consisted of a now fully differentiated membranous area abruptly demarcated from the general roof epithelium and composed of flat cells with narrow elongated nuclei.

Embryo measuring 1:5cm. This specimen presented the same general features as the previous one, but the villi in the ventro-lateral walls of the median fold showed a slight degree of branching, forming an embryonic choroid plexus. Further, a relatively larger number of the epithelial cells in the villi contained a spherical vacuole, still associated with the nucleus.

The small amount of material at our disposal naturally limited the scope of our investigation. Nevertheless, the following facts emerged clearly: (1) Anterior and posterior membranous areas are present in the medullary roof of mouse embryos during development. (2) A well-developed anterior area is present at a stage when the posterior area is only slightly differentiated. (8) When the posterior area is very extensive the anterior area is small. (4) The choroid plexus of the fourth ventricle originates from a median transverse fold in the medullary roof.



Text-fig. 10. Mouse, 1-4 cm.

Owing to insufficiency of data, a detailed history of the membranous areas is not given. The facts recorded above strongly suggest, however, that the sequence of events resembles that found in rabbit, rat and guinea-pig embryos.

==VII. General Summary and Discussion==

Two membranous areas, one anterior and the other posterior, are present in the medullary roof of rabbit, rat, guinea-pig and mouse embryos at different stages of development. In all cases both areas are sharply delimited from the surrounding cuboidal epithelium and consist of extremely flat cells with narrow, elongated nuclei. From their microscopical appearance these areas are identifiable with similar structures recorded by Weed (1917) in pig and in human embryos, and by Cohen & Davies (1937) in the chick. Keegan (1917), investigating rabbit embryos, described two specialized areas in the roof of the fourth ventricle. The findings of the present investigation confirm his observations. Further, the two areas in the mammalian types investigated by us appear to be homologous with the membranous area recorded by Flexner (1929) in the medullary roof of amblystoma embryos. In these latter, however, the area occupies almost the entire roof of the fourth ventricle without differentiation into anterior and posterior regions.

In {{rabbit}}, {{rat}}, {{guinea-pig}} and {{mouse}} embryos the anterior membranous area is a transitory structure and is obliterated early in development. Its disappearance is brought about by encroachment of the developing cerebellum and choroid plexus on the anterior portion of the medullary roof. The posterior area, on the other hand, is persistent and eventually occupies the major portion of the inferior choroid velum. Thus, the sequence of events in the history of the membranous areas in these types resembles very closely that found in {{pig}}, human and {{chick}} embryos.


Weed (1917) found that in pig embryos the first extra- ventricular spread of a 1% double solution of potassium ferrocyanide and iron ammonium citrate injected into the central nervous system was coincident with the development of the choroid plexuses of the brain. Cohen & Davies (1987), using a 0-8 % concentration of the same salts, established a similar correlation in the chick. In pig and in chick embryos the membranous areas in the roof of the fourth ventricle functioned as the sites of fluid escape. The present research shows that the corresponding areas in rabbit, rat and guinea-pig embryos are impermeable to a 1% solution of potassium ferrocyanide and iron ammonium citrate, even at stages when the choroid plexuses are well developed. Keegan (1917) records a similar observation also in rabbit embryos. Thus, it appears from our own findings and from those of other investigators, that the specialized areas in the medullary roof of various:animals show a specific difference in permeability.

After intra-vascular injections of a hypertonic saline solution, the membranous areas in our material were rendered permeable to the 1% double solution, but only at stages when the choroid villi of the brain had developed. Prior to differentiation of the plexuses an extra-ventricular passage of dye was never observed. Therefore, under such experimental conditions, there appears to be a definite relationship between the developing choroid plexuses and the first periaxial spread of fluid injected into the central nervous system. A similar conclusion reported by Weed (1917) in pig embryos and by Cohen & Davies (1937) in the chick is thus substantiated.

After increasing the salt content of the blood in various adult Mammalia Najiagas (1921) and Weed (1922) found that the epithelial cells of the choroid villi take no part in the intra-ventricular absorption of the cerebrospinal fluid. On the other hand, Foley (1928) and Forbes et al. (1928) state that under the influence of intravenous injections of a hypertonic salt solution the choroid plexuses absorb fluid from the ventricles. Our own findings in different embryonic Mammalia agree with those recorded by Najiagas (1921) and by Weed (1922) in adults. In rabbit, rat and guinea-pig embryos subjected to intra-vascular injections of a hypertonic saline solution, a passage of dye through the choroid plexuses was never observed at any stage of development.


During the intra-uterine life of these types the epithelium of the choroid villi under both normal and experimental conditions functions as a physiological barrier to the passage of fluid from the cerebral ventricles.

==VIII. Conclusions==

# Two membranous areas, one anterior and the other posterior, are present in the medullary roof of rabbit, rat, guinea-pig and mouse embryos at different stages of development. In all cases the anterior area is a transitory structure and disappears early in intra-uterine life. The posterior area, on the other hand, persists and eventually forms the major portion of the inferior choroid velum.
# In all stages of development the two areas in rabbit, rat and guinea-pig embryos are impermeable to a 1 % double solution of potassium ferrocyanide and iron ammonium citrate injected under normal conditions into the central nervous system. Under the influence of a hypertonic saline solution in the blood stream, they become permeable and permit of the extra-cerebral passage of dye from the fourth ventricle, but only when the choroid plexuses of the brain are differentiated. Therefore, under such experimental conditions a correlation is established between the development of the choroid villi and the first periaxial spread of a true solution injected into the central nervous system.
# During embryonic life, the epithelium of the choroid plexuses under both normal and experimental conditions is a functional barrier to the passage of a true solution from the cerebral ventricles.



We wish to thank Mr F. Beckwith for taking the photographs. The text figures are the work of Mr Douglas J. Kidd, medical artist.

==References==

Couen, H. & Davizs, S. (1937). ‘‘The development of the cerebrospinal fluid spaces and choroid plexuses in the chick.”’ J. Anat., Lond., vol. txxu, Pt. 1, p. 23.

FLEXNER, L. B. (1929). ‘“‘The development of the meninges in Amphibia: a study of normal and experimental animals.”” Contr. Embryol. Carneg. Instn, vol. xx, Nos. 109-17, p. 33.

Foury, F. E. B. (1923). “Alterations in the currents and absorption of cerebrospinal fluid following salt administration.’ Arch. Surg., Chicago, vol. vi, p. 587.

Forsss, H. §., Fremont-Smiru, F. & Woxrr, H. G. (1928). “Resorption of cerebrospinal fluid through the choroid plexus.” Arch. Neurol. Psychiat., Lond., vol. xx, p. 73.

Kezeaan, J. J. (1917). “A comparative study of the roof of the fourth ventricle.”” Anat. Rec. vol. x1, No. 6, p. 379.

Nafaaas, J. C. (1921). “Experimental studies on hydrocephalus.” Johns Hopk. Hosp. Bull. vol. xxxu, p. 381.

Weep, L. H. (1917). “The development of the cerebrospinal spaces in pig and in Man.” Contr. Embryol. Carneg. Instn, vol. v, No. 14.

—— (1922). “The absorption of cerebrospinal fluid into the venous system.” Amer. J. Anat. vol. xxxI, p. 191. 

==Plate I==

EXPLANATION OF PLATE I

Fig. 1. Photomicrograph of a longitudinal section of the roof of the fourth ventricle in a 13 days’ rabbit embryo. In this specimen an injection of a 1% double solution of potassium ferrocyanide and iron ammonium citrate was made into the central nervous system. (Acid formaldehyde: Ehrlich’s haematoxylin and eosin.) x 27.

Fig. 2. Photomicrograph of a longitudinal section of the roof of the fourth ventricle in a 15 days’ rabbit embryo injected as in specimen illustrated in fig. 1. (Acid formaldehyde: Ehrlich’s haematoxylin and eosin.) x 27.

Fig. 3. Photomicrograph of a longitudinal section of the right cerebral hemisphere in a 17 days’ rabbit embryo injected as in specimen illustrated in fig. 1. (Acid formaldehyde: Ehrlich’s haematoxylin and eosin.) x72.

Fig. 4. Photomicrograph of a longitudinal section of the roof of the fourth ventricle in a rat embryo measuring 2-1 cm. and injected as in specimen illustrated in fig. 1. (Acid formaldehyde: Weigert’s haematoxylin and eosin.) x45.

Fig. 5. Photomicrograph of a longitudinal section of the roof of the fourth ventricle in a guineapig embryo measuring 1-6 cm. and injected as in specimen illustrated in fig. 1. (Acid formaldehyde: Weigert’s haematoxylin and eosin.) x36.

'''Abbreviations''' 

A.M. Anterior membranous area. 

L.V. Lateral ventricle. 

C.B. Cerebellum. 

M.T.F. Median transverse fold.

C.V. Choroid villi. 

P.M. Posterior membranous area.

{{Footer}}