N I ervous system abnormalities comprise the largest single category of human developmental defects, and those resulting from central neuraxis tubular formation are most common among abnormalities associated with neuromotor defects. During periods of prenatal life when organ systems are developing most rapidly, the anläge of the developing system is particularly vulnerable to intrauterine ecological disturbances. The fourth week of human pregnancy is a critical period with reference to the occurrence of meningomyelocele, anencephalus, and related defects.
The objectives of this writing are to consider some dynamic aspects of early embryogenesis of the human central nervous system based upon our studies of 61 human embryos and young fetuses presenting a variety of central neuraxis defects. We also review aspects of neural tube cellular kinetics, as they are concerned with localized areas of abnormal growth in neural tube defects and we relate these data and the types of neural defects found to theories of the pathogenesis of meningomyelocele.
It is well known from study of populations of spontaneous abortions that the incidence of developmental malformations is much higher in this group than it is among full-term liveborn neonates.1'3'3'4 For example, chromosomal abnormalities are about 20 times more frequent, and malformations of the neural tube are six to eight times more frequent in abortuses than in a live newborn population.3'*
THE SPECIMENS STUDIED
Revision of the Japanese Eugenic Protection Law in 1952 allowed abortion for psychosocial and medical reasons. This program enabled one of us (HN) to collect more than 25,000 human embryos and young fetuses accompanied by maternal reproductive and medical data. This collaborative study was made possible by several opportunities for the authors to work together both in New York City and Kyoto, Japan.
From a population of approximately 4,000 specimens, we found 61 with external evidence of myelocele, anencephalus, and hydrocephalus, either singly or in various combinations. The specimens ranged from 30 to 115 days of gestational age (all but two were between 30 and 72 days of gestational age).
DETERMINATION OF GESTATIONAL AGE
The time at which developmental processes occur is an important but exceedingly difficult factor to measure accurately. However, in any study of rapidly changing dynamic processes of embryologie development, time is an important factor. Duration of development, commonly referred to as the gestational or ovulatory age of a developing embryo or fetus, is based on the presumed time of fertilizationgenerally calculated as 14 days after the first day of the last menstrual period. However, this calculation of gestational age assumes that every woman has a regular 28-day menstrual cycle.
Figure 1. Human embryo ovulation age in days related to developmental horizon (Streeter).
Figure 2. Frequency of spontaneous human abortion at various stages of pregnancy and gestational age of 61 specimens with central nervous system defect.
Figure 3. Classification of externally visible malformations of the central nervous system in 61 human embryos and young fetuses (5-16 weeks gestational age).
Another variable not taken into consideration, and one which cannot be eliminated, is the time required for the spermatozoon to reach the ovum (usually in the middle third of the uterine tube) and effect fertilization. This variable may introduce as much as 72 hours' difference in embryonic age. Such variables are not important with reference to total development, but in early embryos and young fetuses a few hours of additional development may produce a visible difference among specimens presumed to be the same age, based on ovulation time.
For many years embryologists have used the remarkable collection developed by Dr. George Streeter at the Carnegie Laboratory of Embryology5'e':·8 as a basis of comparison for their obtained human embryos. The specimens in this collection were derived mostly from spontaneous abortions, and gestational age was based upon menstrual history. Dr. Streeter recognized the difficulties of accurate timing of gestational age with respect to body length (crown-rump length in millimeters) and weight. He therefore tried to verify some stages of development and supplemented the collection with time-matched pregnancies of nonhuman primates. Dr. Streeter also realized that, in view of the time variable and the rapidity with which externally visible changes occurred within a single day during the period of the embryo (up to eight weeks) and young fetus (after eight weeks), recognition of developmental stages would be enhanced by the use of multiple criteria of morphologic development. He therefore referred to identifiable stages of development as "horizons" and described 23 horizons which ranged from the zygote (Horizon I) to a fetus about 48 days old (Horizon XXIII).
Figure 4. Diagrammatic representation of Ine circadian and morphologic aspects of spinal cord formation during the third and Iourth weeks of human embryonic development.
A-D. dorsal, and E-G, lateral views, and A'-G'. a schematic representation of transverse sections through embryos at various stages of development.
A-A', a lale presomite embryo (1.4 mm., 19 days) with a prominent neural plate (np) and paraxial mesoblast (pm). Notochord (nc) and dorsal aorta (da). (Modified after Davies.)
B-B', an early somite stage (20 days) showing three pairs of somites (s). neura! groove (ng), and the somitic mesobiast (srn). (Modified after !ngalls.)
C-C', a seven-somite embryo (2.2 mm., 22 days) with neural tube beginning to close at the levels of the first and second pair of somites. C' is a section through the area iusl in front of the first somite in which the neural folds (nf) are in very close proximity and the tube is not yet closed. (Modified after Payne.) Note the lateral cephalic pericardia! swelling (ps).
D-D'. a 10-somite embryo (23 days). The neural tube (nt) has closed from just behind the 10th somite to the otic placode (ot). (Modilied after Corner.)
E-E', a 14-somite embryo (2.6 mm.. 25 days) in which the neural tube closure is almost complete at ail levels with the exception of the anterior (an) and posterior (pn) neuropore. Intermediate mesoderm (im). (Modified after Heuser.)
F-F1, a 25-somite embryo (28 days) showing the separation and subsequent differentiation of a mesoblastic somite into a sclerotome (st). myotome (ml), and dermatome (dt) portion.
G-G", a 28-somite embryo (30 days) with completely differentiated spinal cord in which the nerve fibers of the ventral motor (vm) and dorsal sensory (ds) roots join to form the trunk of the spinal nerve (sn). which terminates and ramifies in a sensory receptor organ- the muscle (m).
In 1956, Witschi9 compared weight and length of a collection of embryos with Streeter's horizons. One of us10 has compared with Streeter's collection the estimated ovulation age, crown-rump length in millimeters, and body weight in milligrams of 672 embryos and fetuses derived from mothers with a regular cycle. It was concluded that, "crown-rump length and body weight are generally more reliable indicators of developmental state than clinically estimated age."10 Figure 1, derived from this study, shows the estimated ovulation age and developmental horizon in these three collections of human embryos and fetuses.
The specimens from which this report is derived are placed by gestational age under a curve of the percentage frequency of spontaneous human abortions which occur in successive weeks and days through the first 24 weeks of pregnancy (Figure 2). It is apparent from Figure 2 that the majority of the specimens studied ranged from 30 to 60 days of age and were removed from the uterus by dilatation and curettage at a time when the curve of spontaneous abortion frequency was rising rapidly. This implies that a number of our study specimens might have aborted spontaneously had the mother not requested therapeutic abortion. For each of the 61 abnormal specimens studied, an age-matched normal specimen was studied for comparative purposes.
The diagnostic categories of central nervous system disturbances found in the specimens are shown in Figure 3. The four most frequent categories of developmental disturbance in descending order were: spina bifida only, 30 per cent; uncomplicated hydrocephalus, 23 per cent; exencephaly only, 17 per cent; and exencephaly combined with spina bifida, 7 per cent. It is significant that only two examples of spina bifida associated with evidence of increased cerebrospinal fluid pressure were found.
Space limitation precludes detailed description of all the specimens. We have therefore selected several for brief discussion with reference to pathogenesis of meningomyelocele.
Figur· 5. Human embryo with a lumbosacral myelocele (Horizon 20, crown-rump lenglh 18 mm., weight 736 mg,, gestational age 44± days).
NORMAL PROCESS OF NEURAL TUBE CLOSURE
The brain and spinal cord are the first organs to be developed in the embryo. In Figure 4, the sequence of events which occurs in the formation of a tubular brain and spinal cord are illustrated. The most dynamic part of the process occurs during the twentyfirst to twenty-eighth days of gestation. A 21-day human embryo has a total length of about 1.5 mm. The spinal cord is developing rapidly; the cephalic end of the neuraxís has precociously developed and shows evidence of three enlargements which represent the three primary brain vesicles. At this stage of development, there are no cartilaginous formations which are precursors of the axial skeleton. There are, however, seven pairs of condensed masses of mesoderm on either side of the spinal cord, which are known as mesodermal somites. Note that at the level of the second and third mesodermal somite the free edges of the neural plate, which represent the thickened and depressed midline skin of the back, have started to fuse, forming a neural tube with a central canal. This initial area of neural tube closure lies approximately in the midthoracic region.
During the next seven days, the neural tube will close cephalically and caudally as though it were being zippered. During this time, 24 additional pairs of mesodermal somites appear on each side of the tube, the embryo has more than doubled in length (3.6 mm.), and the neuraxis is a tube except for the open anterior and posterior neuropore. The anterior neuropore closes about the twenty-eighth day and the posterior neuropore closes about gestational day 30.1S
Failure of closure or reopening in the brain area results in serious brain malformations (exencephaly, anencephaly, etc.)· Our primary concern is with the lumbosacral area of the spinal cord where failure to close, or reopening after closure, results in meningomyelocele associated with spina bifida. A photograph of a human embryo with a lumbosacral myelocele is shown in Figure 5.
In Figures 5 through 8 it is evident that the myelocele is much larger in volume than the normal spinal cord at the same neurosegmental level. This is the phenomenon of so-called myelocele "overgrowth" which was first described by Patten.13 Following this initial observation, others described localized neural tube overgrowth in animals following surgical procedures14'15 and in human exencephaly.16 In 1962 Kallen17 marshaled abundant evidence to show that this phenomenon of overgrowth is a true hyperplasia of the neural tissue. We will point out later that this occurrence of marked hyperplasia of neural tissue, which is localized to the area of the myelocele, may have importance in connection with some of the neural defects presented by these patients.
Källen called attention to the increased frequency of mitotic figures in the unorganized, hyperplastic, myeloschitic tissue of animals. We have observed more mitotic figures in overgrown myeloschitic tissue of human embryos than in normal spinal cord of the same segment. We have histologie evidence that neuroblasts derived from this mitotic activity are capable of sprouting axons, and we have observed branches of the sciatic nerve reaching and innervating the primitive muscles of the thigh.
It is possible that myeloschitic overgrowth that continues into later fetal life, plus fibrous tissue formation around the anläge of developing vertebrae, which becomes distorted at the level of the defect, precipitate nerve tension or compression that contribute to muscle denervation. The electromyographic studies of Ingberg and Johnson18 and our studies19 indicate that in the newborn or infant with meningomyelocele there is electromyographic evidence of denervation. Since the myelocele occurs prior to the development of axons, this suggests that the myelocele does initially innervate lower limb muscles on a segmentai basis.
CELLULAR KINETICS IN NEURAL TUBE DEVELOPMENT
To produce a thickened plate of ectoderm which becomes depressed below the surface and is rolled into a tube within a one-week period, we can appreciate the fact that repetitive cell division is vital. Also, the evidence for overgrowth of the local area of unclosed neural tube which we have discussed indicates that this process is altered and prolonged at the area of the myelocele.
Figure 6. Human embryo with extensive exencephaly and cervical myelocele (Horizon 20, crown-rump length 18.5 mm., weight 410 mg., gestational age 50^ days).
Although cell division is readily visualized microscopically by viewing developing cells in various stages of mitosis, this is but one phase of kinetic cellular activity that accounts for the dramatic events leading to neural tube formation.
Replication of DNA is a prerequisite for cell mitosis, and thymidine is required for DNA synthesis. Leblond and Messier20 demonstrated that tritium-labeled thymidine (thymidine 3H) was incorporated into cell nuclei during the stage of DNA synthesis and that the labeled cells could be identified by radioautography. Subsequently, Sidman et al.21 showed that in the developing cerebrum of mice fetuses, the undifferentiated neuroepithelial cells synthesize DNA while their nuclei lie near the external part of the cell. They then migrate to the ventricular surface where they divide within a few hours after completion of the DNA synthesis. A number of investigators have contributed to our understanding of the circadian aspects of cell generation cycle time, and a series of excellent papers by Langman and his colleagues22'23'24 have contributed significantly to our understanding of the following cellular kinetics in the developing neural tube.
Briefly, Langman has shown that after the telphase of mitosis, there is a short (0-few minutes) Gl phase, then an S phase (five hours) during which DNA is synthesized, followed by a G2 phase (two hours). The total cell generation cycle time is about seven to eight hours. At the time of DNA synthesis, the nucleus lies near the outer portion of the neural tube. It subsequently migrates toward the lumen of the tube, where mitosis occurs.
It has been estimated that the total number of cells in the neural tube doubles about every eight hours, and this accounts for its extremely rapid growth rate.
A point pertinent to our concern about myelocele formation is that when the neural tube closes, neuroepithelial cell division ceases and cell differentiation begins. When the neural tube fails to close, as indicated in our illustrations of human embryonic neural tube and primitive brain, enormous overgrowth of neural tissue occurs. The amount of overgrowth in the neural tube (Figures 5 through 8) and brain is very significant.
There is little information regarding differences in cell generation cycle time in the myelocele itself. However, a study made in our laboratory25 indicates that in experimentally produced myelocele in the rat, the synthetic stage of the cycle is prolonged and replication of cells persists for a longer period of time. The factors which account for shutting off neuroepithelial cell replication with closure of the tube and local continuation of the process in the unclosed portion are unknown. However, the observation that when the external limiting membrane of a closed chick embryo neural tube is cut, cellular proliferation occurs, suggests that intercellular pressures or cell compactness might have some influence on these dynamic processes.
Figure 7. Human embryo with exencephaly and lumbosacral myelocele (Horizon 22, crown-rump lenglh 19.8 mm., weight 770 mg., gestational age 52± days).
THEORIES OF MYELOCELE PATHOGENESIS
The first medical description of myelocele was given by Tulpius26 more than 300 years ago, and there is general agreement that the typical myelocele represents a segment of the neural tube which is open. However, for many years there has been disagreement about the pathogenetic mechanism by which this open segment of neural tube evolves.
There are two contrasting theories regarding the basic pathogenetic mechanism. More than 200 years ago, Morgagni" concluded that the neural tube closed and was subsequently forced open by increased pressure of cerebrospinal fluid reflected from hydrocephalus. This theory had dominated thinking for more than a century when Von Recklinghausen28 attacked the closure-reopening theory and proposed a failure-to-close theory of pathogenesis.
For the past 87 years, most embryologists who have studied the problem have interpreted the myelocele as did Von Recklinghausen- as an example of localized failure of neural tube closure. 188.8.131.52 On the other hand, in a series of publications beginning in 1959 and culminating in 1973 with a monograph titled, "The Dysraphic States- From Syringomyelia to Anencephaly," W.James Gardner 34, 35, 36, 37 has enthusiastically marshaled some evidence (derived mostly from postnatal observations) that the frequent association of hydrocephalus and syringomyelia with myelocele is evidence for validity of the closure-pressure reopening theory of Morgagni.
Although Gardner37 obtained much of his evidence for the closure-reopening theory from the association of hydrocephalus, syringomyelia, and myelocele in postnatal life, he also accepted conclusions from a series of studies by Padget,38' :ift who reviewed the human and nonhuman primate embryos and young fetuses in the Streeter collection at the Carnegie Institute for Embryology. Gardner37 states, "The first really concrete proof of [the closure-pressure reopening theory] was described by Padget."39 Padget's illustration of the torn edge of a subcutaneous bleb over an unclosed neural tube of a rhesus monkey was interpreted as evidence for pressure reopening of the tube. Gardner concludes his excellent monograph on syringomyelia with the observation, ". . . in the human infant an open neural tube usually represents rupture from overdistention."39
Figure 8. Surface view of the myelocele shown in Figure 7. Note the slit-like opening of the central canal.
INTERPRETATION OF THE STUDY SPECIMENS
In our study, 30 per cent of the specimens presented lumbosacral spina bifida without hydrocephalus or syringomyelia. However, the specimen shown in Figure 5 does show some evidence of disturbed fluid pressure with a lumbosacral hyperplasia. There was no evidence of myelocele. We could not find evidence for syringomyelia or neural tube rupture in this or any other specimens in this category.
All of the specimens illustrated (Figures 5 through 8) exhibit marked neural tissue overgrowth in both the exencephaly and the myelocele. Figure 6 is an extreme example of exencephalic overgrowth. We have seen similar specimens which appeared to be evolving to an anencephalic state as a result of hemorrhage and degeneration. It is possible that the neural hyperplasia could account for some instances of cutaneous membrane rupture.
Serial transverse histologie sections were prepared and studied of the 18 specimens that showed spina bifida without exencephaiy or hydrocephalus, those with exencephaly combined with spina bifida, and several of the specimens which showed only hydrocephalus. In none of the sections was any evidence of syringomyelia found. We uniformly observed subcutaneous bullous lesions in the spina bifida specimens associated with hydrocephalus, but there was no evidence of syringomyelia. Occasionally we found a rupture of the ectoderm near the edge of the myelocele, but there was evidence that this was a section artifact. Bullous lesions were present in all specimens with evidence of disturbed cerebrospinal pressure but without either myelocele or exencephalus.
In interpreting these data, we must realize that in the youngest of our specimens, the lesion is but from a few days to two weeks old at a period when the neuraxis is delicately suspended by a mesenchymal matrix. At this time there are no meninges surrounding the neuraxis and no bony structures. When the arachnoid and dura develop later, the hydrodynamics of disturbed cerebrospinal fluid pressure are quite different, and syringomyelia might arise secondary to the myelocele even though it did not precede it.
Assessment of the possibility of neural tube closure-reopening or failure to close is intimately related to the arcadian aspects of normal closure, choroid plexus formation and activity, and development of meninges which separate the cerebrospinal fluid compartments from the mesenchymal spaces.
There is general agreement that neural tube closure is complete by the twenty-eighth day and that both neuropores are closed by the thirtieth day. In all of our youngest specimens with myelocele (30 days old), the neural tube was closed except for the localized unclosed area which was already showing evidence of hyperplasia. The lateral choroid plexus does not invaginate the ventricles until the thirty-sixth day and appears to achieve its secretory function at approximately the thirty- seventh to fortieth day. The pial membrane is not discrete until approximately day 55. The dura does not encircle the neural tube until the fetus is nearly 9 to 10 weeks of age and does not achieve any degree of compactness until it is approximately 12 weeks of age. This indicates that the cerebrospinal fluid compartment is not closed until near the end of the first trimester.40 Under these circumstances, it is difficult to see how cerebrospinal fluid pressure could build up sufficiently to reopen a closed neural tube when it is evidence that the myelocele occurs at the end of the fourth week of gestation.
We interpret these findings to support the concept of Von Recklinghausen that the majority of myeloceles represent localized failure to close with subsequent extensive neural hyperplasia. This does not preclude the possibility that later in gestation pressure rupture may be a factor in some cases.
On the other hand, the appearance of a myelocele a few days or a week or two after its development is quite different from that after eight additional months of intrauterine development, and postnatal retrospective conjecture about these dynamic phenomena which occur at the end of the first month of gestation presents difficulties.
Detailed pathologic study specimens in the second and third trimester of prenatal development would undoubtedly contribute to our understanding of the prenatal evolution of this serious malformation of the central nervous system.
1. Carr. D. H. Chromosome studies in spontaneous abortion. Obstet. Gynec. 26 (1965), 308.
2. Nelson, T.. Oakley, G. P., Jr., and Shepard. T. H. Collection of Human Embryos and Fetuses. A Centralized Laboratory for Collection of Human Embryos and Fetuses: Seven Years' Experience: II. Ciassification and Tabulation of Conceptual Wastage with Observations on Type of Malformation, Sex Ratio, and Chromosome Studies. E. B. Hook, D. T. Janerich, and I. H. Porter, eds, In: Monitoring, Birth Defects and Environment. The Problem of Surveillance. New York and London: Academic Press, 1971, 45-64.
3. Poland, B, J. Study of developmental anomalies in ine spontaneously aborted fetus. Amer. J. Obstet. Gynec. 100 (Suppl. 4) (1968), 501.
4. Nishimura, H., Takano, K., Tanimura, T., Yasuda, M.. and Uchida, T. High incidence oí several malformations in the early human embryos as compared with infants. Biol. Neonat. 70 (1966), 93.
5. Streeter, G, L. Developmental horizons in human embryos. Description of age group XI. 13 to 20 somites, and age group XII, 21 to 29 somites. Carnegie Contr. Embryol. 30 (1942), 213.
6. Stteeter, G, L Developmental horizons in human embryos. Description of age group XIU, embryos about 4 or 5 millimeters long, and age group XIV, period of indentation of the lens vesicle. Carnegie Contr. Embryol. 37 (1945), 29.
7. Streeter, G. L. Developmental horizons in human embryos. Description of age groups XV, XVI, XVlI and XVIII, being the third issue of a survey of the Carnegie collection. Carnegie Contr. Embryo!. 32 (1948). 133.
8. Streeter, G. L. Developmental horizons in human embryos. Description of age groups XIX, XX, XXI. XX)I and XXDt, being the fifth issue of a survey of the Carnegie collection. Carnegie Contr. Embryol. 34 (1951), 165.
9. Witschi, E. Development of Vertebrates. Philadelphia: W. B. Saunders Co., 1956, 497-498.
10. Nishimura, H., Takano, K., and Tanimura, T. Normal and abnormal development of human embryos: First report of the analysis of 1,213 intact embryos. Teratology 1 (1968), 281.
11. Williams, P. L., Wendell-Smith, C. P., and Treadgold. S. Basic Human Embryology. Philadelphia, Montreal: J. B. Lippincott Co., 1966.
12. Tuchmann-Duplessis, H. Embryologie. Travaux Pratiques et Enseignement Dirige. Fascicule Trois. Paris: Mason & Cie, 1968.
13. Patten. B. M. Overgrowth of the neural tube in young human embryos. Anat. Ree, 773 (1952). 361.
14. Orts-Llorca. F., Genis-Galvez. J. M., and Ruano-Gil, D. Malformations encéphaliques et microphthalmic gauche apres section des vaisseaux vitillins gauches chez l'embryon de poulet. Acta Anal. 33(1959). 1.
15. Jelinek, R. Development of experimenïal exencephalia in the chick. Cisk. Morfologie B (1960), 368.
16. Giroud, A. and Martinet, M. Morphogenise de l'anencephalie. Anat. Micro. Morph. Exp. 46 (1957), 247.
17. Källen, B. Overgrowth malformation and neoplasia in embryonic brain. Conlin. Neurol. 22 (1962). 40.
18. lngberg, H. O. and Johnson, E. W. Electromyographic evaluation of infants with lumbar meningomyelocele. Arch. Phys. Med. & Rehab. 44 (1963), 86.
19. Chantraine, A., Lloyd, K,, and Swinyard, C. A, An electromyographic study of children with spina bifida manifesta. Develop. Med. Chilo. Neurol. 6 (1964), 7.
20. Leblond, C. P. and Messier, B. Renewal of child cells and goblet cells in the small intestine as shown by radioautography after injection of thymidine-HJ in mice. Anat. Ree. 132 (1958), 247.
21. Sidman, R. L., Mìaìe, I. L., and Feder, N. Cell proliferation and migration in the primitive ependymal zone; an autoradiographic study of histogenesis in the nervous system. Exper. Neuro!. 7 (1959), 332.
22. Langman, J., Guerrant, R. L., and Freeman. B. G. Behavior of neuroepithelial cells during closure of the neural tube. J. Comp. Neurol. 127 (1966), 399.
23. Langman, J. and Haden, C. The development and migration of neuroblasts in the spinal cord of the chick embryo. J. Comp. Neurol. 138 (1970), 419.
24. Shimada. M. and Langman, J. Cell proliferation, migration and differentiation in the cerebral cortex of the golden hamster. J. Comp. Neurol. 139 (1970), 227.
25. Murakami, H., Swinyard, C. A-, and Cohlan, S. Q. Unpublished data.
26. TuIp, N. Observations Medicae. Amsterdam: L. Elzevinus, 1652, 403.
27. Morgagni, G. B. De sedibus et causis morborum, 1761. Trans, by B. Alexander. London: A. Miller arid T. Cadeu, 1769.
28. Von Recklinghausen, F. Untersuchungen über die spina bifida. Arch. Path. Anat. 105 (1886), 243.
29. Langman, J. Medical Embryology. Human Development- Normal and Abnormal. Second Edition. Baltimore: Williams & Wilkins Co., 1969.
30. Hamilton, W. J., Boyd, J. D., and Mossman. H. W. Human Embryology. Baltimore: Williams & Wilkins Co., Î959.
31. Lemire. R. J., Shepard, T. H.. and Alvord, E. C., Jr. Caudal myeloschisis (lumbo-sacral spina bifida cystica) in a five millimeter (Horizon XIV) human embryo. Anat. Ree. 152 (1965), 9.
32. Dekaban, A. S. and Bartelmez, G. W. Complete dysrhaphism in a 14-somite human embryo: A contribution to normal and abnormal morphogenesis. Amer. J. Anat. 115 (1964), 27.
33. Gardner, W. J. Anatomic anomalies common to myelomeningocele of infancy and syringomyelia of adulthood suggest common origin. Cleveland Clin. Quart. 26 (1959a), 118.
34. Gardner, W. J. Hydrodynamic mechanism of syringomyelia: Us relationship to nyelocele. J. Neurol. Neurosurg. Psychiat. 28 (1965), 247.
35. Gardner, W. J. Myelocele: Rupture of the neural tube? Clin. Neurosurg. 15 (1968), 57.
36. Gardner, W. J., Abdullah, A. F., and McCormack, L J. The varying expressions of embryonal atresia of the lourlh ventricle in adults. Arnold-Chiari malformation, Dandy-Walker syndrome, "arachnoid" cyst of the cerebellum, and syringomyelia. J. Neurosurg. 14 (1957), 591.
37. Gardner, W. J. The Dysraphic States-From Syringomyelia to Anencephaly. Amsterdam: Excerpta Medica, 1973.
38. Padget, D. H. Spina bifida and embryonic neuroschisis- a causal relationship. Definitions of the postnatal conformations involving a bifid spine. Johns Hopk. Med. J, 123 (1968), 233.
39. Padget, D. H. Neuroschisis and human embryonic maldevelopment; new evidence on anencephaly, spina bifida and diverse mammalian defects. J. Neuropath. Exp. Neuro!. 29 (1970), 10, 185, 192.
40. Sensenig, E. C, The early development of the meninges of the spinal cord in human embryos. Contrib. Embryo!., Carneg. lnstn. 34 (1951). 147.