Difference between revisions of "Paper - The development of the neural folds and cranial ganglia of the rat"

From Embryology
m
 
(2 intermediate revisions by the same user not shown)
Line 62: Line 62:
  
  
I n the hindbrain region of the neural folds Bartelmez ( ’23) recognizes three primary rhombomeres in young human em- bryos. These he designates rhombomeres A, B, and C. Rhombomere A divides first into two secondary rhombomeres- A, and 3- and still later, rhombomere A, divides into rhombomeres 1 and 2. Ultimately, therefore, rhombomere A gives rise to the first three rhombomeres. Rhombomere B (otic rhombomere) is a secondary as well as a primary rhombomere, undergoing no division.<ref>Bartelmez (’25) now believes that rhombomeres 3 and 4 appear at first as a single ‘segment’ (rhombomere B ).</ref> It is the fourth of the series and is recognizable very early as a marked expansion of the neural plate related to the otic placode. Rhombomere C furnishes the last three rhombomeres -5, 6, and 7. For the sake of convenience, the nomenclature employed by Bartelmez will be used in this paper.
+
In the hindbrain region of the neural folds Bartelmez ( ’23) recognizes three primary rhombomeres in young human em- bryos. These he designates rhombomeres A, B, and C. Rhombomere A divides first into two secondary rhombomeres- A, and 3- and still later, rhombomere A, divides into rhombomeres 1 and 2. Ultimately, therefore, rhombomere A gives rise to the first three rhombomeres. Rhombomere B (otic rhombomere) is a secondary as well as a primary rhombomere, undergoing no division.<ref>Bartelmez (’25) now believes that rhombomeres 3 and 4 appear at first as a single ‘segment’ (rhombomere B ).</ref> It is the fourth of the series and is recognizable very early as a marked expansion of the neural plate related to the otic placode. Rhombomere C furnishes the last three rhombomeres -5, 6, and 7. For the sake of convenience, the nomenclature employed by Bartelmez will be used in this paper.
  
  
Line 72: Line 72:
  
  
I now turn to a description of younger stages, in an at- tempt to show the gradual differentiation of the regions above outlined. A brief description of two young embryos, one of 1somite and another of 2 to 3 somites, will serve to introduce the description of a 5-somite embryo which marks the real starting-point of our study. In a 1-somite embryo (fig. 7) the neural folds have just begun to elevate. The folds diverge somewhat caudally. At both extremities they gradually subside to the level of the blastodisc. At the apices of the folds there is, as yet, no delimitation of neural from somatic material, although in the most elevated regions the arrangement of the cells suggests a future line of cleavage.
+
I now turn to a description of younger stages, in an attempt to show the gradual differentiation of the regions above outlined. A brief description of two young embryos, one of 1somite and another of 2 to 3 somites, will serve to introduce the description of a 5-somite embryo which marks the real starting-point of our study. In a 1-somite embryo (fig. 7) the neural folds have just begun to elevate. The folds diverge somewhat caudally. At both extremities they gradually subside to the level of the blastodisc. At the apices of the folds there is, as yet, no delimitation of neural from somatic material, although in the most elevated regions the arrangement of the cells suggests a future line of cleavage.
  
  
In a 2-to-3-somite embryo (fig. 8) marked expansion of the neural plate has occurred. Forward growth and expansion has resulted in the formation of a short head fold into which the foregut extends. The elevation of the anterior end of the embryo is unquestionably in the nature of an overgrowth
+
In a 2-to-3-somite embryo (fig. 8) marked expansion of the neural plate has occurred. Forward growth and expansion has resulted in the formation of a short head fold into which the foregut extends. The elevation of the anterior end of the embryo is unquestionably in the nature of an overgrowth of more ventrally lying parts, initiating a growth process which for a long time continues to characterize the development of the anterior end of the head. In certain sections through this embryo a definite cleft has appeared marking the boundary of neural plate and ectoderm, but as a rule this is not prominent until somewhat later. The two embryos just described should be compared with those figured by Parodi and Widakowich ( ’20).
of more ventrally lying parts, initiating a growth process which for a long time continues to characterize the development of the anterior end of the head. In certain sections through this embryo a definite cleft has appeared marking the boundary of neural plate and ectoderm, but as a rule this is not prominent until somewhat later. The two embryos just described should be compared with those figured by Parodi and Widakowich ( ’20).
 
  
  
Line 82: Line 81:
  
  
The perpendicular anterior face of the neural plate has on each side a very shallow depression (fig. lo), probably the earliest indications of the optic foveae: This is the only one of many 5-somite embryos in which the optic foveae have started their development, and in spite of the fact that transverse sections are not in general favorable for the study of shallow depressions so situated, a careful study of this particular series has convinced me that the optic pits here shown are not artifacts. From the same aspect one can also see a marked deepening of the neural groove anteriorly, but not affecting the extreme cephalic end of the neural plate. It is bounded on each side by eminences which form the summit of the bend of the neural plate. This is the primitive infun- dibulum.
+
The perpendicular anterior face of the neural plate has on each side a very shallow depression (fig. lo), probably the earliest indications of the optic foveae: This is the only one of many 5-somite embryos in which the optic foveae have started their development, and in spite of the fact that transverse sections are not in general favorable for the study of shallow depressions so situated, a careful study of this particular series has convinced me that the optic pits here shown are not artifacts. From the same aspect one can also see a marked deepening of the neural groove anteriorly, but not affecting the extreme cephalic end of the neural plate. It is bounded on each side by eminences which form the summit of the bend of the neural plate. This is the primitive infundibulum.
  
  
Line 120: Line 119:
  
  
In the 9-somite embryo (fig.3) it is possible to establish the cephalic limit of the hindbrain. The embryo from which figure 3 was constructed is cut frontally - a favorable plane for a study of the subdivisions of the neural folds. Figure 20 is a photograph of a section through the hindbrain, showing the relations of rhombomeres 1 to 5. Just caudal to the midbrain there is a broad, shallow depression involving the entire extent of rhombomere A, and comprising secondary sulci related to external rhombomeric swellings. On the right there are two rhombomeres related to it, the smaller caudal one, free from neural crest, is rhombomere 3, while the broader cephalic swelling is rhombomere A,, which later divides into rhombomeres 1and 2. The double nature of this rhombomere is shown on the left, where two swellings are evident anterior to rhombomere 3, and there are indications of two shallow internal sulci corresponding to them. Bartelmez ( ’23) records evidence of subdivision of rhombomere A, in human embryos from the 14-somite stage on. The fifth anlage is attached to rhombomere A, on the right and to hombomeres 1 and 2 on the left. I n this embryo, then, the boundary between hindbrain and midbrain may be definitely determined at the anterior limit of rhombomere A,.
+
In the 9-somite embryo (fig.3) it is possible to establish the cephalic limit of the hindbrain. The embryo from which figure 3 was constructed is cut frontally - a favorable plane for a study of the subdivisions of the neural folds. Figure 20 is a photograph of a section through the hindbrain, showing the relations of rhombomeres 1 to 5. Just caudal to the midbrain there is a broad, shallow depression involving the entire extent of rhombomere A, and comprising secondary sulci related to external rhombomeric swellings. On the right there are two rhombomeres related to it, the smaller caudal one, free from neural crest, is rhombomere 3, while the broader cephalic swelling is rhombomere A,, which later divides into rhombomeres 1and 2. The double nature of this rhombomere is shown on the left, where two swellings are evident anterior to rhombomere 3, and there are indications of two shallow internal sulci corresponding to them. Bartelmez ( ’23) records evidence of subdivision of rhombomere A, in human embryos from the 14-somite stage on. The fifth anlage is attached to rhombomere A, on the right and to hombomeres 1 and 2 on the left. I n this embryo, then, the boundary between hindbrain and midbrain may be definitely determined at the anterior limit of rhombomere A.
  
  
Line 144: Line 143:
  
  
It will be convenient to begin the description of the hind-brain with the preotic rhombomere 3 which is well marked, but which does not extend to the midventral line. Rhombomere 2, the trigeminal rhombomere, is narrow and also dis- appears before the mid-ventral line is reached. Rhombomere 1is difficult to delimit. At 14 somites it is only a slight swelling with a very shallow internal sulcus. While not prominent, it can be demonstrated in favorable frontal and sagittal sections. It will be remembered that it is split off from the anterior portion of rhombomere A,, which early shows its double character. In frontal section it is separated from the midbrain anteriorly and rhombomere 2 posteriorly by shallow grooves. Although it is inconspicuous at 14 somites, it rapidly expands until at 18 somites it is easily recognizable, and I have no doubt that if a fixer with greater shrinking properties than picro-aceto-formol were employed it would be more conspicuous at 14 somites. The trigeminal anlage extends along the side of rhombomeres 1 and 2, but has lost its attachment to rhombomere 1 and is closely applied to the neural tube only in the region of the second hindbrain ‘segment.’ The otic rhombomere to which the VII- VIIT anlage is attached shows a typical midventral swelling and is more extensive ventrally than dorsally. Postotic rhombomere 5 is wedge-shaped, more expanded dorsally than ventrally, and is free of nervous attachments. Rhombomeres G and 7 are apparently present, although a transverse plane is not favorable for the demonstration of such slight swell- ings. However, there are broad, shallow internal sulci cor- responding to the rhombomeres indicated in the figure. Rhombomere 6 (if I am correct in so calling it) is more extensive than 7. In later embryos the last rhombomere expands greatly and becomes most extensive of all. The pro- liferation of the neural crest for the IX-X nerves begins immediately caudal to rhombomere 5 and is continuous with the spinal neural crest. The otic pit has a prominent an- terior lip of thickened ectoderm which overlies the VTI-VIII ganglion. Figure 53, a frontal section of another 14-somite embryo, illustrates many of the above statements.
+
It will be convenient to begin the description of the hindbrain with the preotic rhombomere 3 which is well marked, but which does not extend to the midventral line. Rhombomere 2, the trigeminal rhombomere, is narrow and also dis- appears before the mid-ventral line is reached. Rhombomere 1is difficult to delimit. At 14 somites it is only a slight swelling with a very shallow internal sulcus. While not prominent, it can be demonstrated in favorable frontal and sagittal sections. It will be remembered that it is split off from the anterior portion of rhombomere A,, which early shows its double character. In frontal section it is separated from the midbrain anteriorly and rhombomere 2 posteriorly by shallow grooves. Although it is inconspicuous at 14 somites, it rapidly expands until at 18 somites it is easily recognizable, and I have no doubt that if a fixer with greater shrinking properties than picro-aceto-formol were employed it would be more conspicuous at 14 somites. The trigeminal anlage extends along the side of rhombomeres 1 and 2, but has lost its attachment to rhombomere 1 and is closely applied to the neural tube only in the region of the second hindbrain ‘segment.’ The otic rhombomere to which the VII- VIIT anlage is attached shows a typical midventral swelling and is more extensive ventrally than dorsally. Postotic rhombomere 5 is wedge-shaped, more expanded dorsally than ventrally, and is free of nervous attachments. Rhombomeres G and 7 are apparently present, although a transverse plane is not favorable for the demonstration of such slight swell- ings. However, there are broad, shallow internal sulci cor- responding to the rhombomeres indicated in the figure. Rhombomere 6 (if I am correct in so calling it) is more extensive than 7. In later embryos the last rhombomere expands greatly and becomes most extensive of all. The pro- liferation of the neural crest for the IX-X nerves begins immediately caudal to rhombomere 5 and is continuous with the spinal neural crest. The otic pit has a prominent an- terior lip of thickened ectoderm which overlies the VTI-VIII ganglion. Figure 53, a frontal section of another 14-somite embryo, illustrates many of the above statements.
  
  
Line 165: Line 164:
  
  
One can be less sure of the anterior limit of the hindbrain in tlie 5-somite embryo. Tlie preotic sulcus and otic rhombomere are definite and convenient landmarks. The rostral neural crest (V anlage) is much more extensive than at 8 somites, reaching from a point immediately anterior to the pre-otic sulcus to the rostral flexure of the neural plate.
+
One can be less sure of the anterior limit of the hindbrain in the 5-somite embryo. The preotic sulcus and otic rhombomere are definite and convenient landmarks. The rostral neural crest (V anlage) is much more extensive than at 8 somites, reaching from at point immediately anterior to the pre-otic sulcus to the rostral flexure of the neural plate.
  
  
 +
Just how much of the territory anterior to the preotic sulcus represents rhombomere A1 is a question. There is a slight swelling of the neural plate just anterior to the preotic sulcus, as to the exact significance of which I am somewhat in doubt, because such an appearance might easily be caused by the uneven spreading of a few sections. Since it is such_ a slight swelling, I was unable to determine its validity by a study of sagittal sections of other 5-somite embryos. In determining the boundary line between future hindbrain and midbrain territories of the 5-somite embryo, I was influenced primarily by the fact that the distance between the rostral flexure of the neural plate and the preotic sulcus has not measurably increased between 5 and 8 somites, making it probable that the territory of rhombomere A1 would be of approximately the same size, and this conclusion is lent further weight by the fact that the extent of rhombomere A1 (rhombomeres 1 and 2) in the 14-somite embryo is only slightly greater than that of the 9- or 10-somite embryos. It was found that when a distance equal to the length of rhombomere A1 of the 8-somite embryo was measured off on the neural plate of the 5—somite embryo it coincided in size with the slight expansion noted above anterior to the preotic sulcus.
  
==Text==
 
  
One can be less sure of the anterior limit of the hindbrain
+
It is more difficult to fix the anterior limit of the midbrain. However, a comparison of figures of models of 8-, 9-, and 10somite embryos (figs. 3, 14, 15, 16) will make it clear that the di-mesencephalic boundary lies along the line of the flexure of the neural plate in the 8-somite embryo. The lateral margin of the neural plate is definitely indented on either side at the boundary between the two regions.
in the 5-somite embryo. The preotic sulcus and otic rhombomere are definite and convenient landmarks. The rostral
 
neural crest (V anlage) is much more extensive than at 8
 
somites, reaching from at point immediately anterior to the
 
preotic sulcus to the rostral fiexure of the neural plate.
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
Just how much of the territory anterior to the preotic
 
sulcus represents rhombomere A1 is a question. There is a
 
slight swelling of the neural plate just anterior to the preotic
 
sulcus, as to the exact significance of which I am somewhat
 
in doubt, because such an appearance might easily be caused
 
by the uneven spreading of a few sections. Since it is such_
 
a slight swelling, I was unable to determine its validity by
 
a study of sagittal sections of other 5-somite embryos. In
 
determining the boundary line between future hindbrain and
 
midbrain territories of the 5-somite embryo, I was influenced
 
primarily by the fact that the distance between the rostral
 
flexure of the neural plate and the preotic sulcus has not
 
measurably increased between 5 and 8 somites, making it
 
probable that the territory of rhombomere A1 would be of
 
approximately the same size, and this conclusion is lent
 
further weight by the fact that the extent of rhombomere A1
 
(rhombomeres 1 and 2) in the 14-somite embryo is only
 
slightly greater than that of the 9- or 10-somite embryos. It
 
was found that when a distance equal to the length of rhombomere A1 of the 8-somite embryo was measured off on the
 
neural plate of the 5—somite embryo it coincided in size
 
with the slight expansion noted above anterior to the preotic
 
sulcus.
 
  
It is more difficult to fix the anterior limit of the midbrain.
+
The line marking the di-mesencephalic boundary of the 5-somite embryo begins laterally at the anterior end of the neural-crest proliferation where there is an indentationof the lateral margin of the neural plate and proceeds medially to the mid-line just caudal to the primitive infundibular groove. In the 5-somite embryo there is thus visible in dorsal View anterior to the di-mesencephalic boundary a portion of the forebrain. As a result of the ‘overgrowth’ and expansion of the anterior regions of the neural plate, this material soon comes to form a part of the anterior face of the neural plate and is no longer visible in the dorsal View of the 8-somite embryo. It will be noted that the midbrain has a greater alar than basal extent, which is, of course, to be expected.
However, a comparison of figures of models of 8-, 9-, and 10somite embryos (figs. 3, 14, 15, 16) will make it clear that
 
the di-mesencephalic boundary lies along the line of the
 
flexure of the neural plate in the 8-somite embryo. The lateral margin of the neural plate is definitely indented on either
 
side at the boundary between the two regions.
 
  
The line marking the di-mesencephalic boundary of the
 
5-somite embryo begins laterally at the anterior end of the
 
neural-crest proliferation where there is an indentationof
 
the lateral margin of the neural plate and proceeds medially
 
to the mid-line just caudal to the primitive infundibular
 
groove. In the 5-somite embryo there is thus visible in dorsal
 
View anterior to the di-mesencephalic boundary a portion of
 
38 HOWARD 13. ADELMANN
 
  
the forebrain. As a result of the ‘overgrowth’ and expansion
+
Attempts to carry the delimitation of prospective subdivisions of the neural folds into still earlier stages is deemed neither wise nor profitable, since it must be kept in mind that the brain is formed as the result of the differentiation and enormous expansion of a. relatively small portion of the anterior end of the neural plate, so that attempts to define regions before they are actually differentiated may easily lead to errors similar to those into which His (’74) fell in attempting to outline in mosaic fashion upon the early blastoderm all the parts of the future body.
of the anterior regions of the neural plate, this material soon
 
comes to form a part of the anterior face of the neural plate
 
and is no longer visible in the dorsal View of the 8-somite
 
embryo. It will be noted that the midbrain has a greater
 
alar than basal extent, which is, of course, to be expected.
 
  
Attempts to carry the delimitation of prospective subdivisions of the neural folds into still earlier stages is deemed
 
neither wise nor profitable, since it must be kept in mind that
 
the brain is formed as the result of the differentiation and
 
enormous expansion of a. relatively small portion of the anterior end of the neural plate, so that attempts to define regions before they are actually differentiated may easily lead
 
to errors similar to those into which His (’74) fell in attempting to outline in mosaic fashion upon the early blastoderm
 
all the parts of the future body.
 
  
Unfortunately, the mammal is an especially unfavorable
+
Unfortunately, the mammal is an especially unfavorable form for the determination of the relation of the notochord to the developing brain, since by the time the anterior end of the notochord can be determined, the precocious elevation of the neural folds and the formation of ‘marked bendings of the neural plate and tube disturb the primitive relations of notochord and neural plate. For this reason it was impossible to determine the relation of the floor of the midbrain to the anterior end of the notochord in the rat. However, a study of the growth transformations of the neural plate in the rat seems to indicate that the forebrain and midbrain are differentiated from a very small portion of the anterior end of the neural plate, mainly by marked alar rather than basal expansion, and are therefore primitively much more intimately related than when they are first determinable. As evidence in support of the above statement might be cited the small extent of the medial as compared with the lateral regions in the forebrain and midbrain. Such a conception would aid in understanding the close functional relationships between the optic apparatus and the midbrain, for instance. NEURAL FOLDS AND CBANIAL GANGLIA OF RAT
form for the determination of the relation of the notochord
 
to the developing brain, since by the time the anterior end
 
of the notochord can be determined, the precocious elevation
 
of the neural folds and the formation of ‘marked bendings of
 
the neural plate and tube disturb the primitive relations of
 
notochord and neural plate. For this reason it was impossible to determine the relation of the floor of the midbrain
 
to the anterior end of the notochord in the rat. However, a
 
study of the growth transformations of the neural plate in
 
the rat seems to indicate that the forebrain and midbrain
 
are differentiated from a very small portion of the anterior
 
end of the neural plate, mainly by marked alar rather than
 
basal expansion, and are therefore primitively much more
 
  
intimately related than when they are first determinable. As ‘
 
  
evidence in support of the above statement might be cited
+
This conclusion is in a general way in agreement with Work on other forms (cf. Kingsbury, ’24; Wilson and Hill, ’08; Johnston, ’24). Bartelmez (’23) concludes that the hindbrain is the dominant feature of the brain in early stages of development in human embryos. My studies of the rat embryo agree with his in general.
the small extent of the medial as compared with the lateral
 
regions in the forebrain and midbrain. Such a conception
 
would aid in understanding the close functional relationships
 
between the optic apparatus and the midbrain, for instance.
 
NEURAL FOLDS AND CBANIAL GANGLIA OF RAT
 
  
This conclusion is in a general way in agreement with Work
 
on other forms (cf. Kingsbury, ’24; Wilson and Hill, ’08;
 
Johnston, ’24). Bartelmez (’23) concludes that the hindbrain
 
is the dominant feature of the brain in early stages of development in human embryos. My studies of the rat embryo
 
agree with his in general. .
 
  
Bartelmez states that the midbrain “constitutes the knee
+
Bartelmez states that the midbrain “constitutes the knee of the cranial flexure. The latter can be recognized as the first abrupt bending of the neural axis which we encounter as We pass back from the rostral end.” His figures 3, 4, and 5 seem to indicate, however, that the midbrain really extends for some distance caudal to the ‘knee’ of the rostral flexure’, but perhaps I misinterpret his use of the term. At any rate, the midbrain does extend caudal to the rostral flexure of the neural plate in the rat. Bartelmez’s determination of the midbrain of a 2- and a 4—somite embryo is somewhat puzzling, since the forebrains of these young embryos are not only relatively, but absolutely, larger than they are in the 9—somite embryo figured by him———a condition hardly to be expected. Since the rostral neural crest is not present in the 4—somite human embryo, it is difiicult to make suggestions, but it would seem that in the 4-somite embryo the midbrain might be extended farther anteriorly, which would make it agree more nearly with the 9-somite embryo.“ It seems to me, also, that Bartelmez’s (’23) determination of the midbrain in Veit’s (’18) 8-somite embryo ignores natural boundaries, extending too far cephalad and not extending as far caudally as the anterior boundary of the first hindbrain rhombomere. It appears to the Writer that in Veit’s embryo the di-mesencephalic boundary clearly lies along the line of the rostral flexure just as it does in the rat and the mesencephalon extends caudally from this point as far as rhombomere A1. The summit of thee rostral flexure (knee?) seems to mark the boundary between di- and mesencephalon on the margin of the neural fold. Veit’s embryo then agrees closely with the rat; the mesencephalon lies caudal to the summit of the rostral flexure and has a greater alar than basal extent. It agrees, furthermore, in that the rostral neural crest (Veit’s Kopfganglienleiste) is related to the midbrain and rhombomere A, just as it is in the rat (cf. fig. 1). However, a comparison of Bartelmez’s figures 3, 4, and 5 (with the suggested slight change in fig. 4) reveals in general a happy agreement.
of the cranial flexure. The latter can be recognized as the
 
first abrupt bending of the neural axis which we encounter
 
as We pass back from the rostral end.” His figures 3, 4, and
 
5 seem to indicate, however, that the midbrain really extends
 
for some distance caudal to the ‘knee’ of the rostral flexure’,
 
but perhaps I misinterpret his use of the term. At any rate,
 
the midbrain does extend caudal to the rostral flexure of the
 
neural plate in the rat. Bartelmez’s determination of the
 
midbrain of a 2- and a 4—somite embryo is somewhat puzzling,
 
since the forebrains of these young embryos are not only
 
relatively, but absolutely, larger than they are in the 9—somite
 
embryo figured by him———a condition hardly to be expected.
 
Since the rostral neural crest is not present in the 4—somite
 
human embryo, it is difiicult to make suggestions, but it
 
would seem that in the 4-somite embryo the midbrain might
 
be extended farther anteriorly, which would make it agree
 
more nearly with the 9-somite embryo.“ It seems to me, also,
 
that Bartelmez’s (’23) determination of the midbrain in
 
Veit’s (’18) 8-somite embryo ignores natural boundaries,
 
extending too far cephalad and not extending as far caudally
 
as the anterior boundary of the first hindbrain rhombomere.
 
It appears to the Writer that in Veit’s embryo the di-mesencephalic boundary clearly lies along the line of the rostral
 
flexure just as it does in the rat and the mesencephalon extends caudally from this point as far as rhombomere A1. The
 
  
‘This difficulty has now been eliminated in Bartelmez’s (’25) latest paper.
+
* This difficulty has now been eliminated in Bartelmez’s (’25) latest paper. Before reading the above he had already come to the conclusion that the Eternod embryo, which he knew only from tracings, could not be satisfactorily analyzed, since there is so little to define the midbrain rostrally. He had accordingly omitted it from his final monograph.
Before reading the above he had already come to the conclusion that the Eternod
 
embryo, which he knew only from tracings, could not be satisfactorily analyzed,
 
since there is so little to define the midbrain rostrally. He had accordingly
 
omitted it from his final monograph.
 
-10 HOWARD B. ADELMANN
 
  
summit of tlie rostral flexure (knee?) seems to mark the
 
boundary between di- and mesencephalon on the margin of
 
the neural fold. Veit’s embryo then agrees closely with the
 
rat; the mesencephalon lies caudal to the summit of the
 
rostral flexure and has a greater alar than basal extent. It
 
agrees, furthermore, in that the rostral neural crest (Veit’s
 
Kopfganglienleiste) is related to the midbrain and rhombomere A, just as it is in the rat (cf. fig. 1). However, a comparison of Bartelmez’s figures 3, 4, and 5 (with the suggested
 
slight change in fig. 4) reveals in general a happy agreement.
 
  
The differentiation of the rhombomeres occurs so rapidly
 
in the rat that it is diflicult to be certain of the sequence of
 
their formation. Both the otic rhombomere and the preotic
 
sulcus, marking the position of rhombomere 3, seem to have
 
been formed simultaneously at 3 somites (fig. 22), since they
 
are 11ot yet evident at 2 somites. The localized expansion
 
of the otic rhombomeres probably is responsible for the formation of a sulcus anterior and posterior to it, the anterior
 
(preotic sulcus) being further accentuated by the expansion
 
of material lying ahead of it. Rhombomeres A1 and 3 attain
 
definiteness shortly before rhombomere 5, which is still only
 
imperfectly delimited when rhombomere A, has appeared, at
 
9 somites. At 8 somites the preotic groove marks the site of
 
rhombomere 3, but rhombomere 5 is not recognizable.
 
  
The mesencephalic ‘segments’ are believed to be somewhat
 
inconstant structures in the rat. They are clearly demonstrable at 9 to 10 somites, certainly not marked at 12 to 14
 
somites. They may, perhaps, disappear for a time. Chiarugi
 
( ’22) found that in embryos of the guinea-pig having a closed
 
neural tube he was unable to detect ‘mesomeres’ until the
 
embryos attained a length of 3.5 mm. They can be seen in
 
the guinea-pig until the embryo reaches 7 mm. in length,
 
after which time they disappear. Like Chiarugi, I regard
 
them as being mechanically produced by the marked alar
 
expansion of the neural folds.
 
  
The question of the morphological significance of neuromeres and rhombomeres has been often discussed. Neal’s
+
The differentiation of the rhombomeres occurs so rapidly in the rat that it is diflicult to be certain of the sequence of their formation. Both the otic rhombomere and the preotic sulcus, marking the position of rhombomere 3, seem to have been formed simultaneously at 3 somites (fig. 22), since they are 11ot yet evident at 2 somites. The localized expansion of the otic rhombomeres probably is responsible for the formation of a sulcus anterior and posterior to it, the anterior (preotic sulcus) being further accentuated by the expansion of material lying ahead of it. Rhombomeres A1 and 3 attain definiteness shortly before rhombomere 5, which is still only imperfectly delimited when rhombomere A, has appeared, at 9 somites. At 8 somites the preotic groove marks the site of rhombomere 3, but rhombomere 5 is not recognizable.
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
(’98, ’18) admirable summary of the evidence for and against
+
The mesencephalic ‘segments’ are believed to be somewhat inconstant structures in the rat. They are clearly demonstrable at 9 to 10 somites, certainly not marked at 12 to 14 somites. They may, perhaps, disappear for a time. Chiarugi ( ’22) found that in embryos of the guinea-pig having a closed neural tube he was unable to detect ‘mesomeres’ until the embryos attained a length of 3.5 mm. They can be seen in the guinea-pig until the embryo reaches 7 mm. in length, after which time they disappear. Like Chiarugi, I regard them as being mechanically produced by the marked alar expansion of the neural folds.
tl1e phylogenetic and mechanical theories of their origin obviates the need of an extended discussion here. While Neal
 
recognizes clearly the validity of the arguments advanced
 
to support a mechanical explanation of the origin of neuromeres, he nevertheless believes that the neuromeres are produced in early stages by local thickenings of the lateral walls
 
of the neural tubc——a phenomenon not intelligible as the
 
result of a passive bending or shoving of a tube already
 
formed. Neal regards the rhombomeres as associated in
 
their evolution with the branchial region. They are not, he
 
believes, to be explained as caused by the anlagen of nerves,
 
because rhombomere 3 is Well developed, but lacks nervous
 
connections.
 
Celestino da Costa (’23) has but recently suggested that
 
  
. on peut considérer la neuromerie comme une différenciation provoquée soit par l’émission de la créte
+
The question of the morphological significance of neuromeres and rhombomeres has been often discussed. Neal’s (’98, ’18) admirable summary of the evidence for and against tl1e phylogenetic and mechanical theories of their origin obviates the need of an extended discussion here. While Neal recognizes clearly the validity of the arguments advanced to support a mechanical explanation of the origin of neuromeres, he nevertheless believes that the neuromeres are produced in early stages by local thickenings of the lateral walls of the neural tubc——a phenomenon not intelligible as the result of a passive bending or shoving of a tube already formed. Neal regards the rhombomeres as associated in their evolution with the branchial region. They are not, he believes, to be explained as caused by the anlagen of nerves, because rhombomere 3 is Well developed, but lacks nervous connections. Celestino da Costa (’23) has but recently suggested that
ganglionnaire, soit par le voisinage d’une placode importante
 
et précoce telle que l’auditive.” Rhombomere 3, related to
 
neither neural crest nor placode, again comes to mind as an
 
objection.
 
  
Zimmermann (’91, p. 113) noted the fact that blood vessels
+
. on peut considérer la neuromerie comme une différenciation provoquée soit par l’émission de la créte ganglionnaire, soit par le voisinage d’une placode importante et précoce telle que l’auditive.” Rhombomere 3, related to neither neural crest nor placode, again comes to mind as an objection.
are dispersed inter-rhombomerically in the hindbrain and
 
Graper (’13) found that such inter-rhombomeric vessels are
 
regularly developed in mammals, but are inconstant in the
 
bird. He doubts their segmental significance, “da man an
 
Embryonen in jede Falte des Gehirns Gefasse einsenken
 
sieht.” A similar arrangement of blood vessels has been
 
observed in the rat, but it is questionable if they have any
 
causal significance in the formation of the rhombomeres.
 
  
The Writer believes that in general it is correct to say that
+
Zimmermann (’91, p. 113) noted the fact that blood vessels are dispersed inter-rhombomerically in the hindbrain and Graper (’13) found that such inter-rhombomeric vessels are regularly developed in mammals, but are inconstant in the bird. He doubts their segmental significance, “da man an Embryonen in jede Falte des Gehirns Gefasse einsenken sieht.” A similar arrangement of blood vessels has been observed in the rat, but it is questionable if they have any causal significance in the formation of the rhombomeres.
the rhombomeres are produced by the great growth of the
 
neural tube in a confined space, but believes also that another
 
factor must be taken into consideration. It is quite evident,
 
I believe, that all parts of the neural plate or tube do not
 
expand uniformly and that localized growth of portions of
 
the neural folds are responsible in part for the production of
 
  
£6
+
The Writer believes that in general it is correct to say that the rhombomeres are produced by the great growth of the neural tube in a confined space, but believes also that another factor must be taken into consideration. It is quite evident, I believe, that all parts of the neural plate or tube do not expand uniformly and that localized growth of portions of the neural folds are responsible in part for the production of rhombomeric folds. The hindbrain is predominantly a region of sensory outflow, and associated with this characteristic one might expect a more marked development of the alar regions, accounting for the fact that the rhombomeric grooves do not in general reach the midventral line except in the case of the otic rhombomere where localized expansion is so precocious and marked that the fold affects even the ventral regions. But it must be kept clearly in mind that the recognition of such localized expansion as a part of the developmental pattern of the hindbrain in no way implies that the rhombomeros have any phylogenetic or metameric significance.
42 HOWARD B. ADELMANN
 
  
rhombomeric folds. The hindbrain is predominantly a region
 
of sensory outflow, and associated with this characteristic one
 
might expect a more marked development of the alar regions,
 
accounting for the fact that the rhombomeric grooves do not
 
in general reach the midventral line except in the case of the
 
otic rhombomere where localized expansion is so precocious
 
and marked that the fold affects even the ventral regions. But
 
it must be kept clearly in mind that the recognition of such
 
localized expansion as a part of the developmental pattern
 
of the hindbrain in no way implies that the rhombomeros
 
have any phylogenetic or metameric significance.
 
  
VVilson and Hill (’08) find that in Ornithorynchus welldefined rhombomeres are formed while the neural plate is
+
Wilson and Hill (’08) find that in Ornithorynchus welldefined rhombomeres are formed while the neural plate is perfectly flat, and they believe that this condition excludes a mechanical explanation of the rhombomeres. To the writer’s mind, however, diife es in developmental tempo may account for differences ii a time of appearance of the rhombomeric folds. The anlagen of the cranial ganglia appear very precociously in the monotremes, and it is quite possible that localized expansions may occur early in the open neural-plate stage without having any phylogenetic or segmental significance.
perfectly flat, and they believe that this condition excludes
 
a mechanical explanation of the rhombomeres. To the
 
writer’s mind, however, diife es in developmental tempo
 
may account for differences ii a time of appearance of the
 
rhombomeric folds. The anlagen of the cranial ganglia appear very precociously in the monotremes, and it is quite
 
possible that localized expansions may occur early in the
 
open neural-plate stage without having any phylogenetic or
 
segmental significance.
 
  
THE NEURAL CREST
+
==The Neural Crest==
  
In the rat the cranial neural crest is proliferated from a
+
In the rat the cranial neural crest is proliferated from a strip of ectoderm lying at the margin of the neural plate. In young embryos of 3 somites there is as yet no evidence of neural—crest proliferation and no indication of a differentiation of the neural plate from the ectoderm. At about 4 somites, however, the ectoderm at the summit of the neural folds shows signs of a separation of the neural plate from the ectoderm. The separation is not accomplished, in early embryos at least, by the formation of an obvious cleft between the two regions. The early changes occurring here are diificult to describe, but one gains the impression that a rearrangement of cells takes place, resulting finally in the differentiation of as superficial layer of cells covering the rounded lateral margin of the neural plate and continuous with the ectoderm. This layer of ectodermal cells is illustrated in figure 9. At this time there is no actual cleft separating the superficial layer of ectoderm from the neural ectoderm lying medial to it, but a potential separation is clearly evident from the arrangement of the cells. An actual cleft does exist in some embryos which have undergone marked shrinkage. There is a thickening of the ectoderm at the ventro-lateral margin of the neural plate which is doubtless identical with His’s (’79) ‘Zwischenstrang’ and Neumayer’s (’14) ‘Para-neural-leiste, but which, contrary to the opinion of His, takes no part in neural-crest formation.
strip of ectoderm lying at the margin of the neural plate.
 
In young embryos of 3 somites there is as yet no evidence of
 
neural—crest proliferation and no indication of a differentiation of the neural plate from the ectoderm. At about 4 somites, however, the ectoderm at the summit of the neural
 
folds shows signs of a separation of the neural plate from
 
the ectoderm. The separation is not accomplished, in early
 
embryos at least, by the formation of an obvious cleft between the two regions. The early changes occurring here are
 
diificult to describe, but one gains the impression that a rearrangement of cells takes place, resulting finally in the differentiation of as superficial layer of cells covering the
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
rounded lateral margin of the neural plate and continuous
 
with the ectoderm. This layer of ectodermal cells is illustrated in figure 9. At this time there is no actual cleft separating the superficial layer of ectoderm from the neural
 
ectoderm lying medial to it, but a potential separation is
 
clearly evident from the arrangement of the cells. An actual
 
cleft does exist in some embryos which have undergone
 
marked shrinkage. There is a thickening of the ectoderm at
 
the ventro-lateral margin of the neural plate which is doubtless identical with His’s (’79) ‘Zwischenstrang’ and Neumayer’s (’14) ‘Para-neural-leiste,’ but which, contrary to
 
the opinion of His, takes no part in neural-crest formation.
 
  
Soon, a narrow strip of neural ectoderm, forming the lateral margin of the neural plate and lying just medial to the
+
Soon, a narrow strip of neural ectoderm, forming the lateral margin of the neural plate and lying just medial to the layer of ectoderm above described, begins to thicken and show signs of proliferation. As this material expands, giving rise to a loose mass of neural crest, it causes the elevation of the overlying layer of ectoderm so that a sulcus is formed where thelatter touches the neural plate. This is the condition in the 5-somite embryo (fig. 25). Celestino da Costa’s (’21) figure 1 illustrates identical conditions in the guinea-pig and Veit’s (’18) figures show very similar pictures for man.
layer of ectoderm above described, begins to thicken and show
 
signs of proliferation. As this material expands, giving rise
 
to a loose mass of neural crest, it causes the elevation of the
 
overlying layer of ectoderm so that a sulcus is formed where
 
thelatter touches the neural plate. This is the condition in
 
the 5-somite embryo (fig. 25). Celestino da Costa’s (’21)
 
figure 1 illustrates identical conditions in the guinea-pig and
 
Veit’s (’18) figures show very similar pictures for man.
 
  
In twelve 5-somite embryos studied, the proliferation of
 
  
neural crest extends from a point just behind the rostral
+
In twelve 5-somite embryos studied, the proliferation of neural crest extends from a point just behind the rostral flexure of the neural plate to the anterior border of the preotic sulcus (cf. fig. 1). It is thus being proliferated from the territory of the future midbrain and rhombomere A1. It is evidently identical with the ‘rostral neural crest’ of Bartelmez (’23) and the ‘kraniale Kopfganglienleiste’ of Veit (’18, ’22). The Wave of proliferation passes over this territory extremely rapidly, so that it is difficult to follow its cephalocaudal progress. However, one embryo of 4 to 5 somites shows active proliferation of neural crest for a few sections caudal to the rostral flexure, while sections farther caudally show earlier stages of neuro-ectodermal differentiation, but no actual proliferation of neural crest.
  
flexure of the neural plate to the anterior border of the
 
preotic sulcus (cf. fig. 1). It is thus being proliferated from
 
the territory of the future midbrain and rhombomere A1. It
 
is evidently identical with the ‘rostral neural crest’ of Bartelmez (’23) and the ‘kraniale Kopfganglienleiste’ of Veit (’18,
 
’22). The Wave of proliferation passes over this territory
 
extremely rapidly, so that it is difficult to follow its cephalocaudal progress. However, one embryo of 4 to 5 somites
 
shows active proliferation of neural crest for a few sections
 
caudal to the rostral flexure, while sections farther caudally
 
show earlier stages of neuro-ectodermal differentiation, but
 
no actual proliferation of neural crest.
 
44 HOWARD B. ADELMANN
 
  
Caudal to the anterior margin of preotic sulcus there is no
+
Caudal to the anterior margin of preotic sulcus there is no evidence of neural-crest proliferation in the 5-somite embryo. At 5 somites the rostral neural crest forms a wedge-shaped mass of deeply staining, compacted cells which lies in close contact with the ectoderm (fig. 25). It is fairly easy to distinguish it from the loose mesenchyme, but unless one is careful one is apt to extend it too far ventrally, making it continuous with the more compact mesodermal stalk, lying lateral to the pharynx, which connects the paraxial and lateral mesoderm. Close observation under high power, however, shows that the neural crest is not really continuous with it.
evidence of neural-crest proliferation in the 5-somite embryo.
 
At 5 somites the rostral neural crest forms a wedge-shaped
 
mass of deeply staining, compacted cells which lies in close
 
contact with the ectoderm (fig. 25). It is fairly easy to
 
distinguish it from the loose mesenchyme, but unless one is
 
careful one is apt to extend it too far ventrally, making it
 
continuous with the more compact mesodermal stalk, lying
 
lateral to the pharynx, which connects the paraxial and lateral mesoderm. Close observation under high power, however, shows that the neural crest is not really continuous
 
with it.
 
  
In the 8-somite embryo there are present, in‘ addition to
 
the rostral crest, the anlagen for the VII-VIII and IX-X
 
ganglia. The rostral neural crest (V anlage) (figs. 14 and
 
26), situated just anterior to the preotic sulcus, has lost its
 
attachment to the midbrain and is now related only to rhombomere A1. It forms a loose mass of deeply staining cells
 
situated immediately beneath the ectoderm. It is easily distinguishable from the now greatly condensed mesoderm of
 
the first visceral arch and from the looser paraxial mesoderm
 
medial to it. There is a shallow neuro-ectodermal sulcus
 
related to it on the right side (left in fig. 26) where the
 
overlying ectoderm abuts upon the neural plate.
 
  
The VII-VIII anlage (cf. fig. 14 for its position) is less
+
In the 8-somite embryo there are present, in‘ addition to the rostral crest, the anlagen for the VII-VIII and IX-X ganglia. The rostral neural crest (V anlage) (figs. 14 and 26), situated just anterior to the preotic sulcus, has lost its attachment to the midbrain and is now related only to rhombomere A1. It forms a loose mass of deeply staining cells situated immediately beneath the ectoderm. It is easily distinguishable from the now greatly condensed mesoderm of the first visceral arch and from the looser paraxial mesoderm medial to it. There is a shallow neuro-ectodermal sulcus related to it on the right side (left in fig. 26) where the overlying ectoderm abuts upon the neural plate.
advanced in development. Sections through its anterior portion (cf. right side, fig. 27) show the general features described for the rostral neural crest of younger embryos. The
 
section passes through the anterior limit of the otic placode
 
against which the neural crest is closely applied.
 
  
As one traces the anlage caudally, however, it becomes a
 
wedge-shaped mass, lying just beneath a thin overlying layer
 
of ectoderm above the otic placode (see left side of fig. 27).
 
The caudal part of the anlage is separated from the ectoderm
 
by a distinct cleft, but a mucl1 less well-marked cleft separates it from the neural plate. Vllhile there is a slight neuroectodermal sulcus related to the VII-VIII anlage on the right
 
side of the embryo, none is present on the left.
 
NEURAL rows AND CRANIAL GANGLIA or RAT 45
 
  
The IX-Y anlage is still less advanced in development than
+
The VII-VIII anlage (cf. fig. 14 for its position) is less advanced in development. Sections through its anterior portion (cf. right side, fig. 27) show the general features described for the rostral neural crest of younger embryos. The section passes through the anterior limit of the otic placode against which the neural crest is closely applied.
the VILVIII. It consists of a wedge-shaped mass of cells
 
separated from the neural plate by a very indistinct cleft (fig.
 
28) and covered by a thin layer of ectoderm. There is a very
 
shallow neuro—ectodermal sulcus related to it on each side.
 
  
From the foregoing description it is evident that there
 
is a definite cephalo-caudal sequence inthe differentiation of
 
the anlagen of the cranial ganglia.
 
  
The relation of the IX-X anlage to the spinal neural crest"’
+
As one traces the anlage caudally, however, it becomes a wedge-shaped mass, lying just beneath a thin overlying layer of ectoderm above the otic placode (see left side of fig. 27). The caudal part of the anlage is separated from the ectoderm by a distinct cleft, but a mucl1 less well-marked cleft separates it from the neural plate. While there is a slight neuroectodermal sulcus related to the VII-VIII anlage on the right side of the embryo, none is present on the left.  
is interesting and instructive. As the IX.-X anlage is traced
 
caudally it becomes directly continuous with the spinal neural
 
crest at the anterior edge of the first somite. Curiously
 
enough, at the level of the first somite the spinal neural crest
 
is actively proliferating (fig. 29). Its looser, wandering cells
 
seem to be derived from at small, wedge-shaped mass lying
 
at the summit of the neural folds. Tracing caudally into the
 
regions where the neural folds approximate and finally fuse,
 
one can follow the gradual approach of the neural-crest
 
masses of each side until one finds them fused in the region
 
where the neural folds have fused, at the level of the anterior
 
end of the third somite. Figure 30 shows the mass of neural
 
crest derived from the fusion of the crest of the two sides of
 
the body, lying above the closed neural tube between it and
 
the overlying ectoderm.
 
  
Veit’s (’22) human embryo of 8 somites seems quite comparable with respect to the relation of the IX-X anlage to
 
the spinal neural crest. Veit and Esch (’22, p. 358) did not
 
recognize the existence of the IX-X anlage in their human
 
embryo, but they say: “ . am ikranialen Ende der
 
Ganglienleiste Zellaustritte sich nicht finden, dass da11n eine
 
langere Zone reichlicher Zellaustritte sich anschliesst noch
 
vor dem ersten und in Hohe des 1. und 2. Somiten. . . . . ”
 
The narrow anterior portion of this ganglionic ridge, anterior to the first somite, where no active proliferation is
 
  
5 The term spinal neural crest is used to include both spino—occipit-al and spinal
+
The IX-Y anlage is still less advanced in development than the VILVIII. It consists of a wedge-shaped mass of cells separated from the neural plate by a very indistinct cleft (fig. 28) and covered by a thin layer of ectoderm. There is a very shallow neuro—ectodermal sulcus related to it on each side.
neural crest. This is also the usage of Bartelmez and Evans (’25), justified
 
for many reasons.
 
  
THE JOURNAL OF COMPARATIVE NE-UROLOGY, VOL. 39, N0. 1
 
46 HOWARD B. ADELMANN
 
  
occurring, is evidently identical with what I have recognized
+
From the foregoing description it is evident that there is a definite cephalo-caudal sequence inthe differentiation of the anlagen of the cranial ganglia.
as the IX-X anlage in the rat.
 
  
The spinal neural crest evidently undergoes a cephalocaudal ditferentiationf’ apparently independent of the cranial,
 
since the anterior end of the spinal neural crest shows an
 
advanced stage of proliferation in comparison with the IX-X.
 
anlage immediately anterior to it and with which it is directly continuous. The material available does not allow
 
  
_me to determine at exactly what time the spinal neural crest
+
The relation of the IX-X anlage to the spinal neural crest"’ is interesting and instructive. As the IX.-X anlage is traced caudally it becomes directly continuous with the spinal neural crest at the anterior edge of the first somite. Curiously enough, at the level of the first somite the spinal neural crest is actively proliferating (fig. 29). Its looser, wandering cells seem to be derived from at small, wedge-shaped mass lying at the summit of the neural folds. Tracing caudally into the regions where the neural folds approximate and finally fuse, one can follow the gradual approach of the neural-crest masses of each side until one finds them fused in the region where the neural folds have fused, at the level of the anterior end of the third somite. Figure 30 shows the mass of neural crest derived from the fusion of the crest of the two sides of the body, lying above the closed neural tube between it and the overlying ectoderm.
  
begins to differentiate iii the rat. —I
 
  
Judging by published figures, the process of neural-crest
+
Veit’s (’22) human embryo of 8 somites seems quite comparable with respect to the relation of the IX-X anlage to the spinal neural crest. Veit and Esch (’22, p. 358) did not recognize the existence of the IX-X anlage in their human embryo, but they say: “ . am ikranialen Ende der Ganglienleiste Zellaustritte sich nicht finden, dass da11n eine langere Zone reichlicher Zellaustritte sich anschliesst noch vor dem ersten und in Hohe des 1. und 2. Somiten. . . . . ” The narrow anterior portion of this ganglionic ridge, anterior to the first somite, where no active proliferation is occurring, is evidently identical with what I have recognized as the IX-X anlage in the rat.
formation in the rat seems to be essentially the same as in
 
the cat (Schulte and Tilney, ’15), in the guinea-pig (Chiarugi,
 
’94), and in man (Veit, ’18, ’22), agreeing also in general
 
plan with the elasmobranch (Beard, ’88), the amphibian
 
(Brachet, ’07), the reptile (Neumayer, ’14, p. 444), and the
 
chick (Goronowitsch, ’93, pl. X, fig. 6).
 
  
The description of these early stages has been given in
 
some detail, because they show that the prolonged and heated
 
discussions of His ("(9, ’88), Beard (’87), and others as to
 
whether the neural crest is to be thought of as derived from
 
the neural plate, somatic ectoderm, or from an intermediate
 
zone have really no justification, since the appearances obtained depend entirely upon the precocity of the proliferation
 
of the crest with respect to the dilferentiation and closure of
 
the neural tube. The writer is inclined to regard it as a
 
structure associated primarily with the neural plate. If it
 
differentiates very precociously it may appear as an independent element lying external to the neural plate. VVhen
 
its differentiation is delayed, it may be carried along by the
 
neural folds during their approximation and closure, and so
 
appear to be split off from the neural folds or the roof of
 
the closed neural tube. In the 8-somite rat one sees, in fact,
 
all conditions.
 
  
The neuro—ectodermal sulci which have been described
+
* 5 The term spinal neural crest is used to include both spino—occipit-al and spinal neural crest. This is also the usage of Bartelmez and Evans (’25), justified for many reasons.
above as related to the anlagen of the cranial ganglia and
 
  
“Compare Brachet (’21), pp. 200 and 201.
 
P""'
 
  
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 4(
 
  
which have been noted in tl1e cat by Schulte and Tilney and
+
The spinal neural crest evidently undergoes a cephalocaudal ditferentiationf’ apparently independent of the cranial, since the anterior end of the spinal neural crest shows an advanced stage of proliferation in comparison with the IX-X. anlage immediately anterior to it and with which it is directly continuous. The material available does not allowme to determine at exactly what time the spinal neural crest begins to differentiate in the rat.
in man by Veit, have been thought by Schulte and Tilney ( ’15)
 
to have a phylogenetic and morphological significance. Influenced, no doubt, by the fact that the cranial ganglionic anlagen are hollow in certain cyclostomes (Kupffer, ’95, ’00),
 
they believe that the sulci in question indicate “some small
 
degree of lateral movement of the cells at this point as though
 
to form an evagination. The process is abortive but suggests that in the derivation of the ganglionic crest from the
 
neural tube delamination may have been substituted for evagi~
 
nation and a solid anlage may have replaced a hollow one as
 
elsewhere in the ontogeny of forms rich in cells _ _ . . VVe
 
see in these anlages a series of structures, passing by gradations from the delaminated ganglionic crest through the
 
acoustico—facialis and quintus to the optic vesicle, which is
 
formed by evagination alone.” Veit (’18, p. 312) very justly
 
rejects such an explanation of them. In the rat I have found
 
them to be inconstant, transient structures, not bilaterally
 
symmetrical. They are mechanically produced by the expansion of the neural crest and the consequent elevation of the
 
ectoderm where it adjoins the neural plate.
 
  
The anlagen of the cranial ganglia in the 8—somite rat are
 
quite distinct from one another, a gap separating the V from
 
the VII-VIII and another the VII—VIII from the IX-X an—
 
lage. This differs somewhat from the condition described by
 
Bartelmez (’23) in an 8—somite human embryoH87 in which
 
the rostral neural-crest proliferation extends as far caudally
 
as the acoustico-facial anlage. In a 14-somite human embryo
 
the rostral division of the neural crest has been separated
 
from the VII-VIII by the disappearance of the neural crest
 
from the preotic rhombomere. In the rat no neural crest
 
develops in the region of the preotic sulcus, and, so far as my
 
material allows me to judge, there is from the beginning a
 
gap between the acoustico-facial and IX-X anlagen. Martin
 
(’90) reports that in the eat there is from the beginning a
 
separation between the V and VII anlagen. He found that
 
VIII and IX are continuous at first, but are later separated.
 
48 HOWARD B. ADELMANN
 
  
Tl1e occurrence of a blank rhombomere between the rostral
+
Judging by published figures, the process of neural-crest formation in the rat seems to be essentially the same as in the cat (Schulte and Tilney, ’15), in the guinea-pig (Chiarugi, ’94), and in man (Veit, ’18, ’22), agreeing also in general plan with the elasmobranch (Beard, ’88), the amphibian (Brachet, ’07), the reptile (Neumayer, ’14, p. 444), and the chick (Goronowitsch, ’93, pl. X, fig. 6).
and acoustico-facial neural crest has been recognized in reptiles, birds, mammals, fishes, and Amphibia (for citations cf.
 
Johnston, ’05, p. 192, and Neal, ’98, pp. 212, ff.), but an interruption between the acoustico-facial and IX-X anlagen is
 
present only in reptiles (see, however, Neumayer, ’14, Tafel
 
fig. 12), birds, and mammals, being absent in the Amphibia.
 
Johnston (’05) explains these blank rhombomeres as due to
 
the shifting of the otic vesicle during phylogeny combined
 
with a tendency to concentration of the centers of the V, VII,
 
IX, and X nerves.
 
  
The question of the anterior limit of the neural crest is one
 
of considerable morphological interest and importance. Theoretically, the neural plate might be thought of as possessing
 
the potentialities of neural—crest formation along its entire
 
perimeter. The neural crest, according to this conception,
 
would be proliferated not only throughout the entire forebrain and completely surround the opening of the anterior
 
neuropore; but such a conception is not in accord with the
 
facts.
 
  
In man Bartelmez (’22) describes in an 8—somite embryo
+
The description of these early stages has been given in some detail, because they show that the prolonged and heated discussions of His ("(9, ’88), Beard (’87), and others as to whether the neural crest is to be thought of as derived from the neural plate, somatic ectoderm, or from an intermediate zone have really no justification, since the appearances obtained depend entirely upon the precocity of the proliferation of the crest with respect to the dilferentiation and closure of the neural tube. The writer is inclined to regard it as a structure associated primarily with the neural plate. If it differentiates very precociously it may appear as an independent element lying external to the neural plate. VVhen its differentiation is delayed, it may be carried along by the neural folds during their approximation and closure, and so appear to be split off from the neural folds or the roof of the closed neural tube. In the 8-somite rat one sees, in fact, all conditions.
H87, lateral thickenings of the neural folds extending from
 
the level of Rathke’s pouch (°.?) into the region of the hind
 
brain. He divides this longitudinal thickening into two portions, separated by a constriction in the region of the midbrain. The anterior portion which extends into the forebrain
 
is broader and converges toward its fellow of the opposite
 
side. The posterior portion is narrower and extends caudally
 
into the region of the hindbrain. Since this primordium gives
 
rise to the optic vesicle, neural crest, and head mesenchyme,
 
Bartelmez has termed it the ‘optic crest primordium.’ From
 
its caudal portion there is a typical neural-crest proliferation
 
in the 8—somite embryo. I
 
  
In a 12—somite human embryo H. 197, “The optic crest primordium shows several changes. Its rostral division is wider,
+
The neuro—ectodermal sulci which have been described above as related to the anlagen of the cranial ganglia and which have been noted in tl1e cat by Schulte and Tilney and in man by Veit, have been thought by Schulte and Tilney ( ’15) to have a phylogenetic and morphological significance. Influenced, no doubt, by the fact that the cranial ganglionic anlagen are hollow in certain cyclostomes (Kupffer, ’95, ’00), they believe that the sulci in question indicate “some small degree of lateral movement of the cells at this point as though to form an evagination. The process is abortive but suggests that in the derivation of the ganglionic crest from the neural tube delamination may have been substituted for evagi~ nation and a solid anlage may have replaced a hollow one as elsewhere in the ontogeny of forms rich in cells. We see in these anlages a series of structures, passing by gradations from the delaminated ganglionic crest through the acoustico—facialis and quintus to the optic vesicle, which is formed by evagination alone.” Veit (’18, p. 312) very justly rejects such an explanation of them. In the rat I have found them to be inconstant, transient structures, not bilaterally symmetrical. They are mechanically produced by the expansion of the neural crest and the consequent elevation of the ectoderm where it adjoins the neural plate.
the optic sulcus has deepened so that the optic anlage is delimited on all sides from the rest of the area. Mesial and
 
NEURAL roLDs AND CRANIAL GANGLIA or RAT 4:9
 
  
rostral to it cells are migrating out from the primordium as
 
mesectoderm. . _ . .H197 is the only embryo We have seen
 
which clearly shows mesectoderm formation in the forebrain,
 
nor has it been observed in any mammal.” 1
 
  
Celestino da Costa (’20, ’21, ’23) has recently described
+
“Compare Brachet (’21), pp. 200 and 201. P""'
a proliferation of neural crest in the forebrain region of
 
guinea—pig embryos of 6 and 7 somites. He describes a
 
proliferation of neural crest from the external borders of
 
the optic fossae at the point of neuro-ectodermal reflection.
 
The cell cords so proliferated insinuate themselves between
 
the optic fossa and the overlying ectoderm. The elements
 
are stellate in shape and have a dark protoplasm resembling
 
the elements of the neural plate. These cell cords, according
 
to Celestino da Costa, can be traced caudally, always arising
 
at the point of neuro-ectodermal reflection, but having their
 
origin clearly from the neural plate. As soon as one arrives
 
in the region of the foregut Where the mesenchyme becomes
 
abundant, the cords in question form a typical neural crest
 
such as has been found in other vertebrates. In guinea—pig
 
embryos of 13 to 16 somites the borders of the anterior neuropore seem to be the point of departure of groups of cells
 
identical in appearance with those of the ganglionic crest
 
which extend around the optic vesicles. “Ce n’est plus une
 
créte bien individualisée morphologiquement, mais des groups
 
cellulaires dont les caracteres cytologiques tranchent sur
 
ceux du mésenchyme.
 
  
The fate of this ‘perioptic’ neural crest could not be determined by Celestino da Costa, but he seems to doubt its participation in mesenchyme formation. It will be noted that
 
Chiarugi (’94), who also studied early guinea—pig embryos,
 
did not recognize a perioptic crest, and that Bartelmez’s optic
 
crest arises medial to the optic sulcus, while Celestino da
 
Costa’s arises lateral to the optic fossa. L 1
 
  
A careful study of a large number of rat embryos of 5, 6,
 
7, 8, 9, 10, 11, 12, 13, 14, 16, 17, and 18 somites has failed to
 
reveal the slightest evidence of neural—crest formation in
 
the region of the forebrain. Neither from the neural plate
 
50 HOWARD B. ADELMANN
 
  
lateral or medial to the optic fossa nor from tl1e borders of
+
The anlagen of the cranial ganglia in the 8—somite rat are quite distinct from one another, a gap separating the V from the VII-VIII and another the VII—VIII from the IX-X an— lage. This differs somewhat from the condition described by Bartelmez (’23) in an 8—somite human embryoH87 in which the rostral neural-crest proliferation extends as far caudally as the acoustico-facial anlage. In a 14-somite human embryo the rostral division of the neural crest has been separated from the VII-VIII by the disappearance of the neural crest from the preotic rhombomere. In the rat no neural crest develops in the region of the preotic sulcus, and, so far as my material allows me to judge, there is from the beginning a gap between the acoustico-facial and IX-X anlagen. Martin (’90) reports that in the eat there is from the beginning a separation between the V and VII anlagen. He found that VIII and IX are continuous at first, but are later separated.  
the anterior neuropore was there ever any indication of the
 
proliferation of neural crest. It has been pointed out that
 
the neural crest reaches its greatest anterior extent in embryos of 5 somites when its anterior limit lies at the rostral
 
end of the prospective midbrain, but the extent of its attachment becomes very rapidly reduced, due to factors to be
 
discussed later. Figure 31, which is a transection through
 
the region of the optic vesicles of an 8-somite embryo, shows
 
how ‘clean-cut’ the boundaries of the neural plate are in
 
this region. Sections of 5-, 6-, 7-, and 9-somite embryos give
 
similar pictures in this respect. The mesenchym_e around
 
the expanding optic vesicles becomes compacted as time goes
 
on, and the cells as a natural consequence stain more deeply,
 
due to the loss of their stellate form. They resemble the compact mesenchymal material of the visceral arches far more
 
than the neural-crest cells. There is no evidence of their
 
neural-crest origin. Appearances identical to figure 2 of
 
Celestino da Costa (’22) are found in rat embryos from 12
 
somites on, but the cells surrounding the optic vesicles,
 
though indeed compact, are purely mesenchymal. Celestine
 
da Costa (’21) himself admits that in 13-somite embryos this
 
perioptic tissue is no longer ‘bien individualisée morphologiquement’ and that the cells ‘tranchent sur ceux du mesenchyme.
 
  
The writer believes that the crista neuralis prosencephali
 
of Bartelmez (’25) should be regarded with extreme caution,
 
as he himself clearly appreciates. As the above description
 
shows, no trace of such crest could be found in at large number of young rat embryos. VVith the exception of this single,
 
possibly abnormal embryo (H197), in which Bartelmez describes the above crest, his studies on the extent of the crest
 
agree with the writer’s.
 
  
It seems very probable that the cells surrounding the optic
+
The occurrence of a blank rhombomere between the rostral and acoustico-facial neural crest has been recognized in reptiles, birds, mammals, fishes, and Amphibia (for citations cf. Johnston, ’05, p. 192, and Neal, ’98, pp. 212, ff.), but an interruption between the acoustico-facial and IX-X anlagen is present only in reptiles (see, however, Neumayer, ’14, Tafel fig. 12), birds, and mammals, being absent in the Amphibia. Johnston (’05) explains these blank rhombomeres as due to the shifting of the otic vesicle during phylogeny combined with a tendency to concentration of the centers of the V, VII, IX, and X nerves.
vesicle of a 6-somite guinea-pig embryo which Celestine da
 
Costa (’23, fig. 2) has called perioptic neural crest are merely
 
mesenchymal cells. At least, that is the impression received
 
from the figure.
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 51
 
  
Observations on the rat are supported by Veit’s (’22)
 
study of an 8—somite human embryo, where the rostral neural
 
crest extends along the midbrain, ending a short distance
 
behind its anterior border. As Bartelmez (’23) points out,
 
Veit was obviously in error in assigning this rostral neural
 
crest to the forebrain. Veit’s identification of the eye anlage
 
as related to this crest (cf. his figure 11) is also erroneous, as
 
I believe he would now agree. Schulte and Tilney’s »( ’15)
 
figures show clearly that there is no proliferation of neural
 
crest in the region of the optic vesicles in the cat. In the 4somite cat, the youngest in which a neural crest has appeared,
 
the crest ends some distance caudal to the anterior end of the
 
neural plate.
 
  
Martin’s (’90) statement that the neural crest in the cat
+
The question of the anterior limit of the neural crest is one of considerable morphological interest and importance. Theoretically, the neural plate might be thought of as possessing the potentialities of neural—crest formation along its entire perimeter. The neural crest, according to this conception, would be proliferated not only throughout the entire forebrain and completely surround the opening of the anterior neuropore; but such a conception is not in accord with the facts.
begins “dicht hinter der Ausbuchtung der Augenblasen (beim
 
Embryo von 4 mm. Gesammtlange),” is very indefinite, but
 
excludes a forebrain proliferation, since it must be kept in
 
mind that the early optic vesicles seem to involve practically
 
the entire lateral walls of the forebrain not only in the mammal, but in lower forms as well.
 
  
The absence of a neural-crest proliferation in the forebrain
 
is also generally the rule in lower forms; Neal’s (’98) figures
 
(pl. 3) show no forebrain neural crest in Squalus, but van
 
Wijhe’s (’82) statements are a little hard to understand. On
 
  
page 18 he states, “ . . . . Milncs Marshall zeigte darauf,
+
In man Bartelmez (’22) describes in an 8—somite embryo H87, lateral thickenings of the neural folds extending from the level of Rathke’s pouch (°.?) into the region of the hind brain. He divides this longitudinal thickening into two portions, separated by a constriction in the region of the midbrain. The anterior portion which extends into the forebrain is broader and converges toward its fellow of the opposite side. The posterior portion is narrower and extends caudally into the region of the hindbrain. Since this primordium gives rise to the optic vesicle, neural crest, and head mesenchyme, Bartelmez has termed it the ‘optic crest primordium.’ From its caudal portion there is a typical neural-crest proliferation in the 8—somite embryo. I
dass sie beim Hiinchenembryo bis in den vordersten Theil
 
des Kopfes vorhanden ist. . . . fig. 27 zeigt dass Letzteres
 
  
auch fiir die Salachier gilt.” His figure 27 is a frontal section which does not support his contention, since the upper
+
In a 12—somite human embryo H. 197, “The optic crest primordium shows several changes. Its rostral division is wider, the optic sulcus has deepened so that the optic anlage is delimited on all sides from the rest of the area. Mesial and  rostral to it cells are migrating out from the primordium as mesectoderm. .H197 is the only embryo We have seen which clearly shows mesectoderm formation in the forebrain, nor has it been observed in any mammal.” 1
part of the section lies far caudal to the optic vesicles cut in
 
the lower part of the section.
 
  
Harrison ( ’01) describes the anterior division of the neural
 
crest in Salmo salar as extending from the caudal half of the
 
eye region to shortly in front of the ear pit, but Boeke (’04)
 
found in Muraena that it arose “dicht hinter die Augenblase.” Landacre (’10) finds that in Ameiurus the neural
 
crest begins five or six sections posterior to the optic stalk.
 
52 HOWARD B. ADELMANN
 
  
In Amphibia Brachet (’07) found the V anlage extending as
+
Celestino da Costa (’20, ’21, ’23) has recently described a proliferation of neural crest in the forebrain region of guinea—pig embryos of 6 and 7 somites. He describes a proliferation of neural crest from the external borders of the optic fossae at the point of neuro-ectodermal reflection. The cell cords so proliferated insinuate themselves between the optic fossa and the overlying ectoderm. The elements are stellate in shape and have a dark protoplasm resembling the elements of the neural plate. These cell cords, according to Celestino da Costa, can be traced caudally, always arising at the point of neuro-ectodermal reflection, but having their origin clearly from the neural plate. As soon as one arrives in the region of the foregut Where the mesenchyme becomes abundant, the cords in question form a typical neural crest such as has been found in other vertebrates. In guinea—pig embryos of 13 to 16 somites the borders of the anterior neuropore seem to be the point of departure of groups of cells identical in appearance with those of the ganglionic crest which extend around the optic vesicles. “Ce n’est plus une créte bien individualisée morphologiquement, mais des groups cellulaires dont les caracteres cytologiques tranchent sur ceux du mésenchyme.
far forward as the ‘repli cerebral transverse’ which gives
 
rise, according to him, to the prechordal region of the brain.
 
Stone’s (’22) and Landacre’s (’21) plottings seem to indicate
 
that the proliferation of neural crest does not extend over
 
the forebrain in the amphibian forms they studied. In the
 
chick, however, Marshall (’78), Goronowitsch (’93), and
 
Belogolowy (’10) describe and figure neural-crest prolifera
 
tion in the region of the optic Vesicle. So far as the writer
 
is aware, no one except Celestine da Costa (’21) has ever
 
described a proliferation of neural crest from the lips of the
 
anterior neuropore.
 
  
The absence of neural crest from at least the major portion
 
of the forebrain seems, then, to be a very general condition.
 
It has been commented upon by Schulte and Tilney (’15),
 
who suggest that in the region of the forebrain the neural
 
crest is retained, constituting the dorsal region of the neural
 
plates, forming a zone of material lying ectal to the optic
 
sulcus. According to them, the forebrain must be analyzed
 
not only in terms of alar and basal laminae, but also in terms
 
of this neural-crest element. Such a suggestion is perhaps
 
not without value, but it would perhaps be more correct to
 
say that in the region of the forebrain the neural crest is not
 
formed as such during the process of differential growth and
 
expansion of the forebrain. To say that the ectopic struc~
 
tures are of ganglionic equivalency cannot be sustained by
 
a study of the structure of the cerebrum, and it is extremely
 
doubtful if one can establish the existence of a definite marginal strip of ganglionic material in the forebrain region.
 
There is some evidence, however, that the olfactory sense
 
cells and the bipolar elements of the retina and pineal eye
 
are the equivalent of neural-crest elements. (See discussion
 
in Kingsbury and Adelmann, ’24, pp. 255-268.)
 
NEURAL FOLDS AND CB-ANIAL GANGLIA or RAT 53
 
  
THE DEVELOPMENT OF THE CRANIAL GANGLIA.
+
The fate of this ‘perioptic’ neural crest could not be determined by Celestino da Costa, but he seems to doubt its participation in mesenchyme formation. It will be noted that Chiarugi (’94), who also studied early guinea—pig embryos, did not recognize a perioptic crest, and that Bartelmez’s optic crest arises medial to the optic sulcus, while Celestino da Costa’s arises lateral to the optic fossa.
  
In studying the development of the cranial ganglia,there
 
are three questions of fundamental importance which must
 
be kept in mind. The first is concerned with the part the
 
neural crest plays in their formation. The two extremes of
 
opinion are represented by Goronowitsch (’93), who believes
 
that the neural crest yields only a ‘nervenfiihrendes Gewebe,’
 
and those writers who believe the neural crest to be the sole
 
source of ganglion cells. The second question deals with the
 
relation of the ganglion to the overlying ectoderm and the
 
part which proliferations from localized thickenings of it
 
(placodes) play in the formation of the cranial ganglia in
 
the mammal. The third question is one which has been largely
 
neglected. It has to do with what one might term the gross
 
morphology of the ganglion at different periods of development and its explanation in terms of the growth of the embryo.
 
  
1. The tryigemin/us
+
A careful study of a large number of rat embryos of 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 16, 17, and 18 somites has failed to reveal the slightest evidence of neural—crest formation in the region of the forebrain. Neither from the neural plate lateral or medial to the optic fossa nor from tl1e borders of the anterior neuropore was there ever any indication of the proliferation of neural crest. It has been pointed out that the neural crest reaches its greatest anterior extent in embryos of 5 somites when its anterior limit lies at the rostral end of the prospective midbrain, but the extent of its attachment becomes very rapidly reduced, due to factors to be discussed later. Figure 31, which is a transection through the region of the optic vesicles of an 8-somite embryo, shows how ‘clean-cut’ the boundaries of the neural plate are in this region. Sections of 5-, 6-, 7-, and 9-somite embryos give similar pictures in this respect. The mesenchym_e around the expanding optic vesicles becomes compacted as time goes on, and the cells as a natural consequence stain more deeply, due to the loss of their stellate form. They resemble the compact mesenchymal material of the visceral arches far more than the neural-crest cells. There is no evidence of their neural-crest origin. Appearances identical to figure 2 of Celestino da Costa (’22) are found in rat embryos from 12 somites on, but the cells surrounding the optic vesicles, though indeed compact, are purely mesenchymal. Celestine da Costa (’21) himself admits that in 13-somite embryos this perioptic tissue is no longer ‘bien individualisée morphologiquement’ and that the cells ‘tranchent sur ceux du mesenchyme.’
  
As has been previously described, the neural-crest proliferation which gives rise to the trigeminal ganglion begins its
 
development in the 5—somite embryo as a proliferation from
 
the border of the neural plate throughout the territory of
 
the prospective midbrain and rhombomere A1. The mass of
 
neural-crest cells (fig. 25) lies between the paraxial mesoderm and the ectoderm to which it is closely applied. The
 
paraxial mesoderm is more diffuse than the ganglionic mass
 
and is connected with the pericardial mesoderm by a rather
 
compact column of cells passing lateral to the pharynx. This
 
compact mass of cells connecting the paraxial and lateral
 
mesoderm is the material which upon further condensation
 
becomes the core of the mandibular arch. Upon first sight,
 
it might appear that the ganglion is continuous with the arch
 
mesoderm, but a careful examination of the photograph submitted will show that such is not really the case. An arrow
 
marks the ventral limit of the ganglionic mass at about the
 
level of the dorsum of the pharynx; it is clear that the crest
 
54 HOWARD B. ADELMANN
 
  
has not migrated ventrally into the arch, but ends at the
+
The writer believes that the crista neuralis prosencephali of Bartelmez (’25) should be regarded with extreme caution, as he himself clearly appreciates. As the above description shows, no trace of such crest could be found in at large number of young rat embryos. VVith the exception of this single, possibly abnormal embryo (H197), in which Bartelmez describes the above crest, his studies on the extent of the crest agree with the writer’s.
upper limit of the thickened ectoderm of the arch. Celestine
 
da Costa (’23) describes a similar thickening of the branchial
 
investment of a 6—somite guinea-pig embryo (his fig. 3) which
 
he regards as the epibranchial placode of the mandibular arch.
 
The writer, however, is inclined to regard it simply as branchial ectoderm thickened diffusely, as is usually the case.
 
Tl1e ventral limit of the rostral crest makes contact with,
 
but does not fuse with the dorsal extremity of the thickened
 
branchial ectoderm, but there seems to be no reason for considering it placodal in nature. The inner surface of the
 
branchial ectoderm is perfectly ‘clean’; no proliferation from
 
it could be observed.
 
  
In an 8-somite embryo (figs. 14, '26) the trigeminal anlage
 
has a much less extensive attachment to the neural plate than
 
is the case at 5 somites. It is now confined to the territory
 
of rhombomere A1. And this change has come about with
 
startling rapidity. A 6-somite embryo shows that the rostral
 
neural crest has already lost its attachment to the neural plate
 
anteriorly, extending forward from the preotic sulcus for
 
a distance of only 96 u. A 7—somite specimen is very similar
 
to the 8-somite embryo in the development of the trigeminal
 
anlage. A sagittal section (fig. 24) (somewhat oblique) of
 
an 8-somite embryo is here submitted to Sl10W the form and
 
relations of the ‘V anlage. The blood vessels lying ahead of
 
it should not be mistaken for neural crest. The anterior
 
margin of the anlage seems to be absolutely ‘clean-cut,’ but
 
its ventral limit cannot be made out in this plane. A crosssection of the V anlage at 8 somites (fig. 26) reveals it as
 
a mass of deeply staining cells extending from the margin
 
of the neural plate to the mesenchymal condensation of the
 
mandibular arch, from which it can easily be distinguished.
 
The marked condensation of the arch mesoderm between 5
 
and 8 somites has occurred independent of any neural—crest
 
contribution, I believe, since I was unable to trace any migration of the crest into the arch in 6- and 7-somite embryos.
 
The ventral limit of the ganglion lies at the upper end of the
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 55
 
  
compacted arch mesoderm. The ectoderm is intact, thickened
+
It seems very probable that the cells surrounding the optic vesicle of a 6-somite guinea-pig embryo which Celestine da Costa (’23, fig. 2) has called perioptic neural crest are merely mesenchymal cells. At least, that is the impression received from the figure.  
somewhat over the mesenchymal condensation of the arch, but
 
thin where it comes into contact with the ganglion. Although
 
contact between crest cells and the overlying ectoderm is
 
seen in places, there is clearly no proliferation of ganglionic
 
or mesenchymal elements by the ectoderm.
 
  
The rostral neural crest not only loses its attachment to
 
the midbrain between the 5-and—6-somite stage, but, as far as
 
one is able to judge, it disappears without a trace in that
 
region. Obviously, a change occurring with such rapidity is
 
most difficult to analyze or to follow without a large number
 
of embryos covering the transition period. The possibilities
 
as to the fate of the anterior portion of the rostral neural
 
crest are as follows: 1) The neural crest degenerates, 2)
 
The anterior portion of the rostral neural crest loses its
 
attachment to the midbrain and becomes so diffuse that for
 
a time it is unrecognizable as such, but retaining its identity
 
and specificity, it later recondenses or ‘nebulizes’ to form the
 
ophthalmic branch of the trigeminus. This was the interpretation of Belogolowy (’10), who studied the subject in the
 
chick. 3) The anterior neural crestbecomes diffuse; the cells
 
become indistinguishable from typical mesenchymal elements
 
and, losing their identity and specificity, share the fate of
 
the mesenchymal cells with which they have become indiscriminately intermingled. 4) The anterior portion of the
 
rostral neural crest loses its attachment to the edge of the
 
neural plate as a result of the expansion of the neural plate
 
away from it so that it becomes ‘left behind’ as the strip of
 
neural crest forming the V anlage shown in figure 24.
 
  
The first possibility, namely, that the anterior portion of
+
Observations on the rat are supported by Veit’s (’22) study of an 8—somite human embryo, where the rostral neural crest extends along the midbrain, ending a short distance behind its anterior border. As Bartelmez (’23) points out, Veit was obviously in error in assigning this rostral neural crest to the forebrain. Veit’s identification of the eye anlage as related to this crest (cf. his figure 11) is also erroneous, as I believe he would now agree. Schulte and Tilney’s »( ’15) figures show clearly that there is no proliferation of neural crest in the region of the optic vesicles in the cat. In the 4somite cat, the youngest in which a neural crest has appeared, the crest ends some distance caudal to the anterior end of the neural plate.
the neural crest degenerates, may, I think, be discarded as
 
untrue. The examination of several embryos covering the
 
period in question reveals not the slightest evidence of degeneration of the neural crest in the region of the midbrain.
 
Neither could I find any support for the second possibility.
 
  
Up to and including the 14—somite embryo there is absolutely no trace of an ophthalmic branch of the trigeminus
+
Martin’s (’90) statement that the neural crest in the cat begins “dicht hinter der Ausbuchtung der Augenblasen (beim Embryo von 4 mm. Gesammtlange),” is very indefinite, but excludes a forebrain proliferation, since it must be kept in mind that the early optic vesicles seem to involve practically the entire lateral walls of the forebrain not only in the mammal, but in lower forms as well.
56 HOWARD B. ADELMANN
 
  
ganglion. Belogolowy (’10) observed that in the chick the
+
The absence of a neural-crest proliferation in the forebrain is also generally the rule in lower forms; Neal’s (’98) figures (pl. 3) show no forebrain neural crest in Squalus, but van Wijhe’s (’82) statements are a little hard to understand. On page 18 he states, “ . . . . Milncs Marshall zeigte darauf, dass sie beim Hiinchenembryo bis in den vordersten Theil des Kopfes vorhanden ist. . . . fig. 27 zeigt dass Letzteres auch fiir die Salachier gilt.” His figure 27 is a frontal section which does not support his contention, since the upper part of the section lies far caudal to the optic vesicles cut in the lower part of the section.
ophthalmicus profundus first arose as an intensely staining
 
nebula extending throughout the region of the mesencephalon,
 
but he says (p. 169) that he was unable to follow the process
 
of its formation in detail. However, the fact that the ophthalmicus profundus of the chick arises coincidently with the disappearance of the crest in this region seemed to indicate to
 
him a connection between the two processes. No such gradual
 
‘nebulization’ could be observed in the rat. There is no
 
trace of an ophthalmic ramus of the trigeminus until long
 
after the anterior portion of the rostral neural crest has
 
disappeared. Figure 24 shows clearly that no forward extension of the V anlage can be recognized at 8 somites, and
 
the model of the 14-somite embryo (fig. 17) shows that the
 
ophthalmic branch is not yet formed. The problem of the
 
origin of the ophthalmicus will be returned to later.
 
  
The third possibility must be more seriously entertained.
 
It is quite possible that there is an extremely rapid transformation of these anterior crest cells into mesenchyme, so
 
rapid as to make impossible its detection without a large
 
number of embryos at the age of transition. I was unable
 
to determine from the study of severalembryos of 6 to 9
 
somites in age whether such a transformation of neural crest
 
i11to mesenchyme actually occurs. The possibility must be
 
admitted. Celestino da Costa (’21) also reserves judgment
 
upon this point. .
 
  
In the writer’s opinion, however, the fourth-mentioned possibility expresses the real condition, although the actual
+
Harrison ( ’01) describes the anterior division of the neural crest in Salmo salar as extending from the caudal half of the eye region to shortly in front of the ear pit, but Boeke (’04) found in Muraena that it arose “dicht hinter die Augenblase.” Landacre (’10) finds that in Ameiurus the neural crest begins five or six sections posterior to the optic stalk.  
growth transformations are somewhat diflicult to follow. A
 
comparison of figures 11 and 12 shows that between 5 and 6
 
somites there has been considerable expansion and elevation
 
of the head region, so that it is easily conceivable that during
 
this time the margin of the prospective midbrain area of the
 
neural plate has expanded away from the neural crest, leaving it behind as a band of crest material extending from the
 
anterior margin of the prospective hindbrain ventrally
 
toward the first arch. According to this conception, the anNEURAL FOLDS AND CRANIAL GANGLIA OF RAT 57
 
  
terior edge of the V anlage of the 8-somite embryo (fig. 24)
 
would be thought of as once having been attached to the neural
 
plate. Neal’s (’98) figures 7 to 17 indicate a comparable
 
growth transformation in the shark. In the shark the rostral
 
neural crest retains its attachment to the midbrain until a
 
relatively much later stage than in the rat. Growth and expansion of the midbrain soon result in the loss of this attachment so that neural crest once attached along the midbrain
 
comes to lie along the antero~dorsal margin of the V anlage.
 
The process is complicated in the shark by simultaneous forward growth of the ophthalmicus profundus. In the rat the
 
forward growth effecting the forward extension of the ophthalmic ramus is not coincident with the loss of the attachment of the rostral crest to the neural plate, but occurs somewhat later, as I shall have occasion to show subsequently.
 
  
N eumayer’s (’14) Tafel figure 12 seems to the writer to
+
In Amphibia Brachet (’07) found the V anlage extending as far forward as the ‘repli cerebral transverse’ which gives rise, according to him, to the prechordal region of the brain. Stone’s (’22) and Landacre’s (’21) plottings seem to indicate that the proliferation of neural crest does not extend over the forebrain in the amphibian forms they studied. In the chick, however, Marshall (’78), Goronowitsch (’93), and Belogolowy (’10) describe and figure neural-crest prolifera tion in the region of the optic Vesicle. So far as the writer is aware, no one except Celestine da Costa (’21) has ever described a proliferation of neural crest from the lips of the anterior neuropore.
indicate a similar state of affairs in the reptile.
 
  
The gross form of the V anlage remains much the same
 
from 9 to 14 somites. During this period it remains a roughly
 
oblong mass attached to rhombomere A1. It lies slightly ahead
 
of the hyomandibular cleft and extends ventrally to the mesenchymal condensation of the mandibular arch (see models
 
of 8-, 9-, 10-, 14-somite embryos). Due to the flexure of the
 
neural tube, it willbe appreciated from a study of the models
 
that transverse sections of the mandibular arch passing
 
through the anterior margin of the ganglion would lie immediately behind the eye. Transections passing through the
 
optic vesicle and the trigeminus anlage might, unless a model
 
were made, suggest the existence of an ophthalmic ramus,
 
when really none is present.
 
  
A cross—section of an 11-somite embryo through the trigeminal anlage is given in figure 32. The ganglion is wedgeshaped with the apex directed dorsally. It extends from the
+
The absence of neural crest from at least the major portion of the forebrain seems, then, to be a very general condition. It has been commented upon by Schulte and Tilney (’15), who suggest that in the region of the forebrain the neural crest is retained, constituting the dorsal region of the neural plates, forming a zone of material lying ectal to the optic sulcus. According to them, the forebrain must be analyzed not only in terms of alar and basal laminae, but also in terms of this neural-crest element. Such a suggestion is perhaps not without value, but it would perhaps be more correct to say that in the region of the forebrain the neural crest is not formed as such during the process of differential growth and expansion of the forebrain. To say that the ectopic struc~ tures are of ganglionic equivalency cannot be sustained by a study of the structure of the cerebrum, and it is extremely doubtful if one can establish the existence of a definite marginal strip of ganglionic material in the forebrain region. There is some evidence, however, that the olfactory sense cells and the bipolar elements of the retina and pineal eye are the equivalent of neural-crest elements. (See discussion in Kingsbury and Adelmann, ’24, pp. 255-268.)
margin of the neural plate ventrally where it comes in contact
 
with the condensed mesenchyme of the mandibular arch. In
 
this specimen it is quite possible to determine the boundary
 
between the ganglion and the arch mesoderm. The mesen58 HOWARD B. ADELMANN
 
  
chymal cells are much more compact. Medially, there is a
 
distinct notch at the boundary between the two, due to the
 
fact that the long axis of the ganglion is oblique to the arch,
 
and by tracing the series cephalad and caudad of the ganglion
 
  
one finds that the compact arch mesenchyme extends dorsally ,
+
==The Development of the Cranial Ganglia==
  
to the level indicated by the notch. The integument is, as is
+
In studying the development of the cranial ganglia,there are three questions of fundamental importance which must be kept in mind. The first is concerned with the part the neural crest plays in their formation. The two extremes of opinion are represented by Goronowitsch (’93), who believes that the neural crest yields only a ‘nervenfiihrendes Gewebe,’ and those writers who believe the neural crest to be the sole source of ganglion cells. The second question deals with the relation of the ganglion to the overlying ectoderm and the part which proliferations from localized thickenings of it (placodes) play in the formation of the cranial ganglia in the mammal. The third question is one which has been largely neglected. It has to do with what one might term the gross morphology of the ganglion at different periods of development and its explanation in terms of the growth of the embryo.
usually the case, much thickened over the branchial arch. The
 
epithelium overlying the ganglion is composed of a single
 
layer of cuboidal cells. It becomes much flatter as it passes
 
dorsally. Although the ganglion is in contact with the overlying epithelium, a careful study failed to show any evidence
 
of proliferation on the part of the epithelium. The ganglionic cells are more intimately apposed to the epithelium
 
ventrally than dorsally. The medial boundary of the
 
ganglion is distinct and there seems to be no intermingling
 
of the ganglionic elements with the mesenchyme lying medial
 
to it. The section does not pass through the site of attachment of the V anlage to the neural plate. Conditions are
 
fundamentally the same in 10-somite embryos.
 
  
The trigeminal anlage shows little change between 12 and
+
===1. The Trigeminus===
14 somites. During this time the neural tube closes in the
 
region of rhombomere A1, resulting in a shifting of the contact between ganglion and neural tube. In earlier embryos
 
the ganglion tapers somewhat dorsally and its upper extremity is wedged between the lateral margin of the closing neural
 
tube and the overlying ectoderm. As the neural tube closes
 
the ganglionic anlage maintains contact with its lateral surface, so that at 14 somites, after closure of the neural tube
 
has been effected, the trigeminal anlage has an extensive contact medially with the surface of the hindbrain (fig. 33). Its
 
wedge-shaped dorsal extremity does not quite reach the middorsal line.
 
  
In the 14-somite embryo (fig. 33) the lower limit of the
+
As has been previously described, the neural-crest proliferation which gives rise to the trigeminal ganglion begins its development in the 5—somite embryo as a proliferation from the border of the neural plate throughout the territory of the prospective midbrain and rhombomere A1. The mass of neural-crest cells (fig. 25) lies between the paraxial mesoderm and the ectoderm to which it is closely applied. The paraxial mesoderm is more diffuse than the ganglionic mass and is connected with the pericardial mesoderm by a rather compact column of cells passing lateral to the pharynx. This compact mass of cells connecting the paraxial and lateral mesoderm is the material which upon further condensation becomes the core of the mandibular arch. Upon first sight, it might appear that the ganglion is continuous with the arch mesoderm, but a careful examination of the photograph submitted will show that such is not really the case. An arrow marks the ventral limit of the ganglionic mass at about the level of the dorsum of the pharynx; it is clear that the crest has not migrated ventrally into the arch, but ends at the upper limit of the thickened ectoderm of the arch. Celestine da Costa (’23) describes a similar thickening of the branchial investment of a 6—somite guinea-pig embryo (his fig. 3) which he regards as the epibranchial placode of the mandibular arch. The writer, however, is inclined to regard it simply as branchial ectoderm thickened diffusely, as is usually the case. Tl1e ventral limit of the rostral crest makes contact with, but does not fuse with the dorsal extremity of the thickened branchial ectoderm, but there seems to be no reason for considering it placodal in nature. The inner surface of the branchial ectoderm is perfectly ‘clean’; no proliferation from it could be observed.
ganglionic mass is still marked by a notch medially at the
 
boundary between the compact arch mesoderm and the
 
ganglion. It is separated from the ectoderm by a small
 
amount of looser tissue, presumablymesenchymal in origin.
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 59
 
  
Itewill be observed that the ectoderm related to the ganglion
 
has not thickened appreciably, but that the ectoderm gradually
 
decreases in height as it is followed dorsally. The ganglion
 
makes no contact with it except possibly at its ventral ex
 
tremity and there are no indications of its proliferative
 
  
activity. The position of the primitive head vein medial to
+
In an 8-somite embryo (figs. 14, '26) the trigeminal anlage has a much less extensive attachment to the neural plate than is the case at 5 somites. It is now confined to the territory of rhombomere A1. And this change has come about with startling rapidity. A 6-somite embryo shows that the rostral neural crest has already lost its attachment to the neural plate anteriorly, extending forward from the preotic sulcus for a distance of only 96 u. A 7—somite specimen is very similar to the 8-somite embryo in the development of the trigeminal anlage. A sagittal section (fig. 24) (somewhat oblique) of an 8-somite embryo is here submitted to Sl10W the form and relations of the ‘V anlage. The blood vessels lying ahead of it should not be mistaken for neural crest. The anterior margin of the anlage seems to be absolutely ‘clean-cut,’ but its ventral limit cannot be made out in this plane. A crosssection of the V anlage at 8 somites (fig. 26) reveals it as a mass of deeply staining cells extending from the margin of the neural plate to the mesenchymal condensation of the mandibular arch, from which it can easily be distinguished. The marked condensation of the arch mesoderm between 5 and 8 somites has occurred independent of any neural—crest contribution, I believe, since I was unable to trace any migration of the crest into the arch in 6- and 7-somite embryos. The ventral limit of the ganglion lies at the upper end of the compacted arch mesoderm. The ectoderm is intact, thickened somewhat over the mesenchymal condensation of the arch, but thin where it comes into contact with the ganglion. Although contact between crest cells and the overlying ectoderm is seen in places, there is clearly no proliferation of ganglionic or mesenchymal elements by the ectoderm.
the ganglion should be noted, no venous channel is formed
 
lateral to the trigeminus as in the case of the VII-VIII and
 
IX-X anlagen.
 
  
T Figure 34 is a transection through the body of the trigeminal ganglion of an 18-somite rat embryo. The expansion of
 
the neural tube has resulted in a ventral shifting of the area
 
of contact between ganglion and hindbrain. The area of
 
contact extends dorsoventrally for a little less than the middle
 
third of the lateral surface of the neural tube, its ventral
 
limit lying just above the groove between rhombomeres 2
 
and 3. The -ganglionic mass forms an obtuse angle with the
 
mesoderm of the arch. It is separated from the ectoderm
 
dorsally by a small amount of mesenchyme, but it comes
 
into contact with the ectoderm ventrally. In later stages
 
(fig. 37), as the ganglionic mass expands, the area of contact between ganglion and ectoderm becomes more extensive.
 
In the 18-somite embryo the ectoderm is not specially thickened, being intermediate in height between the thickened
 
integument of the gill arch and the thinner epithelium dorsally. It is composed of a single layer of cuboidal cells. A
 
careful study of the ectoderm in contact with the ganglion
 
revealed no mitoses in this particular embryo and no cells
 
were seen in process of being pinched off from the ectoderm.
 
It is apparently simply a case of intimate contact or adhesion of the compacted ganglionic cells and the adjacent
 
integument.
 
  
The ophthalmic ramus is a direct prolongation of the main
+
The rostral neural crest not only loses its attachment to the midbrain between the 5-and—6-somite stage, but, as far as one is able to judge, it disappears without a trace in that region. Obviously, a change occurring with such rapidity is most difficult to analyze or to follow without a large number of embryos covering the transition period. The possibilities as to the fate of the anterior portion of the rostral neural crest are as follows: 1) The neural crest degenerates, 2) The anterior portion of the rostral neural crest loses its attachment to the midbrain and becomes so diffuse that for a time it is unrecognizable as such, but retaining its identity and specificity, it later recondenses or ‘nebulizes’ to form the ophthalmic branch of the trigeminus. This was the interpretation of Belogolowy (’10), who studied the subject in the chick. 3) The anterior neural crestbecomes diffuse; the cells become indistinguishable from typical mesenchymal elements and, losing their identity and specificity, share the fate of the mesenchymal cells with which they have become indiscriminately intermingled. 4) The anterior portion of the rostral neural crest loses its attachment to the edge of the neural plate as a result of the expansion of the neural plate away from it so that it becomes ‘left behind’ as the strip of neural crest forming the V anlage shown in figure 24.
ganglionic mass. Caudally it lies just above and in contact
 
with the compact mesoderm of the mandibular arch, and as
 
it is traced forward it bears a similar relation to the mesenchymal condensation in the maxillary region (fig. 35). Ex60 HOWARD B. ADELMANN
 
  
ternally, the course of the ophthalmicus is marked by a distinct ridge beneath which it lies. The ophthalmic ramus
 
recedes from the ectoderm as it is traced forward and no contacts between it and the adjacent ectoderm could be noted.
 
Figure 35 shows that the thickened ectoderm over the maxillary region is prolonged dorsally i’,,;,,a very short distance,
 
but the ectoderm is not appreciablgmthickencd along the
 
course of the ophthalmicus. The ram» fiophthalmicus is entirely cellular in nature, more or less_;;‘;,,§tened from side to
 
side and becomes very diffuse as it nears the eye.
 
  
A more extensive area of intimate contact between the
+
The first possibility, namely, that the anterior portion of the neural crest degenerates, may, I think, be discarded as untrue. The examination of several embryos covering the period in question reveals not the slightest evidence of degeneration of the neural crest in the region of the midbrain. Neither could I find any support for the second possibility.
main mass of the trig'ef_(o;_3.¢1al ganglion and the ectoderm above
 
the mandibular arch Wit-5;..?vJOt€d in embryos of 19, 20, 21 (fig.
 
37), and 23 to 24 somites. In all of these stages the characteristics of the overlying epithelium are, first, that it consists
 
of a single layer of cuboidal cells, transitional in height between the integument of the arch and the dorsum; it appears
 
to be a dorsal extension of the branchial investment; secondly, in its antero-posterior extent it coincides with the main
 
ganglionic mass; finally, there seems to be no certain evidence
 
of proliferative activity. F Very few mitotic figures were found
 
in this area in the embryos examined and when the plane of
 
the spindle could be determined it was always, so far as observed, parallel with the surface and therefore not in a position suggesting the proliferation of cells into the subjacent
 
tissue, but providing merely for the expansion of the ectoderm. Furthermore, I was never able to find at cell being
 
pinched off from the ectoderm. I believe it is simply a condition of close contact between an epithelium having no welldefined basement membrane and the ganglionic mass.
 
  
In the interpretation of appearances, the effect of plane of
 
section must always be kept in mind. Figure 36 is an oblique
 
frontal section of an 18-somite embryo which illustrates an
 
especially deceiving condition. Due to the obliquity of the
 
section, the thickness of the integument over the entire surface of the head is exaggerated. Note, for example, the
 
exaggerated thickness of epibranchial placode I. When a
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 61
 
  
closely compacted mass comes into contact witl1 such an
+
Up to and including the 14—somite embryo there is absolutely no trace of an ophthalmic branch of the trigeminus ganglion. Belogolowy (’10) observed that in the chick the ophthalmicus profundus first arose as an intensely staining nebula extending throughout the region of the mesencephalon, but he says (p. 169) that he was unable to follow the process of its formation in detail. However, the fact that the ophthalmicus profundus of the chick arises coincidently with the disappearance of the crest in this region seemed to indicate to him a connection between the two processes. No such gradual ‘nebulization’ could be observed in the rat. There is no trace of an ophthalmic ramus of the trigeminus until long after the anterior portion of the rostral neural crest has disappeared. Figure 24 shows clearly that no forward extension of the V anlage can be recognized at 8 somites, and the model of the 14-somite embryo (fig. 17) shows that the ophthalmic branch is not yet formed. The problem of the origin of the ophthalmicus will be returned to later.
obliquely cut surface, as in the case of the V ganglion illustrated here, it becomes more dilficult than ever to decide
 
whether proliferation is occurring. I could find no direct
 
evidence of it and believe that it is simply a case of contact
 
between the ganglion .1 the epidermis. On the other side
 
of the same embryo 9" l in other 18—somite embryos cut transversely it seems cle: that no proliferation is occurring. The
 
relation of the op ralmicus profundus to the ectoderm is
 
never as intimate as that of the main ganglionic mass; a layer
 
of mesodermal tissue always intervenes (cf. fig. 40). ,
 
  
At 26 to 27 somites (figs. 38 and '7 the ganglionic mass is
 
well circumscribed and no longer in contact with the ectoderm,
 
separated from it by a slight interval. The integument over
 
the ganglion is thi11, showing no thickening. The ophthalmic
 
ramus is more compact and is well separated from the ectoderm. I could find no special thickening of the ectoderm along
 
the course of the ophtlialmic ramus; the epithelium along its
 
course is intermediate in height between that clothing the
 
maxillary region and the flattened epithelium of the dorsum.
 
  
So far as the author is aware, only four investigators definitely commit themselves to the placodal contribution to the
+
The third possibility must be more seriously entertained. It is quite possible that there is an extremely rapid transformation of these anterior crest cells into mesenchyme, so rapid as to make impossible its detection without a large number of embryos at the age of transition. I was unable to determine from the study of severalembryos of 6 to 9 somites in age whether such a transformation of neural crest i11to mesenchyme actually occurs. The possibility must be admitted. Celestino da Costa (’21) also reserves judgment upon this point.
trigeminus in mammals. Chiarugi ( ’94) described the intimate relation of the trigeminus to the ectoderm in the guineapig and noticed the occurrence of ectodermal papillae from
 
which he believed cells were detached to be added to the
 
ganglion. Giglio-Tos (’02), Volker (’22), and Bartelmez
 
(’24) have also described ectodermal contributions to the
 
trigeminus in man and the rat. Celestino da Costa (’23) does
 
not definitely commit himself. He states that the relation of
 
the gasserian ganglion to the ectoderm in the guinea-pig is
 
intimate enough “a permettre une contribution placodiale a
 
la formation du ganglion de Gasser,” but decides that “cette
 
collaboration doit étre peu importante et bien plus réduite
 
. . . ” Froriep (’85), VVeigner (’01), and Neumayer (’06)
 
find no placodes related to the V ganglion in the mammalian
 
forms studied by them. Davis (’23) thus describes the gang—
 
  
THE JOURNAL OF CO)-Il’ARA'l‘I\'E NEUROL0l'.iY, VOL. 39, N0. 1
 
62 nowann B. ADELMANN
 
  
lion of the V nerve in a human embryo of 20 somites, “Laterally the mass is continuous with the condensed mesenchyme
+
In the writer’s opinion, however, the fourth-mentioned possibility expresses the real condition, although the actual growth transformations are somewhat diflicult to follow. A comparison of figures 11 and 12 shows that between 5 and 6 somites there has been considerable expansion and elevation of the head region, so that it is easily conceivable that during this time the margin of the prospective midbrain area of the neural plate has expanded away from the neural crest, leaving it behind as a band of crest material extending from the anterior margin of the prospective hindbrain ventrally toward the first arch. According to this conception, the anterior edge of the V anlage of the 8-somite embryo (fig. 24) would be thought of as once having been attached to the neural plate. Neal’s (’98) figures 7 to 17 indicate a comparable growth transformation in the shark. In the shark the rostral neural crest retains its attachment to the midbrain until a relatively much later stage than in the rat. Growth and expansion of the midbrain soon result in the loss of this attachment so that neural crest once attached along the midbrain comes to lie along the antero~dorsal margin of the V anlage. The process is complicated in the shark by simultaneous forward growth of the ophthalmicus profundus. In the rat the forward growth effecting the forward extension of the ophthalmic ramus is not coincident with the loss of the attachment of the rostral crest to the neural plate, but occurs somewhat later, as I shall have occasion to show subsequently.
of the first branchial arch, the two tissues being c.lewrly different-iatcd ftromeach other. (The italics are mine.) The
 
ectoderm covering this arch exhibits a distinct general thickening, but anything approaching a true placodal structure
 
is wanting.
 
  
It has been mentioned that the growth of the ophthalmicus
 
occurs in the rat between 14 and 18 somites. The question
 
of its origin may now be considered. In many forms the
 
ophthalmicus profundus has been described as arising in part
 
or entirely by placodal proliferation. Miss Platt (’96) believed both the gasserian ganglion and the ramus ophthalmicus profundus of Necturus to be derived from cells migrating from the primitive supraorbital ridge. Landacre
 
(’12, p. 10) found a ‘pronounced contact’ of the ophthalmicus
 
profundus with the epidermis in Lepidosteus, but did not
 
determine its placodal nature. Stone (’22) believes that in
 
Amblystoma punctatum the ophthalmicus profundus is composed entirely from cells proliferated from a supraorbital
 
placode. Wlien the supraorbital placode is experimentally
 
removed, the ophthalmicus fails to develop. In Chrysemys
 
Brachet (’14) found that the maxillo-mandibular ganglion is
 
entirely of neural—crest origin, but that the ophthalmic placode is transformed entirely and directly into the ophthalmic
 
ganghonf _
 
  
Belogolowy (’10) concludes (p. 169) that the formation
+
Neumayer’s (’14) Tafel figure 12 seems to the writer to indicate a similar state of affairs in the reptile.
of the ophthalmicus profundus of the chick must be ascribed
 
exclusively to the condensation of ganglionic-crest elements.
 
He could recognize no supplementary factors of any importance. However, he did find on each side, in relation to the
 
ophthalmicus, one or two ectodermal papillae which were
 
very different from the epibranchial placodes, but which resembled in their structure ectodermal papillae which he found
 
‘iiberall gleichm‘.ei.ssig verstreuten.’ The papillae associated
 
with the ophthalmicus were two to three times as large as
 
  
7 In this connection, I desire to call attention to a recent paper by de Beer ( ’2-1). _
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
those found elsewhere, and he found that they constantly
+
The gross form of the V anlage remains much the same from 9 to 14 somites. During this period it remains a roughly oblong mass attached to rhombomere A1. It lies slightly ahead of the hyomandibular cleft and extends ventrally to the mesenchymal condensation of the mandibular arch (see models of 8-, 9-, 10-, 14-somite embryos). Due to the flexure of the neural tube, it willbe appreciated from a study of the models that transverse sections of the mandibular arch passing through the anterior margin of the ganglion would lie immediately behind the eye. Transections passing through the optic vesicle and the trigeminus anlage might, unless a model were made, suggest the existence of an ophthalmic ramus, when really none is present.
invaded the layer of mesenchyme lying between the ectoderm
 
and the ganglion, and at their summit proliferated cells.
 
These papillae are seen for only a short time and soon disappear without trace. He concludes that, since they yield
 
so small an amount of material, their importance in ganglion
 
formation must be minimal. Chiarugi (’94) and Griglio-Tos
 
(’02) describe an ectodermal contribution to the ophthalmicus in the guinea-pig and man. Chiarugi noted the presence of ectodermal papillae from which he believed cells were
 
detached to be added to the ganglion. Celestine da Costa
 
(’23, p. 518) found in the guinea-pig that “Avant ou aprés
 
la courte période entre 16 et 18 myotomes l’épibla’ste qui recouvre la créte du trijumeau et son long prolongement antérieur qui va jusqu’auX vésicules optiques réagit a ce
 
voisinage, en 's’épaississant d’une facon diffuse. La possibilité n’est pas exclue d’une contribution de ce genre, telle
 
que Chiarugi l’avait supposée.” It is my impression, however, that Celestino da Costa has misinterpreted the sections
 
presented as evidence (see his figs. 7, 10). They do not show
 
the ophthalmic ramus of the trigeminus, but pass through
 
the caudal wall of the optic vesicle and the main trigeminal
 
mass (cf. my 14-somite model). The right side of his figure
 
10 shows the thickening of the branchial integument with subjacent mesenchyme, not a dorsolateral placode and the ophthalmic ramus.
 
  
In my study of the rat several places were noted where the
+
A cross—section of an 11-somite embryo through the trigeminal anlage is given in figure 32. The ganglion is wedgeshaped with the apex directed dorsally. It extends from the margin of the neural plate ventrally where it comes in contact with the condensed mesenchyme of the mandibular arch. In this specimen it is quite possible to determine the boundary between the ganglion and the arch mesoderm. The mesenchymal cells are much more compact. Medially, there is a distinct notch at the boundary between the two, due to the fact that the long axis of the ganglion is oblique to the arch, and by tracing the series cephalad and caudad of the ganglion one finds that the compact arch mesenchyme extends dorsally , to the level indicated by the notch. The integument is, as is usually the case, much thickened over the branchial arch. The epithelium overlying the ganglion is composed of a single layer of cuboidal cells. It becomes much flatter as it passes dorsally. Although the ganglion is in contact with the overlying epithelium, a careful study failed to show any evidence of proliferation on the part of the epithelium. The ganglionic cells are more intimately apposed to the epithelium ventrally than dorsally. The medial boundary of the ganglion is distinct and there seems to be no intermingling of the ganglionic elements with the mesenchyme lying medial to it. The section does not pass through the site of attachment of the V anlage to the neural plate. Conditions are fundamentally the same in 10-somite embryos.
ectoderm showed slight papillary thickenings to the inner
 
surfaces of which cells adhered in a manner to suggest a
 
possible proliferation. Then again, where the ganglion approaches the ectoderm closely, ganglion cells were found
 
adherent to the overlying ectoderm which was not especially
 
thickened at the place of contact. In places where the ganglionic mass was separated from the ectoderm some of the
 
intervening mesenchymal cells might appear to adhere to
 
the overlying ectoderm. These appearances, however, are
 
not confined to the ectoderm overlying the trigeminal gang64 HOWARD B. ADELMANN
 
  
lion or to the i11tegument of the head, but may be observed
+
The trigeminal anlage shows little change between 12 and 14 somites. During this time the neural tube closes in the region of rhombomere A1, resulting in a shifting of the contact between ganglion and neural tube. In earlier embryos the ganglion tapers somewhat dorsally and its upper extremity is wedged between the lateral margin of the closing neural tube and the overlying ectoderm. As the neural tube closes the ganglionic anlage maintains contact with its lateral surface, so that at 14 somites, after closure of the neural tube has been effected, the trigeminal anlage has an extensive contact medially with the surface of the hindbrain (fig. 33). Its wedge-shaped dorsal extremity does not quite reach the middorsal line.
all over the body, as botl1 Belogolowy (’10) and Veit (’22)
 
have already observed. None of Veit’s illustrations show a
 
cell actually being pinched off from the ectoderm, but merely
 
fully formed cells adhering to it. I interpret these appearances as cases of adhesion. VV. H. Lewis (’22) has commented upon the marked adhesive properties of mesenchymal
 
cells, and of cells in tissue cultures, such cells frequently
 
becoming so adherent to glass that they may even be centrifugalized without dislodgment. He also points out the
 
important role which this adhesive property plays in early
 
stages of development. Cells in culture were found to migrate
 
out from the explant along certain solid objects (coverglasses, fibrin filaments, etc.) for which they seem to display
 
a special adhesive tendency. I think that this adhesive property of embryonic cells is responsible for the fact that the
 
V and VII (and the early IX-X) ganglionic proliferations
 
grow ventrally in such close approximation to the overlying
 
ectoderrn which possibly exercises a directive influence, and
 
is responsible also for the fact that the vagus (cf. figs. 83,
 
88) hugs so closely the outer wall of the vena capitis lateralis
 
as it grows ventrally. It explains why it is so difficult to
 
draw a boundary line between branchial ectoderm and brancl1ial mesoderm and accounts for numerous descriptions of
 
mesectoderm formation by the branchial integument. Brachet (’21, p. 385) has noted this peculiar and intimate relationship between the ganglionic anlagen and the deep face of
 
the ectoderm; “il semble que ce dernier (the epiblast) exerce,
 
sur ses couche superficielles, un tactisme particulier,” he
 
says. If the proliferation of mesenchymal or ganglionic elements is so general a phenomenon, one would expect to find
 
abundant examples of cells being actually pinched off from
 
the ectoderm into the subjacent tissue. I
 
  
The only constant ectodermal thickening related to the
+
In the 14-somite embryo (fig. 33) the lower limit of the ganglionic mass is still marked by a notch medially at the boundary between the compact arch mesoderm and the ganglion. It is separated from the ectoderm by a small amount of looser tissue, presumablymesenchymal in origin.  
eye is illustrated in figure 41. It lies just caudal to and above
 
the eye and the underlying mesenchymal condensation comes
 
intoclose contact with it. Its under surface is intact. This
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 65
 
  
is, however, not a placode. It is simply a thickening of the
 
ectoderm above the cleft separating the mandibular arch
 
from the anterior part of the head and it is continuous with
 
the thickened integument of the mandibular arch and a less
 
marked diffuse thickening of the ectoderm extending forward
 
over the eye. It can be observed for a long period, but never
 
comes into contact with the ophthalmic ramus of V. This
 
thickening was, I believe, noticed by Froriep (’85, p. 44) in
 
older embryos of the calf, but he says “ . . . . dieselbe
 
steht aber in keiner Beziehung zum Trigeminus. Sie ist vielleicht ein Rudiment der Seitenorgananlage, welche bei
 
Selachierembryonen an den vorderen Dorsalasten des Facialis
 
sich bildet.” A placodal origin for the ramus ophthalmicus
 
profundus trigemini in the mammal is, I believe, untenable.
 
  
The possibility of a placodal origin of the ophthalmicus
+
It will be observed that the ectoderm related to the ganglion has not thickened appreciably, but that the ectoderm gradually decreases in height as it is followed dorsally. The ganglion makes no contact with it except possibly at its ventral ex tremity and there are no indications of its proliferative activity. The position of the primitive head vein medial to the ganglion should be noted, no venous channel is formed lateral to the trigeminus as in the case of the VII-VIII and IX-X anlagen.
being excluded, how, then, does the ramus ophthalmicus of
 
the V ganglion arise in the rat between the ages of 14 and
 
18 somites? The answer lies in a study of the growth transformations of the region during that period. Up to the 14somite stage, growth has proceeded but slowly, since the
 
embryo requires some time to becomeproperly established
 
in the uterus. At about this time, however, environmental
 
conditions more favorable to rapid growth have apparently
 
been established and the embryo consequently ‘speeds up’
 
in growth, the increase in bulk between 14 and 18 somites
 
being almost incredible. Coincidently with this enormous
 
growth, significant shiftings of parts have occurred which
 
account, I believe, for the growth of the ophthalmic ramus
 
of the V ganglion.
 
  
Externally, one notes the following changes. In embryos
 
of 14 somites and younger, the mandibular arch lies imme
 
diately caudal to that part of the head composed mainly of
 
  
forebrain. Since in early embryos the optic vesicle is relatively larger than subsequently, constituting almost the entire lateral wall of the forebrain, the core of the mandibular
+
Figure 34 is a transection through the body of the trigeminal ganglion of an 18-somite rat embryo. The expansion of the neural tube has resulted in a ventral shifting of the area of contact between ganglion and hindbrain. The area of contact extends dorsoventrally for a little less than the middle third of the lateral surface of the neural tube, its ventral limit lying just above the groove between rhombomeres 2 and 3. The -ganglionic mass forms an obtuse angle with the mesoderm of the arch. It is separated from the ectoderm dorsally by a small amount of mesenchyme, but it comes into contact with the ectoderm ventrally. In later stages (fig. 37), as the ganglionic mass expands, the area of contact between ganglion and ectoderm becomes more extensive. In the 18-somite embryo the ectoderm is not specially thickened, being intermediate in height between the thickened integument of the gill arch and the thinner epithelium dorsally. It is composed of a single layer of cuboidal cells. A careful study of the ectoderm in contact with the ganglion revealed no mitoses in this particular embryo and no cells were seen in process of being pinched off from the ectoderm. It is apparently simply a case of intimate contact or adhesion of the compacted ganglionic cells and the adjacent integument.
arch lies just caudal to the eye. Consequently, no maxillary
 
region can be identified. The above facts are most strikingly
 
66 HOWARD B. ADELMANN
 
  
G. V G. VII-VIII 01. pl. 0. 1x-x S. 1
 
  
+
The ophthalmic ramus is a direct prolongation of the main ganglionic mass. Caudally it lies just above and in contact with the compact mesoderm of the mandibular arch, and as it is traced forward it bears a similar relation to the mesenchymal condensation in the maxillary region (fig. 35). Externally, the course of the ophthalmicus is marked by a distinct ridge beneath which it lies. The ophthalmic ramus recedes from the ectoderm as it is traced forward and no contacts between it and the adjacent ectoderm could be noted. Figure 35 shows that the thickened ectoderm over the maxillary region is prolonged dorsally i’,,;,,a very short distance, but the ectoderm is not appreciablgmthickencd along the course of the ophthalmicus. The ram» fiophthalmicus is entirely cellular in nature, more or less_;;‘;,,§tened from side to side and becomes very diffuse as it nears the eye.
  
R. op. v G. V. G. vn-vm
 
  
5
+
A more extensive area of intimate contact between the main mass of the trig'ef_(o;_3.¢1al ganglion and the ectoderm above the mandibular arch Wit-5;..?vJOt€d in embryos of 19, 20, 21 (fig. 37), and 23 to 24 somites. In all of these stages the characteristics of the overlying epithelium are, first, that it consists of a single layer of cuboidal cells, transitional in height between the integument of the arch and the dorsum; it appears to be a dorsal extension of the branchial investment; secondly, in its antero-posterior extent it coincides with the main ganglionic mass; finally, there seems to be no certain evidence of proliferative activity. F Very few mitotic figures were found in this area in the embryos examined and when the plane of the spindle could be determined it was always, so far as observed, parallel with the surface and therefore not in a position suggesting the proliferation of cells into the subjacent tissue, but providing merely for the expansion of the ectoderm. Furthermore, I was never able to find at cell being pinched off from the ectoderm. I believe it is simply a condition of close contact between an epithelium having no welldefined basement membrane and the ganglionic mass.
  
Figs. 4: t.o 6 A series of reconstructions from models of 14-, 18-, and 26-somite
 
rats, respectively (X 50), to illustrate the formation of the ophthalmic ramus
 
of the trigeminus, coincident with the growth shiftings involved in the establishment of the maxillary region. The figures also illustrate the relation of the
 
ganglia to the branchial arches. See text, pages 65-69, for discussion. In
 
figure 4 the extent of the thickened ectoderm surrounding the otic pit is indicated
 
by a broken line. In figure 5 the otic vesicle has not yet been completely closed
 
off from the ectoderm (fig. 18).
 
  
A.br., branchial arch; D.c.c., dorsal cornmunicating cord of ganglionic material
+
In the interpretation of appearances, the effect of plane of section must always be kept in mind. Figure 36 is an oblique frontal section of an 18-somite embryo which illustrates an especially deceiving condition. Due to the obliquity of the section, the thickness of the integument over the entire surface of the head is exaggerated. Note, for example, the exaggerated thickness of epibranchial placode I. When a closely compacted mass comes into contact witl1 such an obliquely cut surface, as in the case of the V ganglion illustrated here, it becomes more dilficult than ever to decide whether proliferation is occurring. I could find no direct evidence of it and believe that it is simply a case of contact between the ganglion .1 the epidermis. On the other side of the same embryo 9" l in other 18—somite embryos cut transversely it seems cle: that no proliferation is occurring. The relation of the op ralmicus profundus to the ectoderm is never as intimate as that of the main ganglionic mass; a layer of mesodermal tissue always intervenes (cf. fig. 40).
between the IX and X ganglia; G., ganglion; 013.19., optic vesicle; 0t.pl., otic
 
placode; Ot.'u., otic vesicle; R.op.V., ophthalmic ramus of the trigeminus; SJ,
 
the first somite; V.c.c., ventral communicating cord of ganglionic material between
 
the IX and X ganglia.
 
NEURAL LFOLDS AND ORAN-IAL GANGLIA or RAT 67
 
  
illustrated in embryos of 8 to 14 somites (figs. 14, 4, 17). The
 
external aspect of the 10—somite embryo shows that between
 
8 and 10 somites some forward growth of dorsally lying
 
parts has occurred, and this is reflected in the oblique direction of the long axis of the mandibular arch, the dorsal ex
 
G. V. G. VII-VIII 0t. V.
 
  
tremity of which lies considerably cranial to its Ventral end,
+
At 26 to 27 somites (figs. 38 and '7 the ganglionic mass is well circumscribed and no longer in contact with the ectoderm, separated from it by a slight interval. The integument over the ganglion is thi11, showing no thickening. The ophthalmic ramus is more compact and is well separated from the ectoderm. I could find no special thickening of the ectoderm along the course of the ophtlialmic ramus; the epithelium along its course is intermediate in height between that clothing the maxillary region and the flattened epithelium of the dorsum.
  
i which is more or less fixed. A forward extension ofthe mesenchymal condensation of the first arch actually rests against
 
the postero-superior border of the optic vesicle. The position
 
of the trigeminus with respect to the optic Vesicle has already
 
been commented upon.
 
68 HOWARD _B. ADELMANN
 
  
At 12 and 14 somites (figs. 4 and 17) the eye is still in
+
So far as the author is aware, only four investigators definitely commit themselves to the placodal contribution to the trigeminus in mammals. Chiarugi ( ’94) described the intimate relation of the trigeminus to the ectoderm in the guineapig and noticed the occurrence of ectodermal papillae from which he believed cells were detached to be added to the ganglion. Giglio-Tos (’02), Volker (’22), and Bartelmez (’24) have also described ectodermal contributions to the trigeminus in man and the rat. Celestino da Costa (’23) does not definitely commit himself. He states that the relation of the gasserian ganglion to the ectoderm in the guinea-pig is intimate enough “a permettre une contribution placodiale a la formation du ganglion de Gasser,” but decides that “cette collaboration doit étre peu importante et bien plus réduite . . . ” Froriep (’85), VVeigner (’01), and Neumayer (’06) find no placodes related to the V ganglion in the mammalian forms studied by them. Davis (’23) thus describes the ganglion of the V nerve in a human embryo of 20 somites, “Laterally the mass is continuous with the condensed mesenchyme of the first branchial arch, the two tissues being c.lewrly differentiatcd ftromeach other. (The italics are mine.) The ectoderm covering this arch exhibits a distinct general thickening, but anything approaching a true placodal structure is wanting.
intimate contact with the mandibular-arch territory. The
 
mesenchymal condensation above and behind the eye shown
 
i11 figure 41 is a continuation forward of the mesenchymal
 
condensation of the mandibular arch. The long axis of the
 
arch is still oblique. During the interval between 14 and
 
18 somites at significant change has occurred. The maxillary
 
region has been established by the forward growth of material
 
at thedorsal end of the mandibular arch so that the optic
 
vesicle is now some distance removed from the mandibular
 
arch proper, the maxillary region intervening. The mai11
 
mass of the trigeminal ganglion lies over the caudal region
 
of the first arch and the ophthalmic ramus extends forward
 
from it, traversing the maxillary region to reach the optic
 
vesicle (cf. figs. 4 to 6).
 
  
An analysis of the growth shiftings of internal structures
+
It has been mentioned that the growth of the ophthalmicus occurs in the rat between 14 and 18 somites. The question of its origin may now be considered. In many forms the ophthalmicus profundus has been described as arising in part or entirely by placodal proliferation. Miss Platt (’96) believed both the gasserian ganglion and the ramus ophthalmicus profundus of Necturus to be derived from cells migrating from the primitive supraorbital ridge. Landacre (’12, p. 10) found a ‘pronounced contact’ of the ophthalmicus profundus with the epidermis in Lepidosteus, but did not determine its placodal nature. Stone (’22) believes that in Amblystoma punctatum the ophthalmicus profundus is composed entirely from cells proliferated from a supraorbital placode. Wlien the supraorbital placode is experimentally removed, the ophthalmicus fails to develop. In Chrysemys Brachet (’14) found that the maxillo-mandibular ganglion is entirely of neural—crest origin, but that the ophthalmic placode is transformed entirely and directly into the ophthalmic ganglion
will perhaps allow us to understand more fully what has
 
occurred. Between 10 and 14 somites considerable expansion has occurred in the ectopic region of the telencephalon,
 
and the diencephalon has enlarged a11d become more clearly
 
separable from it. The midbrain has elongated somewhat
 
and there has been some further bending of the neural tube.
 
So far as the relation of the optic vesicle to the trigeminal
 
ganglion is concerned, however, the growth shiftings occur
 
ring during this period have been, in the main, compensatory,
 
so that no significant change has occurred in the relation of
 
the two structures, the trigeminal ganglionic mass lying in
 
the transverse plane immediately dorsal to the caudal half
 
of the optic vesicle.
 
  
Between 14 and 18 somites, the cranial neural tube has
+
Belogolowy (’10) concludes (p. 169) that the formation of the ophthalmicus profundus of the chick must be ascribed exclusively to the condensation of ganglionic-crest elements. He could recognize no supplementary factors of any importance. However, he did find on each side, in relation to the ophthalmicus, one or two ectodermal papillae which were very different from the epibranchial placodes, but which resembled in their structure ectodermal papillae which he found ‘iiberall gleichm‘.ei.ssig verstreuten.The papillae associated with the ophthalmicus were two to three times as large as those found elsewhere, and he found that they constantly invaded the layer of mesenchyme lying between the ectoderm and the ganglion, and at their summit proliferated cells. These papillae are seen for only a short time and soon disappear without trace. He concludes that, since they yield so small an amount of material, their importance in ganglion formation must be minimal. Chiarugi (’94) and Griglio-Tos (’02) describe an ectodermal contribution to the ophthalmicus in the guinea-pig and man. Chiarugi noted the presence of ectodermal papillae from which he believed cells were detached to be added to the ganglion. Celestine da Costa (’23, p. 518) found in the guinea-pig that “Avant ou aprés la courte période entre 16 et 18 myotomes l’épibla’ste qui recouvre la créte du trijumeau et son long prolongement antérieur qui va jusqu’auX vésicules optiques réagit a ce voisinage, en 's’épaississant d’une facon diffuse. La possibilité n’est pas exclue d’une contribution de ce genre, telle que Chiarugi l’avait supposée.” It is my impression, however, that Celestino da Costa has misinterpreted the sections presented as evidence (see his figs. 7, 10). They do not show the ophthalmic ramus of the trigeminus, but pass through the caudal wall of the optic vesicle and the main trigeminal mass (cf. my 14-somite model). The right side of his figure 10 shows the thickening of the branchial integument with subjacent mesenchyme, not a dorsolateral placode and the ophthalmic ramus.
grown enormously. The prospective cerebral region has expanded greatly. The optic vesicle has grown slightly, but
 
its relative size with respect to the telencephalon has been
 
much reduced. The diencephalon, too, has expanded con—
 
siderably, but by far the most marked expansion has occurred
 
in the region of the midbrain which has practically doubled
 
i11 length from the di-mesencephalic boundary to the anterior
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
margin of the cerebellar rhombomere. The le11gthening and
+
7 In this connection, I desire to call attention to a recent paper by de Beer ( ’2-1).  
accentuation of the isthmus is especially noticeable. The
 
cerebellar rhombomere has also lengthened. Now, while in
 
the 14-somite embryo the compensatory flexing of the neural
 
tube is suflicient to retain the earlier relation of optic vesicle
 
and trigeminal anlage, in the period between 14: and 18 somites flexure of the neural tube has not kept pace with its
 
rapid expansion, and as a result the eye suffers considerable
 
forward displacement with respect to the main mass of the
 
trigeminus and the dorsal portion of the mandibular arch.
 
The lengthening of this interval is accompanied by the shifting forward of the material dorsal to the mandibular arch
 
to establish the maxillary region.
 
  
The important fact from the standpoint of this discussion
 
is the fact that the trigeminal anlage has taken part in the
 
growth shiftings above analyzed and as the optic vesicle has
 
been displaced forward the trigeminal anlage has kept pace
 
with that shifting, the ophthalmic ramus representing material which has grown forward from the main ganglionic
 
mass simultaneously with the elongation of the maxillary
 
region. A study of 16- and 17-somite embryos is confirmatory. The growth of the ophthalmic ramus forward keeps
 
pace with the shifting of the optic vesicle, which occurs with
 
startling suddenness between 14 and 18 somites. There are
 
no placodes to account for its formation and there is no evidence from a study of the rat to support Be1ogolowy’s (’10)
 
supposition that it arises by the condensation of crest elements formed along the extent of the midbrain. Neal (’98,
 
pl. 3) illustrates a comparable forward growth of the ophthalmicus profundus from the main mass of the trigeminus ganglion in the shark, and I believe that both Landacre’s (’21,
 
figs. 7 to 10) and Stone’s (’22, figs. 1 to 8) plottings of the
 
neural crest in Amphibia may be similarly interpreted.
 
  
N 0 great change in the gross morphology of the ganglionic
 
anlage of the V nerve has occurred in the 26—somite embryo
 
(cf. figs. 6 and 19). The cephalic fiexure is more acute than
 
in the 18-somite embryo and the optic vesicle has shifted pos70 HOWARD B. ADELMANN
 
  
teriorly for a short distance so that its caudal margin lies
 
opposite the posterior limit of rhombomere 2 in a plane
 
transverse to the long axis of the neural tube. In the 18somite embryo the caudal margin of the optic vesicle lies
 
relatively much farther forward, in the plane of the anterior
 
border of the first rhombomere. The ophthalmic ramus is
 
still entirely cellular in nature, but much less diffuse than at
 
18 somites, extending now for some distance over the optic
 
vesicle, but too tenuous to model. There is as yet no evidence
 
of a maxillary ramus and only a short ventro-caudal projection represents the mandibular branch. Tho ganglion is
 
now separated somewhat from the condensed mesenchyme of
 
the arch. In earlier embryos (figs. 34, 37) the ventral extremity of the V ganglion makes a verylbroad contact with the
 
condensed mesenchyme of the mandibular arch, and the two
 
tissues are sometimes difiicult to distinguish. However, after
 
careful study, it is not impossible to differentiate the ganglionic mass from the mesoderm, the two tissues differing
 
somewhat in staining properties. Davis (’23) found in a
 
20-somite human embryo “the mass is continuous with the
 
condensed mesenchyme of the first branchial arch, the two
 
tissues being clearly differentiated from each other.” I
 
find that there is a distinct medial notch constantly present
 
at the boundary between ganglion and arch mesenchyme in
 
embryos of 10 to 24 somites, and when the ganglion separates from the mesenchyme at 26 to 27 somites, the line
 
of separation lies at the level of the notch in younger embryos (cf. figs. 37, 38). Another feature which serves
 
to indicate the boundary between ganglion and mesenchyme is the presence of considerable numbers of degenerating cells at the boundary between the two, commencing with embryos about 17 somites in age. All stages of
 
degeneration may be observed, and there are in addition
 
some cells having irregular nuclei and containing a large
 
number of coarse acid-staining granules. These cells apparently act as phagocytes. Some degenerating cells and some
 
phagocytes may be found among the ganglion cells, but they
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
occur predominantly in the arch mesenchyme, becoming very
+
In my study of the rat several places were noted where the ectoderm showed slight papillary thickenings to the inner surfaces of which cells adhered in a manner to suggest a possible proliferation. Then again, where the ganglion approaches the ectoderm closely, ganglion cells were found adherent to the overlying ectoderm which was not especially thickened at the place of contact. In places where the ganglionic mass was separated from the ectoderm some of the intervening mesenchymal cells might appear to adhere to the overlying ectoderm. These appearances, however, are not confined to the ectoderm overlying the trigeminal ganglion or to the i11tegument of the head, but may be observed all over the body, as botl1 Belogolowy (’10) and Veit (’22) have already observed. None of Veit’s illustrations show a cell actually being pinched off from the ectoderm, but merely fully formed cells adhering to it. I interpret these appearances as cases of adhesion. VV. H. Lewis (’22) has commented upon the marked adhesive properties of mesenchymal cells, and of cells in tissue cultures, such cells frequently becoming so adherent to glass that they may even be centrifugalized without dislodgment. He also points out the important role which this adhesive property plays in early stages of development. Cells in culture were found to migrate out from the explant along certain solid objects (coverglasses, fibrin filaments, etc.) for which they seem to display a special adhesive tendency. I think that this adhesive property of embryonic cells is responsible for the fact that the V and VII (and the early IX-X) ganglionic proliferations grow ventrally in such close approximation to the overlying ectoderrn which possibly exercises a directive influence, and is responsible also for the fact that the vagus (cf. figs. 83, 88) hugs so closely the outer wall of the vena capitis lateralis as it grows ventrally. It explains why it is so difficult to draw a boundary line between branchial ectoderm and brancl1ial mesoderm and accounts for numerous descriptions of mesectoderm formation by the branchial integument. Brachet (’21, p. 385) has noted this peculiar and intimate relationship between the ganglionic anlagen and the deep face of the ectoderm; “il semble que ce dernier (the epiblast) exerce, sur ses couche superficielles, un tactisme particulier,” he says. If the proliferation of mesenchymal or ganglionic elements is so general a phenomenon, one would expect to find abundant examples of cells being actually pinched off from the ectoderm into the subjacent tissue.
abundant in later stages. Such cells were found also in relation to the developing VII, IX, and X nerves, a few scattered
 
among the ganglion cells, but mostly confined to the mesenchyme. They occur first in the mandibular, next in the hyoid,
 
and finally in the mesenchyme of the third and fourth arches.
 
Davis (’23) has described a large number of small polymorphonuclear cells, many of which exhibit vacuolization, indicating “possibly a degeneration of neuroblasts or an invasion
 
of the ganglia by macrophages.” Since I have found these
 
cells to be constantly present in the rat in connection with
 
the developing ganglia, I can hardly regard them as abnormal,
 
but I am inclined to consider them as cells degenerating as
 
a result of growth pressure or tension at the junction of the
 
two tissues. Similar cells have been found in other situations of the body. A more complete discussion of them is
 
reserved for a later communication.
 
  
At 29 somites the ophthalmic ramus is becoming more
+
The only constant ectodermal thickening related to the eye is illustrated in figure 41. It lies just caudal to and above the eye and the underlying mesenchymal condensation comes intoclose contact with it. Its under surface is intact. This is, however, not a placode. It is simply a thickening of the ectoderm above the cleft separating the mandibular arch from the anterior part of the head and it is continuous with the thickened integument of the mandibular arch and a less marked diffuse thickening of the ectoderm extending forward over the eye. It can be observed for a long period, but never comes into contact with the ophthalmic ramus of V. This thickening was, I believe, noticed by Froriep (’85, p. 44) in older embryos of the calf, but he says “ . . . . dieselbe steht aber in keiner Beziehung zum Trigeminus. Sie ist vielleicht ein Rudiment der Seitenorgananlage, welche bei Selachierembryonen an den vorderen Dorsalasten des Facialis sich bildet.” A placodal origin for the ramus ophthalmicus profundus trigemini in the mammal is, I believe, untenable.
fibrous distally. The mandibular branch is short and a few
 
motor fibers accompany it. The maxillary ramus is represented by a few scattered fiber bundles projecting into the
 
maxillary region.
 
  
In a 34-somite embryo the three branches of the ganglion
 
are well developed. The ophthalmicus is very cellular proximally, but becomes more fibrous distally. The mandibular
 
division can be traced into the mandibular arch for a short
 
distance. It also consists of a cellular extension from the
 
main ganglionic mass, sending a few fibers into the arch
 
accompanied by the fibers of the motor root. The maxillary
 
division is most diffuse, consisting of a few scattered fiber
 
bundles springing directly from the ganglion into the mesen—
 
chyme of the maxillary region. Since the further history
 
of the trigeminal ganglion is so well known, it is omitted
 
here. g
 
  
Returning now to the study of the main ganglionic mass
+
The possibility of a placodal origin of the ophthalmicus being excluded, how, then, does the ramus ophthalmicus of the V ganglion arise in the rat between the ages of 14 and 18 somites? The answer lies in a study of the growth transformations of the region during that period. Up to the 14somite stage, growth has proceeded but slowly, since the embryo requires some time to becomeproperly established in the uterus. At about this time, however, environmental conditions more favorable to rapid growth have apparently been established and the embryo consequently ‘speeds up’ in growth, the increase in bulk between 14 and 18 somites being almost incredible. Coincidently with this enormous growth, significant shiftings of parts have occurred which account, I believe, for the growth of the ophthalmic ramus of the V ganglion.
of the trigeminus of the 18-somite embryo, we find at that
 
time a spur-like projection from the antero-ventral aspect
 
72 HOWARD B. ADELMANN
 
  
of the base of the ganglionic mass. The cells resemble those
 
of the ganglion in their histological appearance and are without doubt derived from the main ganglionic mass, growing
 
ventrally in close contact with the wall of the neural tube.
 
In later stages there is evidenced the separation of a medial
 
mass of ganglionic cells, of which the spur is a continuation,
 
from at more lateral group which makes contact with the
 
neural tube just caudal to the medial mass so separated. It
 
is well shown in a 21—somite embryo (fig. 37).
 
  
The spur is well defined at 27 somites (fig. 38); its cells
+
Externally, one notes the following changes. In embryos of 14 somites and younger, the mandibular arch lies imme diately caudal to that part of the head composed mainly of
are still very similar in appearance to those of the main
 
ganglionic mass, but there is a distinct difference in their
 
arrangement. At 31 somites (fig. 42) there is a great difference in the appearance of the cells of the spur and those of
 
the main ganglionic mass. Those of the latter are spindleshaped with cloudy, deeply staining cytoplasm; the nucleus
 
is oval and the cells are compactly arranged with their long
 
axes parallel with the long axis of the ganglionic mass. The
 
cells of the spur, on the other hand, are irregular in shape;
 
the nuclei are round or slightly oval and the arrangement of
 
the cells is irregular. At 34 somites (fig. 43) the mass is
 
invaded by the outgrowth of motor fibers from the neural
 
tube, and finally the mass is obliterated when the motor root
 
is well established. Even at 31 somites there are delicate
 
fibrils to be detected extending from the neural tube into the
 
spur. A similar spur with an identical fate has been observed
 
in the case of the acoustico-facial anlage. It will be described,
 
and a discussion of its significance will be given later.
 
  
Other than the slight temporary cleavage of the ganglion
+
forebrain. Since in early embryos the optic vesicle is relatively larger than subsequently, constituting almost the entire lateral wall of the forebrain, the core of the mandibular arch lies just caudal to the eye. Consequently, no maxillary region can be identified. The above facts are most strikingly illustrated in embryos of 8 to 14 somites (figs. 14, 4, 17). The external aspect of the 10—somite embryo shows that between 8 and 10 somites some forward growth of dorsally lying parts has occurred, and this is reflected in the oblique direction of the long axis of the mandibular arch, the dorsal extremity of which lies considerably cranial to its Ventral end, which is more or less fixed. A forward extension ofthe mesenchymal condensation of the first arch actually rests against the postero-superior border of the optic vesicle. The position of the trigeminus with respect to the optic Vesicle has already been commented upon.  
noted above, there is no evidence, in the rat, supporting the
 
view that the trigeminal ganglion of the mammal is a composite structure. Giglio-Tos (’02) concludes from the study
 
of a single 15-somite human embryo that there is a complete
 
correspondence between the development of the trigeminus
 
in man and the lamprey. He identifies nine parts in the
 
trigeminal ganglion of man--three primitive neural proganglia (ophthalmic, maxillary, a11d mandibular), three mesoNEURAL FOLDS AND CRANIAL GANGLIA or RAT 73
 
  
cephalic (epibranchial) proganglia, and three branchial pronerves. Giglio-Tos’ figures show clearly that he has
 
confused mcsenchymal condensations with ganglionic anlagen.
 
  
A careful search was made for evidence of the existence
 
of a separate profundus division of the trigeminus, but none
 
was found. Schulte and Tilney have described a profundus
 
anlage in the cat and have figured it in 10-, 12-, 14-, 16-, 17-, 19-,
 
and 21-somite embryos. I have become convinced that the
 
structure consistently termed by them the profundus ganglion
 
in the figures given on plates 32 to 38 is in reality the quintal
 
anlage and that the structure termed by them the quintal
 
anlage is really the facial. What they have called the facial
 
is really the IX-X anlage continuous with the spinal neural
 
crest. In a 12-somite cat which I have examined the trigeminal ganglion is quite small, as Schulte and Tilney’s model
 
of the so-called profundus anlage shows, and this may be
 
the reason for the confusion. A study of their figures (pl. 34,
 
  
for instance) shows the so—called quintal anlage attached to '
 
  
the otic rhombomere which may be identified by its ventral
 
swelling. Ahead of it lies rhombomere 3, free of ganglionic
 
attachment, and then their so-called profundus anlage attached to the caudal portion of rhombomere A1. Caudal to
 
the otic rhombomere lies rhombomere V, and still farther
 
caudally is situated their so-called acoustico-facial continuous
 
with the spinal crest. This figure alone quite plainly indicates that their profundus equals the trigeminus, their quintal anlage is really the facial, and that their acoustico-facial
 
is the IX-X anlage with its characteristic relation to the spinal
 
crest. Schulte and Tilney have apparently ignored the rostral neural crest of younger embryos. The cat is apparently
 
not alone in having a quintal anlage of smaller size than the
 
acoustico-facial, since Davis (’23, p. 19) mentions the same
 
condition in man. Celestine da Costa (’23) and I have independently come to the same conclusion regarding Schulte and
 
Tilney’s profundus anlage.
 
74 HOWARD B. ADELMANN
 
  
,2. The acoustico-facial ganglion i
+
Figs. 4: t.o 6 A series of reconstructions from models of 14-, 18-, and 26-somite rats, respectively (X 50), to illustrate the formation of the ophthalmic ramus of the trigeminus, coincident with the growth shiftings involved in the establishment of the maxillary region. The figures also illustrate the relation of the ganglia to the branchial arches. See text, pages 65-69, for discussion. In figure 4 the extent of the thickened ectoderm surrounding the otic pit is indicated by a broken line. In figure 5 the otic vesicle has not yet been completely closed off from the ectoderm (fig. 18).
  
The position and relations of the acoustico-facial anlage
+
A.br., branchial arch; D.c.c., dorsal cornmunicating cord of ganglionic material between the IX and X ganglia; G., ganglion; 013.19., optic vesicle; 0t.pl., otic placode; Ot.'u., otic vesicle; R.op.V., ophthalmic ramus of the trigeminus; SJ, the first somite; V.c.c., ventral communicating cord of ganglionic material between the IX and X ganglia.  
of the 8-somite embryo have been briefly described above and
 
are illustrated in figure 27. It is shown on both sides of the
 
figure. The plane is very slightly oblique, so that the section
 
passes through the anterior end of the ganglion on the right
 
and through the caudal portion of the left ganglion. The
 
proliferation is more advanced in the anterior than in the
 
caudal part of the ganglion, but only two or three sections
 
on each side show an active ventral migration of cells. The
 
anlage extends through thirteen 10 u sections on the right
 
and twelve on the left. The relations of the mesoderm should
 
be observed. The loose mesenchymal material forming the
 
paraxial strand is connected with the pericardial mesoderm
 
by an intermediate column of cells which passes lateral to the
 
pharynx and which in subsequent stages becomes greatly
 
condensed to form the mesodermal core of the second gill
 
arch when that structure is delimited by the development of
 
the second pouch. One has no difiiculty in defining the boundaries of the anlage even in sections where proliferation is
 
most advanced.
 
  
The otic placode is extensive at 8 somites, and it is impossible to determine its exact boundaries. It is about 260 u in
 
length on the right and somewhat less extensive on the left
 
side. It is probable that the slight invagination of the placode illustrated in figure 27 is an artifact. The acoustico—
 
facial anlage rests closely against the medial surface of the
 
cephalic extremity of the placode; there is no intervening
 
mesenchyme.
 
  
A very early stage of the acoustico-facial anlage can be
 
recognized in a 7-somite embryo. Its position with respect
 
to the neuralplate is as described. for the 8-somite embryo.
 
  
Figure 44 is a section through the acoustico-facial anlage
 
of a 9-somite embryo. The section is somewhat horizontal
 
in direction, the lower portions of the section being caudal to
 
the upper. The left side of the section is anterior to the right.
 
NEURAL FOLDS AND CRANIAL GANGLIA on RAT 75
 
  
The ganglion has grown considerably in size since the 8-somite
+
At 12 and 14 somites (figs. 4 and 17) the eye is still in intimate contact with the mandibular-arch territory. The mesenchymal condensation above and behind the eye shown i11 figure 41 is a continuation forward of the mesenchymal condensation of the mandibular arch. The long axis of the arch is still oblique. During the interval between 14 and 18 somites at significant change has occurred. The maxillary region has been established by the forward growth of material at thedorsal end of the mandibular arch so that the optic vesicle is now some distance removed from the mandibular arch proper, the maxillary region intervening. The mai11 mass of the trigeminal ganglion lies over the caudal region of the first arch and the ophthalmic ramus extends forward from it, traversing the maxillary region to reach the optic vesicle (cf. figs. 4 to 6).
stage. It consists of a column of densely packed rounded or
 
oval cells rich in cytoplasm, which adheres closely to the inner
 
surface of the otic placode. The latter, however, takes no
 
part in the formation of the ganglion;absolutely no indication of proliferation could be observed. The ganglion can
 
be traced ventrally to the dorsal level of the first gill pouch
 
where it is in intimate contact with the mesodermal material
 
of the second arch, but the two tissues can be differentiated.
 
In transverse sectionsithe ganglionic mass is fusiform in
 
shape; in frontal sections (fig. 20) it is round or slightly
 
oval. One notes that the ganglionic mass is well, even sharply
 
delimited from the mesenchyme medial to it. As in the previous stage, the primitive head vein lies just medial to it.
 
The mesodermal stalk lateral to the pharynx has not grown
 
appreciably. Neither is it more compactthan previously.
 
  
The extent of the otic placode and its relations to the acoustico-facial ganglion are well shown in figure 20. The thickening of the placode is most pronounced opposite the fourth
+
An analysis of the growth shiftings of internal structures will perhaps allow us to understand more fully what has occurred. Between 10 and 14 somites considerable expansion has occurred in the ectopic region of the telencephalon, and the diencephalon has enlarged a11d become more clearly separable from it. The midbrain has elongated somewhat and there has been some further bending of the neural tube. So far as the relation of the optic vesicle to the trigeminal ganglion is concerned, however, the growth shiftings occur ring during this period have been, in the main, compensatory, so that no significant change has occurred in the relation of the two structures, the trigeminal ganglionic mass lying in the transverse plane immediately dorsal to the caudal half of the optic vesicle.
and fifth rhombomeres. Its invagination commences first
 
caudal to the acoustico-facial anlage, but for some time the
 
latter lies beneath a pronounced anteriorlip of the otic pit
 
(fig. 53).
 
  
The most noticeable change in the 11-somite embryo (fig.
+
Between 14 and 18 somites, the cranial neural tube has grown enormously. The prospective cerebral region has expanded greatly. The optic vesicle has grown slightly, but its relative size with respect to the telencephalon has been much reduced. The diencephalon, too, has expanded considerably, but by far the most marked expansion has occurred in the region of the midbrain which has practically doubled i11 length from the di-mesencephalic boundary to the anterior margin of the cerebellar rhombomere. The le11gthening and accentuation of the isthmus is especially noticeable. The cerebellar rhombomere has also lengthened. Now, while in the 14-somite embryo the compensatory flexing of the neural tube is suflicient to retain the earlier relation of optic vesicle and trigeminal anlage, in the period between 14: and 18 somites flexure of the neural tube has not kept pace with its rapid expansion, and as a result the eye suffers considerable forward displacement with respect to the main mass of the trigeminus and the dorsal portion of the mandibular arch. The lengthening of this interval is accompanied by the shifting forward of the material dorsal to the mandibular arch to establish the maxillary region.
45) is the great growth of the mesoderm of the second visceral
 
arch. It has not only increased considerably in thickness,
 
but has become much compacted. The individual cells have
 
consequently lost their stellate shape, and as a result resemble
 
very closely the cells of the ganglionic anlage. Ten-somite
 
embryos show the arch mesoderm in stages of compaction
 
intermediate between the 9- and 11-somite embryos. The
 
acoustico-facial anlage has changed but slightly. It is still
 
a fusiform mass extending ventrally to the level of the dorsum of the first gill pouch, which extends dorsally to the
 
level of the ventral surface of the neural tube. In spite of
 
the similarity between the compacted mesodermal cells of
 
the hyoid arch and those of the ganglionic mass, the ventral
 
76 HOWARD B. ADELMANN
 
  
limit of the acoustico—facial anlage can be approximately
 
fixed. The long axis of the ganglion is oblique to the axis of
 
the hyoid arch, so that there is a medial notch where the
 
ganglion comes into contact with the arch mesoderm. The
 
notch lies at the upper level of the dorsum of the first gill
 
pouch, to which height the mesoderm ca11 be seen to extend
 
in sections just in front of or caudal to the ganglion (fig. 46),
 
and it will be recalled that the ganglion extends ventrally
 
to that level iii 9- and 10—somite embryos. A study of 10somite embryos yields convincing evidence that the ganglionic
 
cells take no part in the formation of the mesodermal core
 
of the branchial arch.
 
  
A very thin lamina of looser cells (probably mesenchymal)
+
The important fact from the standpoint of this discussion is the fact that the trigeminal anlage has taken part in the growth shiftings above analyzed and as the optic vesicle has been displaced forward the trigeminal anlage has kept pace with that shifting, the ophthalmic ramus representing material which has grown forward from the main ganglionic mass simultaneously with the elongation of the maxillary region. A study of 16- and 17-somite embryos is confirmatory. The growth of the ophthalmic ramus forward keeps pace with the shifting of the optic vesicle, which occurs with startling suddenness between 14 and 18 somites. There are no placodes to account for its formation and there is no evidence from a study of the rat to support Be1ogolowy’s (’10) supposition that it arises by the condensation of crest elements formed along the extent of the midbrain. Neal (’98, pl. 3) illustrates a comparable forward growth of the ophthalmicus profundus from the main mass of the trigeminus ganglion in the shark, and I believe that both Landacre’s (’21, figs. 7 to 10) and Stone’s (’22, figs. 1 to 8) plottings of the neural crest in Amphibia may be similarly interpreted.
separates the upper portion of the ganglion from the ectoderm in the 11-somite embryo. It can be recognized by its
 
more diffuse character and tl1e presence of a few blood cells.
 
The blood cells have deeply staining nuclei in the figure.
 
  
Tl1e otic placode becomes thicker, but the invagination
+
N 0 great change in the gross morphology of the ganglionic anlage of the V nerve has occurred in the 26—somite embryo (cf. figs. 6 and 19). The cephalic fiexure is more acute than in the 18-somite embryo and the optic vesicle has shifted posteriorly for a short distance so that its caudal margin lies opposite the posterior limit of rhombomere 2 in a plane transverse to the long axis of the neural tube. In the 18somite embryo the caudal margin of the optic vesicle lies relatively much farther forward, in the plane of the anterior border of the first rhombomere. The ophthalmic ramus is still entirely cellular in nature, but much less diffuse than at 18 somites, extending now for some distance over the optic vesicle, but too tenuous to model. There is as yet no evidence of a maxillary ramus and only a short ventro-caudal projection represents the mandibular branch. Tho ganglion is now separated somewhat from the condensed mesenchyme of the arch. In earlier embryos (figs. 34, 37) the ventral extremity of the V ganglion makes a verylbroad contact with the condensed mesenchyme of the mandibular arch, and the two tissues are sometimes difiicult to distinguish. However, after careful study, it is not impossible to differentiate the ganglionic mass from the mesoderm, the two tissues differing somewhat in staining properties. Davis (’23) found in a 20-somite human embryo “the mass is continuous with the condensed mesenchyme of the first branchial arch, the two tissues being clearly differentiated from each other.” I find that there is a distinct medial notch constantly present at the boundary between ganglion and arch mesenchyme in embryos of 10 to 24 somites, and when the ganglion separates from the mesenchyme at 26 to 27 somites, the line of separation lies at the level of the notch in younger embryos (cf. figs. 37, 38). Another feature which serves to indicate the boundary between ganglion and mesenchyme is the presence of considerable numbers of degenerating cells at the boundary between the two, commencing with embryos about 17 somites in age. All stages of degeneration may be observed, and there are in addition some cells having irregular nuclei and containing a large number of coarse acid-staining granules. These cells apparently act as phagocytes. Some degenerating cells and some phagocytes may be found among the ganglion cells, but they occur predominantly in the arch mesenchyme, becoming very abundant in later stages. Such cells were found also in relation to the developing VII, IX, and X nerves, a few scattered among the ganglion cells, but mostly confined to the mesenchyme. They occur first in the mandibular, next in the hyoid, and finally in the mesenchyme of the third and fourth arches. Davis (’23) has described a large number of small polymorphonuclear cells, many of which exhibit vacuolization, indicating “possibly a degeneration of neuroblasts or an invasion of the ganglia by macrophages.” Since I have found these cells to be constantly present in the rat in connection with the developing ganglia, I can hardly regard them as abnormal, but I am inclined to consider them as cells degenerating as a result of growth pressure or tension at the junction of the two tissues. Similar cells have been found in other situations of the body. A more complete discussion of them is reserved for a later communication.
has progressed but slightly between 9 and 11 somites.
 
  
In the interval between 12 and 14 somites (figs. 47 to 50),
 
the acoustico—facial anlago keeps pace in growth with the
 
embryo and its relations to surrounding parts remain much
 
the same as in earlier stages. During this interval the neural
 
tube closes in the region of the otic rhombomere and the
 
figures show the attendant ventral shifting of the attachment
 
of the ganglion to the neural tube. The fusion of the lips
 
of the neural folds is not very intimate, so that they separate
 
somewhat in the 14-somite embryo as a result of handling.
 
At 12 somites the dorsal end of the ganglion is in contact
 
medially with the dorsal surface of the closing neural tube,
 
lying between it and the overlying ectoderm (fig. 47), but at
 
14 somites it has shifted ventrally, so that it is now in contact with the middle third of the neural tube.
 
  
The compact mesoderm of the hyoid arch still extends
+
At 29 somites the ophthalmic ramus is becoming more fibrous distally. The mandibular branch is short and a few motor fibers accompany it. The maxillary ramus is represented by a few scattered fiber bundles projecting into the maxillary region.
dorsally to the upper level of the first gill pouch, i.e., approximately to the level of the ventral surface of the neural tube
 
(fig. 47 and cf. the right and left sides of figs. 48, 50), where
 
NEURAL rows AND CRANIAL GANGLIA or RAT 77
 
  
it comes into contact with the ventral extremity of the ganglion. The long axis of the acoustico-facial ganglionic mass
 
is even more markedly oblique to the axis of the hyoid arch
 
than previously, and consequently there is a more pronounced
 
medial notch at the boundary between the two structures
 
than in younger embryos. Furthermore, one notices a distinct
 
difference in the arrangement of the cells of the mesoderm
 
and the ganglionic anlage, more pronounced in some specimens than in others. In some embryos the differentiation of
 
ganglionic and mesodermal tissues is particularly striking
 
(for instance, figs. 48, 49). ‘In embryos of 12 to 14 somites,
 
therefore, one is justified, I believe, in placing the ventral
 
limit of the ganglionic anlage at a point corresponding approximately to the dorsal level of the first pharyngeal pouch,
 
which is situated just anterior to the acoustico-facial anlage.
 
The exact level varies but slightly. It is impossible to determine whether there is an intermingling of ganglionic elements with the mesoderm at the place of contact.
 
  
Figures 47 to 50 also show the gradual separation of the
+
In a 34-somite embryo the three branches of the ganglion are well developed. The ophthalmicus is very cellular proximally, but becomes more fibrous distally. The mandibular division can be traced into the mandibular arch for a short distance. It also consists of a cellular extension from the main ganglionic mass, sending a few fibers into the arch accompanied by the fibers of the motor root. The maxillary division is most diffuse, consisting of a few scattered fiber bundles springing directly from the ganglion into the mesen— chyme of the maxillary region. Since the further history of the trigeminal ganglion is so well known, it is omitted here.
  
acoustico—facial anlage from the overlying ectoderm by the I
+
Returning now to the study of the main ganglionic mass of the trigeminus of the 18-somite embryo, we find at that time a spur-like projection from the antero-ventral aspect of the base of the ganglionic mass. The cells resemble those of the ganglion in their histological appearance and are without doubt derived from the main ganglionic mass, growing ventrally in close contact with the wall of the neural tube. In later stages there is evidenced the separation of a medial mass of ganglionic cells, of which the spur is a continuation, from at more lateral group which makes contact with the neural tube just caudal to the medial mass so separated. It is well shown in a 21—somite embryo (fig. 37).
  
invasion of mesenchyme during this period. The separation
 
of the ganglion from the ectoderm is more marked dorsally
 
than ventrally where the ganglion approaches the surface
 
more closely. In figures 47, 48, the separation appears more
 
marked on the right than on the left, but that is because the
 
plane passes more caudally on the left, passing immediately
 
(fig. 48) in front of the otic-pit against which the ganglion
 
rests. The integrity of the inner surface of the epithelium
 
both in relation to the ganglion and the branchial arch is
 
unquestionable. Frontal sections of the ganglion in the 14somite embryo show it as a perfectly discrete cell column (fig.
 
53) separated widely from the ectoderm dorsally, but approaching it ventrally. A new venous channel, the vena
 
capitis lateralis, has been formed in the mesenchyme lateral
 
to the ganglion between 12 and 14,somites. In the 12-somite
 
embryo sprouts from the primitive head vein are beginning
 
to work their way around the ganglionic mass, and the process
 
is completed at 13 somites.
 
  
THE JOURNAL or COMPARATIVE NEUROLOGY, voL. 39, NO. 1
+
The spur is well defined at 27 somites (fig. 38); its cells are still very similar in appearance to those of the main ganglionic mass, but there is a distinct difference in their arrangement. At 31 somites (fig. 42) there is a great difference in the appearance of the cells of the spur and those of the main ganglionic mass. Those of the latter are spindleshaped with cloudy, deeply staining cytoplasm; the nucleus is oval and the cells are compactly arranged with their long axes parallel with the long axis of the ganglionic mass. The cells of the spur, on the other hand, are irregular in shape; the nuclei are round or slightly oval and the arrangement of the cells is irregular. At 34 somites (fig. 43) the mass is invaded by the outgrowth of motor fibers from the neural tube, and finally the mass is obliterated when the motor root is well established. Even at 31 somites there are delicate fibrils to be detected extending from the neural tube into the spur. A similar spur with an identical fate has been observed in the case of the acoustico-facial anlage. It will be described, and a discussion of its significance will be given later.
78 HOWARD B. ADELMANN
 
  
Figure 51 is a section through the ventral tip of the
 
acoustico-facial anlage of an 18-somite embryo. Its attachme11t to the neural tube is shown in more caudal sections.
 
Figure 51 is the third section caudal to the first branchial
 
pouch. A study of it shows that while the embryo has grown
 
markedly there has been no change in the relative positions
 
of the parts concerned. The ventral extremity of the ganglion can be easily determined in figure 51, resting directly
 
upon the mesodermal core of the hyoid arch. It can be
 
traced forward to the section immediately caudal to the
 
first branchial pouch where it makes contact with the epibranchial placode, but there is no continuity between the two
 
structures. A line drawn laterally from the ventral surface of
 
the neural tube coincides approximately with the ventral end
 
of the anlage and serves as a convenient reference line for
 
comparison with younger embryos. Figure 52, passing
 
through the attachment of the VILVIII ganglion to the brain,
 
three sections caudal to figure 531, shows that the compact
 
core of the hyoid arch has the same relative dorsal extent as
 
in younger embryos. Study of the series shows that the upper
 
limit of the mesodermal core of the hyoid arch lies at the
 
level of the dorsal surface of the first branchial pouch—
 
a fact well illustrated also in figure 67, a sagittal section of
 
an embryo of 19 somites. It will be appreciated from the
 
above description that the ventral extremity of the acousticofacial anlage of the 18-somite embryo lies in the same relative
 
position as in the younger stages—namely, immediately
 
caudal to the dorsal end of the first pouch.
 
  
The most important development of this’ stage is the placodal thickening of the ectoderm at the upper level of the
+
Other than the slight temporary cleavage of the ganglion noted above, there is no evidence, in the rat, supporting the view that the trigeminal ganglion of the mammal is a composite structure. Giglio-Tos (’02) concludes from the study of a single 15-somite human embryo that there is a complete correspondence between the development of the trigeminus in man and the lamprey. He identifies nine parts in the trigeminal ganglion of man--three primitive neural proganglia (ophthalmic, maxillary, a11d mandibular), three mesocephalic (epibranchial) proganglia, and three branchial pronerves. Giglio-Tos’ figures show clearly that he has confused mcsenchymal condensations with ganglionic anlagen.
arch. ‘It is illustrated in figures 51 and 52, and no doubt is
 
identical with Froriep’s epibranchial sense organ. The thickening begins immediately caudal a11d dorsal to the first cleft
 
and extends caudally and dorsally as far as the anterior
 
border of the otic vesicle. The extreme ventral end of the
 
acoustico—facial anlage rests lightly against it, but there is
 
no cellular continuity between the placode and the ganglion;
 
NEURAL FOLDS AND CR-ANIAL GANGLIA OF RAT
 
  
the inner surface of the placode is everywhere intact. Four
 
mitoses, tl1e planes of which were not determinable, were
 
  
found in the entire left placode which may be traced through
+
A careful search was made for evidence of the existence of a separate profundus division of the trigeminus, but none was found. Schulte and Tilney have described a profundus anlage in the cat and have figured it in 10-, 12-, 14-, 16-, 17-, 19-, and 21-somite embryos. I have become convinced that the structure consistently termed by them the profundus ganglion in the figures given on plates 32 to 38 is in reality the quintal anlage and that the structure termed by them the quintal anlage is really the facial. What they have called the facial is really the IX-X anlage continuous with the spinal neural crest. In a 12-somite cat which I have examined the trigeminal ganglion is quite small, as Schulte and Tilney’s model of the so-called profundus anlage shows, and this may be the reason for the confusion. A study of their figures (pl. 34, for instance) shows the so—called quintal anlage attached to the otic rhombomere which may be identified by its ventral swelling. Ahead of it lies rhombomere 3, free of ganglionic attachment, and then their so-called profundus anlage attached to the caudal portion of rhombomere A1. Caudal to the otic rhombomere lies rhombomere V, and still farther caudally is situated their so-called acoustico-facial continuous with the spinal crest. This figure alone quite plainly indicates that their profundus equals the trigeminus, their quintal anlage is really the facial, and that their acoustico-facial is the IX-X anlage with its characteristic relation to the spinal crest. Schulte and Tilney have apparently ignored the rostral neural crest of younger embryos. The cat is apparently not alone in having a quintal anlage of smaller size than the acoustico-facial, since Davis (’23, p. 19) mentions the same condition in man. Celestine da Costa (’23) and I have independently come to the same conclusion regarding Schulte and Tilney’s profundus anlage. 74 HOWARD B. ADELMANN
  
fifteen sections. It is evidently not growing very actively.
+
===2. The Acoustico-Facial Ganglion===
A few cells could be found in it which showed unmistakable
 
evidence of degeneration, and these degenerating cells increase greatly in number as time goes on.
 
  
From the base of the ganglion where it comes into contact
+
The position and relations of the acoustico-facial anlage of the 8-somite embryo have been briefly described above and are illustrated in figure 27. It is shown on both sides of the figure. The plane is very slightly oblique, so that the section passes through the anterior end of the ganglion on the right and through the caudal portion of the left ganglion. The proliferation is more advanced in the anterior than in the caudal part of the ganglion, but only two or three sections on each side show an active ventral migration of cells. The anlage extends through thirteen 10 u sections on the right and twelve on the left. The relations of the mesoderm should be observed. The loose mesenchymal material forming the paraxial strand is connected with the pericardial mesoderm by an intermediate column of cells which passes lateral to the pharynx and which in subsequent stages becomes greatly condensed to form the mesodermal core of the second gill arch when that structure is delimited by the development of the second pouch. One has no difiiculty in defining the boundaries of the anlage even in sections where proliferation is most advanced.
with the neural tube, one finds a spur of cells which seems to
 
be growing ventrally from the base of the ganglion, keeping
 
in close contact with the neural tube. 'Tl1is spur of ganglionic
 
cells is small at 18 somites, but is rather prominent at 21
 
somites (fig. 54). Its possible fate will be discussed later.
 
  
The changes are slight in 19- and 20-somite embryos. In
 
specimens of these ages the epibranchial placode is still extensive, reaching caudally as far as the otic vesicle. The
 
acoustico-facial anlage extends laterally as it proceeds ventrally, its ventral extremity just touching the epibranchial
 
placode, but not fusing with it. i
 
  
A series of sections through the acoustieo—facial anlage of
+
The otic placode is extensive at 8 somites, and it is impossible to determine its exact boundaries. It is about 260 u in length on the right and somewhat less extensive on the left side. It is probable that the slight invagination of the placode illustrated in figure 27 is an artifact. The acoustico—facial anlage rests closely against the medial surface of the cephalic extremity of the placode; there is no intervening mesenchyme.
a 21-somite embryo are given in figures 54 to 56. Figure 56
 
is a section through the caudal edge of the ganglion. It shows
 
the anterior surface of the otic vesicle. The attachment of
 
the ganglion to the brain is shown, but its ventral extremity
 
is not included in the section. The figure illustrates the dorsal
 
extent of the arch mesoderm, which is relatively the same as
 
in younger embryos. Figure 55 is five sections anterior,
 
revealing the full extent of the ganglion, whose ventral extremity approaches the placode. Three sections farther anteriorly (fig. 54) the ventral tip of the ganglion comes to lie
 
in the angle between the epibranchial placode a11d the first
 
branchial pouch, and in this embryo there is only a very
 
restricted contact, but no fusion of ganglion and placode at
 
this point. The ventralspur extending from the base of
 
the ganglion along the side of the neural tube is well shown.
 
  
In this,21-somite embryo the epibranchial placode is no
 
less extensive than in the 18- and 19-somite embryos described.
 
80 HOWARD B. ADELMANN
 
  
It has not thickened appreciably and no evidences of proliferative activity could be discerned. Its inner surface seems
+
A very early stage of the acoustico-facial anlage can be recognized in a 7-somite embryo. Its position with respect to the neuralplate is as described. for the 8-somite embryo.
intact, and while mesenchymal cells lie close against it, there
 
seemed to be no continuity between the two. Sagittal and
 
frontal sections of 21-somite embryos confirm the evidence
 
of the transverse sections as to the extent and relations of
 
the acoustico-facial anlage. At 21 somites a few degenerating cells and phagocytes may be observed in the dorsal part
 
of the mesoderm of the second arch. They are not numerous
 
and seem to lie chiefly caudal to the ganglion.
 
  
Three sections through the acoustico-facial anlage of a 26»
 
somite embryo are presented in figures 57 to 59. Figure 57 is
 
a section through the tip of the ganglion which extends into
 
the angle between the dorsal surface of the first branchial
 
pouch and the epibranchial placode, where it makes contact
 
with the placode, but does not fuse with it. The relation of
 
ganglion to placode is well shown in figure 59—a section
 
through the VII-VIII ganglion on the opposite side of the
 
same embryo. The contact of ganglion and placode extends
 
through only two sections on both sides of the embryo. Elsewhere ganglion and placode are separated by mesenchyme
 
which is closely applied to the placode. In sections immediately caudal to the pouch where the curvature of the arch is
 
steepest and in which consequently the placodes are most
 
obliquely cut, this close adhesion of mesenchyme and placode
 
may be interpreted as proliferative activity on the part of
 
the placode. Figures 57, 58 may serve as an illustration of
 
this. It will be observed that the obliquely cut wall of the
 
adjacent pouch shows an epithelio—mesenchymal relationship
 
similar to the placode, so that one must grant an entodermal
 
proliferation of mesenchyme if the placode is interpreted as
 
proliferating. In this connection, one recalls Weigner’s comment: “ . . . . die Beurtheilung eines Zusammenhanges
 
gerade an dieser Stelle mit fast uniiberwindlichen Schwierigkeiten zu kampfen hat. Es ist eben jeder der Schlundbogen
 
nach allen Richtungen convex und wird deshalb immer seine
 
Epithel schief getroffen. Liegt nun eine Zellgruppe diesem
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
Epithel nahe so kiinnen sehr leicht solche Schiefschnitte eine
+
Figure 44 is a section through the acoustico-facial anlage of a 9-somite embryo. The section is somewhat horizontal in direction, the lower portions of the section being caudal to the upper. The left side of the section is anterior to the right.  
Verbindung vortéiuschen auch an Stellen, an denen keine
 
existiert.” However, the difficulty is solved when frontal
 
sections of embryos of the same age are consulted. These
 
show no evidence of proliferation from the placodes, although
 
mesenchyme may be in close contact with them.
 
  
In some of the eight 26-somite embryos examined, a number of degenerating cells were found in the placodal epithelium, giving evidence of early retrogressive change. Mitoses
 
are relatively few, those found are superficial in position,
 
with the spindle axis parallel to the free surface when determinable. Considerable numbers of degenerating cells and
 
phagocytes are found in the mesenchyme near the dorsal end
 
of the second arch in these embryos.
 
  
The acoustico-facial anlage shows no essential change in
 
eight 27 —somite embryos studied. In 29- and 30-somite embryos the anlage is somewhat larger, and for the first time a
 
few fibers can be seen extending from the ganglion for a
 
short distance into the hyoid arch.
 
  
A 30-somite embryo, favorably sectioned, shows interesting developments. The ganglion has been still further separated from the ectoderm by the increase of the intervening
+
The ganglion has grown considerably in size since the 8-somite stage. It consists of a column of densely packed rounded or oval cells rich in cytoplasm, which adheres closely to the inner surface of the otic placode. The latter, however, takes no part in the formation of the ganglion;absolutely no indication of proliferation could be observed. The ganglion can be traced ventrally to the dorsal level of the first gill pouch where it is in intimate contact with the mesodermal material of the second arch, but the two tissues can be differentiated. In transverse sectionsithe ganglionic mass is fusiform in shape; in frontal sections (fig. 20) it is round or slightly oval. One notes that the ganglionic mass is well, even sharply delimited from the mesenchyme medial to it. As in the previous stage, the primitive head vein lies just medial to it. The mesodermal stalk lateral to the pharynx has not grown appreciably. Neither is it more compactthan previously.
mesenchyme. The ventral extremity of the ganglionic anlage
 
is now perfectly distinct, and from it two branches may be
 
followed. A short cellular cord proceeds from the ventral
 
extremity of the ganglion laterally to the placode to which
 
it adheres, but there still seems to be no intimate fusion of
 
the two tissues. This placodal ramus is formed, I believe,
 
during the separation of the ganglion from the ectoderm by
 
the expansion of the hyoid arch. The original contact of the
 
ventral extremity of the ganglion and the placode persisting,
 
the recession of the ganglionic mass from the ectoderm results in the ‘drawing out’ of a cord of cells leading from the
 
ventral extremity of the ganglion to the site of original contact with the placode. The second branch of the ganglion
 
can be followed into the hyoid arch for some distance.
 
82 HOWARD B. ADELMANN
 
  
A similar but somewhat stouter placodal ramus may be
 
identified in embryos of 31 a11d 34 somites (figs. 60, 61). Its
 
caliber varies considerably, but its point of attachment to
 
the placode is remarkably constant in position, namely, just
 
above and behind the first branchial pouch in the angle between pouch and placode. This point, it will be recalled,
 
has been described as marking the ventral limit of the ganglion in earlier stages. The relations just described are well
 
shown in figure 61, which is a frontal section of a 34-somite
 
embryo. In this embryo three terminal branches of the ganglion may be followed—one passes antero—medially between
 
the dorsal aorta a11d the pharynx (the greater superficial
 
petrosal nerve), one continues ventrally into the hyoid arch,
 
and the third is the placodal ramus above described.
 
  
The placodal branch of the geniculate ganglion is still well
+
The extent of the otic placode and its relations to the acoustico-facial ganglion are well shown in figure 20. The thickening of the placode is most pronounced opposite the fourth and fifth rhombomeres. Its invagination commences first caudal to the acoustico-facial anlage, but for some time the latter lies beneath a pronounced anteriorlip of the otic pit (fig. 53).
developed in some embryos of thirteen days, but at this time
 
it seems to be in process of disappearance, since in some
 
specimens of this age it can be recognized upon one side only
 
and in others it can be recognized with difficulty. In some it
 
is still prominent.
 
  
In an embryo of thirteen days four hours there is no
 
longer any trace of an epibranchial connection of the genicu—
 
late ganglion. On one side the placodal region takes part
 
in the formation of a short ductus branchialis I with a very
 
small lumen. On the other side the adjacent surfaces of the
 
first cleft have fused, forming a short cord of cells leading
 
from the pouch to the surface, but there is no lumen.
 
  
One could never be surer of the perfect integrity of the
+
The most noticeable change in the 11-somite embryo (fig. 45) is the great growth of the mesoderm of the second visceral arch. It has not only increased considerably in thickness, but has become much compacted. The individual cells have consequently lost their stellate shape, and as a result resemble very closely the cells of the ganglionic anlage. Ten-somite embryos show the arch mesoderm in stages of compaction intermediate between the 9- and 11-somite embryos. The acoustico-facial anlage has changed but slightly. It is still a fusiform mass extending ventrally to the level of the dorsum of the first gill pouch, which extends dorsally to the level of the ventral surface of the neural tube. In spite of the similarity between the compacted mesodermal cells of the hyoid arch and those of the ganglionic mass, the ventral limit of the acoustico—facial anlage can be approximately fixed. The long axis of the ganglion is oblique to the axis of the hyoid arch, so that there is a medial notch where the ganglion comes into contact with the arch mesoderm. The notch lies at the upper level of the dorsum of the first gill pouch, to which height the mesoderm ca11 be seen to extend in sections just in front of or caudal to the ganglion (fig. 46), and it will be recalled that the ganglion extends ventrally to that level in 9- and 10—somite embryos. A study of 10somite embryos yields convincing evidence that the ganglionic cells take no part in the formation of the mesodermal core of the branchial arch.
inner surface of the epibranchial placode of the VII ganglion
 
than i11 these embryos of 30 to 34 somites and beyond. Mesenchymal cells are at times found closely applied to the placode, but there is no indication of their being proliferated
 
from it. At the place of fusion of ganglionic ramus and
 
placode there is, of course, some doubt. The branch itself
 
is, I am confident, of ganglionic origin. VVhether cells pass
 
into the ganglion from the placode by way of this branch it
 
is impossible to say. There is, however, no direct evidence
 
NEURAL FOLDS AND CRANIAL GANGLIA or BAT 83
 
  
of it. Surprisingly few mitoses are present in the placode,
 
and these are superficial with spindle axes parallel to the
 
surface when determinable. After 31 somites the epibranchial placodal of the facial rapidly thins out and is soon not
 
to be separated from the surrounding ectoderm.
 
  
In the case of the acoustico-facial, as in the case of the
+
A very thin lamina of looser cells (probably mesenchymal) separates the upper portion of the ganglion from the ectoderm in the 11-somite embryo. It can be recognized by its more diffuse character and tl1e presence of a few blood cells. The blood cells have deeply staining nuclei in the figure.
trigeminus, there is a well-defined, spur—like downgrowth
 
from the base of the ganglion along the side of the neural tube.
 
It is, however, more pronounced in the case of the acousticofacial than the trigeminus. This spur from the acousticofacial ganglion first becomes well defined in embryos of 17
 
and 18 somites. Like that of the fifth nerve, it seems to be
 
derived from the main ganglionic mass by a ventral growth
 
of cells along the side of the neural tube, and one soon recognizes a median ganglionic mass of which the spur is a projection as distinguished from the more lateral mass. It is well
 
marked at 27 somites in figure 62, and is still well defined at
 
31 somites. At 34 somites (fig. 63) it constitutes a well~
 
defined compact mass lying just ventral to the place of attachment of the acoustico-facial ganglion to the neural tube. At
 
its ventral extremity one now finds a well-developed tuft of
 
motor fibers growing out from the neural tube. As time goes
 
on, the spur becomes more and more completely invaded,
 
until at thirteen days (fig. 64) it is only a tiny mass crowded
 
between the now greatly enlarged motor root and the base of
 
the ganglion. At thirteen days and four hours (fig. 65) it has
 
disappeared, its place having been taken by the enlarged
 
motor root.
 
  
In the case of both the trigeminus and the acoustico-facial,
 
then, the motor fibers seem to grow out into the substance
 
of the spur, finally obliterating it. The fate of its cells is
 
a matter of some interest. Harrison (’06, ’24) has shown
 
very clearly for the spinal nerves that the sheath cells of the
 
motor root are derived from the ganglionic crest in Amphibia,
 
and he has also found (’24) that “after removal of the ganglion crest and placodes of the head, cranial motor nerves
 
without sheath cells may develop, but iii the case of these
 
84 HOWARD B. ADELMANN
 
  
medullary cells evidently begin to migrate along the ventral
+
The otic placode becomes thicker, but the invagination has progressed but slightly between 9 and 11 somites.
roots at a relatively early period, giving rise to some sheath
 
cells, even though the other sources are removed.
 
  
According to Neal (’14), the neural crest seems to be the
 
chief source of origin for the sheath cells in reptiles, birds,
 
and mammals. In Selachii, however, the sheath cells have
 
been recorded as being exclusively medullary in origin.
 
  
In the spinal region of the rat (fig. 66) the ganglionic crest
+
In the interval between 12 and 14 somites (figs. 47 to 50), the acoustico—facial anlago keeps pace in growth with the embryo and its relations to surrounding parts remain much the same as in earlier stages. During this interval the neural tube closes in the region of the otic rhombomere and the figures show the attendant ventral shifting of the attachment of the ganglion to the neural tube. The fusion of the lips of the neural folds is not very intimate, so that they separate somewhat in the 14-somite embryo as a result of handling. At 12 somites the dorsal end of the ganglion is in contact medially with the dorsal surface of the closing neural tube, lying between it and the overlying ectoderm (fig. 47), but at 14 somites it has shifted ventrally, so that it is now in contact with the middle third of the neural tube.
forms a spindle-shaped mass closely adherent to the sides of
 
the neural tube. The apex of the ganglionic mass comes in
 
contact with the developing motor fibers as soon as they
 
leave the neural tube, probably, according to the most reliable evidence of Harrison, supplying these fibers with sheath
 
cells. The ganglionic spur described in connection with the
 
anlagen of the trigeminus and acoustico-facial ganglia bears
 
a relationship to the developing motor roots of those ganglia
 
which is entirely comparable to the relation of the ganglionic
 
crest to the motor root in the case of the spinal nerve, and the
 
suggestion is strong that the spurs, invaded by the developing
 
motor fibers, furnish the neurilemma cells of the motor portions of these cranial nerves. It is significant that the anlagen of the IX-X nerves develop no such spurs.
 
  
The author does 11ot wish to infer the homology of the spur
 
with the spinal ganglionic crest in its entirety, nor does he
 
consider the sp11r and the main ganglionic crest as representing different ‘orders’ of crest. The entire cranial ganglion
 
plus the spur are probably the equivalent of the spinal neural
 
crest.
 
  
The acoustic gawngltirm. The present study was directed
+
The compact mesoderm of the hyoid arch still extends dorsally to the upper level of the first gill pouch, i.e., approximately to the level of the ventral surface of the neural tube (fig. 47 and cf. the right and left sides of figs. 48, 50), where it comes into contact with the ventral extremity of the ganglion. The long axis of the acoustico-facial ganglionic mass is even more markedly oblique to the axis of the hyoid arch than previously, and consequently there is a more pronounced medial notch at the boundary between the two structures than in younger embryos. Furthermore, one notices a distinct difference in the arrangement of the cells of the mesoderm and the ganglionic anlage, more pronounced in some specimens than in others. In some embryos the differentiation of ganglionic and mesodermal tissues is particularly striking (for instance, figs. 48, 49). ‘In embryos of 12 to 14 somites, therefore, one is justified, I believe, in placing the ventral limit of the ganglionic anlage at a point corresponding approximately to the dorsal level of the first pharyngeal pouch, which is situated just anterior to the acoustico-facial anlage. The exact level varies but slightly. It is impossible to determine whether there is an intermingling of ganglionic elements with the mesoderm at the place of contact.
mainly toward a solution of the problem of the origin of the
 
acoustic ganglionic mass, and no special attention was paid
 
to its subsequent differentiation-—-a subject to which Streeter
 
(’06) has already devoted a special paper. Concerning the
 
origin of the ganglion cells of the acoustic nerve in man,
 
Streeter (’12) says: “They are evidently not derived from
 
neural crest; but Whether they migrate out from the brain
 
wall or the walls of the developing ear vesicle, or are derived
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
from the ectoderm immediately adjacent to the auditory pit,
+
Figures 47 to 50 also show the gradual separation of the acoustico—facial anlage from the overlying ectoderm by the invasion of mesenchyme during this period. The separation of the ganglion from the ectoderm is more marked dorsally than ventrally where the ganglion approaches the surface more closely. In figures 47, 48, the separation appears more marked on the right than on the left, but that is because the plane passes more caudally on the left, passing immediately (fig. 48) in front of the otic-pit against which the ganglion rests. The integrity of the inner surface of the epithelium both in relation to the ganglion and the branchial arch is unquestionable. Frontal sections of the ganglion in the 14somite embryo show it as a perfectly discrete cell column (fig. 53) separated widely from the ectoderm dorsally, but approaching it ventrally. A new venous channel, the vena capitis lateralis, has been formed in the mesenchyme lateral to the ganglion between 12 and 14,somites. In the 12-somite embryo sprouts from the primitive head vein are beginning to work their way around the ganglionic mass, and the process is completed at 13 somites.
remains to be determined.
 
  
In the rat the last possibility is ruled out by a study of
 
frontal sections of the acoustico-facial mass. Until the V11somite stage the acoustico-facial anlage does lie in close
 
contact with the otic placode, but at this time one can find
 
no evidence of ectodermal proliferation. After .12 somites
 
the ganglion is a perfectly discrete cell column separated
 
from the prominent anterior lip of the otic pit and from the
 
anterior wall of the pit itself by considerable mesenchyme.
 
It never subsequently comes into contact with the epithelium
 
overlying the otic vesicle.
 
  
There was no evidence in the rat of a proliferation of
+
Figure 51 is a section through the ventral tip of the acoustico-facial anlage of an 18-somite embryo. Its attachme11t to the neural tube is shown in more caudal sections. Figure 51 is the third section caudal to the first branchial pouch. A study of it shows that while the embryo has grown markedly there has been no change in the relative positions of the parts concerned. The ventral extremity of the ganglion can be easily determined in figure 51, resting directly upon the mesodermal core of the hyoid arch. It can be traced forward to the section immediately caudal to the first branchial pouch where it makes contact with the epibranchial placode, but there is no continuity between the two structures. A line drawn laterally from the ventral surface of the neural tube coincides approximately with the ventral end of the anlage and serves as a convenient reference line for comparison with younger embryos. Figure 52, passing through the attachment of the VILVIII ganglion to the brain, three sections caudal to figure 531, shows that the compact core of the hyoid arch has the same relative dorsal extent as in younger embryos. Study of the series shows that the upper limit of the mesodermal core of the hyoid arch lies at the level of the dorsal surface of the first branchial pouch— a fact well illustrated also in figure 67, a sagittal section of an embryo of 19 somites. It will be appreciated from the above description that the ventral extremity of the acousticofacial anlage of the 18-somite embryo lies in the same relative position as in the younger stages—namely, immediately caudal to the dorsal end of the first pouch.
acoustic ganglion cells from the wall of the otic vesicle. From
 
12 to 18 somites the acoustico-facial ganglion is distinctly
 
separated from the anterior wall of the otic pit or vesicle by
 
mesenchyme, but shortly thereafter there is evidence of the
 
migration of cells from the caudal face of the ganglion, giving
 
rise to a mass of cells closely applied to the anterior wall
 
of the otic vesicle and not at first separated from the facial
 
ganglion. A series of photographs is presented to show the
 
formation of the acusticus. At 19 somites (fig. 67) there is
 
still no evidence of the presence of an acoustic ganglion. In
 
the 21-somite embryo there is only a small caudal projection
 
of the ganglion toward the otic vesicle, but in a 24—somite
 
embryo (fig. 68) there is a well-developed mass of cells between the facial mass proper and the otic vesicle. To what
 
extent this compact mass is composed of mesenchymal cells
 
could not be determined. The photograph does not give an
 
adequate idea of its size, since in other sections it could be
 
followed along the anterior face of the otic vesicle and to a
 
slight extent along its ventral surface.
 
  
At 27 somites (fig. 69) the proliferation of ganglion cells
 
from its caudal face has brought the acoustico-facial mass
 
into close contact with the anterior face of the otic vesicle.
 
At 28 to 29 somites (fig. 70) the two components of the ganglion are distinct. A proliferation of cells of the acoustic
 
86 HOWARD B. ADELMANN
 
  
ganglion from the walls of the otic vesicle could not be observed, since all through the period of formation of the acoustic mass the walls of the otic vesicle are absolutely intact. I
+
The most important development of this’ stage is the placodal thickening of the ectoderm at the upper level of the arch. ‘It is illustrated in figures 51 and 52, and no doubt is identical with Froriep’s epibranchial sense organ. The thickening begins immediately caudal a11d dorsal to the first cleft and extends caudally and dorsally as far as the anterior border of the otic vesicle. The extreme ventral end of the acoustico—facial anlage rests lightly against it, but there is no cellular continuity between the placode and the ganglion; the inner surface of the placode is everywhere intact. Four mitoses, tl1e planes of which were not determinable, were found in the entire left placode which may be traced through fifteen sections. It is evidently not growing very actively. A few cells could be found in it which showed unmistakable evidence of degeneration, and these degenerating cells increase greatly in number as time goes on.
therefore conclude that the acoustic mass is a derivative of
 
the main ganglionic mass and hence derived from the neural
 
crest.
 
  
In summarizing the development of the acoustico-facial
 
anlage, the following facts may be emphasized. The anlage,
 
a mass of neural crest attached to the otic rhombomere, grows
 
ventrally between the paraxial mesoderm and the overlying
 
thickened ectoderm of the otic placode with which it is in
 
intimate contact, but from which it receives no contribution
 
of cells. The ganglionic mass quickly reacl1es its ventral limit
 
(10 to 11 somites) at the upper level of the first pouch where
 
it rests against the condensed Inesenchyme of the second
 
arch. The condensation of arch mesenchyme occurs independently of neural crest or other ectodermal contributions
 
and no migration of neural-crest elements into the arch could
 
be observed. The ganglion gradually becomes separated from
 
the overlying ectoderm by the invasion of mesenchyme lateral to it, but its ventral extremity maintains contact with
 
the ectoderm just behind the dorsal extremity of the first
 
branchial pouch, where at 16 to 18 somites a pronounced
 
thiel<ening marks the site of the epibranchial. placode.
 
  
The contact of the ganglion with the ectoderm above and
+
From the base of the ganglion where it comes into contact with the neural tube, one finds a spur of cells which seems to be growing ventrally from the base of the ganglion, keeping in close contact with the neural tube. 'Tl1is spur of ganglionic cells is small at 18 somites, but is rather prominent at 21 somites (fig. 54). Its possible fate will be discussed later.
behind the first pouch is, then, of long standing, existing from
 
the time the ganglion reaches its ventral limit at 9 to 10
 
somites, and the relative position of the ventral extremity
 
of the acoustico-facial anlage remains the same until by
 
growth the anlage comes to be separated from the mesenchymal condensation of the arch (of. figs. 67 to 70). The
 
ganglionic and mesodermal tissues are always tolerably well
 
differentiated, so that it is not difiicult to ascertain the boundary between them.
 
  
At about 29 to 31 somites the expansion of the hyoid arch
 
results in the separation of even the ventral extremity of the
 
ganglion from the ectoderm, but, as the ganglion recedes,
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
it maintains its placodal contact by the drawing out of a cord
+
The changes are slight in 19- and 20-somite embryos. In specimens of these ages the epibranchial placode is still extensive, reaching caudally as far as the otic vesicle. The acoustico-facial anlage extends laterally as it proceeds ventrally, its ventral extremity just touching the epibranchial placode, but not fusing with it.
of ganglionic cells extending from the lower pole of the ganglion to the ectoderm which persists until shortly after the
 
thirteen—day stage. There is no doubt in the writer’s mind
 
that this placodal ramus is ganglionic in origin.
 
  
In determining the relation of ganglion and placode one
 
must rely on good transections through the long axis of the
 
ganglion or on frontal sections. The latter, however, are
 
not favorable for determining the ventral extremity of the
 
ganglion, but allow one to judge of the integrity of the inner
 
surface of the placode. Oblique sections of any kind are
 
apt to be misleading and the curvature of the arch surface
 
must be kept in mind even in the interpretation of good
 
transections. While the lower extremity of the ganglionic
 
mass is almost from the beginning in co11tact with the pla
 
codal ectoderm, no intimate fusion of parts appears to occur
 
  
until about 31 somites. It is impossible to determine whether
+
A series of sections through the acoustieo—facial anlage of a 21-somite embryo are given in figures 54 to 56. Figure 56 is a section through the caudal edge of the ganglion. It shows the anterior surface of the otic vesicle. The attachment of the ganglion to the brain is shown, but its ventral extremity is not included in the section. The figure illustrates the dorsal extent of the arch mesoderm, which is relatively the same as in younger embryos. Figure 55 is five sections anterior, revealing the full extent of the ganglion, whose ventral extremity approaches the placode. Three sections farther anteriorly (fig. 54) the ventral tip of the ganglion comes to lie in the angle between the epibranchial placode a11d the first branchial pouch, and in this embryo there is only a very restricted contact, but no fusion of ganglion and placode at this point. The ventralspur extending from the base of the ganglion along the side of the neural tube is well shown.
there is any migration of cells from the placode to the ganglion by way of the placodal ramus, but appearances do not
 
favor such an interpretation.
 
  
3. The g£o.9sophary/ngcal and vagus ganglia
+
In this,21-somite embryo the epibranchial placode is no less extensive than in the 18- and 19-somite embryos described.  
  
So far as the writer is aware, the early history of the
 
glossopharyngeal and vagus ganglia has never been completely followed in the mammal. The work of Martin (’91)
 
on the IX-XII nerves of the cat and the excellent studies of
 
Chiarugi (’89, ’90, ’94) on the rabbit and guinea—pig, and
 
of Streeter (’O4) on the human embryo deal with later phases
 
of their development.
 
  
The IX—X anlage is first discernible i11 the rat of 8 somites.
+
It has not thickened appreciably and no evidences of proliferative activity could be discerned. Its inner surface seems intact, and while mesenchymal cells lie close against it, there seemed to be no continuity between the two. Sagittal and frontal sections of 21-somite embryos confirm the evidence of the transverse sections as to the extent and relations of the acoustico-facial anlage. At 21 somites a few degenerating cells and phagocytes may be observed in the dorsal part of the mesoderm of the second arch. They are not numerous and seem to lie chiefly caudal to the ganglion.
It has been described previously as being at that stage the
 
least advanced in development of the anlagen of the cranial
 
ganglia. Attention may again be called to figure 28, which
 
is a section through the IX-X anlage of the 8—somite embryo.
 
At this early period, active proliferation has not yet begun
 
and the anlage consists simply of a conical mass of cells fitted
 
between the lateral edge of the neural plate and the ectoderm.
 
88 HOWARD B. ADELMANN
 
  
Externally, there is a shallow groove on each side marking
 
the boundary between the neural plate and the ganglionic
 
anlage. As yet the ectoderm cannot be traced over the ganglionic anlage to the edge of the neural plate. The IX-X anlarge may be traced caudally to the anterior margin of the
 
first somite, where it becomes directly continuous with the
 
actively proliferating spinal neural crest. The relations of
 
the vas primitivum rhombencephali (Sabin) should be noted.
 
It lies just below the IX-X anlage in close proximity to the
 
neural fold.
 
  
At 9 somites (fig. 71) the proliferation of the IX—X anlage
+
Three sections through the acoustico-facial anlage of a 26» somite embryo are presented in figures 57 to 59. Figure 57 is a section through the tip of the ganglion which extends into the angle between the dorsal surface of the first branchial pouch and the epibranchial placode, where it makes contact with the placode, but does not fuse with it. The relation of ganglion to placode is well shown in figure 59—a section through the VII-VIII ganglion on the opposite side of the same embryo. The contact of ganglion and placode extends through only two sections on both sides of the embryo. Elsewhere ganglion and placode are separated by mesenchyme which is closely applied to the placode. In sections immediately caudal to the pouch where the curvature of the arch is steepest and in which consequently the placodes are most obliquely cut, this close adhesion of mesenchyme and placode may be interpreted as proliferative activity on the part of the placode. Figures 57, 58 may serve as an illustration of this. It will be observed that the obliquely cut wall of the adjacent pouch shows an epithelio—mesenchymal relationship similar to the placode, so that one must grant an entodermal proliferation of mesenchyme if the placode is interpreted as proliferating. In this connection, one recalls Weigner’s comment: “ . . . . die Beurtheilung eines Zusammenhanges gerade an dieser Stelle mit fast uniiberwindlichen Schwierigkeiten zu kampfen hat. Es ist eben jeder der Schlundbogen nach allen Richtungen convex und wird deshalb immer seine Epithel schief getroffen. Liegt nun eine Zellgruppe diesem Epithel nahe so kiinnen sehr leicht solche Schiefschnitte eine Verbindung vortéiuschen auch an Stellen, an denen keine existiert.” However, the difficulty is solved when frontal sections of embryos of the same age are consulted. These show no evidence of proliferation from the placodes, although mesenchyme may be in close contact with them.
has begun, the cells migrating laterally and ventrally to form
 
a compact mass lying lateral to the vas primitivum rhombencephali and in contact with the overlying ectoderm which
 
extends over the neural crest to the edge of the neural fold.
 
The thickness of the ectoderm overlying the anlage is exaggerated in figure 28, due to the curvature of the embryo. It
 
is impossible to fix the exact ventral limit of the crest. In
 
some sections it appears to extend to the ventral border of
 
the primitive rhombencephalic vessel. The extent of the IX-X
 
anlage of the 9-somite embryo is somewhat greater than at
 
8 somites, due to the expansion of the territory between the
 
anterior margin of the first somite and the otic placode. As
 
one traces the IX-X crest caudally it becomes somewhat more
 
diffuse, but maintains its position lateral to the primitive
 
rhombencephalic vessel. At the anterior boundary of the
 
first somite the vas primitivum rhombencephali divides into
 
two slenderer branches; one passes medial to the somite and
 
the other ventro-laterally into the somatopleure. The crest
 
proliferation for the IX-X ganglia accompanies the medial
 
branch of the vessel and becomes directly continuous with
 
the spinal neural crest medial to the somite. A
 
  
In figures 72 to 74 a series of photographs of an embryo in
 
which the twelfth somite is just forming is presented to
 
illustrate the essential features of the IX-X anlage at this
 
stage. Figure 72 lies three sections caudal to the otic pit on
 
the left. At this level the crest is a definite compact mass
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
lying lateral to the Vas primitivum rhombencephali. The
+
In some of the eight 26-somite embryos examined, a number of degenerating cells were found in the placodal epithelium, giving evidence of early retrogressive change. Mitoses are relatively few, those found are superficial in position, with the spindle axis parallel to the free surface when determinable. Considerable numbers of degenerating cells and phagocytes are found in the mesenchyme near the dorsal end of the second arch in these embryos.
primitive rhombencephalic vein makes an abrupt Ventro—
 
lateral bend just in front of the first somite to make its way
 
into the somatopleure. A branch passing medial to the somite
 
could not be found. Part of its course may be followed in
 
the figures. As the crest is traced caudally it becomes somewhat more diffuse (fig. 73) and, passing around the caudal
 
border of the Vas primitivum rhombencephali (anterior cardinal vein?) as the latter bends ventrally, becomes continuous
 
with the spinal neural crest. Figure 74 is a section just
 
caudal to the upper end of the anterior cardinal Vein showing
 
the place of transition between the IX-X and the spinal
 
neural crest. The diffuse character of the IX-X crest at
 
most levels makes it impossible to determine accurately its
 
ventral limit.
 
  
The IX—X anlage of '13—somite embryos is Very similar to
 
that of 11- and 12-somite specimens, but one finds some
 
changes in the relation of the crest to the venous channels of
 
the head. In the l3—somite embryo the anterior portion of
 
the postotic neural crest in the region of the future IX proliferation is just beginning to be separated from the overlying ectoderm by the establishment of a new Venous channel
 
lateral to it, the beginnings of which are illustrated in figure 75. One finds many places where the neural crest seems
 
to have migrated ventrally medial to the Vas primitivum
 
rhombencephali as well as lateral to it, so that in some sections (of. the left side of fig. 76) that Vessel appears to be
 
surrounded by neural-crest cells. The relations of the posterior portion of the IX-Y crest remainas described for the
 
IX-X crest of earlier embryos. The anterior cardinal Vein
 
is cut longitudinally on the right side of figure 76, which
 
shows the neural crest lateral to that Vessel. The neural
 
crest passes around the caudal wall of the Vein to become
 
continuous with the neural crest related to the first somite.
 
  
Sections just caudal to the otic pit of the 14-somite embryo
+
The acoustico-facial anlage shows no essential change in eight 27 —somite embryos studied. In 29- and 30-somite embryos the anlage is somewhat larger, and for the first time a few fibers can be seen extending from the ganglion for a short distance into the hyoid arch.
(fig. 77) show that the new Venous channel lateral to the anterior portion of the IX—X anlage is now well established;
 
90 HOWARD B. AIJELMANN
 
  
the old vessel is probably represented by a more or less collapsed channel embedded in the substance of the neural crest,
 
but which communicates (fig. 78) frequently by anastomoses
 
with the new vena capitis lateralis. The latter tends to lie
 
more and more ventral to the crest (fig. 79) as it is traced
 
caudally. I was unable to determine whether a new venous
 
channel has been established in the caudal region of the IX-X
 
anlage or whether there has simply been some ventral and
 
lateral shifting of the old vas primitivum rhombencephali in
 
this region. The limits of the neural crest constituting the
 
IX—X anlage at this stage are better defined for a short distance caudal to the ear vesicle in the region of the prospective
 
IX downgrowth where the crest is more compact and abundant than in the more caudal portion. In the territory of
 
the future IX proliferation the crest lies medial to the vena
 
capitis lateralis, but as the sections approach the first somite
 
the crest becomes Very diffuse and shifts dorsally somewhat
 
so that it tends to lie above the lateral head vein and closer
 
to the overlying ectoderm. In this region, on account of its
 
diffuse character, it becomes impossible to accurately delimit
 
the crest from the mesenchyme (fig. 79), but careful study
 
shows it to be continuous with the spinal crest at the level of
 
the first somite (fig. 80). In embryos of this age the relations
 
of the IX-X anlage to the early venous channels of the head
 
are extremely variable, as one would expect from the plastic
 
nature of the neural crest and the endothelial tubes, so that
 
the description given above applies in detail only to the right
 
side of the 14:—somite embryo here illustrated.
 
  
In a 16-somite embryo the IX-X anlage is widely separated
+
A 30-somite embryo, favorably sectioned, shows interesting developments. The ganglion has been still further separated from the ectoderm by the increase of the intervening mesenchyme. The ventral extremity of the ganglionic anlage is now perfectly distinct, and from it two branches may be followed. A short cellular cord proceeds from the ventral extremity of the ganglion laterally to the placode to which it adheres, but there still seems to be no intimate fusion of the two tissues. This placodal ramus is formed, I believe, during the separation of the ganglion from the ectoderm by the expansion of the hyoid arch. The original contact of the ventral extremity of the ganglion and the placode persisting, the recession of the ganglionic mass from the ectoderm results in the ‘drawing out’ of a cord of cells leading from the ventral extremity of the ganglion to the site of original contact with the placode. The second branch of the ganglion can be followed into the hyoid arch for some distance.
from the ectoderm by diffuse mesenchyme in which the now
 
rather capacious vena capitis lateralis courses. Beginning
 
immediately caudal to the otic vesicle, the IX-X crest forms
 
a lens—shaped mass closely applied to the lateral surface of
 
the neural tube. It is largest in its anterior portion and
 
thins out, becoming more diffuse as it proceeds caudally. An»
 
teriorly, the anlage extends ventrally to the middle of the
 
lateral wall of the neural tube, but does not reach the midline
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
dorsally. It shifts dorsally, however, as it is traced caudally,
 
and in its more posterior regions a thin li11e of crest cells
 
overlies the dorsal surface of the neural tube. In some sections through the caudal portion of the IX-X anlage the crest
 
is so diffuse that it is impossible to be positive of its presence.
 
However, such appearances are confined to single sections
 
and one has no difficulty in determining the continuity of the
 
crest of the IX and X ganglia with the spinal neural crest. It
 
is quite probable that such isolated sections where the crest
 
appears to be absent are due to localized rarefactions of the
 
crest of no significance. In later stages the caudal portion
 
of the IX-X crest is more compact and there is never any
 
doubt of its continuity with the spinal crest.
 
  
Wliile the IX—X crest is more extensive, more compact, and
+
A similar but somewhat stouter placodal ramus may be identified in embryos of 31 a11d 34 somites (figs. 60, 61). Its caliber varies considerably, but its point of attachment to the placode is remarkably constant in position, namely, just above and behind the first branchial pouch in the angle between pouch and placode. This point, it will be recalled, has been described as marking the ventral limit of the ganglion in earlier stages. The relations just described are well shown in figure 61, which is a frontal section of a 34-somite embryo. In this embryo three terminal branches of the ganglion may be followed—one passes antero—medially between the dorsal aorta a11d the pharynx (the greater superficial petrosal nerve), one continues ventrally into the hyoid arch, and the third is the placodal ramus above described.
more prominent just caudal to the otic pit than in more
 
posterior portions, there is as yet no more definite indication
 
of the position of the anlage of the glossopharyngeus. A
 
definite ventral extension of the IX-X crest marking the
 
position of the ninth nerve is first observed in embryos of 17
 
and 18 somites. In the 18-somite embryo (fig. 18), three
 
sections caudal to the ear, a slender cord of cells, visible in
 
two sections 7.5 u thick, extends ventrally from the crest as
 
far as the ventral border of the vena capitis lateralis, where
 
it comes into contact with the mcsenchymal condensation of
 
the third branchial arch. It lies medial to the vena capitis
 
lateralis, a position which it retains.
 
  
A series of sections through the IX-X anlage of a 21-somite
 
embryo are presented in figures 81 and 82. The section illustrated in figure 81 lies eight sections caudal to the second
 
gill pouch and five sections caudal to the otic vesicle. Figure 82 passes through the region of the ultimobranchial body.
 
The IX anlage shown in figure 81 is a compact cord of crest
 
cells extending through four 7.5 p sections. It lies medial
 
to the vena capitis lateralis, to the lower border of which
 
the ganglionic anlage extends. The anlage comes intocontact with the mesenchymal condensation of the third arch,
 
where it seems to be insinuating itself between the compact
 
92 HOWARD B. ADELMANN
 
  
arch mesenchyme and the lower border of the lateral head
+
The placodal branch of the geniculate ganglion is still well developed in some embryos of thirteen days, but at this time it seems to be in process of disappearance, since in some specimens of this age it can be recognized upon one side only and in others it can be recognized with difficulty. In some it is still prominent.
vein. There is a slight thickening of the ectoderm above and
 
behind the second gill pouch, which probably represents the
 
epibranchial placode of the ninth nerve, but the thickening
 
is not marked nor is it extensive. The IX anlage does not
 
come into contact with it. The anlage of the glossopl1aryngeus is continuous caudally with the lens-shaped mass of
 
neural-crest cells (fig. 82) that finally becomes continuous
 
with the spinal crest. 8 It should be noted that at this stage the
 
condensed arch mesenchyme caudal to the third pouch extends
 
dorsally as far as the ventral border of the lateral head vein,
 
and in more caudal sections it extends still farther dorsally
 
around the lateral border of the vena capitis lateralis.
 
  
Four sections caudal to figure 82 as few neural-crest cells
 
appeared to be moving ventrally over the lateral surface of
 
the vena _capitis lateralis, probably the earliest indication of
 
the vagus anlage. Twenty—four—somite embryos were the first
 
to show a definite vagal proliferation. In the 24-somite embryo there are two well-defined downgrowths from the postotic neural crest—-—the glossopharyngeal and the vagal. The
 
former lies medial to the lateral head vein, while the latter is
 
situated lateral to it. The glossopharyngeal proliferation has
 
increased considerably in bulk, its ventral extremity is somewhat enlarged and seems to be insinuating itself between the
 
lower surface of the lateral head vein and the compact arch
 
mesoderm. The mesenchyme above the level of the dorsal
 
aorta shows numerous degenerating cells, and subsequent
 
stages show that the compact mesenchyme disappears in large
 
part from this situation (cf. figs. 83, 89). The vagus (fig. 83),
 
proliferating lateral to the vena capitis lateralis, also seems
 
to be insinuating itself between the lateral head vein and the
 
compacted arch mesenchyme, but in this region no degenerating cells are found in the mesenchyme at this stage. It
 
is almost impossible to determine the exact limits of the
 
vagal anlage. I11 the figure of the 24-somite embryo here
 
reproduced, it seems to be easily distinguishable from the
 
mesenchyme, but such is not the case in other embryos of the
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 93
 
  
same age. It is also difficult to determine the caudal extent
+
In an embryo of thirteen days four hours there is no longer any trace of an epibranchial connection of the genicu— late ganglion. On one side the placodal region takes part in the formation of a short ductus branchialis I with a very small lumen. On the other side the adjacent surfaces of the first cleft have fused, forming a short cord of cells leading from the pouch to the surface, but there is no lumen.
of the ventral extremity of the vagus, because just caudal to
 
the vagal proliferation the compact mesenchyme extends dorsally lateral to the vena capitis lateralis. In the interval
 
between the IX and X proliferations there is a short extent
 
of neural crest where the proliferation is less active, but
 
nevertheless occurs with sulficient intensity to allow one to
 
trace crest cells ventrally along the inner side of the vena
 
capitis lateralis; the cells so proliferated form in later stages
 
a definite cord of ganglionic cells extending from the glossepl1a.1'y11gea.l to the vagal anlage.
 
  
About twenty-five embryos cut in various planes were
 
available for the study of the neural crest of the glossepharyngeus a11d vagus in the period between l.'W€11t}"—fiV€ and
 
tl1irty—one somites. Great variatiionwas noted in embryos
 
of the same age and many features which could be clearly
 
observed in some embryos were not so obvious in others.
 
For instance, in some embryos it was possible to delimit the
 
ganglionic anlage easily from the adjacent mesenchyme, in
 
others such a delimitation was not possible, due to individual
 
variations in the compactness of the surrounding mesenchyme
 
or of the ganglionic anlage itself. The following descriptions
 
must therefore apply in fine detail only to the embryos illustrated; attention will be called to important individual deviations.
 
  
Five pliotograplis of sections of a 26—somite embryo are
+
One could never be surer of the perfect integrity of the inner surface of the epibranchial placode of the VII ganglion than i11 these embryos of 30 to 34 somites and beyond. Mesenchymal cells are at times found closely applied to the placode, but there is no indication of their being proliferated from it. At the place of fusion of ganglionic ramus and placode there is, of course, some doubt. The branch itself is, I am confident, of ganglionic origin. VVhether cells pass into the ganglion from the placode by way of this branch it is impossible to say. There is, however, no direct evidence of it. Surprisingly few mitoses are present in the placode, and these are superficial with spindle axes parallel to the surface when determinable. After 31 somites the epibranchial placodal of the facial rapidly thins out and is soon not to be separated from the surrounding ectoderm.
reproduced in figures 84 to 88. The first section passes
 
through the caudal wall of the second branchial pouch and
 
is the most anterior section through the glossopharyngeal
 
anlage. A stout cord of ganglionic crest cells constituting the
 
IX anlage could be traced through nine sections of 7.5 u.
 
Caudal to that point the neural crest becomes very attenuate
 
in its middle portion, so that extending between the glossepharyngeal and vagal proliferations there are two cords of
 
neu1‘al-crest cells (fig. 87), a dorsal one lying close against
 
the lateral surface of the neural tube and a ventral less
 
compact cord of crest cells which lies just medial to the vena
 
  
THE JOURNAL or C()1\[I’ARA'l.‘I\'}£ NEUROLOGY, vor.-. 39, N0. 1
 
9-1 HOWARD B. ADELMANN
 
  
capitis lateralis and just dorsal to the dorsal aorta. As this
+
In the case of the acoustico-facial, as in the case of the trigeminus, there is a well-defined, spur—like downgrowth from the base of the ganglion along the side of the neural tube. It is, however, more pronounced in the case of the acousticofacial than the trigeminus. This spur from the acousticofacial ganglion first becomes well defined in embryos of 17 and 18 somites. Like that of the fifth nerve, it seems to be derived from the main ganglionic mass by a ventral growth of cells along the side of the neural tube, and one soon recognizes a median ganglionic mass of which the spur is a projection as distinguished from the more lateral mass. It is well marked at 27 somites in figure 62, and is still well defined at 31 somites. At 34 somites (fig. 63) it constitutes a well~ defined compact mass lying just ventral to the place of attachment of the acoustico-facial ganglion to the neural tube. At its ventral extremity one now finds a well-developed tuft of motor fibers growing out from the neural tube. As time goes on, the spur becomes more and more completely invaded, until at thirteen days (fig. 64) it is only a tiny mass crowded between the now greatly enlarged motor root and the base of the ganglion. At thirteen days and four hours (fig. 65) it has disappeared, its place having been taken by the enlarged motor root.
latter cord is traced caudally it passes laterally beneath the
 
vena capitis lateralis to join the ventral extremity of the
 
vagus anlage. There are thus two communicating strands,
 
a dorsal and a ventral, between the ganglionic anlagen of the
 
IX and X nerves at this time (of. fig. 19). The dorsal and
 
ventral communicating cords owe their origin to a continuation of crest proliferation caudal to the IX a11lage. A study
 
of embryos of 24 to 27 somites confirms fully this statement.
 
In some embryos there is a profuse proliferation of crest
 
cells medial to the vena capitis lateralis extending as far
 
caudally as the vagal proliferation, while in others one finds
 
the two communicating cords as described above with a delicate strand of crest cells extending between them. I have
 
therefore been drawn to the conclusion that this ventral communicating ramus is formed by the piling up of crest cells
 
medial to the vena capitis lateralis in the region between
 
the IX and X proliferations. As time goes on, the dorsal
 
communicating ramus becomes somewhat reduced in size.
 
In this embryo many degenerating cells were found in it,
 
possibly some degeneration being a factor in its reduction
 
in size. The cells of the ventral communicating ramus are
 
somewhat diffuse and in some embryos of 26 to 27 somites
 
it could not be traced with certainty as far caudally as the
 
vagus.
 
  
In this particular embryo it is somewhat difficult to fix the
 
ventral limit of the glossopharyngeal anlage in the first few
 
sections caudal to the second pouch. However, careful study
 
of several embryos of the same period as well as the evidence
 
presented by this embryo makes me feel fairly certain in
 
fixing its ventral limit at about the dorsal boundary of the
 
dorsal aorta. In the first place, reference to figures of 21~
 
somite embryos will show that the dorsal limit of the compact arch mesoderm of the third arch lies slightly above the
 
dorsal border of the dorsal aorta. In some embryos in which
 
the ventral end of the IX anlage is less compact than the
 
arch mesoderm one l1as no difficulty in distinguishing the
 
NEURAL FOLDS AND OB-ANIAL GANGLIA or RAT 95
 
  
limits of the two tissues. The same is true also of embryo
+
In the case of both the trigeminus and the acoustico-facial, then, the motor fibers seem to grow out into the substance of the spur, finally obliterating it. The fate of its cells is a matter of some interest. Harrison (’06, ’24) has shown very clearly for the spinal nerves that the sheath cells of the motor root are derived from the ganglionic crest in Amphibia, and he has also found (’24) that “after removal of the ganglion crest and placodes of the head, cranial motor nerves without sheath cells may develop, but iii the case of these medullary cells evidently begin to migrate along the ventral roots at a relatively early period, giving rise to some sheath cells, even though the other sources are removed.
figured in sections through slightly more caudal portions of
 
the glossopharyngeal anlage. Secondly, one finds in embryos
 
of 25 to 27 somites a large number of degenerating cells at
 
the boundary between the two tissues, the degeneration affecting the compact mesoderm lying just lateral to the dorsal
 
aorta both in the region of the IX anlage proper and extending caudal to this point beneath the ventral communicating strand. The degeneration tends to become slightly less
 
marked in this embryo as the vagus is approached, but there
 
is considerable individual variation in the amount of degeneration observed. In some degree, however, it can be found
 
in all embryos of this age (24 to 27 somites). It is probable
 
that this degeneration of the mesenchyme may be induced by
 
the growth pressure of the ganglionic anlage, but while appearances indicate that the degeneration affects the mesoderm
 
most profoundly some degenerating cells could also be found
 
among the cells of the ganglionic anlage itself. The vagus
 
anlage at 26 somites lies lateral to the vena capitis lateralis
 
(fig. 88). It extends ventrally to the lower margin of the vein
 
and then curves medially to join the ventral communicating
 
strand. Due to the compactness of the mesoderm of the embryo figured (figs. 84 to 87), these relations of the vagus could
 
not be well made out, but another embryo of the same age
 
(fig. 88) as well as older embryos (fig. 89, 29 somites) show
 
clearly that the vagal proliferation curves medially around
 
the ventro-lateral margin of the vena capitis lateralis to join
 
the ventral communicating ramus. In embryos of 26 and 27
 
somites it is impossible to follow the vagus farther ventrally
 
and caudally. Posterior to and continuous with the vagal
 
anlage a strip of diffuse neural crest extends along the side
 
of the neural tube to join t-he spinal crest. In a 27—somite
 
embryo, i11 sections immediately caudal to the second gill
 
pouch, one finds that the glossopharyngeal anlage extends
 
somewhat farther ventral than the upper border of the dorsal
 
aorta which seemed to mark its lower limit in younger embryos. In this specimen one finds clear evidence of the ventral
 
96 1-IOVVABD B. ADELMANN
 
  
extension of the IX ganglionic anlage lateral to the dorsal
 
aorta. The advancing mass of ganglionic cells seems to be
 
pushing ahead of it a mass of degenerating mesodermal cells.
 
It is quite impossible, however, to determine the ventral
 
limit of the IX anlage in a 27-somite embryo sectioned frontally; one must accept the evidence afforded by favorable
 
transections as more reliable. The ventral extension of the
 
IX anlage lateral t.o the dorsal aorta and third aortic arch
 
never progresses far, however. The ganglionic aulage
 
reaches its definitive ventral limit in embryos of 29 and 31
 
somites when it reaches (figs. 90, 91) a point slightly ventral
 
to the dorsal level of the second pouch, so that it is ‘tucked’
 
into the angle between the caudal wall of the second gill
 
pouch and the ectodermal thickening which constitutes the
 
epibranchial placode. In figures 90, 91 the thickness of the
 
epibranchial placode of the glossopharyngeal is exaggerated,
 
due to the curvature of the caudal wall of the second cleft.
 
The ganglionic anlage rests lightly against the placode and
 
there is no fusion of the two structures nor is there any evidence of placodal proliferation. At 31 somites the glossepliaryngcal anlage still shows no evidence of division into
 
superior and inferior ganglionic masses. It has remained up
 
to this time a purely cellular structure and, while the cells
 
of which it is composed are much more elongated or spindleshaped than inprevious stages, exhibiting delicate fibrous
 
processes, no fibrous processes could be traced into the substance of the third arch. Reference to the figures shows that
 
a ‘spur’ of ganglionic cells such as was described for the V
 
and VII ganglia cannot be detected in the case of the IX or
 
the X ganglia.
 
  
The glossopliaryngeal anlage of 29- and 31—somite embryos
+
According to Neal (’14), the neural crest seems to be the chief source of origin for the sheath cells in reptiles, birds, and mammals. In Selachii, however, the sheath cells have been recorded as being exclusively medullary in origin.
is continued caudally by a lamina of neural-crest cells lying
 
medial to the vena capitis lateralis; as this is traced caudally
 
  
it divides into two communicating cords, one passing laterally
 
  
along the lower surface of the vena capitis lateralis to join.
+
In the spinal region of the rat (fig. 66) the ganglionic crest forms a spindle-shaped mass closely adherent to the sides of the neural tube. The apex of the ganglionic mass comes in contact with the developing motor fibers as soon as they leave the neural tube, probably, according to the most reliable evidence of Harrison, supplying these fibers with sheath cells. The ganglionic spur described in connection with the anlagen of the trigeminus and acoustico-facial ganglia bears a relationship to the developing motor roots of those ganglia which is entirely comparable to the relation of the ganglionic crest to the motor root in the case of the spinal nerve, and the suggestion is strong that the spurs, invaded by the developing motor fibers, furnish the neurilemma cells of the motor portions of these cranial nerves. It is significant that the anlagen of the IX-X nerves develop no such spurs.
the nodosal region of the vagus anlage and the other continuing caudally along the side of the neural tube to join the root
 
NEURAL FOLDS AND CBANIAL GANGLIA OF RAT
 
  
of the vagus anlage (cf. fig. 92). The vagus anlage of -the
 
31-somite embryo (fig. 93) still lies lateral to the old vena
 
capitis lateralis, but a new venous channel has become established lateral to it. Three or four small venous channels
 
may be recognized lateral to the vagus anlage in the 31-somite
 
embryo, but in most embryos this new channel is single. In
 
slightly older embryos (fig. 96, 34: somites), the venous channels lateral and medial to the vagus have enlarged so that
 
the nerve is surrounded by a lake of venous blood, the vascular endothelium being closely applied to the entire circumference of the nerve between tl1e ganglion radicis and the
 
ganglion nodosum. I
 
  
In sections of the 31-somite embryo through the region of
+
The author does 11ot wish to infer the homology of the spur with the spinal ganglionic crest in its entirety, nor does he consider the sp11r and the main ganglionic crest as representing different ‘orders’ of crest. The entire cranial ganglion plus the spur are probably the equivalent of the spinal neural crest.
the attachment of the vagus to the neural tube one finds the
 
beginning of an enlargement of the anlage lateral to the vena
 
capitis lateralis, constituting the nodosal swelling, but there
 
is still no well—defined division into root and nodosal ganglia.
 
In the 26- and 27-somite embryo no caudal extension of the
 
vagus anlage could be traced. Now, however (in 29- and 31somite embryos), as prominent cellular extension of the ganglionic mass lying lateral and slightly ventral to the vena
 
capitis lateralis can be tracedlfor some distance caudally
 
(fig. 94). It curves ventrally and medially a few sections
 
posterior to the ultimobranchial body and soon ends in the
 
mesenchyme lateral to the pharynx.
 
  
This caudal extension of the vagus anlage between the ages
 
of 27, 29, and 31 somites is without doubt associated with
 
the growth of parts to which the vagus is related. In the
 
26-somite embryo (fig. 19) the vagus anlage lies directly
 
above the third (the last true pouch in the rat) pharyngeal
 
pouch and the ultimobranchial body which arises as a caudal
 
extension of the posterior wall of the third pouch. In 29and 31-somite embryos, however, great growth and expansion
 
of the caudal region of the pharynx has occurred, resulting
 
in the separation of the ultimobrancl1ial body from the third
 
pouch, so that it comes to open independently into the cavity
 
of the pharynx. In the 31-somite embryo the ultimobranchial
 
98 HOWARD B. ADELMANN
 
  
body lies about twelve sections caudal to the third pouch.
+
The acoustic gawnglion. The present study was directed mainly toward a solution of the problem of the origin of the acoustic ganglionic mass, and no special attention was paid to its subsequent differentiation-—-a subject to which Streeter (’06) has already devoted a special paper. Concerning the origin of the ganglion cells of the acoustic nerve in man, Streeter (’12) says: “They are evidently not derived from neural crest; but Whether they migrate out from the brain wall or the walls of the developing ear vesicle, or are derived from the ectoderm immediately adjacent to the auditory pit, remains to be determined.
Now the vagus apparently takes part in the growth shiftings
 
of this region, growing caudally as the territory of the
 
pharynx to which it is related expands, and in keeping pace
 
with the growth shiftings the ganglion nodosum becomes extended caudally and slightly ventrally along the territory
 
of the fourth arch (and more caudally in some cases). This
 
caudal extension of the vagus is much slenderer and farther
 
removed from the ectoderm in the 29-somite embryo than in
 
the 31-somite specimen. In the latter it has attained some
 
size and lies nearer the surface ectoderm, but, so far as could
 
be determined, makes no contact with the extensive area of
 
thickened ectoderm caudal to thethird cleft.
 
  
There is a caudal extension of the neural crest back of the
 
X anlage of the 31—somite embryo, representing a prolongation of the ganglion radicis nervi vagi which becomes continuous with the neural crest related to the occipital myotomes, of which there seem to be three. Among its cells
 
numerous fibers of the spinal accessory nerve can be recognized, extending as far caudally as the sixth myotome (probably the third cervical, since the fourth myotome is the first
 
to which a definite ganglionic swelling is related). The fibers
 
of the spinal accessory form a well-defined bundle along the
 
caudal border of the vagus anlage. In the 29-somite embryo
 
a few fibers of the spinal accessory can be found among the
 
cells of the ganglion radicis vagi, but a distinct bundle of
 
accessory nerve fibers could not be identified associated with
 
the main ganglionic mass———a condition similar to what
 
Streeter has described in a 4-mm. human embryo.
 
  
Superior and inferior ganglionic masses could be recognized in both the glossopharyngeal and vagal anlagen of a
+
In the rat the last possibility is ruled out by a study of frontal sections of the acoustico-facial mass. Until the V11somite stage the acoustico-facial anlage does lie in close contact with the otic placode, but at this time one can find no evidence of ectodermal proliferation. After .12 somites the ganglion is a perfectly discrete cell column separated from the prominent anterior lip of the otic pit and from the anterior wall of the pit itself by considerable mesenchyme. It never subsequently comes into contact with the epithelium overlying the otic vesicle.
34-somite embryo (figs. 95, 96). In the case of the IX nerve,
 
however, the division into the ganglion of the root (Ehrenritter’s) and the petrosal ganglion was not nearly so distinct
 
as the separation of the vagus into a ganglion of the root and
 
the massive ganglion nodosum. The root ganglia are con
 
nected by a wide cellular lamina and the petrosal and nodosal
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 99
 
  
ganglia are joined by a cellular cord passing under the vena
 
capitis lateralis. The venous channel lateral to the vagus
 
should be noted.
 
  
The petrosal ganglion extends to the angle between the
+
There was no evidence in the rat of a proliferation of acoustic ganglion cells from the wall of the otic vesicle. From 12 to 18 somites the acoustico-facial ganglion is distinctly separated from the anterior wall of the otic pit or vesicle by mesenchyme, but shortly thereafter there is evidence of the migration of cells from the caudal face of the ganglion, giving rise to a mass of cells closely applied to the anterior wall of the otic vesicle and not at first separated from the facial ganglion. A series of photographs is presented to show the formation of the acusticus. At 19 somites (fig. 67) there is still no evidence of the presence of an acoustic ganglion. In the 21-somite embryo there is only a small caudal projection of the ganglion toward the otic vesicle, but in a 24—somite embryo (fig. 68) there is a well-developed mass of cells between the facial mass proper and the otic vesicle. To what extent this compact mass is composed of mesenchymal cells could not be determined. The photograph does not give an adequate idea of its size, since in other sections it could be followed along the anterior face of the otic vesicle and to a slight extent along its ventral surface.
caudal wall of the dorsal end of the second gill pouch and the
 
ectoderm of the second cleft where its lower pole comes into
 
contact withthe tip of the short ductus branchialis II, the
 
caudal wall of which, slightly thickened, represents the epibranchial placode of the IX nerve. This relationship is best
 
shown in a frontal section of an embryo of the same age
 
(fig. 97). Study of the figure reveals that the fusion of the
 
IX nerve with the ectoderm is neither very intimate nor
 
extensive. The ductus branchialis II becomes much elongated
 
in older embryos. Its medial end occupies a characteristic
 
position dorsal to the third pouch and it extends laterally
 
and somewhat caudally, its lumen communicating with the
 
cervical sinus. The tip of ductus branchialis II remains in
 
contact with the lower pole of the ganglion petrosum (fig. 97)
 
until the thirteen—day stage. In embryos of thirteen days
 
four hours it is represented by a vesicle connected with the
 
ectoderm by a cord of cells, the lumen of its outer portion
 
having been obliterated. The vesicle, however, is no longer
 
in contact with the lower pole of the ganglion, but rests
 
against the nerve a short distance below the ganglion petrosum. In thirteen—day-twelve-hour embryos the cord connecting the vesicle with ectoderm is degenerating and in a
 
fourteen-day embryo only a solid mass fused with the dorsal
 
side of the second pouch remains to mark its presence. There
 
is never the slightest evidence, in the writer’s opinion, that
 
the ectoderm contributes any cells to the substance of the
 
petrosal ganglion.
 
  
The ganglion nodosumof the 34—somite embryo extends
 
caudally over the territory of the fourth arch and bends ventrally back of the ultimobranchial body. A cellular extension of it can be traced to the root of the lung buds. There
 
is an extensive area of thickened ectoderm lying caudal to
 
the third gill cleft, representing a continuation of the bran100 H()WAT.{I') B. ADELMANN
 
  
chial ectoderm caudal to the last cleft and including in its
+
At 27 somites (fig. 69) the proliferation of ganglion cells from its caudal face has brought the acoustico-facial mass into close contact with the anterior face of the otic vesicle. At 28 to 29 somites (fig. 70) the two components of the ganglion are distinct. A proliferation of cells of the acoustic ganglion from the walls of the otic vesicle could not be observed, since all through the period of formation of the acoustic mass the walls of the otic vesicle are absolutely intact. I therefore conclude that the acoustic mass is a derivative of the main ganglionic mass and hence derived from the neural crest.
dorsal portio11 the vagus placode, which is 11ot separable from
 
it. The ganglion nodosum is now a bulky structure which
 
skirts along the upper margin of the thickened epithelium
 
back of the third cleft. In the 34-somite embryo of figure 96,
 
in one or two sections the ganglion lies closely against the
 
ectoderm, but there was no fusion of the two structures, the
 
inner surface of the ectoderm being everywhere quite ‘clean.’
 
On the other hand, another 34—somite embryo (fig. 98), evidently somewhat farther advanced in development, shows two
 
well-defined contacts of the vagus with the ectoderm. A
 
cellular extension of the ganglion running closely against
 
the caudal wall of the third pouch is probably the superior
 
laryngeal nerve, first recognizable in this embryo. It comes
 
into contact and fuses with the thickened ectoderm just candal to the third cleft. Somewhat farther dorsally and caudally the ganglion nodosum itself is fused with the ectoderm.
 
The ectoderm with which the vagus is fused now forms a
 
part of the caudal and dorsal wall of the cervical sinus. More
 
extensive contacts were noted in twelve—day~eighteen-hour
 
(fig. 99) and »tl1irteen—day embryos. At thirteen days it becomes included in the cervical vesicle which becomes cut off
 
from the exterior mainly by the expansion of the hyoid arch
 
and the postbranehial region. Figure 100 illustrates the formation of the cervical vesicle by the expansion of postbranchial region and the second arch. The material of the third
 
arch, not undergoing expansion, is becoming buried and overgrown by the rapidly expanding material lying on either side
 
of it. The vagus thus maintains its contact with the cervical
 
vesicle which lies at the lower pole of the ganglion at the
 
level of the superior laryngeal nerve. Wit-l'1 the continued
 
expansion of the postbranchial and hyoid-arch material and
 
the consequent obliteration of the cervical sinus, the
 
cervical vesicle becomes more deeply buried a11d for a
 
time is connected with the ectoderm by a cellular cord which
 
soon ruptures. Its attachment to the lower pole of the
 
ganglion nodosum at the level of the superior laryngeal nerve
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 101
 
  
persists for a long time, but in embryos of fourteen days
 
twelve hours, it seems to have separated from the lower pole
 
of the ganglion. i
 
  
The relation of the vesicle to the ganglion nodosum is nicely
+
In summarizing the development of the acoustico-facial anlage, the following facts may be emphasized. The anlage, a mass of neural crest attached to the otic rhombomere, grows ventrally between the paraxial mesoderm and the overlying thickened ectoderm of the otic placode with which it is in intimate contact, but from which it receives no contribution of cells. The ganglionic mass quickly reacl1es its ventral limit (10 to 11 somites) at the upper level of the first pouch where it rests against the condensed Inesenchyme of the second arch. The condensation of arch mesenchyme occurs independently of neural crest or other ectodermal contributions and no migration of neural-crest elements into the arch could be observed. The ganglion gradually becomes separated from the overlying ectoderm by the invasion of mesenchyme lateral to it, but its ventral extremity maintains contact with the ectoderm just behind the dorsal extremity of the first branchial pouch, where at 16 to 18 somites a pronounced thiel<ening marks the site of the epibranchial. placode.
shown in a thirteen-day—four-hour embryo (fig. 101). Careful study of the vagus placode and the cervical vesicle failed
 
to convince me that any contribution is made by it to the
 
vagus ganglion. It is, I believe, a case of adhesion, which, I
 
think, is borne out by a study of the relations of the vagus
 
to the vesicle in embryos of thirteen days, thirteen days four
 
hours (fig. 101), thirteen days twelve hours, and fourteen
 
days.
 
  
The caudal prolongation of the ganglion radicis vagi
 
joining the spinal neural crest is still prominent in the 34.somite embryo. It has been almost entirely replaced by
 
fibers of the accessory nerve i11 thirteen—day embryos.
 
  
Attention l1as already been called to the fact that beginning
+
The contact of the ganglion with the ectoderm above and behind the first pouch is, then, of long standing, existing from the time the ganglion reaches its ventral limit at 9 to 10 somites, and the relative position of the ventral extremity of the acoustico-facial anlage remains the same until by growth the anlage comes to be separated from the mesenchymal condensation of the arch (of. figs. 67 to 70). The ganglionic and mesodermal tissues are always tolerably well differentiated, so that it is not difiicult to ascertain the boundary between them.
with embryos of 29 somites, a new venous channel is established lateral to the vagus nerve, the latter thereby coming
 
to lie in the midst of a venous circle. The ninth and tenth
 
nerves are thus separated by the medial branch of this venous
 
ring, but connected with one another by dorsal and ventral
 
communicating rami which form a ring of ganglionic material
 
around the medial vein. Later, with the expansion of the
 
caudal pharyngeal and postpharyngeal regions, the descent
 
of the heart and the attendant descent of the caudal pharyngeal complexes related to these two nerves, onefinds that the
 
two come to approach one another closely as the logical result
 
of longitudinal tension. VVith the approximation of the two
 
nerves the medial half of the venous ring around the vagus
 
becomes compressed, attenuate, and finally ruptures. In a
 
fourteen~day embryo there is an interval between the place
 
of emergence from the brain of the root fibers of the IX and
 
X-‘H nerves, but there is an intimate fusion between the root
 
ganglia of the two nerves and a fusion also between the
 
inferior part of the petrosal and the superior portion of the
 
nodosal ganglia which lie side by side.
 
102 HOWARD B. ADELMANN
 
  
Streeter (’04) found a communication between the ganglion petrosum and nodosum in a human embryo of 7 mm., but
 
“in other embryos of this stage, and younger, they are completely separated. Later, following the relative change in
 
position of adjacent parts which succeeds their unequal
 
growth, these structures are gradually brought together, and
 
secondary communications are established between them.”
 
  
Ohiarugi (’90) studied the development of the vagus in
+
At about 29 to 31 somites the expansion of the hyoid arch results in the separation of even the ventral extremity of the ganglion from the ectoderm, but, as the ganglion recedes, it maintains its placodal contact by the drawing out of a cord of ganglionic cells extending from the lower pole of the ganglion to the ectoderm which persists until shortly after the thirteen—day stage. There is no doubt in the writer’s mind that this placodal ramus is ganglionic in origin.
rabbit embryos. In 4.5-mm. embryos the vagal and glossopharyngeal “semblent avoir une origine independante.” His
 
description is not detailed enough to allow me to correlate
 
his 4.5-mm embryos with rat embryos, but it would appear
 
that they are fairly advanced in development. Chiarugi,
 
however, apparently recognized the continuity of the neural
 
crest, because he says of embryos of 6.5 mm.: “ . . .
 
sur les points 011 elle (the neural crest) s’est conservé, elle est
 
plus nettement accentuée et a pris la forme de cordon commissural entre les productions ganglionnaires que se sont
 
formée a ses dépens. ” He also describes the continuity of the
 
neural crest of the vagus with that related to the myotomes.
 
  
In the rat the fusion of the ganglia of the roots and the
 
ganglion petrosum and nodosum of the ninth and tenth nerves
 
was found to be constant. Moreover, the connections were
 
found not to be secondary, but primary connections, since
 
the ninth and tenth anlagen are derived from an essentially
 
continuous postotic proliferation of neural crest which is
 
molded more or less by the mechanical influence of related
 
structures. There are two especially well-developed downgrowths of the postotic neural crest; one just caudal to the
 
otic vesicle back of the second pouch, situated medial to the
 
vena capitis lateralis, is the glossopharyngeal anlage and
 
another occurring just above and behind the third pouch and
 
lying lateral to the vena capitis lateralis is the vagus anlage.
 
Between these two prominent proliferations, however, there
 
is a continuous if less-marked proliferation of neural crest
 
giving rise to dorsal and ventral communicating cords between the ninth and tenth ganglia. This agrees with findings
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT
 
  
in lower groups where the IX-X are ‘einheitlich’ (cf. Neumayer, ’06). With the approximation of the ganglia due to
+
In determining the relation of ganglion and placode one must rely on good transections through the long axis of the ganglion or on frontal sections. The latter, however, are not favorable for determining the ventral extremity of the ganglion, but allow one to judge of the integrity of the inner surface of the placode. Oblique sections of any kind are apt to be misleading and the curvature of the arch surface must be kept in mind even in the interpretation of good transections. While the lower extremity of the ganglionic mass is almost from the beginning in co11tact with the pla codal ectoderm, no intimate fusion of parts appears to occur until about 31 somites. It is impossible to determine whether there is any migration of cells from the placode to the ganglion by way of the placodal ramus, but appearances do not favor such an interpretation.
the growth shiftings described on page 101, the communicating cords become less prominent.
 
  
Streeter ( ’04, p. 102) thought that the petrosal and nodosal
+
===3. The Glossopharyngeal and Vagus Ganglia===
ganglia of the nint:h and tenth nerves might possibly have an
 
origin different from that of the ganglia of the roots of these
 
two nerves, and suggests that the former may arise in situ
 
rather than by a subdivision from the rest of the anlage, but
 
if I interpret him correctly, advances the idea purely as a.
 
supposition.
 
  
The question raised by Streeter (’04) can be answered
+
So far as the writer is aware, the early history of the glossopharyngeal and vagus ganglia has never been completely followed in the mammal. The work of Martin (’91) on the IX-XII nerves of the cat and the excellent studies of Chiarugi (’89, ’90, ’94) on the rabbit and guinea—pig, and of Streeter (’O4) on the human embryo deal with later phases of their development.
only by a study of these ganglia in embryos younger than
 
were available for his study. In the rat, due to the ease of
 
collecting material, I have been fortunate enough to be able
 
to follow completely their early development. The youngest
 
specimen studied by Streeter, a 4-mm. human embryo, corresponds approximately to a 34-somite rat embryo in its stage
 
of development, preceding which stage both nerves have had
 
long and interesting histories. Examination of the complete
 
series of rat embryos available for this study shows convincingly, I believe, that the petrosal and nodosal ganglia
 
do not arise in situ, but that they attain their positions relative to the gill arches as a result of the migration ventrally
 
of the ganglionic crest from which they are derived. The
 
migration can be followed in detail in a close series of embryos. An examination of the series of photographs presented shows clearly that the division of the anlagen of the
 
IX and X nerves into ganglia of the root and ganglion petrosum and nodosum, respectively, occurs relatively late in the
 
development of these ganglia, i.e., not until about 34 somites.
 
  
DISCUSSION
 
  
In the foregoing pages the development of the sensory
+
The IX—X anlage is first discernible i11 the rat of 8 somites. It has been described previously as being at that stage the least advanced in development of the anlagen of the cranial ganglia. Attention may again be called to figure 28, which is a section through the IX-X anlage of the 8—somite embryo. At this early period, active proliferation has not yet begun and the anlage consists simply of a conical mass of cells fitted between the lateral edge of the neural plate and the ectoderm.  
ganglia of the V, VII-VIII, IX, and X nerves has been described. They have been found to be derived from proliferations of the cranial neural crest which move ventrally in each
 
IO4 HOWARD B. ADELMAN-N
 
  
case approximately to the level of the corresponding branchial pouch. The growth of the ganglia and the nerves derived from them has been found to be influenced largely by
 
the growth of parts to which they are related. In the case
 
of the trigeminus ganglion, for instance, we have found that
 
the growth of the ophthalmic ramus accompanies the shifting in the position of the optic vesicle which occurs during
 
the forward growth of material in the anterior region of the
 
head (of. discussion, pp. 65-69). No placodes were found which
 
could possibly account for its formation. The growth of the
 
ma.xilla1*y and mandibular rami undoubtedly accompanies
 
the downgrowth of the maxillary process and the downgrowth of the mandibular arches to form the lower jaw (of.
 
Streeter, ’22, figs. 3 and 4).
 
  
The growth of the vagus nerve affords a striking illustration of the same principle, showing a marked response to the
+
Externally, there is a shallow groove on each side marking the boundary between the neural plate and the ganglionic anlage. As yet the ectoderm cannot be traced over the ganglionic anlage to the edge of the neural plate. The IX-X anlarge may be traced caudally to the anterior margin of the first somite, where it becomes directly continuous with the actively proliferating spinal neural crest. The relations of the vas primitivum rhombencephali (Sabin) should be noted. It lies just below the IX-X anlage in close proximity to the neural fold.
growth of related parts. It was found (of. p. 101) to be
 
influenced in its growth by the rapid growth transformations
 
and shiftings of the caudal pharyngeal and postpharyngeal
 
regions. During the rapid expansion of the caudal region
 
of the pharynx, and the growth shiftings of the caudal pharyngeal complexes attending the descent of the heart, the ganglion nodosum becomes greatly elongated caudally and
 
ventrally posterior to the ultimobranchial body, and it might
 
be suggested that the problem of the distribution of the vagus
 
to the viscera might be explained upon the basis of the expansion and shifting of parts, relatively very restricted in
 
extent inthe early embryo. One has only to note in the
 
models here reproduced the relatively short extent of the
 
digestive tract between the last pharyngeal pouch and the
 
liver diverticulum to appreciate the tremendous elongation
 
which the region must subsequently undergo. Jackson (’09)
 
and Carey (’20) have already commented on the great elongation of the esophagus in the embryo.
 
  
The VII nerve has the least interesting growth history,
+
At 9 somites (fig. 71) the proliferation of the IX—X anlage has begun, the cells migrating laterally and ventrally to form a compact mass lying lateral to the vas primitivum rhombencephali and in contact with the overlying ectoderm which extends over the neural crest to the edge of the neural fold. The thickness of the ectoderm overlying the anlage is exaggerated in figure 28, due to the curvature of the embryo. It is impossible to fix the exact ventral limit of the crest. In some sections it appears to extend to the ventral border of the primitive rhombencephalic vessel. The extent of the IX-X anlage of the 9-somite embryo is somewhat greater than at 8 somites, due to the expansion of the territory between the anterior margin of the first somite and the otic placode. As one traces the IX-X crest caudally it becomes somewhat more diffuse, but maintains its position lateral to the primitive rhombencephalic vessel. At the anterior boundary of the first somite the vas primitivum rhombencephali divides into two slenderer branches; one passes medial to the somite and the other ventro-laterally into the somatopleure. The crest proliferation for the IX-X ganglia accompanies the medial branch of the vessel and becomes directly continuous with the spinal neural crest medial to the somite. A
since it is related to a territory situated at the boundary
 
between two opposing growth tendencies, namely, the forNEURAL FOLDS AND CRANIAL GANGLIA or RAT 105
 
  
ward growth and expansion of the anterior part of the head
 
and pharynx and the relative caudal movement of the posterior regions accompanying the descent of the heart
 
(cf. rKingsbury’s, ’15, analysis of the growth of the
 
pharynx). At first, therefore, the seventh nerve merely
 
grows into the substance of the hyoid arch and has a somewhat negative history. Its subsequent migrations with the
 
forward spreading of the facial musculature are of course
 
well known.
 
  
The early history of the growth of the glossopharyngeal
+
In figures 72 to 74 a series of photographs of an embryo in which the twelfth somite is just forming is presented to illustrate the essential features of the IX-X anlage at this stage. Figure 72 lies three sections caudal to the otic pit on the left. At this level the crest is a definite compact mass lying lateral to the Vas primitivum rhombencephali. The primitive rhombencephalic vein makes an abrupt Ventro— lateral bend just in front of the first somite to make its way into the somatopleure. A branch passing medial to the somite could not be found. Part of its course may be followed in the figures. As the crest is traced caudally it becomes somewhat more diffuse (fig. 73) and, passing around the caudal border of the Vas primitivum rhombencephali (anterior cardinal vein?) as the latter bends ventrally, becomes continuous with the spinal neural crest. Figure 74 is a section just caudal to the upper end of the anterior cardinal Vein showing the place of transition between the IX-X and the spinal neural crest. The diffuse character of the IX-X crest at most levels makes it impossible to determine accurately its ventral limit.
nerve is also simple; it undergoes at first but little caudal
 
shifting, due, no doubt, to the more or less negative role of
 
the material of the III arch, which has been found to undergo
 
but little growth, and to be buried during the obliteration of
 
the cervical sinus by the growth and expansion of the material
 
of the hyoid arch and postbranchial regions.
 
  
The most interesting question relative to the development
 
of the cranial ganglia concerns the r6le which the epibranchial
 
placodes play in their formation. In lower forms——cyclostomes, elasmobranchs, teleosts, and Amphibia-——two series
 
  
, of placodes are described, a dorso—lateral series and an epi
+
The IX—X anlage of '13—somite embryos is Very similar to that of 11- and 12-somite specimens, but one finds some changes in the relation of the crest to the venous channels of the head. In the l3—somite embryo the anterior portion of the postotic neural crest in the region of the future IX proliferation is just beginning to be separated from the overlying ectoderm by the establishment of a new Venous channel lateral to it, the beginnings of which are illustrated in figure 75. One finds many places where the neural crest seems to have migrated ventrally medial to the Vas primitivum rhombencephali as well as lateral to it, so that in some sections (of. the left side of fig. 76) that Vessel appears to be surrounded by neural-crest cells. The relations of the posterior portion of the IX-Y crest remainas described for the IX-X crest of earlier embryos. The anterior cardinal Vein is cut longitudinally on the right side of figure 76, which shows the neural crest lateral to that Vessel. The neural crest passes around the caudal wall of the Vein to become continuous with the neural crest related to the first somite.
branchial group of placodes. Tl1e former are concerned with
 
the development of the lateral-line nerves. VVith regard to
 
the lower forms, it has now come to be quite generally agreed
 
that cells proliferated from both series of placodes are incorporated in the cranial ganglia. There is, of course, considerable variation in the extent to which proliferation from
 
the two series of placodes participate in the formation of a
 
particular ganglion in the different forms. A review of the
 
literature of the subject relating to non-mammalian forms
 
may be foregone here, since many excellent summaries have
 
  
already been presented.
 
  
VVith regard to the placodes of the mammal, it would seem
+
Sections just caudal to the otic pit of the 14-somite embryo (fig. 77) show that the new Venous channel lateral to the anterior portion of the IX—X anlage is now well established; the old vessel is probably represented by a more or less collapsed channel embedded in the substance of the neural crest, but which communicates (fig. 78) frequently by anastomoses with the new vena capitis lateralis. The latter tends to lie more and more ventral to the crest (fig. 79) as it is traced caudally. I was unable to determine whether a new venous channel has been established in the caudal region of the IX-X anlage or whether there has simply been some ventral and lateral shifting of the old vas primitivum rhombencephali in this region. The limits of the neural crest constituting the IX—X anlage at this stage are better defined for a short distance caudal to the ear vesicle in the region of the prospective IX downgrowth where the crest is more compact and abundant than in the more caudal portion. In the territory of the future IX proliferation the crest lies medial to the vena capitis lateralis, but as the sections approach the first somite the crest becomes Very diffuse and shifts dorsally somewhat so that it tends to lie above the lateral head vein and closer to the overlying ectoderm. In this region, on account of its diffuse character, it becomes impossible to accurately delimit the crest from the mesenchyme (fig. 79), but careful study shows it to be continuous with the spinal crest at the level of the first somite (fig. 80). In embryos of this age the relations of the IX-X anlage to the early venous channels of the head are extremely variable, as one would expect from the plastic nature of the neural crest and the endothelial tubes, so that the description given above applies in detail only to the right side of the 14:—somite embryo here illustrated.
that there are several questions at issue:
 
1. To which cranial ganglia are placodes related’?
 
2. Are dorso—lateral placodes present in the mammal‘?
 
106 HOWARD B. ADELMANN
 
  
Froriep (’85), who made the pioneer observations on the
 
epibranchial placodes, described them in relation to the VII,
 
IX, and X ganglia in cow and sheep embryos, but was unable
 
to find a placode for the V nerve.
 
  
Chiarugi (’94) studied the development of the trigeminal
+
In a 16-somite embryo the IX-X anlage is widely separated from the ectoderm by diffuse mesenchyme in which the now rather capacious vena capitis lateralis courses. Beginning immediately caudal to the otic vesicle, the IX-X crest forms a lens—shaped mass closely applied to the lateral surface of the neural tube. It is largest in its anterior portion and thins out, becoming more diffuse as it proceeds caudally. An» teriorly, the anlage extends ventrally to the middle of the lateral wall of the neural tube, but does not reach the midline dorsally. It shifts dorsally, however, as it is traced caudally, and in its more posterior regions a thin li11e of crest cells overlies the dorsal surface of the neural tube. In some sections through the caudal portion of the IX-X anlage the crest is so diffuse that it is impossible to be positive of its presence. However, such appearances are confined to single sections and one has no difficulty in determining the continuity of the crest of the IX and X ganglia with the spinal neural crest. It is quite probable that such isolated sections where the crest appears to be absent are due to localized rarefactions of the crest of no significance. In later stages the caudal portion of the IX-X crest is more compact and there is never any doubt of its continuity with the spinal crest.
ganglion in the guinea-pig, and described clearly the broad
 
and intimate contact of the V ganglion with the epidermis in
 
embryos of 3 to 3.6 mm. He believes that cells proliferated
 
from the epidermis take part in its formation.
 
  
Weigner (’01), who studied the development of the V and
 
VII ganglia of the ground-squirrel (a few pig and human
 
embryos were also studied), was unable to confirm Chiarugi’s
 
observations on the fusion of the trigeminus with the
 
ectoderm.
 
  
Griglio—Tos (’02) describes three dorso-lateral and three
+
Wliile the IX—X crest is more extensive, more compact, and more prominent just caudal to the otic pit than in more posterior portions, there is as yet no more definite indication of the position of the anlage of the glossopharyngeus. A definite ventral extension of the IX-X crest marking the position of the ninth nerve is first observed in embryos of 17 and 18 somites. In the 18-somite embryo (fig. 18), three sections caudal to the ear, a slender cord of cells, visible in two sections 7.5 u thick, extends ventrally from the crest as far as the ventral border of the vena capitis lateralis, where it comes into contact with the mcsenchymal condensation of the third branchial arch. It lies medial to the vena capitis lateralis, a position which it retains.
epibranchial thickenings related to the trigeminus ganglion
 
in a 15-somite human embryo, all of which proliferate ganglionic material. He recognizes nine subdivisions of the trigeminus and finds a complete correspondence between the
 
development of the trigeminus ganglion of man and the
 
lamprey. And all this from the study of a single embryo!
 
Giglio-Tos has obviously confused mesenchymal condensations with ganglionic material. He also describes (’02) dorsolateral and epibranchial placodes of the VII-VIII ganglion.
 
  
Celestine da Costa (’20, ’21, ’23) recognizes dorso-lateral
 
and epibranchial placodes in connection with the V and VII
 
nerves. The epibranchial placode of the V nerve is present
 
at 6 somites in the guinea-pig, but the dorso-lateral placode
 
does not appear until 16 to 18 somites, and is visible only
 
during this time, after which it disappears suddenly. The
 
most intimate relations exist between this placode and the
 
trigeminal ganglion, making a placodal contribution to it
 
possible. “En tout cas cette collaboration doit étre peu importante et bien plus réduite que celle qu’a décrite Brachet
 
chez les Reptiles” (’23, p. 518).
 
  
Volker (’22) states that in the ground-squirrel the ganglia
+
A series of sections through the IX-X anlage of a 21-somite embryo are presented in figures 81 and 82. The section illustrated in figure 81 lies eight sections caudal to the second gill pouch and five sections caudal to the otic vesicle. Figure 82 passes through the region of the ultimobranchial body. The IX anlage shown in figure 81 is a compact cord of crest cells extending through four 7.5 p sections. It lies medial to the vena capitis lateralis, to the lower border of which the ganglionic anlage extends. The anlage comes intocontact with the mesenchymal condensation of the third arch, where it seems to be insinuating itself between the compact arch mesenchyme and the lower border of the lateral head vein. There is a slight thickening of the ectoderm above and behind the second gill pouch, which probably represents the epibranchial placode of the ninth nerve, but the thickening is not marked nor is it extensive. The IX anlage does not come into contact with it. The anlage of the glossopl1aryngeus is continuous caudally with the lens-shaped mass of neural-crest cells (fig. 82) that finally becomes continuous with the spinal crest. 8 It should be noted that at this stage the condensed arch mesenchyme caudal to the third pouch extends dorsally as far as the ventral border of the lateral head vein, and in more caudal sections it extends still farther dorsally around the lateral border of the vena capitis lateralis.
of the V, VII, IX, and X nerves fuse with the ectoderm and
 
“erhalten von ihm Zuschuss zu ihren Zellen.
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 107
 
  
Davis (’23) finds no true placode related to the trigeminus
 
in a 20-somite human embryo.
 
  
The Writer has found that, in the case of each of the early
+
Four sections caudal to figure 82 as few neural-crest cells appeared to be moving ventrally over the lateral surface of the vena capitis lateralis, probably the earliest indication of the vagus anlage. Twenty—four—somite embryos were the first to show a definite vagal proliferation. In the 24-somite embryo there are two well-defined downgrowths from the postotic neural crest—-—the glossopharyngeal and the vagal. The former lies medial to the lateral head vein, while the latter is situated lateral to it. The glossopharyngeal proliferation has increased considerably in bulk, its ventral extremity is somewhat enlarged and seems to be insinuating itself between the lower surface of the lateral head vein and the compact arch mesoderm. The mesenchyme above the level of the dorsal aorta shows numerous degenerating cells, and subsequent stages show that the compact mesenchyme disappears in large part from this situation (cf. figs. 83, 89). The vagus (fig. 83), proliferating lateral to the vena capitis lateralis, also seems to be insinuating itself between the lateral head vein and the compacted arch mesenchyme, but in this region no degenerating cells are found in the mesenchyme at this stage. It is almost impossible to determine the exact limits of the vagal anlage. I11 the figure of the 24-somite embryo here reproduced, it seems to be easily distinguishable from the mesenchyme, but such is not the case in other embryos of the same age. It is also difficult to determine the caudal extent of the ventral extremity of the vagus, because just caudal to the vagal proliferation the compact mesenchyme extends dorsally lateral to the vena capitis lateralis. In the interval between the IX and X proliferations there is a short extent of neural crest where the proliferation is less active, but nevertheless occurs with sufficient intensity to allow one to trace crest cells ventrally along the inner side of the vena capitis lateralis; the cells so proliferated form in later stages a definite cord of ganglionic cells extending from the glossopharyngeal to the vagal anlage.
ganglionic anlagen in the rat, the ganglion proliferates at
 
first lateral to the paraxial mesoderm in close contact with
 
the overlying ectoderm. This contact of the early ganglionic
 
proliferations with the ectoderm is constant and characteristic. It is due, I believe, to two factors: First, it appears,
 
as I have already pointed out (p. 64), that the ectoderm
 
exercises a directive influence upon the rapidly proliferating
 
ganglion cells, which migrate ventrally along its inner surface, a parallel case being the directive influence of solid
 
objects upon the growth of tissues in vitro. Secondly, the
 
rapid lateral expansion of the paraxial mesoderm tends to
 
keep the ganglion in contact with the ectoderm. This expan~
 
sion is most marked and continues longest in the region of
 
the mandibular arch. Consequently, the contact of the V
 
ganglion with the ectoderm is most prolonged. The VII-VIII
 
anlage soon becomes separated from the ectoderm by the
 
invasion of mesenchyme lateral to it, its ventral extremity
 
keeping contact with the epibranchial placode. The entire
 
early IX—X anlage becomes separated from the ectoderm in
 
a similar fashion. Wheii the vagus proliferation from the
 
postotic crest occurs at 24 to 25 somites, it hugs the lateral
 
surface of the vena capitis lateralis, which seems to exercise
 
a directive or tactic influence.
 
  
The question as to what constitutes a true placode is a
 
puzzling one, considerable confusion having arisen due to the
 
fact that placodal thickenings have not always been differentiated from the diffuse and general thickening of the ectoderm over the entire branchial regionf‘ The thickening of
 
  
3 In an abstract appearing after the completion of the above text and too late
+
About twenty-five embryos cut in various planes were available for the study of the neural crest of the glossepharyngeus a11d vagus in the period between l.'W€11t}"—five and thirty—one somites. Great variatiionwas noted in embryos of the same age and many features which could be clearly observed in some embryos were not so obvious in others. For instance, in some embryos it was possible to delimit the ganglionic anlage easily from the adjacent mesenchyme, in others such a delimitation was not possible, due to individual variations in the compactness of the surrounding mesenchyme or of the ganglionic anlage itself. The following descriptions must therefore apply in fine detail only to the embryos illustrated; attention will be called to important individual deviations.
for proper notice in the body of this paper, Bartelmez (’24) has described the
 
extent of the thickened branchial ectoderm in young human embryos. He finds
 
“clear evidences of cell migration from the area. V as well as the more caudal
 
visceral arch areas. ’ ’
 
  
In his complete monograph Bartelmez (’25) believes he has evidence that the
 
placodes contribute la.rg'el_v to the ganglia. of the V and VII nerves and also
 
108 HOWARD B. ADELMANN
 
  
the branchial ectoderm begins first iii the region of the mandibular areh, extending forward over the mouth clef t, thinning
+
Five pliotograplis of sections of a 26—somite embryo are reproduced in figures 84 to 88. The first section passes through the caudal wall of the second branchial pouch and is the most anterior section through the glossopharyngeal anlage. A stout cord of ganglionic crest cells constituting the IX anlage could be traced through nine sections of 7.5 u. Caudal to that point the neural crest becomes very attenuate in its middle portion, so that extending between the glossopharyngeal and vagal proliferations there are two cords of neu1‘al-crest cells (fig. 87), a dorsal one lying close against the lateral surface of the neural tube and a ventral less compact cord of crest cells which lies just medial to the vena capitis lateralis and just dorsal to the dorsal aorta. As this latter cord is traced caudally it passes laterally beneath the vena capitis lateralis to join the ventral extremity of the vagus anlage. There are thus two communicating strands, a dorsal and a ventral, between the ganglionic anlagen of the IX and X nerves at this time (of. fig. 19). The dorsal and ventral communicating cords owe their origin to a continuation of crest proliferation caudal to the IX a11lage. A study of embryos of 24 to 27 somites confirms fully this statement. In some embryos there is a profuse proliferation of crest cells medial to the vena capitis lateralis extending as far caudally as the vagal proliferation, while in others one finds the two communicating cords as described above with a delicate strand of crest cells extending between them. I have therefore been drawn to the conclusion that this ventral communicating ramus is formed by the piling up of crest cells medial to the vena capitis lateralis in the region between the IX and X proliferations. As time goes on, the dorsal communicating ramus becomes somewhat reduced in size. In this embryo many degenerating cells were found in it, possibly some degeneration being a factor in its reduction in size. The cells of the ventral communicating ramus are somewhat diffuse and in some embryos of 26 to 27 somites it could not be traced with certainty as far caudally as the vagus.
somewhat as it is traced forward. The thickening of the
 
branchial integument proceeds caudally with the differentiation of the branchial pouches. At 14 somites the ectoderm
 
over the branchial region caudal to the second branchial
 
pouch is thickened as far caudally as the anterior border of
 
the first somite, and the thickening extends ventrally over
 
the pericardial region for a short distance (cf. figs. 72 to 79).
 
After the formation of the last branchial pouch (the third in
 
the rat), one finds a marked thickening of the ectoderm caudal
 
to it, which does not, howe\>*er, extend as far caudally as the
 
cranial border of the first somite as it did earlier.
 
  
As I have described, thetrigeminal ganglion has for as considerable time a broad area of contact with the ectoderm
 
above the first branchial arch, but the ectoderm related to
 
the ganglion is so slightly thickened and the slight thickening
 
so diffuse that it is extremely improbable that it constitutes
 
a true placode. The ectoderm over the ganglion is transitional in thickness between that covering the mandibular arch
 
and that over the dorsum of the embryo, forming a natural
 
transition between the two. The ectodermal thickening figured by Celestine da Costa (’ 3, fig. 3), and interpreted by
 
him as an epibranchial placode related to the trigeminal
 
ganglion, is regarded by the writer as merely thickened
 
branchial ectoderm.
 
  
The epibranchial placode related to the VII nerve is the
+
In this particular embryo it is somewhat difficult to fix the ventral limit of the glossopharyngeal anlage in the first few sections caudal to the second pouch. However, careful study of several embryos of the same period as well as the evidence presented by this embryo makes me feel fairly certain in fixing its ventral limit at about the dorsal boundary of the dorsal aorta. In the first place, reference to figures of 21~ somite embryos will show that the dorsal limit of the compact arch mesoderm of the third arch lies slightly above the dorsal border of the dorsal aorta. In some embryos in which the ventral end of the IX anlage is less compact than the arch mesoderm one l1as no difficulty in distinguishing the limits of the two tissues. The same is true also of embryo figured in sections through slightly more caudal portions of the glossopharyngeal anlage. Secondly, one finds in embryos of 25 to 27 somites a large number of degenerating cells at the boundary between the two tissues, the degeneration affecting the compact mesoderm lying just lateral to the dorsal aorta both in the region of the IX anlage proper and extending caudal to this point beneath the ventral communicating strand. The degeneration tends to become slightly less marked in this embryo as the vagus is approached, but there is considerable individual variation in the amount of degeneration observed. In some degree, however, it can be found in all embryos of this age (24 to 27 somites). It is probable that this degeneration of the mesenchyme may be induced by the growth pressure of the ganglionic anlage, but while appearances indicate that the degeneration affects the mesoderm most profoundly some degenerating cells could also be found among the cells of the ganglionic anlage itself. The vagus anlage at 26 somites lies lateral to the vena capitis lateralis (fig. 88). It extends ventrally to the lower margin of the vein and then curves medially to join the ventral communicating strand. Due to the compactness of the mesoderm of the embryo figured (figs. 84 to 87), these relations of the vagus could not be well made out, but another embryo of the same age (fig. 88) as well as older embryos (fig. 89, 29 somites) show clearly that the vagal proliferation curves medially around the ventro-lateral margin of the vena capitis lateralis to join the ventral communicating ramus. In embryos of 26 and 27 somites it is impossible to follow the vagus farther ventrally and caudally. Posterior to and continuous with the vagal anlage a strip of diffuse neural crest extends along the side of the neural tube to join t-he spinal crest. In a 27—somite embryo, i11 sections immediately caudal to the second gill pouch, one finds that the glossopharyngeal anlage extends somewhat farther ventral than the upper border of the dorsal aorta which seemed to mark its lower limit in younger embryos. In this specimen one finds clear evidence of the ventral extension of the IX ganglionic anlage lateral to the dorsal aorta. The advancing mass of ganglionic cells seems to be pushing ahead of it a mass of degenerating mesodermal cells. It is quite impossible, however, to determine the ventral limit of the IX anlage in a 27-somite embryo sectioned frontally; one must accept the evidence afforded by favorable transections as more reliable. The ventral extension of the IX anlage lateral t.o the dorsal aorta and third aortic arch never progresses far, however. The ganglionic aulage reaches its definitive ventral limit in embryos of 29 and 31 somites when it reaches (figs. 90, 91) a point slightly ventral to the dorsal level of the second pouch, so that it is ‘tucked’ into the angle between the caudal wall of the second gill pouch and the ectodermal thickening which constitutes the epibranchial placode. In figures 90, 91 the thickness of the epibranchial placode of the glossopharyngeal is exaggerated, due to the curvature of the caudal wall of the second cleft. The ganglionic anlage rests lightly against the placode and there is no fusion of the two structures nor is there any evidence of placodal proliferation. At 31 somites the glossepliaryngcal anlage still shows no evidence of division into superior and inferior ganglionic masses. It has remained up to this time a purely cellular structure and, while the cells of which it is composed are much more elongated or spindleshaped than inprevious stages, exhibiting delicate fibrous processes, no fibrous processes could be traced into the substance of the third arch. Reference to the figures shows that a ‘spur’ of ganglionic cells such as was described for the V and VII ganglia cannot be detected in the case of the IX or the X ganglia.
only one which, in the rat, is distinctly separable from the
 
difl’use thickening of the ectoderm over the branchial region.
 
The epibranchial placode related to the ninth nerve was found
 
to be less clearly delimited from the branchial ectoderm than
 
the epibranchial placode related to the VII ganglion. The
 
epibranchial placode, with which the superior laryngeal nerve
 
and ganglion nodosum fuse, is not separable from the thickstates that “ . . . . we are convi1'1ce(l that in Homo, as in many Icht.hyopsid:-'1,
 
  
the rostral neural crest furnishes a large contribution to the supporting elenients
 
of the visceral region of the head.”
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 109
 
  
ened ectoderm caudal to the third cleft which becomes involved in the formation of the cervical vesicle. The fusion of
+
The glossopliaryngeal anlage of 29- and 31—somite embryos is continued caudally by a lamina of neural-crest cells lying medial to the vena capitis lateralis; as this is traced caudally it divides into two communicating cords, one passing laterally along the lower surface of the vena capitis lateralis to join. the nodosal region of the vagus anlage and the other continuing caudally along the side of the neural tube to join the root of the vagus anlage (cf. fig. 92). The vagus anlage of -the 31-somite embryo (fig. 93) still lies lateral to the old vena capitis lateralis, but a new venous channel has become established lateral to it. Three or four small venous channels may be recognized lateral to the vagus anlage in the 31-somite embryo, but in most embryos this new channel is single. In slightly older embryos (fig. 96, 34: somites), the venous channels lateral and medial to the vagus have enlarged so that the nerve is surrounded by a lake of venous blood, the vascular endothelium being closely applied to the entire circumference of the nerve between tl1e ganglion radicis and the ganglion nodosum.
the ninth ganglion with its placode was found to be less intimate than that of the VII or X ganglia—-—a condition somewhat different than has been described in other mammalian
 
forms. In the case of the VII ganglion, a cord of cells connects the ganglion with its epibranchial placode; no such
 
placodal branches of the petrosal or nodosal ganglia were
 
noted. The only structure which may possibly be interpreted as a dorso-lateral placode in the mammal is the otic
 
placode. In early embryos the VII ganglion makes intimate
 
contact with the otic placode, but it was impossible to detect
 
the migration of any cells from it into the ganglionic mass.
 
  
3. Do the placodes contribute to the formation of ganglionic material?
 
  
4. If the placodes do proliferate material incorporated in
+
In sections of the 31-somite embryo through the region of the attachment of the vagus to the neural tube one finds the beginning of an enlargement of the anlage lateral to the vena capitis lateralis, constituting the nodosal swelling, but there is still no well—defined division into root and nodosal ganglia. In the 26- and 27-somite embryo no caudal extension of the vagus anlage could be traced. Now, however (in 29- and 31somite embryos), as prominent cellular extension of the ganglionic mass lying lateral and slightly ventral to the vena capitis lateralis can be tracedlfor some distance caudally (fig. 94). It curves ventrally and medially a few sections posterior to the ultimobranchial body and soon ends in the mesenchyme lateral to the pharynx.
the ganglia, what part does this material play?
+
 
 +
 
 +
This caudal extension of the vagus anlage between the ages of 27, 29, and 31 somites is without doubt associated with the growth of parts to which the vagus is related. In the 26-somite embryo (fig. 19) the vagus anlage lies directly above the third (the last true pouch in the rat) pharyngeal pouch and the ultimobranchial body which arises as a caudal extension of the posterior wall of the third pouch. In 29and 31-somite embryos, however, great growth and expansion of the caudal region of the pharynx has occurred, resulting in the separation of the ultimobrancl1ial body from the third pouch, so that it comes to open independently into the cavity of the pharynx. In the 31-somite embryo the ultimobranchial body lies about twelve sections caudal to the third pouch. Now the vagus apparently takes part in the growth shiftings of this region, growing caudally as the territory of the pharynx to which it is related expands, and in keeping pace with the growth shiftings the ganglion nodosum becomes extended caudally and slightly ventrally along the territory of the fourth arch (and more caudally in some cases). This caudal extension of the vagus is much slenderer and farther removed from the ectoderm in the 29-somite embryo than in the 31-somite specimen. In the latter it has attained some size and lies nearer the surface ectoderm, but, so far as could be determined, makes no contact with the extensive area of thickened ectoderm caudal to thethird cleft.
 +
 
 +
 
 +
There is a caudal extension of the neural crest back of the X anlage of the 31—somite embryo, representing a prolongation of the ganglion radicis nervi vagi which becomes continuous with the neural crest related to the occipital myotomes, of which there seem to be three. Among its cells numerous fibers of the spinal accessory nerve can be recognized, extending as far caudally as the sixth myotome (probably the third cervical, since the fourth myotome is the first to which a definite ganglionic swelling is related). The fibers of the spinal accessory form a well-defined bundle along the caudal border of the vagus anlage. In the 29-somite embryo a few fibers of the spinal accessory can be found among the cells of the ganglion radicis vagi, but a distinct bundle of accessory nerve fibers could not be identified associated with the main ganglionic mass———a condition similar to what Streeter has described in a 4-mm. human embryo.
 +
 
 +
 
 +
Superior and inferior ganglionic masses could be recognized in both the glossopharyngeal and vagal anlagen of a 34-somite embryo (figs. 95, 96). In the case of the IX nerve, however, the division into the ganglion of the root (Ehrenritter’s) and the petrosal ganglion was not nearly so distinct as the separation of the vagus into a ganglion of the root and the massive ganglion nodosum. The root ganglia are connected by a wide cellular lamina and the petrosal and nodosal ganglia are joined by a cellular cord passing under the vena capitis lateralis. The venous channel lateral to the vagus should be noted.
 +
 
 +
 
 +
The petrosal ganglion extends to the angle between the caudal wall of the dorsal end of the second gill pouch and the ectoderm of the second cleft where its lower pole comes into contact withthe tip of the short ductus branchialis II, the caudal wall of which, slightly thickened, represents the epibranchial placode of the IX nerve. This relationship is best shown in a frontal section of an embryo of the same age (fig. 97). Study of the figure reveals that the fusion of the IX nerve with the ectoderm is neither very intimate nor extensive. The ductus branchialis II becomes much elongated in older embryos. Its medial end occupies a characteristic position dorsal to the third pouch and it extends laterally and somewhat caudally, its lumen communicating with the cervical sinus. The tip of ductus branchialis II remains in contact with the lower pole of the ganglion petrosum (fig. 97) until the thirteen—day stage. In embryos of thirteen days four hours it is represented by a vesicle connected with the ectoderm by a cord of cells, the lumen of its outer portion having been obliterated. The vesicle, however, is no longer in contact with the lower pole of the ganglion, but rests against the nerve a short distance below the ganglion petrosum. In thirteen—day-twelve-hour embryos the cord connecting the vesicle with ectoderm is degenerating and in a fourteen-day embryo only a solid mass fused with the dorsal side of the second pouch remains to mark its presence. There is never the slightest evidence, in the writer’s opinion, that the ectoderm contributes any cells to the substance of the petrosal ganglion.
 +
 
 +
 
 +
The ganglion nodosumof the 34—somite embryo extends caudally over the territory of the fourth arch and bends ventrally back of the ultimobranchial body. A cellular extension of it can be traced to the root of the lung buds. There is an extensive area of thickened ectoderm lying caudal to the third gill cleft, representing a continuation of the branchial ectoderm caudal to the last cleft and including in its dorsal portio11 the vagus placode, which is 11ot separable from it. The ganglion nodosum is now a bulky structure which skirts along the upper margin of the thickened epithelium back of the third cleft. In the 34-somite embryo of figure 96, in one or two sections the ganglion lies closely against the ectoderm, but there was no fusion of the two structures, the inner surface of the ectoderm being everywhere quite ‘clean.’ On the other hand, another 34—somite embryo (fig. 98), evidently somewhat farther advanced in development, shows two well-defined contacts of the vagus with the ectoderm. A cellular extension of the ganglion running closely against the caudal wall of the third pouch is probably the superior laryngeal nerve, first recognizable in this embryo. It comes into contact and fuses with the thickened ectoderm just candal to the third cleft. Somewhat farther dorsally and caudally the ganglion nodosum itself is fused with the ectoderm. The ectoderm with which the vagus is fused now forms a part of the caudal and dorsal wall of the cervical sinus. More extensive contacts were noted in twelve—day~eighteen-hour (fig. 99) and »tl1irteen—day embryos. At thirteen days it becomes included in the cervical vesicle which becomes cut off from the exterior mainly by the expansion of the hyoid arch and the postbranehial region. Figure 100 illustrates the formation of the cervical vesicle by the expansion of postbranchial region and the second arch. The material of the third arch, not undergoing expansion, is becoming buried and overgrown by the rapidly expanding material lying on either side of it. The vagus thus maintains its contact with the cervical vesicle which lies at the lower pole of the ganglion at the level of the superior laryngeal nerve. Wit-l'1 the continued expansion of the postbranchial and hyoid-arch material and the consequent obliteration of the cervical sinus, the cervical vesicle becomes more deeply buried a11d for a time is connected with the ectoderm by a cellular cord which soon ruptures. Its attachment to the lower pole of the ganglion nodosum at the level of the superior laryngeal nerve persists for a long time, but in embryos of fourteen days twelve hours, it seems to have separated from the lower pole of the ganglion.
 +
 
 +
 
 +
The relation of the vesicle to the ganglion nodosum is nicely shown in a thirteen-day—four-hour embryo (fig. 101). Careful study of the vagus placode and the cervical vesicle failed to convince me that any contribution is made by it to the vagus ganglion. It is, I believe, a case of adhesion, which, I think, is borne out by a study of the relations of the vagus to the vesicle in embryos of thirteen days, thirteen days four hours (fig. 101), thirteen days twelve hours, and fourteen days.
 +
 
 +
 
 +
The caudal prolongation of the ganglion radicis vagi joining the spinal neural crest is still prominent in the 34.somite embryo. It has been almost entirely replaced by fibers of the accessory nerve in thirteen—day embryos.
 +
 
 +
 
 +
Attention l1as already been called to the fact that beginning with embryos of 29 somites, a new venous channel is established lateral to the vagus nerve, the latter thereby coming to lie in the midst of a venous circle. The ninth and tenth nerves are thus separated by the medial branch of this venous ring, but connected with one another by dorsal and ventral communicating rami which form a ring of ganglionic material around the medial vein. Later, with the expansion of the caudal pharyngeal and postpharyngeal regions, the descent of the heart and the attendant descent of the caudal pharyngeal complexes related to these two nerves, onefinds that the two come to approach one another closely as the logical result of longitudinal tension. VVith the approximation of the two nerves the medial half of the venous ring around the vagus becomes compressed, attenuate, and finally ruptures. In a fourteen~day embryo there is an interval between the place of emergence from the brain of the root fibers of the IX and X-‘H nerves, but there is an intimate fusion between the root ganglia of the two nerves and a fusion also between the inferior part of the petrosal and the superior portion of the nodosal ganglia which lie side by side.
 +
 
 +
 
 +
Streeter (’04) found a communication between the ganglion petrosum and nodosum in a human embryo of 7 mm., but “in other embryos of this stage, and younger, they are completely separated. Later, following the relative change in position of adjacent parts which succeeds their unequal growth, these structures are gradually brought together, and secondary communications are established between them.”
 +
 
 +
 
 +
Ohiarugi (’90) studied the development of the vagus in rabbit embryos. In 4.5-mm. embryos the vagal and glossopharyngeal “semblent avoir une origine independante.” His description is not detailed enough to allow me to correlate his 4.5-mm embryos with rat embryos, but it would appear that they are fairly advanced in development. Chiarugi, however, apparently recognized the continuity of the neural crest, because he says of embryos of 6.5 mm.: “ . . . sur les points 011 elle (the neural crest) s’est conservé, elle est plus nettement accentuée et a pris la forme de cordon commissural entre les productions ganglionnaires que se sont formée a ses dépens. ” He also describes the continuity of the neural crest of the vagus with that related to the myotomes.
 +
 
 +
 
 +
In the rat the fusion of the ganglia of the roots and the ganglion petrosum and nodosum of the ninth and tenth nerves was found to be constant. Moreover, the connections were found not to be secondary, but primary connections, since the ninth and tenth anlagen are derived from an essentially continuous postotic proliferation of neural crest which is molded more or less by the mechanical influence of related structures. There are two especially well-developed downgrowths of the postotic neural crest; one just caudal to the otic vesicle back of the second pouch, situated medial to the vena capitis lateralis, is the glossopharyngeal anlage and another occurring just above and behind the third pouch and lying lateral to the vena capitis lateralis is the vagus anlage. Between these two prominent proliferations, however, there is a continuous if less-marked proliferation of neural crest giving rise to dorsal and ventral communicating cords between the ninth and tenth ganglia. This agrees with findings in lower groups where the IX-X are ‘einheitlich’ (cf. Neumayer, ’06). With the approximation of the ganglia due to the growth shiftings described on page 101, the communicating cords become less prominent.
 +
 
 +
 
 +
Streeter ( ’04, p. 102) thought that the petrosal and nodosal ganglia of the nint:h and tenth nerves might possibly have an origin different from that of the ganglia of the roots of these two nerves, and suggests that the former may arise in situ rather than by a subdivision from the rest of the anlage, but if I interpret him correctly, advances the idea purely as a. supposition.
 +
 
 +
 
 +
The question raised by Streeter (’04) can be answered only by a study of these ganglia in embryos younger than were available for his study. In the rat, due to the ease of collecting material, I have been fortunate enough to be able to follow completely their early development. The youngest specimen studied by Streeter, a 4-mm. human embryo, corresponds approximately to a 34-somite rat embryo in its stage of development, preceding which stage both nerves have had long and interesting histories. Examination of the complete series of rat embryos available for this study shows convincingly, I believe, that the petrosal and nodosal ganglia do not arise in situ, but that they attain their positions relative to the gill arches as a result of the migration ventrally of the ganglionic crest from which they are derived. The migration can be followed in detail in a close series of embryos. An examination of the series of photographs presented shows clearly that the division of the anlagen of the IX and X nerves into ganglia of the root and ganglion petrosum and nodosum, respectively, occurs relatively late in the development of these ganglia, i.e., not until about 34 somites.
 +
 
 +
==Discussion==
 +
 
 +
In the foregoing pages the development of the sensory ganglia of the V, VII-VIII, IX, and X nerves has been described. They have been found to be derived from proliferations of the cranial neural crest which move ventrally in each case approximately to the level of the corresponding branchial pouch. The growth of the ganglia and the nerves derived from them has been found to be influenced largely by the growth of parts to which they are related. In the case of the trigeminus ganglion, for instance, we have found that the growth of the ophthalmic ramus accompanies the shifting in the position of the optic vesicle which occurs during the forward growth of material in the anterior region of the head (of. discussion, pp. 65-69). No placodes were found which could possibly account for its formation. The growth of the ma.xilla1*y and mandibular rami undoubtedly accompanies the downgrowth of the maxillary process and the downgrowth of the mandibular arches to form the lower jaw (of. Streeter, ’22, figs. 3 and 4).
 +
 
 +
 
 +
The growth of the vagus nerve affords a striking illustration of the same principle, showing a marked response to the growth of related parts. It was found (of. p. 101) to be influenced in its growth by the rapid growth transformations and shiftings of the caudal pharyngeal and postpharyngeal regions. During the rapid expansion of the caudal region of the pharynx, and the growth shiftings of the caudal pharyngeal complexes attending the descent of the heart, the ganglion nodosum becomes greatly elongated caudally and ventrally posterior to the ultimobranchial body, and it might be suggested that the problem of the distribution of the vagus to the viscera might be explained upon the basis of the expansion and shifting of parts, relatively very restricted in extent inthe early embryo. One has only to note in the models here reproduced the relatively short extent of the digestive tract between the last pharyngeal pouch and the liver diverticulum to appreciate the tremendous elongation which the region must subsequently undergo. Jackson (’09) and Carey (’20) have already commented on the great elongation of the esophagus in the embryo.
 +
 
 +
 
 +
The VII nerve has the least interesting growth history, since it is related to a territory situated at the boundary between two opposing growth tendencies, namely, the forward growth and expansion of the anterior part of the head and pharynx and the relative caudal movement of the posterior regions accompanying the descent of the heart (cf. rKingsbury’s, ’15, analysis of the growth of the pharynx). At first, therefore, the seventh nerve merely grows into the substance of the hyoid arch and has a somewhat negative history. Its subsequent migrations with the forward spreading of the facial musculature are of course well known.
 +
 
 +
 
 +
The early history of the growth of the glossopharyngeal nerve is also simple; it undergoes at first but little caudal shifting, due, no doubt, to the more or less negative role of the material of the III arch, which has been found to undergo but little growth, and to be buried during the obliteration of the cervical sinus by the growth and expansion of the material of the hyoid arch and postbranchial regions.
 +
 
 +
 
 +
The most interesting question relative to the development of the cranial ganglia concerns the r6le which the epibranchial placodes play in their formation. In lower forms——cyclostomes, elasmobranchs, teleosts, and Amphibia-——two series of placodes are described, a dorso—lateral series and an epi branchial group of placodes. Tl1e former are concerned with the development of the lateral-line nerves. VVith regard to the lower forms, it has now come to be quite generally agreed that cells proliferated from both series of placodes are incorporated in the cranial ganglia. There is, of course, considerable variation in the extent to which proliferation from the two series of placodes participate in the formation of a particular ganglion in the different forms. A review of the literature of the subject relating to non-mammalian forms may be foregone here, since many excellent summaries have already been presented.
 +
 
 +
 
 +
With regard to the placodes of the mammal, it would seem that there are several questions at issue: 1. To which cranial ganglia are placodes related’? 2. Are dorso—lateral placodes present in the mammal?
 +
 
 +
Froriep (’85), who made the pioneer observations on the epibranchial placodes, described them in relation to the VII, IX, and X ganglia in cow and sheep embryos, but was unable to find a placode for the V nerve.
 +
 
 +
Chiarugi (’94) studied the development of the trigeminal ganglion in the guinea-pig, and described clearly the broad and intimate contact of the V ganglion with the epidermis in embryos of 3 to 3.6 mm. He believes that cells proliferated from the epidermis take part in its formation.
 +
 
 +
Weigner (’01), who studied the development of the V and VII ganglia of the ground-squirrel (a few pig and human embryos were also studied), was unable to confirm Chiarugi’s observations on the fusion of the trigeminus with the ectoderm.
 +
 
 +
Griglio—Tos (’02) describes three dorso-lateral and three epibranchial thickenings related to the trigeminus ganglion in a 15-somite human embryo, all of which proliferate ganglionic material. He recognizes nine subdivisions of the trigeminus and finds a complete correspondence between the development of the trigeminus ganglion of man and the lamprey. And all this from the study of a single embryo! Giglio-Tos has obviously confused mesenchymal condensations with ganglionic material. He also describes (’02) dorsolateral and epibranchial placodes of the VII-VIII ganglion.
 +
 
 +
Celestine da Costa (’20, ’21, ’23) recognizes dorso-lateral and epibranchial placodes in connection with the V and VII nerves. The epibranchial placode of the V nerve is present at 6 somites in the guinea-pig, but the dorso-lateral placode does not appear until 16 to 18 somites, and is visible only during this time, after which it disappears suddenly. The most intimate relations exist between this placode and the trigeminal ganglion, making a placodal contribution to it possible. “En tout cas cette collaboration doit étre peu importante et bien plus réduite que celle qu’a décrite Brachet chez les Reptiles” (’23, p. 518).
 +
 
 +
Volker (’22) states that in the ground-squirrel the ganglia of the V, VII, IX, and X nerves fuse with the ectoderm and “erhalten von ihm Zuschuss zu ihren Zellen.”
 +
 
 +
Davis (’23) finds no true placode related to the trigeminus in a 20-somite human embryo.
 +
 
 +
The Writer has found that, in the case of each of the early ganglionic anlagen in the rat, the ganglion proliferates at first lateral to the paraxial mesoderm in close contact with the overlying ectoderm. This contact of the early ganglionic proliferations with the ectoderm is constant and characteristic. It is due, I believe, to two factors: First, it appears, as I have already pointed out (p. 64), that the ectoderm exercises a directive influence upon the rapidly proliferating ganglion cells, which migrate ventrally along its inner surface, a parallel case being the directive influence of solid objects upon the growth of tissues in vitro. Secondly, the rapid lateral expansion of the paraxial mesoderm tends to keep the ganglion in contact with the ectoderm. This expan~ sion is most marked and continues longest in the region of the mandibular arch. Consequently, the contact of the V ganglion with the ectoderm is most prolonged. The VII-VIII anlage soon becomes separated from the ectoderm by the invasion of mesenchyme lateral to it, its ventral extremity keeping contact with the epibranchial placode. The entire early IX—X anlage becomes separated from the ectoderm in a similar fashion. Wheii the vagus proliferation from the postotic crest occurs at 24 to 25 somites, it hugs the lateral surface of the vena capitis lateralis, which seems to exercise a directive or tactic influence.
 +
 
 +
The question as to what constitutes a true placode is a puzzling one, considerable confusion having arisen due to the fact that placodal thickenings have not always been differentiated from the diffuse and general thickening of the ectoderm over the entire branchial regionf‘ The thickening of
 +
 
 +
3 In an abstract appearing after the completion of the above text and too late for proper notice in the body of this paper, Bartelmez (’24) has described the extent of the thickened branchial ectoderm in young human embryos. He finds “clear evidences of cell migration from the area. V as well as the more caudal visceral arch areas.
 +
 
 +
In his complete monograph Bartelmez (’25) believes he has evidence that the placodes contribute largely to the ganglia. of the V and VII nerves and also
 +
 
 +
the branchial ectoderm begins first in the region of the mandibular arch, extending forward over the mouth clef t, thinning somewhat as it is traced forward. The thickening of the branchial integument proceeds caudally with the differentiation of the branchial pouches. At 14 somites the ectoderm over the branchial region caudal to the second branchial pouch is thickened as far caudally as the anterior border of the first somite, and the thickening extends ventrally over the pericardial region for a short distance (cf. figs. 72 to 79). After the formation of the last branchial pouch (the third in the rat), one finds a marked thickening of the ectoderm caudal to it, which does not, howe\>*er, extend as far caudally as the cranial border of the first somite as it did earlier.
 +
 
 +
As I have described, thetrigeminal ganglion has for as considerable time a broad area of contact with the ectoderm above the first branchial arch, but the ectoderm related to the ganglion is so slightly thickened and the slight thickening so diffuse that it is extremely improbable that it constitutes a true placode. The ectoderm over the ganglion is transitional in thickness between that covering the mandibular arch and that over the dorsum of the embryo, forming a natural transition between the two. The ectodermal thickening figured by Celestine da Costa (’ 3, fig. 3), and interpreted by him as an epibranchial placode related to the trigeminal ganglion, is regarded by the writer as merely thickened branchial ectoderm.
 +
 
 +
The epibranchial placode related to the VII nerve is the only one which, in the rat, is distinctly separable from the difl’use thickening of the ectoderm over the branchial region. The epibranchial placode related to the ninth nerve was found to be less clearly delimited from the branchial ectoderm than the epibranchial placode related to the VII ganglion. The epibranchial placode, with which the superior laryngeal nerve and ganglion nodosum fuse, is not separable from the thickstates that “ . . . . we are convi1'1ce(l that in Homo, as in many Icht.hyopsid:-'1,
 +
 
 +
the rostral neural crest furnishes a large contribution to the supporting elenients of the visceral region of the head
 +
 
 +
ened ectoderm caudal to the third cleft which becomes involved in the formation of the cervical vesicle. The fusion of the ninth ganglion with its placode was found to be less intimate than that of the VII or X ganglia—-—a condition somewhat different than has been described in other mammalian forms. In the case of the VII ganglion, a cord of cells connects the ganglion with its epibranchial placode; no such placodal branches of the petrosal or nodosal ganglia were noted. The only structure which may possibly be interpreted as a dorso-lateral placode in the mammal is the otic placode. In early embryos the VII ganglion makes intimate contact with the otic placode, but it was impossible to detect the migration of any cells from it into the ganglionic mass.
  
Chiarugi (’94), Griglio-Tos (’02), and Volker (’22) definitely
+
3. Do the placodes contribute to the formation of ganglionic material?
commit themselves to the occurrence of placodal additions to
 
the cranial ganglia in mammals.
 
  
Froriep (’85) concludes that the participation of the ectoderm in ganglion formation, if it occurs at all (‘Wenn sie fiberhaupt besteht’), must be confined to early stages and that
+
4. If the placodes do proliferate material incorporated in the ganglia, what part does this material play?
such participationdoes not constitute the ultimate significance of the fusion of ganglion and placode.
 
  
Celestino da Costa (’23, p. 518) thinks that for the trigeminus a placodal contribution is possible, but in any case
+
Chiarugi (’94), Griglio-Tos (’02), and Volker (’22) definitely commit themselves to the occurrence of placodal additions to the cranial ganglia in mammals.
unimportant; in the case of the acoustic ganglion, ‘plus que
 
probable,’ and ‘n’est guere niable’ for the geniculate ganglion.
 
  
Weigne1' (’01) was unable to be certain of ectodermal contacts of the V and VII ganglia and is inclined to believe that
+
Froriep (’85) concludes that the participation of the ectoderm in ganglion formation, if it occurs at all (‘Wenn sie fiberhaupt besteht’), must be confined to early stages and that such participationdoes not constitute the ultimate significance of the fusion of ganglion and placode.
appearances which have been described as placodal proliferations are in most, if not all, cases due to oblique sectioning.
 
  
Streeter (’04) could find “no indication of the interchange
+
Celestino da Costa (’23, p. 518) thinks that for the trigeminus a placodal contribution is possible, but in any case unimportant; in the case of the acoustic ganglion, ‘plus que probable,’ and ‘n’est guere niable’ for the geniculate ganglion.
of cells between ganglion and epidermis” in the epibranchial
 
placodes of the ganglion petrosum and nodosum of the human
 
embryo.
 
  
THE JOURNAL OF C01\-£PARA'I‘1VE NEUROLOGY, VOL. 39, NO. 1
+
Weigne1' (’01) was unable to be certain of ectodermal contacts of the V and VII ganglia and is inclined to believe that appearances which have been described as placodal proliferations are in most, if not all, cases due to oblique sectioning.
110 HOWARD B. ADELMANN
 
  
Turning to lower forms, Landacre (’10) found that in
+
Streeter (’04) could find “no indication of the interchange of cells between ganglion and epidermis” in the epibranchial placodes of the ganglion petrosum and nodosum of the human embryo.
Ameiurus the visceral ganglion of the IX nerve is derived
 
entirely from cells split off en masse from the epibranchial
 
placode of IX. Since Herrick (’01, ’07) had found only
 
special visceral fibers arising from the visceral ganglion of
 
the IX nerve in that form, Landacre concluded that the “epibranchial placodes give rise to those ganglionic cells from
 
wliich gustatory fibers arise.” According to Landacre, only
 
general visceral and general cutaneous components are de—
 
rived from the neural crest. Stone (’22) found that the neural
 
crest yields only the general visceral component of the cranial
 
ganglia in Amblystoma.
 
  
Tello ( ’23), studying the development of the cranial ganglia
 
of the chick by use of the Cajal—pyridine-silver technique,
 
came to tl1e conclusion that the real ganglionic elements are
 
derived from the neural crest alone.
 
  
There is a complete lack of evidence concerning the fate of
+
Turning to lower forms, Landacre (’10) found that in Ameiurus the visceral ganglion of the IX nerve is derived entirely from cells split off en masse from the epibranchial placode of IX. Since Herrick (’01, ’07) had found only special visceral fibers arising from the visceral ganglion of the IX nerve in that form, Landacre concluded that the “epibranchial placodes give rise to those ganglionic cells from wliich gustatory fibers arise.” According to Landacre, only general visceral and general cutaneous components are de— rived from the neural crest. Stone (’22) found that the neural crest yields only the general visceral component of the cranial ganglia in Amblystoma.
any cells which may be added to the ganglionic anlagen by
 
placodal proliferation in the mammal. Viilker (’22) assumed
 
the transformation of such cells into ganglion cells in the
 
spermophile “auf grund von Beobachtungen von andern
 
Forschern. ’ ’ I
 
  
In the rat I have found surprisingly few mitoses in the pla~
+
Tello ( ’23), studying the development of the cranial ganglia of the chick by use of the Cajal—pyridine-silver technique, came to tl1e conclusion that the real ganglionic elements are derived from the neural crest alone.
codes at any time-—~—far too few to account for any extensive
 
addition of cells to the ganglia. Furthermore, if such proliferation were occurring, it would seem that in the examination of a large number of embryos one would be certain to
 
find at least some numbers of mitosesiwith spindles arranged
 
so that division would result in the addition of cells to the
 
underlying ganglion and in addition some cells being pinched
 
off from the ectoderm. The few mitoses found in the placodes were superficial in position (see, for instance, fig. 57)
 
and only one or two were found with spindle axes perpendicular to the surface. Deceptive appearances were often
 
caused by oblique sections. It was found that whenever the
 
section was obviously true one had no hesitancy in deciding
 
against proliferation and that when the section was obviously
 
NEURAL FOLDS AND CRANIAL GANGLIA OF RAT 111
 
  
oblique then it became diflicult to be certain. It seems justifiable to conclude that it is improbable that ganglionic elements are contributed by the epibranchial placodes in the
+
There is a complete lack of evidence concerning the fate of any cells which may be added to the ganglionic anlagen by placodal proliferation in the mammal. Viilker (’22) assumed the transformation of such cells into ganglion cells in the spermophile “auf grund von Beobachtungen von andern Forschern. ’ ’ I
mammal. The absence of placodal contributions to the cranial ganglia in the mammalia would not exclude the possibility of such occurring in lower forms, and, conversely, it is
 
dangerous to argue that, since the placodes appear to yield
 
ganglionic elements in the Amphibia, they must do likewise
 
in the bird, reptile, or mammal. Embryological work on the
 
higher forms has been too often distorted by phylogenetic
 
influences. Beard (’85), for instance, although he had not
 
investigated the Mammalia, nevertheless expressed “a firm
 
conviction that such rudiments (i.e., sense organs in connection with the ciliary and Gasserian ganglia) exist at some
 
stage or other in mammalian development. This conviction
 
rests upon a two-fold basis, an a, priori one (the italics are
 
mine) that in Elasmobranchii the sense organs of the ciliary
 
and Gasserian ganglia are Well developed,” etc. Giglio-Tos’
 
papers are another case in point. And What shall we say of
 
Vi)'lker’s assumptions based upon others’ observations?
 
  
It may not be inappropriate to remark that the actual transformation of placodal cells into ganglion cells has not been
+
In the rat I have found surprisingly few mitoses in the pla~ codes at any time-—~—far too few to account for any extensive addition of cells to the ganglia. Furthermore, if such proliferation were occurring, it would seem that in the examination of a large number of embryos one would be certain to find at least some numbers of mitosesiwith spindles arranged so that division would result in the addition of cells to the underlying ganglion and in addition some cells being pinched off from the ectoderm. The few mitoses found in the placodes were superficial in position (see, for instance, fig. 57) and only one or two were found with spindle axes perpendicular to the surface. Deceptive appearances were often caused by oblique sections. It was found that whenever the section was obviously true one had no hesitancy in deciding against proliferation and that when the section was obviously oblique then it became diflicult to be certain. It seems justifiable to conclude that it is improbable that ganglionic elements are contributed by the epibranchial placodes in the mammal. The absence of placodal contributions to the cranial ganglia in the mammalia would not exclude the possibility of such occurring in lower forms, and, conversely, it is dangerous to argue that, since the placodes appear to yield ganglionic elements in the Amphibia, they must do likewise in the bird, reptile, or mammal. Embryological work on the higher forms has been too often distorted by phylogenetic influences. Beard (’85), for instance, although he had not investigated the Mammalia, nevertheless expressed “a firm conviction that such rudiments (i.e., sense organs in connection with the ciliary and Gasserian ganglia) exist at some stage or other in mammalian development. This conviction rests upon a two-fold basis, an a, priori one (the italics are mine) that in Elasmobranchii the sense organs of the ciliary and Gasserian ganglia are Well developed,” etc. Giglio-Tos’ papers are another case in point. And What shall we say of Vi)'lker’s assumptions based upon others’ observations?
observed in any form. Furthermore, it would seem that the
 
study of Tello (’23), employing the Cajal-pyridine-silver
 
technique, definitely rules out the transformation of placodal elements into ganglion cells in at least one form, the
 
chick.
 
  
Landacre (’12) found in Ameiurus “a close correspondence
+
It may not be inappropriate to remark that the actual transformation of placodal cells into ganglion cells has not been observed in any form. Furthermore, it would seem that the study of Tello (’23), employing the Cajal-pyridine-silver technique, definitely rules out the transformation of placodal elements into ganglion cells in at least one form, the chick.
between the size of the epibranchial placode and the number
 
of gustatory fibers to which the ganglion gives rise in the
 
adult,” and further states that, “In view of the reduced
 
character of the gustatory system in many forms as compared with the fishes, it is not surprising that the actual contribution of cells by the epibrancliial placode should not be
 
large and might take place during the period of contact and
 
still be diificult to demonstrate.” If one grants the contribution of ganglionic elements by the placodes in the mammal,
 
112 HOWARD B. ADELMANN
 
  
still no such correspondence between the size of the contact
+
Landacre (’12) found in Ameiurus “a close correspondence between the size of the epibranchial placode and the number of gustatory fibers to which the ganglion gives rise in the adult,” and further states that, “In view of the reduced character of the gustatory system in many forms as compared with the fishes, it is not surprising that the actual contribution of cells by the epibrancliial placode should not be large and might take place during the period of contact and still be diificult to demonstrate.” If one grants the contribution of ganglionic elements by the placodes in the mammal, still no such correspondence between the size of the contact and the number of gustatory fibers in the adult nerve exists, so far as our present knowledge of the distribution of gustatory nerves in the mammal allows us to judge. The epibranchial contact of the vagus is most extensive in the mammal (cf. Froriep, ’85; Kingsbury, ’15), and yet of the three nerves (VII, IX, X) distributed to the taste buds, the vagus has the most restricted territory (of. Parker, ’22, and Zander, "97).
and the number of gustatory fibers in the adult nerve exists,
 
so far as our present knowledge of the distribution of gustatory nerves in the mammal allows us to judge. The epibranchial contact of the vagus is most extensive in the mammal (cf. Froriep, ’85; Kingsbury, ’15), and yet of the three
 
nerves (VII, IX, X) distributed to the taste buds, the vagus
 
has the most restricted territory (of. Parker, ’22, and Zander, "97).
 
  
 
5. What is the fate of the placodes themselves?
 
5. What is the fate of the placodes themselves?
  
6. Does the fusion of ganglion and placode exercise any
+
6. Does the fusion of ganglion and placode exercise any mechanical influence in development?
mechanical influence in development?
 
  
In the rat epibranchial placode I becomes much reduced in
+
In the rat epibranchial placode I becomes much reduced in thickness as development proceeds and finally disappears. Its contact with the VII ganglion is lost shortly after thirteen days of age. Epibranchial placode II becomes involved in the formation of ductus branchialis II and epibranchial pla code III becomes included in the cervical sinus, sharing the
thickness as development proceeds and finally disappears.
 
Its contact with the VII ganglion is lost shortly after thirteen
 
days of age. Epibranchial placode II becomes involved in
 
the formation of ductus branchialis II and epibranchial pla
 
code III becomes included in the cervical sinus, sharing the
 
  
fate of those structures, but it is hardly likely that the fusion
+
fate of those structures, but it is hardly likely that the fusion of ganglion and placode is responsible in any mechanical sense for the formation of either ductus branchialis II or the cervical sinus, botl1 of which structures can be more adequately accounted for by the expansion of the surrounding arcl1 material.
of ganglion and placode is responsible in any mechanical sense
 
for the formation of either ductus branchialis II or the cervical sinus, botl1 of which structures can be more adequately
 
accounted for by the expansion of the surrounding arcl1
 
material.
 
  
 
7. What is the significance of the placodes?
 
7. What is the significance of the placodes?
  
This last question is a puzzling one. They have been regarded as rudimentary sense organs (Froriep, ’85; Johnson
+
This last question is a puzzling one. They have been regarded as rudimentary sense organs (Froriep, ’85; Johnson and Sheldon, ’86; Beard, ’86), as “différentiations provoquée par le voisinage du -tube neural, ou de ses émanations, vésicule optique ou crétes ganglionnaires” (Celestine da Costa, ’23, p. 519), or as “eines Uebergreifens des VVucherungsreizes von den ihnen engst benachbarten Ganglienanlagen auf ihren Zellen” (Viilker, ’22, p. 197). Johnston (’05) believes that the dorso-lateral placodes “represent material comparable to the neural crest which has separated more completely from the brain than has the remainder of the crest.”  
and Sheldon, ’86; Beard, ’86), as “différentiations provoquée
 
par le voisinage du -tube neural, ou de ses émanations,
 
vésicule optique ou crétes ganglionnaires” (Celestine da
 
Costa, ’23, p. 519), or as “eines Uebergreifens des VVucherungsreizes von den ihnen engst benachbarten Ganglienanlagen auf ihren Zellen” (Viilker, ’22, p. 197). Johnston (’05)
 
believes that the dorso-lateral placodes “represent material
 
comparable to the neural crest which has separated more
 
completely from the brain than has the remainder of the
 
crest.”
 
NEURAL FOLDS AND CRANIAL GANGLIA or RAT 113
 
  
To the writer it seems possible that the placodes arise dur
+
To the writer it seems possible that the placodes arise dur ing development as a result of a tissue interaction (possibly
ing development as a result of a tissue interaction (possibly
 
  
reciprocal) between the epidermis and the ganglion, or that
+
reciprocal) between the epidermis and the ganglion, or that they represent rudimentary sense organs developed under the stimulus of nervous contact, the placode undergoing retrogression when the contact is broken as a result of growth shiftings. In this connection one is reminded of the degeneration of certain end organs upon the severance of the nerves supplying them (viz., the taste buds). The pl.acodes need never have functioned as specific sense organs at any time during ontogeny or phylogeny. Any attempt to homologize them would be futile.
they represent rudimentary sense organs developed under
 
the stimulus of nervous contact, the placode undergoing retrogression when the contact is broken as a result of growth
 
shiftings. In this connection one is reminded of the degeneration of certain end organs upon the severance of the
 
nerves supplying them (viz., the taste buds). The pl.acodes
 
need never have functioned as specific sense organs at any
 
time during ontogeny or phylogeny. Any attempt to homologize them would be futile.
 
  
We come finally to a consideration of mesectoderm formation in the mammal. Veit (’18) found it impossible to delimit
+
We come finally to a consideration of mesectoderm formation in the mammal. Veit (’18) found it impossible to delimit the diffuse neural crest in an 8-somite human embryo, and concluded that neural-crest cells were transformed into mesenchymal cells. Appearances similar to those described by Veit are observable in the rat. In most cases where the neural crest is diffuse one can draw no definite boundary between neural crest and mesenchyme. Now it is possible that cells may become detached from the neural crest, become a part of the mesenchyme, lose their ectodermal potentialities, and differentiate into some mesenchymal derivative, as Stone’s (’22) experiments seem to indicate for the Amphibia. Mangold (’24) has demonstrated that ectodermal cells transplanted into the mesoderm may differentiate into mesodermal structures, and it is quite possible that some such process may occur in the case of the neural crest. However, as Brachet (’07, ’21, p. 386), Celestino da Costa (’23, p. 519), and Keibel (’24, p. 38) point out, there is nothing to prove that cells detached from the crest do not differentiate into nervous structures.
the diffuse neural crest in an 8-somite human embryo, and
 
concluded that neural-crest cells were transformed into mesenchymal cells. Appearances similar to those described by
 
Veit are observable in the rat. In most cases where the
 
neural crest is diffuse one can draw no definite boundary
 
between neural crest and mesenchyme. Now it is possible
 
that cells may become detached from the neural crest, become
 
a part of the mesenchyme, lose their ectodermal potentialities, and differentiate into some mesenchymal derivative, as
 
Stone’s (’22) experiments seem to indicate for the Amphibia.
 
Mangold (’24) has demonstrated that ectodermal cells transplanted into the mesoderm may differentiate into mesodermal
 
structures, and it is quite possible that some such process
 
may occur in the case of the neural crest. However, as Brachet (’07, ’21, p. 386), Celestino da Costa (’23, p. 519), and
 
Keibel (’24, p. 38) point out, there is nothing to prove that
 
  
_ cells detached from the crest do not differentiate into nervous
 
  
structures.
+
There is, furthermore, no evidence, in the rat, of an organized contribution of the neural crest to the mesoderm, such as has been described in Amphibia by Platt, Landacre, and Stone. In the case of the rostral neural crest, which gives rise to the trigeminal ganglion, one could distinguish it fairly definitely from the surrounding mesenchyme as a lozenge-shaped mass of deeply staining cells even at 5 somites, and while cells might occasionally be found which seemed to be in process of being added to the mesenchyme, there was no evidence of a migration of the crest into the branchial arch. The acoustico—facial ganglionic mass is always compact, and it was possible to observe the condensation of the mesenchyme of the second arch independent of any neural—crest contribution in a close series of embryos from 8 to 12 somites. The postotic neural crest. is very diffuse in embryos of 9, 16, and 17 somites and no definite boundary between crest and mesenchyme can be set. The mesenchyme of the third and fourth arches becomes very compact some time before the neural crest has migrated ventrally to their upper levels, the condensation occurring independent of any neural-crest contribution. When the neural crest has finally reached its ventral limit at the dorsal levelsof the arches in question, the definite ganglionic masses appear for the IX and X nerves, the arch mesoderm is too compact to permit of any further ventral migration of the neural crest, and no such migration could be observed. Celestine da Costa (’23, p. 519) comes to a similar conclusion from his study of the guinea-pig.
  
There is, furthermore, no evidence, in the rat, of an organized contribution of the neural crest to the mesoderm, such
 
as has been described in Amphibia by Platt, Landacre, and
 
Stone. In the case of the rostral neural crest, which gives
 
rise to the trigeminal ganglion, one could distinguish it fairly
 
114 HOVVABD B. ADELMANN
 
  
definitely from the surrounding mesenchyme as a lozengeshaped mass of deeply staining cells even at 5 somites, and
+
Veit (’f2.2) has also expressed the belief that the mesectoderm is proliferated from the epidermis over the entire body. Appearances such as are pictured in his figures (7, 7 a, 8, 8 a) were observed in the rat, but are, I believe, to be differently interpreted. The first indication of the formation of a gill arch is a thickening of the overlying ectoderm. The condensation of the mesenchyme proceeds rapidly and the compacted mesenchyme comes to lie closely against the thickened branchial ectoderm——-so closely, in fact, that the line of separation is often obliterated, as one might expect when a closely compacted mass of cells comes in contact with an epithelium with no well-defined basement membrane. The difficulties of interpretation, as \Veigner (’O1) has observed, are increased by the curvature of an arch, so that sections are more than likely to be oblique. One gains the impression that possibly the mesenchyme is condensing under the influence of the ectoderm, or, conversely, that the condensation and expansion of the arch mesenchyme are responsible for the thickening of the overlying epithelium, as Harrison (’18) found to be the case in the limb bud. A striking illustration of the above was seen in rat embryos of 11 to 14 somites. Here the territory of the prospective third and fourth arches is indicated by an extensive area of thickened ectoderm which extends caudal to the second gill cleft almost to the border of the first somite. It extends ventrally a short distance over the pericardial region. Immediately beneath this area of thickened brancl1iogenetic ectoderm one finds the mesenchyme condensing and in close contact with the overlying epidermis, the condensation extending over the pericardium. Now the question arises, are these mesenchymal cells being proliferated from the ectoderm? I do not believe so. Very few mitoses could be found —-—not nearly enough to account for so extensive and profuse a proliferation. Those mitoses Whose spindle plane could be determined were parallel with the surface and no cells could be observed being pinched off from the ectoderm. The thickening of the ectoderm is accounted for by its rapid expansion, and it is to be expected that the rapid proliferation and condensation of the mesenchyme should bring it into contact with the ectoderm. Later, in embryos of 16 to 18 somites, the boundary between ectoderm and underlying mesenchyme becomes absolutely definite, except in the case of oblique sections, when it becomes difficult to delimit the ectoderm of the gill arch from the mesenchyme. Appearances figured by Veit and Esch (’22, fig. 7 as) are believed to be due simply to the adhesion of mesenchyme to the epidermis, although it is conceivable and not unlikely that occasionally cells may be detached from the ectoderm to be added to the mesenchyme.  
while cells might occasionally be found which seemed to be
 
in process of being added to the mesenchyme, there was no
 
evidence of a migration of the crest into the branchial arch.
 
The acoustico—facial ganglionic mass is always compact, and
 
it was possible to observe the condensation of the mesenchyme
 
of the second arch independent of any neural—crest contribution in a close series of embryos from 8 to 12 somites. The
 
postotic neural crest. is very diffuse in embryos of 9, 16, and
 
17 somites and no definite boundary between crest and mesenchyme can be set. The mesenchyme of the third and fourth
 
arches becomes very compact some time before the neural
 
crest has migrated ventrally to their upper levels, the condensation occurring independent of any neural-crest contribution. When the neural crest has finally reached its ventral
 
limit at the dorsal levelsof the arches in question, the definite
 
ganglionic masses appear for the IX and X nerves, the arch
 
mesoderm is too compact to permit of any further ventral
 
migration of the neural crest, and no such migration could
 
be observed. Celestine da Costa (’23, p. 519) comes to a
 
similar conclusion from his study of the guinea-pig.
 
  
. Veit (’f2.2) has also expressed the belief that the mesecto—
+
==Summary And Conclusions==
derm is proliferated from the epidermis over the entire body.
 
Appearances such as are pictured in his figures (7, 7 a, 8, 8 a)
 
were observed in the rat, but are, I believe, to be differently
 
interpreted. The first indication of the formation of a gill
 
arch is a thickening of the overlying ectoderm. The condensation of the mesenchyme proceeds rapidly and the compacted
 
mesenchyme comes to lie closely against the thickened branchial ectoderm——-so closely, in fact, that the line of separation
 
is often obliterated, as one might expect when a closely compacted mass of cells comes in contact with an epithelium with
 
no well-defined basement membrane. The difficulties of interpretation, as \Veigner (’O1) has observed, are increased
 
by the curvature of an arch, so that sections are more than
 
likely to be oblique. One gains the impression that possibly
 
NEURAL FOLDS AND CRANIAL GANGLIA on RAT 115
 
  
the mesenchyme is condensing under the influence of the ectoderm, or, conversely, that the condensation and expansion of
+
1. The development of the neural folds of the rat is de~ scribed with the following results:
the arch mesenchyme are responsible for the thickening of
 
the overlying epithelium, as Harrison (’18) found to be the
 
case in the limb bud. A striking illustration of the above was
 
seen in rat embryos of 11 to 14 somites. Here the territory
 
of the prospective third and fourth arches is indicated by an
 
extensive area of thickened ectoderm which extends caudal
 
to the second gill cleft almost to the border of the first somite.
 
It extends ventrally a short distance over the pericardial region. Immediately beneath this area of thickened brancl1iogenetic ectoderm one finds the mesenchyme condensing and
 
in close contact with the overlying epidermis, the condensation
 
extending over the pericardium. Now the question arises, are
 
these mesenchymal cells being proliferated from the ectoderm? I do not believe so. Very few mitoses could be found
 
—-—not nearly enough to account for so extensive and profuse
 
a proliferation. Those mitoses Whose spindle plane could
 
be determined were parallel with the surface and no cells
 
could be observed being pinched off from the ectoderm. The
 
thickening of the ectoderm is accounted for by its rapid expansion, and it is to be expected that the rapid proliferation
 
and condensation of the mesenchyme should bring it into contact with the ectoderm. Later, in embryos of 16 to 18 somites,
 
the boundary between ectoderm and underlying mesenchyme
 
becomes absolutely definite, except in the case of oblique sections, when it becomes difficult to delimit the ectoderm of the
 
gill arch from the mesenchyme. Appearances figured by
 
Veit and Esch (’22, fig. 7 as) are believed to be due simply
 
to the adhesion of mesenchyme to the epidermis, although it
 
is conceivable and not unlikely that occasionally cells may be
 
detached from the ectoderm to be added to the mesenchyme.
 
116 HOWARD B. ADELMANN
 
  
SUMMARY AND CONCLUSIONS
+
a. The rat ultimately develops seven typical rhombomeres. Rhombomere 1 is the cerebellar rhombomere. Rhombomere 2 is related to the trigeminal nerve, but rhombomere 3 is free of nervous connections. The facial nerve is attached to rhombomere 4 and the otic placode corresponds to rhombo— mere 5. Rhombomeres 6 and 7 are related to the IX-X anlage.
  
1. The development of the neural folds of the rat is de~
+
b. The expansion for the otic rhombomere appears earliest, i.e., at 3 somites.
scribed with the following results:
 
  
a. The rat ultimately develops seven typical rhombomeres.
+
c. A prominent sulcus (preotic sulcus)‘ lying in front of the otic rhombomeric expansion marks the site of rhombomere 3.
Rhombomere 1 is the cerebellar rhombomere. Rhombomere
 
2 is related to the trigeminal nerve, but rhombomere 3 is free
 
of nervous connections. The facial nerve is attached to
 
rhombomere 4 and the otic placode corresponds to rhombo—
 
mere 5. Rhombomeres 6 and 7 are related to the IX-X
 
anlage.
 
  
b. The expansion for the otic rhombomere appears earliest,
+
d. Rhombomere A1 appears anterior to rhombomere 3 at about 9 somites. Its anterior edge marks the cephalic limit of the hindbrain. It early gives evidence of its double value and soon divides into rhombomeres 1 and 2.
i.e., at 3 somites.
 
  
c. A prominent sulcus (preotic sulcus)‘ lying in front of the
+
e. Rhombomeres 5, 6, and 7 differentiate in order caudal to rhombomere 4. It is doubtful if a primary rhombomere C can be recognized.
otic rhombomeric expansion marks the site of rhombomere 3.
 
  
d. Rhombomere A1 appears anterior to rhombomere 3 at
+
f. The recognition of primary rhombomeres is of doubtful value morphologically, but convenient for descriptive purposes.
about 9 somites. Its anterior edge marks the cephalic limit
 
of the hindbrain. It early gives evidence of its double value
 
and soon divides into rhombomeres 1 and 2.
 
  
c. Rhombomeres 5, 6, and 7 differentiate in order caudal
+
g. The midbrain territory lies just caudal to the pronounced rostral flexure of the neural plate. Its floor is more restricted in extent than the more lateral expanded portions.
to rhombomere 4. It is doubtful if a primary rhombomere
 
C can be recognized.
 
  
f. The recognition of primary rhombomeres is of doubtful
+
h. While it is impossible to carry the delimitation of regions into earlier stages than 4 or 5 somites, the conclusion that forebrain and midbrain are differentiated from a relatively small anterior portion of the neural plate seems justified.
value morphologically, but convenient for descriptive purposes.
 
  
g. The midbrain territory lies just caudal to the pronounced
+
i. Two more or less variable mesencephalic ‘segments’ are recognized.
rostral flexure of the neural plate. Its floor is more restricted
 
in extent than the more lateral expanded portions.
 
  
12. While it is impossible to carry the delimitation of regions
+
j. Rhombomeres are regarded as expressions of a combination of growth factors——~rapid forward growth in a confined space and localized expansion looking forward to the development of the sensory mechanisms of the hindbrain. They probably have no phylogenetic import.  
into earlier stages than 4 or 5 somites, the conclusion that
 
forebrain and midbrain are differentiated from a relatively
 
small anterior portion of the neural plate seems justified.
 
  
2'. Two more or less variable mesencephalic ‘segments’ are
 
recognized.
 
  
j. Rhombomeres are regarded as expressions of a combination of growth factors——~rapid forward growth in a confined
+
2. The neural crest is associated primarily with the neural plate, but whether it proliferates from the edges of the neural folds or from the roof of the neural tube seems to depend entirely upon the time of its appearance with respect to the differentiation and closure of the neural folds. Both conditions are found in the same embryo.
space and localized expansion looking forward to the development of the sensory mechanisms of the hindbrain. They
 
probably have no phylogenetic import.
 
NEURAL rows AND CRANIAL GANGLIA or BAT 117
 
  
2. The neural crest is associated primarily with the neural
+
3. In the head region there are three neural—crest proliferations which give rise to the ganglia of the V, VILVIII, IX—X nerves. The proliferation for the fifth ganglion (rostral neural crest) extends along the territory of the prospective midbrain and rhombomere A1 of the 4- to 5~somi’te embryo. The proliferation for the VILVIII ganglion appears at 7 to 8 somites and is separated from the rostral neural crest in the territory of the preotic sulcus, and there is also a break in the neural crest between it and the IX-X anlage. The IX—X anlage can first be recognized at 8 somites, but it is not then in a state of active proliferation. It is from the first continuous with the spinal neural crest, which is actively proliferating at 8 somites.
plate, but whether it proliferates from the edges of the neural
 
folds or from the roof of the neural tube seems to depend
 
entirely upon the time of its appearance with respect to the
 
differentiation and closure of the neural folds. Both conditions are found in the same embryo.
 
  
3. In the head region there are three neural—crest proliferations which give rise to the ganglia of the V, VILVIII,
+
4. The following conclusions are reached in regard to the development of the trigeminus ganglion:
IX—X nerves. The proliferation for the fifth ganglion (rostral neural crest) extends along the territory of the prospective midbrain and rhombomere A1 of the 4- to 5~somi’te
 
embryo. The proliferation for the VILVIII ganglion appears at 7 to 8 somites and is separated from the rostral
 
neural crest in the territory of the preotic sulcus, and there
 
is also a break in the neural crest between it and the IX-X
 
anlage. The IX—X anlage can first be recognized at 8 somites,
 
but it is not then in a state of active proliferation. It is
 
from the first continuous with the spinal neural crest, which
 
is actively proliferating at 8 somites.
 
  
4. The following conclusions are reached in regard to the
+
a. The rostral neural crest (V anlage) soon loses its attachment to the midbrain and at 8 somites forms a mass attached to the neural plate for a short distance ahead of the preotic sulcus. The anterior portion of the rostral neural crest does not degenerate. There are three remaining possibilities as to its fate, viz.: 1) that it very quickly becomes diffuse, mingles with the mesenchyme, loses its specificity and individuality, and shares the fate of the mesenchymal cells in that region; 2) that it loses its attachment to the midbrain, becomes diffuse, and later recondenses to form the ramus ophthalmicus, or, 3) that the neural folds expand away from it, leaving it behind as the V ganglionic mass. The last is deemed most probable.
development of the trigeminus ganglion:
 
  
a. The rostral neural crest (V anlage) soon loses its attachment to the midbrain and at 8 somites forms a mass attached
+
b. The ophthalmic ramus is not placodal in origin, since no placodal contacts can be found. It grows forward from the main ganglionic mass of trigeminus, keeping pace with the growth shiftings coincident with the establishment of the maxillary region, following closely the consequent shiftings of the optic vesicle.
to the neural plate for a short distance ahead of the preotic
 
sulcus. The anterior portion of the rostral neural crest does
 
not degenerate. There are three remaining possibilities as
 
to its fate, viz.: 1) that it very quickly becomes diffuse,
 
mingles with the mesenchyme, loses its specificity and individuality, and shares the fate of the mesenchymal cells in
 
that region; 2) that it loses its attachment to the midbrain,
 
becomes diffuse, and later recondenses to form the ramus ophthalmicus, or, 3) that the neural folds expand away from it,
 
leaving it behind as the V ganglionic mass. The last is deemed
 
most probable.
 
  
b. The ophthalmic ramus is not placodal in origin, since no
+
0. The main ganglionic mass of the trigeminus has a broad area of contact with the epidermis, but apparently receives no contributions from it. No definite placodal structure is found at the place of contact and no cells could be observed actually proliferating from the ectoderm in a large number of embryos.
placodal contacts can be found. It grows forward from the
 
main ganglionic mass of trigeminus, keeping pace with the
 
118 HOWARD B. ADELMANN
 
  
growth shiftings coincident with the establishment of the
+
5. There is a well-developed epibranchial placode related to the VII ganglion.
maxillary region, following closely the consequent shiftings
 
of the optic vesicle.
 
  
0. The main ganglionic mass of the trigeminus has a broad
+
6. The VIII ganglion is a derivative of the common acoustico-facial ganglionic mass, owing its origin entirely to neuralcrest proliferation.
area of contact with the epidermis, but apparently receives
 
no contributions from it. No definite placodal structure is
 
found at the place of contact and no cells could be observed
 
actually proliferating from the ectoderm in a large number
 
of embryos.
 
  
5. There is a well-developed epibranchial placode related
+
7. The IX-X ganglia are derived from a continuous postotic neural-crest proliferation continuous with the spinal neural crest.
to the VII ganglion.
 
  
6. The VIII ganglion is a derivative of the common acoustico-facial ganglionic mass, owing its origin entirely to neuralcrest proliferation.
+
8. The IX ganglion establishes light contact with the ectoderm, but its epibranchial placode is less clearly delimited than that of VII. The ductus branchialis II is for a long time in contact with its lower pole.
  
7. The IX-X ganglia are derived from a continuous postotic
+
9. The ganglion nodosum makes contact with the ectoderm on the dorsal edge of the cervical sinus and the cervical vesicle remains in contact with its lower pole for some time. The growth of the ganglion nodosum caudally over the territory caudal to the third branchial pouch is thought to be influenced by the expansion of the caudal pharyngeal complex.
neural-crest proliferation continuous with the spinal neural
 
crest.
 
  
8. The IX ganglion establishes light contact with the ectoderm, but its epibranchial placode is less clearly delimited
+
10. Since no placodal contributions could be made out, the conclusion is reach ed that the ganglia of the V, VII-VIII, and IX-X nerves are entirely of neural-crest origin.
than that of VII. The ductus branchialis II is for a long
 
time in contact with its lower pole.
 
  
9. The ganglion nodosum makes contact with the ectoderm
+
11. No organized contribution of ectodermal cells to the mesoderm occurs in the rat. .
on the dorsal edge of the cervical sinus and the cervical vesicle
 
remains in contact with its lower pole for some time. The
 
growth of the ganglion nodosum caudally over the territory
 
caudal to the third branchial pouch is thought to be influenced
 
by the expansion of the caudal pharyngeal complex.
 
  
10. Since no placodal contributions could be made out, the
+
12. The writer is inclined to regard the placodes as rudimentary sense organs differentiated under the influence of the ganglionic contact with the ectoderm. This does not imply that they have ever been functional in the course of phylogeny.  
conclusion is reach ed that the ganglia of the V, VII-VIII, and
 
IX-X nerves are entirely of neural-crest origin.
 
  
11. N o organized contribution of ectoderm