ABSTRACT
AT the time when Peter reviewed the subject of secondary mammalian palate embryology (Peter, 1924), it had been established that the secondary palate is derived from two shelves of tissue, which originate in the dorsolateral wall of the oral cavity and grow downward toward the lower jaw, passing lateral to the tongue. The shelves subsequently come to lie in the transverse plane dorsal to the tongue, where they fuse with the nasal septum and with each other, thus forming the roof of the mouth. The problem in this sequence is how the shelves manage to reach a position dorsal to the tongue after having had the tongue tightly wedged between them. Various mechanisms postulated to explain this change of position were reviewed by Lazzaro in 1940. The theories fell into three categories: (1) external forces such as muscular pressure by the tongue, (2) growth changes involving regression of the ventral portion and an outgrowth in the horizontal plane; and (3) a rapid rotation of the shelves due to some intrinsic force. Lazzaro considered the evidence to be in favour of rapid movement, and cited the cases of embryos with one shelf vertical and the other shelf horizontal as examples of a rapid transitional stage in shelf rotation. He considered his own work to indicate a swelling in the shelves due to a considerable increase of intercellular substance in the embryo’s connective tissue. The result was visualized as a type of erection which would cause the shelves to rise when the obstacle of the tongue was removed. Lazzaro considered it universally recognized that the release of the tongue took place before anything else, even if the mechanism of release was ‘slightly obscure’. The mechanisms suggested were: (1) a lowering of the mandible and tongue, (2) a forward displacement of the tongue, (3) a lifting of the roof of the oral cavity, (4) changes in form of the tongue due to muscular development, and (5) muscular movements of the tongue.
INTRODUCTION
Thus evidence on the physiology of palate closure is scant, and even the morphology of closure is incompletely described. A further study of normal palate development in mammals is desirable to provide a standard for comparison with spontaneous and experimentally induced palate abnormalities. The work presented here was carried out in conjunction with a study on cortisone-induced cleft palate, and we expect to publish the embryology of this abnormality in the near future.
MATERIALS AND METHODS
The mice used in this experiment were of strains A/Jax, C57BL/Fr, and DBA/1. They were fed Purina Fox Chow or Derwood Farms Mouse Pellets. Female mice were mated overnight when their vaginal smears indicated estrus, and were isolated the next morning if a vaginal plug was present. When calculating the age of embryos, fertilization was assumed to take place at 2 a.m. (Snell et al., 1940). This system of timing is probably accurate only to within 8 hours (Lewis & Wright, 1935; Braden & Austin, 1954). Females were killed at various ages, and the embryos were placed in Bouin’s fixative while still within the uterus, stored in 70 per cent, alcohol, then dissected out for gross observation or histological sectioning. The palate region can be seen readily through a dissecting microscope by removing the embryo’s head and cutting away its lower jaw (see Plates 1,2, and 3).
Live embryos were studied while clamped between two foam-rubber pads, being attached to their anaesthetized mother by their umbilical cords. The palatine shelves were observed by opening the mouth and moving the tongue with glass instruments.
RESULTS
When embryos are fixed around the time of palate closure, the embryos from one uterus will display a variety of morphological states of the palate, depending on the point at which each embryo is halted during the continuous process of palate development. These states have an obvious sequence from open palate to closed palate. For convenience, the various morphological states have been grouped into seven stages which are presented below in sequence. The stages have been artificially imposed on a continuous process, and their limits are of course entirely arbitrary. Figs. A and B in Plate 1 have been included to show the tongue in its normal position. Although the tongue is missing in the other figures, its position was observed during dissection. In the following descriptions of palatine shelf position, ‘vertical’ refers to the sagittal plane, and ‘horizontal’ refers to a plane parallel to a transverse section.
The palate shelves originate anteriorly at the point where the nasal pits join the upper part of the oral cavity and they terminate posteriorly as free knobs.
Stage 1. At the beginning of the period covered by our observations the primary palate and alveolus have already formed. The medial portion of each palatine shelf lies in the vertical plane, while laterally the shelf lies in the horizontal plane (Plate 1, figs. C and D). The tongue at first lies completely between the shelves, but later an increase in width of the posterior part of the tongue causes it to spread out across the shelves so that an observer studying the area from a ventral position could not see the posterior ends of the shelves (Plate 1, fig. A) unless the tongue were removed (Plate 1, fig. C). Thus the tongue lies ventral to the shelves posteriorly and medial to the shelves anteriorly. At the place where it goes from ventral to medial it makes a grooved impression on the sides of the shelves. When the tongue is dissected out, the grooved impression can be seen running dorso-anteriorly on the side of each shelf (Plate 1, fig. C; Plate 2, fig. E; and right shelf in Plate 2, fig. G).
The medial sides of the shelves are not perfectly vertical, but slope somewhat. Posterior to the above-mentioned groove, the shelves slope at an obtuse angle away from the roof of the nasal cavity, so that the medial surfaces are visible when viewed from a ventral position. Anterior to the groove the shelves slope at an acute angle and thus tend to cup the tongue between them; and when the tongue is removed, the medial sides of the shelves are not fully visible to an observer looking towards the roof of the mouth (Plate 1, fig. C; Plate 2, fig. E; and right shelf in Plate 2, fig. G).
Stage 2. In Plate 2, fig. E, the groove has moved to a point midway along the shelves. This is stage 2. Plate 2, fig. F and the right shelf in Plate 4, fig. O show cross-sections through the grooves. The medial bulges of shelf tissue are the anterior ends of the obtuse-angled slopes, and the ventral bulges are the posterior ends of the acute-angled slopes.
Stage 3. Shelf activity is not necessarily synchronized bilaterally, and a condition can be seen (Plate 1, fig. B and Plate 2, fig. G) where one shelf lies completely dorsal to the tongue while the other shelf does not. This is defined as stage 3. It should be clear from Plate 2, fig. H that the tongue has been depressed sufficiently on one side to make room for the horizontal shelf, but it seems unlikely that any actual ‘rotation’ of the shelf from the vertical to the horizontal position could have taken place.
Stage 4. When both shelves have assumed a horizontal position and lie dorsal to the tongue (Plate 3, figs. I, J), they are at first separated by a small space which is soon bridged by further flattening of the shelves (compare flattening left shelf with flattened right shelf in Plate 3, fig. I).
Stage 5 is the stage at which fusion of shelf epithelium begins.
Stages 6 and 7. Epithelial fusion spreads anteriorly and posteriorly (stage 6, Plate 3, figs. K, L) until the shelves are fused throughout their length (stage 7).
Preliminary to further analysis it seemed desirable to establish (1) the time in gestation at which these changes occur, (2) how long they take, and (3) whether the assigned stages do represent a chronological sequence.
The first method used to estimate the age of the embryo at palate closure was based on the calculated time of conception, and the results are shown in Table 1. In strain A/Jax, palates were open (stage 1) in all embryos examined from day 13 / 8 ( = 13 days, 8 hours after conception) to day 14/16. However, at 14 /18, nine embryos had open palates and one embryo had a palate with one shelf in the horizontal position (stage 3). At 14/20, a litter is listed in which six embryos had closed palates, thus tending to upset the trend suggested by the other figures in the table, and raising the problem of whether or not this litter was mistimed. In general, it can be seen that, for strain A/Jax, closing stages (stages 2–6) appear in Table 1 with increasing frequency after the beginning of the 15th day. Thus the palate tends to close during late day 14 and early day 15 in strain A/Jax embryos. Similarly, the palate appears to close in C57BL embryos from early to late day 14, and in DBA embryos during late day 14 (Table 1). The great variability shown in this table was matched by variability in the developmental stages of embryos within the same uterus and from uteri of animals thought to have been killed at the same stage of pregnancy. (Rarely, all the embryos in one uterus were at least a day younger or older, according to Grüneberg’s (1943) criteria, than they should have been according to the calculated time of conception, in which case they were not included in Tables 1 and 2.)
Because of the irregularities in the apparent chronological age of these embryos, it was thought that a more satisfactory approach would be to estimate age by the developmental features of the embryo as described by Grüneberg (1943), modified by a further subdivision of the morphological states (e.g. webbed paws were classed as webbed,
webbed,
webbed). A ‘morphological age’ for an embryo was calculated by recording the condition of its front paws, hind paws, hair follicles, ears, and eyes, and then adding up the values that had been assigned to these conditions (Walker, 1954).
Morphological age is compared with palate stage in Table 2. The positive correlation between the progressing morphology of the embryo and the successive palate stages (Table 2) is much greater than the correlation of palate stage with assumed chronological age (Table 1). The existence of such a correlation suggests that the stages of palate closure as defined do indeed form a chronological sequence. With regard to time of closure, palate stages 2–6 start in strain A/Jax at morphological rating 10–11, whereas in strain C57BL closure starts as early as at morphological rating 5–6. In general, the palate closes sooner (by morphological age) in C57BL embryos than in A/Jax embryos, with DBA embryos being intermediate, although overlapping occurs between all three strains. Thus the strain differences suggested by chronological timing are clearer when morphological timing is used.
Having established a probable sequence of palate conditions during closure, the next step was to find the force behind this tissue movement. Mitoses can be seen in shelf tissue, but they occur too infrequently to play a significant role during the few hours (see p. 185) occupied by the closing process. Nevertheless, the head and palatine shelves were measured in fifty-two embryos; and the results were in harmony with the interpretation of tissue movement without appreciable growth. There was still the question of whether the force resided in the shelves or, for instance, in the tongue. This question was settled by the following experiments.
While examining living embryos it was noticed that the palatine shelves would often change position if the tongue was lifted. A number of genetically heterogeneous and C57BL embryos were examined on the 14th day post-conception, and it was found that the shelves often proceeded from the vertical to the horizontal position within anywhere from several seconds to a minute after the tongue was lifted. The actual movement was rapid, requiring 1 or 2 seconds; and the lack of contrast in the living embryo’s mouth tissues made it very difficult to follow the movement visually. When a technique was finally evolved that produced minimal damage to the embryo and allowed clear observation of the palate region, a study was done in which observations on the living embryos were checked by fixing and re-examining these embryos. A discrepancy between the observations in vivo and on fixed material was found to have been caused by reversibility of shelf movement. Sometimes, shelves that had become horizontal when the tongue was lifted, reverted to the vertical position when the mouth closed again, as would happen when the embryo was transferred to fixative.
It was found that the anterior portions of the shelves developed the ability to change shape sooner than the posterior portions. This agrees with the impression gained from morphological studies that during closure (especially stage 4) the shelves flatten out completely in the horizontal plane sooner anteriorly than posteriorly. Yet the posterior portions of the shelves are normally the first to slide dorsal to the tongue despite having less potential for movement than the anterior portions. They apparently do so because the tongue crosses the extreme posterior ends of the shelves ventrally, thus producing the arched condition that later spreads like a wave along the shelves and carries them around the tongue.
It is clear that the shelves have some potentiality for changing shape long before they would normally do so; for example, shelf movement was seen in an A/Jax embryo with a morphological rating of 5 at day 14/10, whereas palate closure is not normally expected until an A/Jax embryo has achieved a morphological rating of 10 or higher on late day 14.
Two embryos were put into 70 per cent, alcohol and examined 30 minutes later to observe the effects of alcohol on shelf tissue. Unexpectedly, it was discovered that the shelves still went from the vertical to the horizontal position after the tongue was lifted, despite having been immersed in alcohol. Furthermore, the shelves could be returned to the vertical position by pressing the mouth closed, and then caused to become horizontal again by lifting the tongue. Considering the resistance of shelf movement to alcohol, it is not surprising that there was no evidence of shelf movement having been affected by interrupted blood-supply or by mutilation of the embryo. Of course, continued treatment with alcohol eventually hardened the tissue and halted all shelf movement.
It was noticed, both with alcohol treated and untreated embryos, that the shelves could often be induced to spring back and forth between their two customary positions (horizontal and vertical), but that they would not maintain an intermediate position. This suggested a quality of elasticity; and also that two alternative stable positions existed during a certain stage of palate development, this stage falling between an early palate condition in which only the vertical position could exist and a final condition in which the shelves would force their way into a horizontal position. However, this procedure (used on embryos of appropriate age) would work only a few times with each shelf, after which the shelf would tend to remain in one of the two possible positions.
Examination of the palatine shelves in hematoxylin and eosin sections offers little to clarify the problem of what produces the force responsible for shelf movement. The palatine shelves consist of a core of loose connective tissue covered by epithelium (Plate 4, fig. N). Bone formation is occurring in the tissue proximal to the base of the shelves, and this area of bone formation parallels the shelves for a considerable distance antero-posteriorly by the 15th day. No bony projections enter the palatine shelves until about the 17th day, and even then the bone has only entered along a short portion of the palate, and has not extended completely across it.
In an effort to discover substances in shelf connective tissue that could account for shelf movement, various stains were tried. Toluidine blue gave a metachromatic reaction, but instability of the metachromasia made the study of its distribution difficult. Orcein and Verhoeff’s hematoxylin both stained fibres throughout the connective tissue of the head. Of the other staining procedures used (Masson’s trichrome, periodic acid-Schiff, and aldehyde-fuchsin), the only one that gave a reaction with any degree of specificity for shelf tissue was Gomori’s aldehyde-fuchsin stain (Gomori, 1950). The aldehyde-fuchsin appeared to be staining a network of fibres throughout the shelves (Plate 4, fig. M). Areas of cartilage and bone formation were stained, as were connective tissue fibres in some areas other than the palate; but in many regions of the head the connective tissue fibres did not stain at all with aldehyde-fuchsin (Plate 4, fig. O). Incubation in hyaluronidase (Wyeth Laboratories) for 18 hours (Bunting, 1950) rendered the shelf connective tissue unstainable with aldehyde-fuchsin, so the reacting material can be tentatively identified as an acid mucopolysaccharide.
DISCUSSION
Morphology
Published information on the embryonic development of the secondary mammalian palate appears reliable except for the phase when the palate shelves change from a vertical to a horizontal position. Enough data have been assembled (Peter, 1924; Lazzaro, 1940) to indicate that the transition is too rapid to be due primarily to growth, but no satisfactory account has been offered of the mechanism by which the shelves move. In the present paper, a sequence of morphological conditions of the palate has been presented to describe how the palatine shelves can change position. Briefly, this sequence starts with the anterior portion of the tongue lying between the palatine shelves, which are hanging vertically from the roof of the mouth, and with the broad posterior portion of the tongue crossing the posterior ends of the shelves ventrally (Text-fig. 1). The change from a vertical to a horizontal position appears to start posteriorly by a bulging of the medial wall in over the tongue and an accompanying retraction of the ventral portion of the shelf (Text-fig. 2). This process proceeds in a wavelike manner anteriorly, until the whole shelf lies dorsal to the tongue (Text-fig. 3). Sometimes (perhaps always) the wave of closure passes along one shelf more rapidly than the other, thus giving rise to the condition seen by other authors (e.g. Lazzaro, 1940) of one shelf being horizontal while the other is still vertical (Text-fig. 3). When the shelves have become horizontal (Text-fig. 4) they flatten until in contact with each other (Text-fig. 5) and then fuse (Text-fig. 6).
In each figure the drawing on the left is a schematic representation of an embryo head with lower jaw and tongue removed; the drawings on the right (A and B) represent crosssections taken from levels indicated by the lines connecting with the gross specimens. The arrows in Text-fig. 2B indicate the direction of tissue movement. These diagrams show the process of palate closure from a condition where the shelves are vertical (Text-fig. 1) through shelf movement (Text-figs. 2, 3, and 4) to a condition where the shelves are horizontal and fusing (Text-figs. 5 and 6).
In each figure the drawing on the left is a schematic representation of an embryo head with lower jaw and tongue removed; the drawings on the right (A and B) represent crosssections taken from levels indicated by the lines connecting with the gross specimens. The arrows in Text-fig. 2B indicate the direction of tissue movement. These diagrams show the process of palate closure from a condition where the shelves are vertical (Text-fig. 1) through shelf movement (Text-figs. 2, 3, and 4) to a condition where the shelves are horizontal and fusing (Text-figs. 5 and 6).
The passive role played by the tongue is emphasized by its position in histological sections of palates in which one shelf is horizontal and the other vertical (Plate 2, fig. H). There is no sign of the tongue having dropped to allow the shelves to become horizontal, as had been suggested by numerous authors (Peter, 1924; Lazzaro, 1940). On the contrary, the general position of the tongue remains constant, while only its shape changes, apparently in response to the change in shelf position. The side of the tongue (Plate 2, fig. H) adjacent to the vertical shelf is thick and compressed laterally, whereas the part of the tongue ventral to the horizontal shelf is more extended laterally and consequently not as thick.
Physiology
When the tongue is experimentally displaced in a living embryo of appropriate age, the shelves will move from a vertical to a horizontal position. The force within the shelves which produces this movement is presumably the same force that drives the shelves dorsal to the tongue. This shifting of shelf tissue begins posteriorly and proceeds anteriorly and is responsible for the forward movement of the grooved condition of the shelves. Finally, all shelf tissue becomes horizontal and the grooved condition disappears.
In the 20-hour period from day 14/8 to day 15/4 during which C57BL embryos were killed (Table 1), the number of closing palates seen should have been proportional to the length of time required for closing, if all palates underwent closure sometime during this period, and if the embryos were collected in equal numbers at equally spaced periods. Since these conditions were not met, the following estimation of palate closure duration is necessarily a very rough approximation of the true rate. Based on Table 1 the time required for shelf movement (stages 2, 3, and 4) is about 3 hours, and the time required for fusion (stages 5 and 6) is about 6 hours. The relatively lengthy period required for shelf movement during normal development is probably caused by the tongue’s resistance to displacement, since the shelves can change position within a minute when the tongue is experimentally displaced. Experimentally induced shelf movement is reversible during a developmental stage preceding palate closure, indicating that a gradual shift of maximum stability from vertical to horizontal alignment occurs.
The idea of a rotation of the shelves (Peter, 1924; Lazzaro, 1940) is unsatisfactory, since histological study shows that the relationship of tongue to shelves does not suggest such a movement (Plate 2, fig. H), and because an intermediate shelf condition not involving rotation has been found (Plate 2, fig. F). Also, the proponents of a growth mechanism (e.g. Pons-Tortella, 1937) were incorrect regarding the force (i.e. growth and resorption) bringing about the change in shape, but they were close to being correct in their theory of how the shelves by-passed the tongue (i.e. a ‘resorption’ of the vertically aligned portion and an outward ‘growth’ of the medial wall).
Pons-Tortella (1937) reported a human embryo with palatine shelves that were in a horizontal plane anteriorly but still vertical posteriorly. He interpreted this to be a part of the palate closure process in which a transformation of the shelves started anteriorly and proceeded posteriorly. Lazzaro (1940) discounted this finding because the embryo was damaged. Judging from Pons-Tortella’s photographs, the embryo was certainly in bad condition. Yet the position of the palatine shelves was comparable to shelf position in mouse embryos whose tongues had been experimentally moved at a certain developmental stage when only the anterior portions of the shelves would stay horizontal. Thus Pons-Tortella’s human embryo, which had probably received rough treatment before it could be preserved, may have had its tongue dislodged from between the shelves, allowing precocious shelf movement. The mechanism postulated by Pons-Tortella to explain shelf-transformation was resorption and growth. Although growth is presumably necessary for the building up of the force causing palate shelf movement, it certainly need not be operative during the actual movement, since the shelves have been observed in living embryos to go from a vertical to a horizontal position within one minute, with the shelves then touching each other throughout most of their length.
The metachromasia and the affinity for aldehyde-fuchsin displayed by shelf connective tissue suggests two possible mechanisms for shelf mobility. Firstly, Lazzaro’s theory of tissue turgor could implicate hyaluronic acid as a water barrier, and this acid could account for the metachromasia. However, the ineffectiveness of 70 per cent, alcohol’s dehydrating action on shelf movement is evidence against a turgor theory. In contrast, elastic fibres are relatively resistant to alcohol, and aldehyde-fuchsin is considered to be an elastic tissue stain (Gomori, 1950); also the presence of metachromasia is consistent with an interpretation of elastic fibres (Scott & Clayton, 1953). This suggests the second theory, namely, shelf movement is due to the tensions of a developing network of elastic fibres. The results of in vivo experiments on palatine shelves correlates well with a theory of an elastic fibre network which, due to changes in growth patterns, could gradually become placed under a tension. Such a system could be utilized by the embryo on other occasions. For example, the aldehyde-fuchsin staining of the maxillary and nasal processes of a day 11/10 embryo is again suggestive of an elastic fibre network (or even an elastic membrane in some areas). The fibres could serve to stiffen the processes and cause the maxillary process to bend towards the nasal process and press tightly against it in preparation for fusion.
Strain differences
A strain difference between A/Jax and C57BL embryos appears to exist in the embryonic age (measured from assumed time of conception) at which closure takes place, with C57BL palates closing about 10–12 hours before A/Jax palates (Table 1) and DBA embryos being intermediate in this respect. Because of the variability in relation of palate stage to chronological age, this difference is hard to demonstrate statistically. The less variable relation of morphological age and palate stage makes this a more suitable basis for analyzing strain differences.
When morphological ratings are tabulated with palate stages, again the trend seen for A/Jax embryos differs from the one for C57BL. (A statistical comparison of these trends is complicated by the fact that stage 1 extends back indefinitely, and stage 7 extends onward in development indefinitely, so the only significant points in these stages are where stage 1 ends and where stage 7 begins.) The most satisfactory test of the difference would be one that utilized all the characteristics of the trends; but for convenience, a much simpler test can be used. In Table 2 it is seen that very few cases of stage 5 (shelves touching) occur, so if this stage is eliminated from calculations, two distinct groups are obtained: palates open, stages 1–4; and palates fusing or closed, stages 6 and 7. If C57BL palates usually close at an earlier developmental age than A/Jax palates, then at some particular morphological rating more palates should be closed in the one strain than in the other. It appears that by morphological rating 11, most C57BL palates have started fusing, whereas most A/Jax palates are still open. When the number of open palates and the number of closed palates up to rating 11 in A/Jax embryos (25:1) are compared with the same statistics in C57BL embryos (29:22) the two strains show a significant difference (χ2= 12 ·69, d.f.= 1, p < 0 ·001). In this test only positive ratings were used, to eliminate samples from early age groups (e.g. day 13). It is true that a difference in the number of young embryos in the Stage 1 group could bias the results; but since this bias would be in favour of open palates, and since the C57BL group carried more young embryos (11 below a rating of 5) than the A/Jax group (2 below 5), it is obvious that this does not contradict the conclusion. If a similar test is performed to see whether a larger proportion of the embryos have open palates than closed palates at morphological rating 11, and later, in A/Jax (11:24) than in C57BL embryos (0 28), a significant difference is again shown to exist (p = 0 · 0007 by Fisher’s exact method). In this latter test, old embryos in stage 7 were slightly more frequent in A / Jax, which would have favoured a nonsignificant result.
CONCLUSIONS
The secondary palate in the mouse closes by a rapid movement of the palatine shelves from an initial sagittal to a final transverse plane. This movement consists of a bulging of the medial wall and a regression of the ventral wall of each shelf, with the transformation proceeding in a wave-like motion from the posterior ends of the shelves to the anterior ends. The flow of palatine shelf tissue into the bulge of the dorsal-medial wall carries the shelf dorsal to the tongue and forces the latter into a more ventral position.
The shelves move by means of an internal force, which increases until it is sufficiently strong to drive the shelves dorsal to the tongue. Evidence is presented which suggests that this force resides in a network of elastic fibres in the connective tissue of the shelves.
The palate closes at an earlier developmental age in C57BL embryos than in A/Jax embryos, with time of closure in DBA embryos being intermediate.
ACKNOWLEDGEMENTS
The financial assistance of the Kate E. Taylor Fund of the Banting Research Foundation, and of the National Research Council of Canada, and of the Faculty of Medicine, McGill University (Blanche Hutchison Fund) is gratefully acknowledged.
REFERENCES
EXPLANATION OF PLATES
Figs. A and B. C57BL embryo heads with lower jaws removed showing palate stages 1 and 3 respectively. Magnification × 10.
Fig. C. A/Jax embryo head with lower jaw and tongue removed showing stage 1. Magnification × 10.
Fig. D. Cross-section at the level of the nasal septum showing the palatine shelves in a vertical position. Magnification × 33.
Figs. E and G. A/Jax embryo heads with lower jaws and tongues removed showing palate stages 2 and 3 respectively. Magnification × 33.
Fig. F. Cross-section at a level well back of the nasal septum (comparable to Text-fig. 2B), with lower jaw and tongue removed, showing the palatine shelves at stage 2. Magnification × 33.
Fig. H. Cross-section at the level of the nasal septum showing the palate at stage 3. Magnification × 33.
Figs. I and K. A/Jax strain embryo heads with lower jaw and tongue removed showing palate stages 4 and 6. Magnification × 10.
Figs. J and L. Cross-sections at the level of the nasal septum showing palates at stages 4 and 6 respectively (tongue and lower jaw removed in Fig. L). Magnification × 33.
Fig. M. Section of palate shelf stained with Gomori’s aldehyde-fuchsin. Magnification × 460. Compare with Fig. N.
Fig. N. Section of palatine shelf stained with hematoxylin and eosin. Magnification × 460.
Fig. O. Section of palatine shelf in the process of moving from the vertical to the horizontal position. Gomori’s aldehyde-fuchsin stain. Magnification × 56.