THE ANATOMICAL RECORD 228:191-210 (1990)

Myotome and Early Neurogenesis in Chick Embryos EARL D. KING AND BRYCE L. MUNGER Department of Anatomy, T h e Milton S . Hershey Medical Center, T h e Pennsylvania State University, Hershey, Pennsylvania

ABSTRACT The present study was undertaken in order to verify the identification of profiles of presumptive growth cones in vivo. The developing spinal nerves of chick embryos were studied by light and electron microscopy. We traced the onset of efferent and afferent innervation of the myotome in 2- to 4-day-old chick embryos in order to be sure that we were examining the growing tips of axons. In the process of studying these growing axons, we were able to observe some unique relationships of neural tube, myotome, and differentiating spinal nerves. The neural tube tightly abuts the myotome in Hamburger and Hamilton’s (HH) stage 14 chick embryos and cytoplasmic projections from the myotome directly abut the neural tube. The first ventral roots could be identified in HH stage 15 embryos and dorsal roots in HH stage 16 embryos, both under 2% days of age. The advancing spinal nerve courses toward the anterior or cranial half of the myotome, and growth cones directly contact the medial wall of the myotome. The spinal nerves continue to abut tightly the myotome during the succeeding day of embryonic life, and growth cones enter the substance of the myotome by 3 days, or HH stage 19 embryos. These dorsolaterally directed axons will form the dorsal ramus of the spinal nerves and the ventral ramus continues to be contiguous with the myotome. Invasion of the myotome by axons (putative innervation), and thus innervation of myotomal cells in the 3-day chick embryos, was a totally unexpected finding. The myotome and its potential derivatives thus have extensive neural contact by 3 days of embryonic life in the chick. These findings document a parallel differentiation of afferent and efferent elements of the nervous system and confirm previous accounts identifying growth cones in a n intact organism. These findings suggest that afferent as well as efferent nerves may have critical roles in the differentiation of the mesodermal a s well a s ectodermal derivatives. During the course of studies on the differentiation of hairs (Bressler and Munger, 1983) and the ridges of glabrous skin (Dell and Munger, 1986; Moore and Munger, 1989), axons and profiles identified a s portions of growth cones were identified earlier in gestation than expected. The identification of axonal growth cones was based on the fact that axons could be traced to the epidermis from the dermis. Thus, at the dermoepiderma1 junction, growth cones could be expected to be present if the appropriate stage in gestation was studied. Dell and Munger (1986) were limited in the availability of first-trimester embryos due to difficulty in determining pregnancy in our open colony of monkeys. Moore and Munger (1989) began their study while the present study was underway in order to evaluate the early differentiation of ridges in glabrous skin in human first-trimester specimens. The present study thus began a s another test of our ability to recognize axonal and growth cone profiles in vivo. We chose the growing spinal nerves of chick embryos a s a relatively accessible and well-studied system. We have been able to verify the identification of axonal and growth cone profiles and in addition have observed some unique relationships of spinal nerves ic, 1990 WILEY-LISS,

INC.

and myotomes. We were surprised by the extent of neural contact with myotomal cells in 2- to 3-day embryos. We can confirm many aspects of the study by Tosney and Landmesser (1985) that appeared while the present study was under way. We have also been able to confirm the series of papers that have noted differences in the cranial and caudal halves of the myotome (Pannattoni and Sisto Daneo, 1979; Keynes and Stern, 1984;, Rickman et al., 1985; Stern et al., 1986; Loring and Erickson, 1987; Stern and Keynes, 1987). The primary objective of the present study is perhaps of less biological importance than the conclusions that can be reached regarding early myotomal and neural differentiation. We will use the present study a s a baseline for a subsequent evaluation of the early onset of cuta-

Received J u n e 12, 1989; accepted February 2, 1990. Address reprint requests to Bryce L. Munger, M.D., Department of Anatomy, The Milton S. Hershey Medical Center, The Pennsylvania State University, P.O. Box 850, Hershey, PA 17033. Dr. King is now a t the Department of Internal Medicine, Temple University School of Medicine, 3400 North Broad Street, Philadelphia, PA 19140.

192

E.D. KING AND B.L. MUNGER

Fig. 1a.b.

MYOTOME AND EARLY NEUKOGENESIS

193

Fig. 1c. Fig.l a d . These micrographs consist of a semithin (a)and adjacent thin section (b) of a stage 14 embryo and demonstrate the intimate relationship of the central neural tube and adjacent myotomes (arrowheads). At this stage, the neural crest ( N ) can be identified migrating into position between the myotomes and neural tube. The myotome has numerous slender processes (arrow) extending toward the neural tube. This myotomal process is illustrated a t higher magnification ( c ) .The loosely arranged cells ventral to the myotome and surrounding the notochord represent the sclerotome (S). The area contained in the circle in b is the myocoele, i.e., the central cavity of the

myotome. The myotomal cells a t this point have junctional complexes as they abut the central lumen (c) and these are depicted at a higher magnification in c . In c and d (on overleaf) the neural tube is to the left and the myotome to the right, and numerous cytoplasmic processes are present in this compartment between neural tube and myotome. The arrow in c is a higher magnification view of the area of the arrow in b. The arrowhead in c indicates a process that directly abuts the basal lamina of the neural tube. The basal lamina of the myotome is seen as discontinuous masses of filamentous material (arrows)in d. a, x 175; b, x 800; c, x 6,900; and d, x 8,900.

194

E.D. KING A N D B.L. MUNGER

Fig. Id.

MYOTOME AND EARLY NEUKOGENESIS

Fig. 2. The specimen in this photograph illustrates the ventral root (arrows) in a HH stage 15 embryo. Sequentially thick and thin sections were searched attempting to identify dosal roots, but none could be found. The dorsal root ganglion ( G )contains relatively sparse cells, and some cells are angular, suggesting the intiation of differentiation and the formation of axonal processes. The myotome is thinned in the

neous innervation in chick embryos (Munger et al., 1990).

195

middle (XI of the site where the dorsal ramus of the spinal nerve will penetrate the myotome. No axons were present in the dorsolateral funiculus or anterior commissure. The plane of section was somewhat oblique throughout the myotome, extending its profile in a dorsal ventral axis. x 710.

For light microscopy, 15 embryos were impregnated with silver using the staining sequence described by Cajal (1960), which consists of fixation in alcohol folMATERIALS AND M E T H O D S lowed by direct immersion in 1.5% silver nitrate for Red leghorn chicken embryos were obtained from a 5-7 days, washing in water, reduction in hydroquilocal hatchery. A total of 47 embryos were used, with none, and subsequent serial paraffin section. Our best 12 prepared for electron microscopy and the remainder preparations resembled those depicted by Cajal in for light microscopy. The embryos were staged accord- which axons of dorsal root ganglia are impregnated in ing to the series of Hamburger and Hamilton (19511, addition to ventral roots in the 3-day chicks (HH stage hereafter referred to as HH stages. For electron micros- 171, which, in our experience, were easier to impregcopy, embryos were fixed in Karnovsky’s glutaralde- nate. Photomicrographs of the light microscopic prephyde-paraformaldehyde mixture and, following osmifi- arations were taken using Kodak Technical Pan film cation, were embedded in araldite (Durcapan Fluka). that was developed in POTA developer in order to proThe embryos were sectioned in a coronal plane to ap- vide contrast enhancement of blackened axons (Munproach the cervical spinal cord in a nearly cross-sec- ger and Rice, 1986). tional plane. The semithin sections were stained with Twenty embryos were fixed in 10% neutral buffered the sequence described by Laczko and Levai (1975). formalin, paraffin embedded, and serially sectioned

196

E.D. KING AN11 R.L.M U N G E K

Fig. 3. These two photomicrographs are silver-stained, sequential serial parafin sections from a stage 16 embryo and illustrate the first signs of axonal differentiation in dorsal root ganglia as well as axons coursing within the neural tube as presumptive commissural axons (arrow) as identified by Cajal (1960).The ventral root ( v ) contains axons that course toward the lower edge of the myotome just to the

right and above the figure numbers. The area of the dorsal root entry zone contains thin processes that may represent presumptive axons. Similar fine filaments iaxons?) (black on white small arrows) could also be seen in the substance of the ganglion iG), especially in b. The ganglion is more cellular as compared with Figure 2. The border between the myotome iM)and the ganglion is more distinct in b. x 350.

through the entire embryo. Serial paraffin sections were stained using the Sevier-Munger ( 1965) silver method. Our evaluation of a good impregnation was the same a s above; that is, when fine dorsal root fibers could be visualized, the impregnation was deemed a success. The serial paraffin sets were consulted to determine the appropriate plane of section in the semithin sections. In all of the results described below, we have selected a low cervical level encompassing the dorsal root ganglia (DRG) and spinal nerve, avoiding areas of the myotomes lacking DRGs. Once we had determined the appropriate level and plane of section in the semithin section, thin sections were cut, mounted on copper grids, stained with lead and uranyl salts, and examined in Philips 400 electron microscope.

RESULTS Two-Day Chick Embryo: HH Stage 14

The lower cervical spinal cord of stage 14 embryos is oval in profile, being elongated dorsoventrally (Fig. 1). The roof plate is closed and consists of a thin layer of cells. Neural crest cell migration has begun, and neural crest cells could be seen between the spinal cord and the dorsal portion of the myotome a s in Figure 1. The myotomes closely abut the neural tube (Figs. l a , b). The cells of the lateral portions of the myotomes are not tightly aggregated into a n epithelial arrangement and constitute migrating mesenchymal elements. The cavity of the myotome or myocele is surrounded by the myotomal cells joined a t their apices with junctional

complexes (Fig. lc). The relatively loose cells of the sclerotome are ventral to the compact myotome surrounding the notochord (Fig. l a , b) (Gasser, 1979). A discontinuous and delicate basal lamina is present on the side of the myotome abutting the neural tube (Fig. Id), but no basal lamina is present on the cutaneous border of the myotome. The base of the myotomal cells would thus be the surface abutting the neural tube. Most of the cells of the myotome were not differentiated, but a few cells contained scattered filaments, presumably representing the first signs of myofilament formation (Holtzer et al., 1957; Fishman, 1986). The myotomal cells were intimately related to the cells of the neural tube with small cytoplasmic extensions extending toward the neural tube and in some cases directly abutting the basal lamina of the neural tube (Fig. lc, d) as described previously by Filogamo and Sisto Daneo (1977) and Filogamo (1981). The matrix between the neural tube and myotome was also specialized, containing numerous fine filaments. The migrating neural crest cells (Fig. l a , b) were loosely connected to one another as they leave the neural tube and will migrate into this specialized matrix between neural tube and myotome. Two-Plus Day Chick Embryo: HH Stage 15

The earliest sign of root formation was observed in HH stage 15 chicks. The specimen illustrated in Figure 2 was taken from a series of serial semithin and thin sections and the ventral root that was verified by electron microscopy. The neural crest migration had occurred, forming a loose mass of developing DRG cells separating the neural tube from the myotome and associated dermamyotome (Nathanson, 1989). Mitotic figures were frequently encountered. Many of these developing DRG neuroblasts were angular in shape, indicating the onset of axonal outgrowth. The condensation of developing DRG cells was restricted to the anterior half of the dermamyotome. The myotome is medial to the more dense cells of the dermatome (Nathanson, 1989). The medial or neural border of the dermamyotome is thus triangular in cranial-caudal plane, as noted originally by Pannattoni and Sisto Daneo (1979). The dermamyotome was thinned (Fig. 2) at the site of the future dorsal ramus of the spinal nerve. We could not identify any fibers crossing the basal lamina of the neural tube in the position of the dorsal root, as depicted in Figure 3. Two and a Half Day Chicks: HH Stage 16

By stage 16 we were able to identify both dorsal and ventral roots in silver-stained paraffin sections (Fig. 3) a s well a s in semithin and thin sections by electron microscopy (Fig. 4). In silver-stained sections, as in Figure 3, we not only could identify dorsal and ventral roots, but also the axons of the earliest commissural cells (Cajal, 1960). The ventral roots are larger and contain more axons than dorsal roots, but axons from dorsal root ganglion cells join with the ventral roots, forming a spinal nerve. As noted above, the spinal nerves are restricted to the anterior half of the dermamyotome, a s are the DRGs. Delicate axons could also be identified within the substance of the DRGs (Fig. 3a). Similar profiles representing dorsal and ventral

roots are identified in semithin and thin sections in Figure 4. The pale, uniform cytoplasmic processes in Figure 4a are clusters of axons in Figure 4b and 4c. The supporting cells have ribosomes in their cytoplasm and thus stain more intensely with methylene blue. The variable appearance of profiles of growth cones and axons can be confused with other cytoplasmic elements present in the tissues under study, but the presence of ribosomes and rough endoplasmic reticulum (ER) in all cells other than axons, a s discussed subsequently, reduced the probability of errors in identification. The electron micrographs of Figure 4 represent a similar region of the alar plate, a s depicted in Figure 3b (arrows) and represent the dorsal root entry zone. Numerous axonal and growth cone profiles were admixed with processes of supporting (presumptive Schwann) cells, and they can be traced across the basal lamina of the neural tube in Figure 4b and c. Axons also were identified within the spinal cord (Fig. 4b) in the position of the commissural and association axons, a s noted in Figure 3. Axons and growth cone profiles can also be identified in branches of the spinal nerve as i t courses toward the dermamyotome in Figure 4d. The numerous axonal and growth cone profiles in Figure 4d can be contrasted to the polyribosome and ribosome containing cytoplasmic profiles of the presumptive Schwann cells. Bundles of filaments are a ubiquitous finding in the matrix between the developing dorsal root ganglion and myotome (Fig. 4d). One remarkable characteristic of these developing axons and growth cones is the absence of identifiable Schwann cells and any sign of a basal lamina. In large nerve trunks, a cell surrounding bundles of axons would be a presumptive perineurial cell. Typical Schwann cells with well-developed basal lamina are simply not present up to 6 days of age in chick embryos. As axons leave the nerve bundles, they enter the con-

Fig. 4 (continued onto t h e next two pages). These micrographs a r e sequential semithin and thin sections through a stage 16 embryo; similar findings were observed in stage 17 embryos. a: Two dorsal roots (arrows)can be traced from the ganglion (GI to enter the neural tube to the left. The roof of the neural tube is present a t t h e top of the micrograph. The myotome and skin surface a r e present to t h e right. b: Sequential thin section a t low magnification; c: t h e same field a t higher magnification. These micrographs illustrate several profiles of axons (arrows),some of which may be profiles of growth cones such a s the large expanse of axoplasm; central arrow indicates this axoplasm in c . Axons can be traced directly into t h e neural tube a t t h e top. This is t h e same area illustrated in a. The axons all lack ribosomes, whereas other cells in t h e field have prominent ribosomes. The filamentous matrix material between t h e axons and the neural tube is distincitive. d: Taken from the junction of t h e myotome and DRG and illustrates a bundle of axons (arrows) and associated presumptive Schwann or perineuraial cells. The axons lack ribosomes, polyribosomes, and granular ER t h a t a r e prominent in the associated cells. The variety of axonal profiles is identical to the profiles in c ; and although the axonal profiles a r e variable in appearance, t h e homogeneous appearance of the axoplasm is distinctive. These electron micrographs thus verify the identification of axonal profiles in semithin sections such a s a. The axonal profiles t h a t a r e relatively empty presumably represent the lamellipodia of growth cones. The identity of' the bundle of filaments indicated by the arrow could not be verified; but such bundles of filaments a r e consistently present near bundles of growth cones and axons. a: Axons can also be identified within the neural tube (arrowhead) as the descending association and commissural axons noted in Figures 3 and 5. a. x 690; b, x 2,500; c, x 11,500: d. x 11,500.

Fig. 4a,b.

200

E.D. KING AND B.1,. MUNGER

Fig. 5. These sections were taken from a Cajal silver impregnation of a stage 18 embryo. By stage 19 the accumulated axons in the ventrolateral funiculus produce a lateral bulge in the profile of the neural tube just below the ventral root entry zone. The silver impregnation obscures the dorsal root entry zone ( r ) ,but the axons within the (GI course dorsally toward the root and ventrally toward the spinal nerve(S). The higher magnification in b is not from the same field

illustrated in a but it is similar to either the right or left field in a. The axons leave the fused spinal nerve and some course dorsolaterally, as indicated by arrows in b, and form the dorsal ramus of the spinal nerves. The axons that continue to abut the myotome are t.he ventral ramus (v).The spinal cord contains association and commissural axons. a, x 250; b, x 500.

nective tissue grouped in twos and threes, but also in many cases they are totally isolated from other cells. When axons do tightly abut a cell in the mesenchyme, identification of these cells is speculative. We have previously conjectured that many cells within the developing dermis may indeed be immature Schwann cells (Bressler and Munger, 1983; Dell and Munger, 1986; Jones and Munger, 1985).

Three-Day Chick: HH Stage 18-19

By stage 18 the spinal cord in cross-section is almost as broad as it is long, and the roof plate is thicker a s compared with younger stages (Fig. 5). The pseudostratified appearance of the spinal cord is lost with the formation of a distinct mantle layer of differentiated motor neurons. A small anterior commissure becomes

20 1

MYOTOME AND EARLY NEUROGENESIS

progressively larger a s day 3 progresses. The commissural fibers, instead of coursing directly ventral a s in stage 16 embryos, course a t a n angle toward the central canal and hence through the developing motor neuron pool. An acellular marginal layer has developed in the dorsolateral and ventrolateral portions of the spinal cord consisting of fiber tracts of ascending dorsal root fibers in the dorsolateral funiculus and descending fibers in the ventrolateral funiculus. In silver-stained paraffin sections, small axons can be seen emerging from the DRGs to merge with the ventral motor root, forming a spinal nerve. The spinal nerve courses toward, and axonal growth cones directly abut, the myotome. Some axons penetrate the basal lamina of the myotome and come into direct contact with myotomal cells. The axons that course into the substance of the myotome dorsolaterally are the first signs of the dorsal rami of the spinal nerve (Fig. 5b) The ventral rami continue to abut the myotome tightly. More DRG cells had silver positive axons by HH stage 18, and axons coursing ventrally joined the spinal nerve and those that coursed dorsally will form the dorsal root (Fig. 5a). The dorsal root entry zone was larger and contained many axons a s well a s developing neurons and presumptive Schwann cells (Fig. 6a). The relatively homogenous areas within the ganglion and root are bundles of axons depicted at higher magnification in Figure 6b. The basal lamina of the DRG is present in the upper right of Figure 6b and can be compared with the absence of basal lamina in all of the branches of peripheral nerves illustrated in the present report. The variability of axonal and growth cone profiles resembles that noted previously. By stage 19, the DRGs are more compact and the spinal nerves a r e larger (Fig. 7a). The spinal nerves course toward the inferior margin of the myotome and dermamyotome and will enter the dermis, as noted subsequently (Munger et al., 1989). The ventral rami of each spinal nerve terminate as branches directed toward the myotome, a s illustrated in the components of Figures 7 and 8. As noted previously in Fig. 5, a portion of the ventral ramus tightly abuts the myotome; this relationship is illustrated in Figure 7c. The various axonal and growth cone profiles identified in Figure 7a require higher magnification, a s in Figure 7 M , in order to verify the identification. Some growth cone profiles in Figure 7b contain considerable amounts of vesicular material, and others are relatively empty (arrowhead to the right). The higher-magnification electron micrographs in Figures 7 and 8 not only illustrate the variety of profiles of axons and growth cones commented on previously, but they also illustrate the unique relationship of the components of the ventral ramus of a spinal nerve and its associated myotome. Figure 7a would be typical of the area removed from the ramus. A tight bundle of axons (lower right of Fig. 7a) is also illustrated a t higher magnification (Fig. 7b). Profiles identified a s growth cones are indicated by arrowheads. The associated cells have numerous ribosomes. The zone where the ventral ramus tightly abuts the myotome is illustrated in Figure 7c. No basal lamina is present between the axons and the myotomal cells that are conclusively identified as such by their content of

sarcomeres. Figure 7d shows the same field as the circle in Figure 7a and is included to verify the extent of the innervation of the myotome. A distinct basal lamina is present a t the junction of the myotome and the connective tissue compartment. Figure 8 is a fortuitous section depicting a n axon crossing the basal lamina of the myotome to become directly in contact with the myotomal cells. DISCUSSION

Although the present study was undertaken to characterize growth cones abutting early chick somites, the results have significant implications about early neural and somitic differentiation. The prsence of sensory a s well as motor components by HH stage 15 implies a relatively parallel differentiation of afferent and efferent components of the nervous system. As reviewed by Jacobson (1978), studies on the differentiation of the central nervous system and associated sensory ganglia, frequently citing the classic studies of Cajal, have usually concluded that motor components of the nervous system differentiate precociously a s compared with sensory components. Because our conclusions regarding afferent a s compared with efferent neural differentiation rely on the admittedly still tentative identification of growth cones in vivo, we will first review our findings on growth cones and developing axons in vivo and then turn to the question of temporal sequences of neural differentiation and myotomal development. We will close with a n attempt to explain the historical discrepancy between our findings and the vast bulk of the past literature. Growth cones, as reviewed in detail by Bunge (1973), is a term coined by Cajal prior to 1900 based on observations on developing and regenerating neural tissue. Cajal(l960) uses the term several times in the series of papers written between 1885 and 1895 contained in his Studies of Vertebrate Neurogenesis (Cajal, 1960). The term took on new meaning when Harrison (1910) was able to observe axons growing in cultures due to activity of growth cones. In three-dimensional terms, growth cones are shaped like a human hand with fingers extended and thumb somewhat apposed. The palm would represent folds or lamellipodia, and the fingers would represent the microspikes or filopodia (Bunge, 1973). Growth cones have been identified by electron microscopy both in vitro by Yamada et al. (1971); Bunge (1973, 1977), Bunge et al. (1983), Luduena and Wessells (19731, Bartlett and Banker (1984a, 1984b), and in vivo by Kelly and Zacks (1969), Tennyson (19701, Al-Ghaith and Lewis (1982), Kawana and Akert (19711, Bastiani e t al. (1985), Norlander and Singer (1982), Skoff and Hamburger (1975), and Tosney and Landmesser (1985). As reviewed by Bunge (1973) and Landis (1983), the in vivo studies provide conflicting accounts a s to what constitutes a growth cone. A consensus can be found in the papers by Tennyson (1970) and Yamada e t al. (1970, 1971) and the series of papers from Bunge’s laboratory (Bunge, 1973, 1977; Rees et al., 1976; Bunge et al., 1983) and more recently by Bartlett and Banker (1984a, b). This consensus is based largely on in vitro studies, a s the problem in vivo is knowing where indeed the growing tip of a n axon is in a complex intact organism (Munger, 1971). The present study confirms previous accounts

MYOTOME AND EARLY NEUROCENESIS

that had selected the dermoepidermal junction as a site where growing axons should be present (Dell and Munger, 1986; Moore and Munger, 1989). The sources cited above contain some important discrepancies. Kelly and Zacks (1969) did not illustrate typical growth cones. Carry e t al. (1983) focused on muscle and to only a minor extent dealt with growth cones. Al-Ghaith and Lewis’s (1982) conclusions that axons were rarely, if ever, free in the connective tissue matrix and that they rather were typically enveloped by other cells are quite at variance with the results of the present study and the previous accounts of Dell and Munger (1986) and Moore and Munger (1989). As noted in the studies of Cajal (1960), the 3-day chick embryo is difficult to impregnate with silver, and silver-stained sections of younger embryos to our knowledge have not been described in the literature. We were fortunate to obtain successful impregnations in both our en bloc Cajal-type impregnation a s well a s our silver method used on formalin-fixed paraffin sections in embryos from stage 16 on. In both cases, we could identify the masses of growth cones directly abutting the somite, which thus represent a site where growth cones should be identifiable by electron microscopy. The present results confirm Dell and Munger (1986) and clearly agree with the in vitro characterization as described by Yamada and coworkers (19711, Bunge (1973), Bartlett and Banker (1984), and the in vivo results of Tennyson (1970). As discussed in these studies and reviewed more recently by Johnson and Wessells (1980), Landis (1983), and Bunge et al. (19831, growth cones have a variable appearance, depending on the particular region that has been sampled in a given section. Growth cones typically contain membranous profiles of agranular ER, scattered vesicles, scant mitochondria, occasional lysosomal elements, and scattered filaments. Bundles of filaments are found in microspikes, and filopodia have been shown to contain actin (Letourneau, 1982). Filopodia are thus one of the most distinctive areas of a growth cone in terms of cytological characteristics. Regions of growth cones that lack organelles are the most difficult to identify, as they can be confused with areas of poor fixation; but in both the present study and others cited relatively empty profiles were surrounded by areas of good fixation and permit the tentative identification a s such profiles of growth cones. Ribonucleoprotein (RNP) particles have not been observed in profiles we identify as growth cones. RNP particles or polysomes were noted “on occasion” by Ten-

203

nyson (1970) and confirmed a t least on “rare occasions” by Bunge (1973) and by Bunge et al. (1983). We have traced numerous axons in continuity with growth cones and have never been convinced of the identification of polysomes, isolated ribosomes, or granular ER. Bartlett and Banker (1984a, b) have provided convincing quantitative evidence that axons lack ribosomes in contrast to dendrites in hippocampal neurons in culture. We conclude that the absence of ribosomes and granular ER are major criteria that can be used in the putative identification of growth cones in vivo. The surrounding mesenchymal and Schwann cells have prominent granular ER and numerous free ribosomes. Axons can even be identified in semithin sections stained with toluidine blue due to the lack of affinity for the dye as compared with the surrounding mesenchymal and neural crest cells (Fig. 4a). Basal lamina cannot be used as a criteria for identification of presumptive Schwann cells or perineural epithelium. As noted by Gamble (1966) and Webster and Billings (1972) and confirmed by recent studies from our laboratory (Bressler and Munger, 1983; Dell and Munger, 1986; Jones and Munger, 1985; Moore and Munger, 1989), embryonic Schwann cells have only focal patches of basal lamina, and over large expanses of cytoplasm there is no sign of basal lamina whatsoever. One of the most intriguing aspects of the present study is the possible role of axons in myotomal development. Invasion of the myotomal basal lamina by axons, to our knowledge, has not been observed before. Filogamo (1981) and Filogamo and Sisto Daneo (1977) first noted the intimate relationship of myotome and neural tube prior to neural crest migration and segmentation of the dorsal root ganglia. Our results would indicate that this intimate relationship continues until the sensory nerves reach the skin a t day 4. Somites have clearly formed long before neural crest migration, as noted by Tam et al. (1982) and Packard and Meier (1984), and thus have a critical role in establishing the metameric pattern of DRGs and the segmental spinal innervation of the skin a s originally proposed by Lehman (1927) and Detweiler (1934). An extremely important point of both Dell and Munger’s (1986) study, a s well as the present study is the identification of growth cones free in the mesenchyme without any sign of a cellular sheath quite in contrast to the observation of Al-Ghaith and Lewis (19821, who stated that growth cones never lacked a cellular investment as axons grow into a developing limb bud tissue. They further stated that they could “find no sign of a n association between the pioneer growth cones and fibrils of the extracellular matrix.” Tosney and Landmesser (1985) also noted a lack of filamentous strucFig. 6. These micrographs are taken from the dorsal root entry zone tures in association with bundles of axons. Isolated (Rl of an HH stage 19 embryo. The neural tube IS present to the right growth cones cannot be identified not only free in the and the DRG (Gl to the left. a: The root entry zone contains numerous cellular profiles that are developing neuroblasts as well a s presump- mesenchymal matrix both in the chick and primate, tive satellite and Schwann cells. In addition, the root entry zone con- but they are often associated with small bundles of tains electron-lucent, homogenous-appearing areas; b: At higher filaments. Furthermore, they are present in great magnification, these areas can be identified as bundles of axons. Nunumbers, a s Cajal(1960)stated clearly in his summary merous axons are also present in the neural tube immediately inside written in 1921. As noted recently by Jones and Munthe dorsal root entry zone coursing toward the basal plate. All profiles that were identified a s axonal, although extremely variable in axo- ger 1985) in developing opossum skin, many more axplasmic content, consistently lack ribosomes. Axons are present to the ons are present in electron micrographs than can be left, and developing neuroblasts of the ganglion (Gl are present to the accounted for by light microscopy. This discrepancy is right. The ganglion has a distinct basal lamina (arrow)in contrast to all of the peripheral nerves illustrated in this report (Fig. 71. a , also typical for the developing somite. Many axons seemingly fail to impregnate with any of the silver x 1,600; b, x 8,900.

204

E.1). KING AND

techniques, thus resulting in a significant underestimation of peripheral axons by light microscopy. The usual cytologic element implicated in silver positivity are microtubules and neurofilaments, and the former is conspicuously absent in growth cones. We conclude that all previous estimates of numbers of axons in developing systems by light microscopy, including Cajal (19601, have failed to appreciate the density of peripheral innervation in both skin and muscle. The fact that axons are growing without cellular ensheathment raises two important questions: What is the substrate upon which growth cones grow? What attracts them to the somite or skin? The first question can be answered only in part, a s axons grow on elements of the extracellular matrix and are focally closely associated with cells of the mesenchyme. Axons clearly have a special affinity for cells of skin and muscle. Elements of the extracellular matrix have been implicated in the migration of neural crest (Lofberg and Ahlfore, 1978; Yamada, 1980; Brauer et al., 1983; Le Douarin, 1981; Sanes, 1983) a s well a s in nerve growth in vivo (Swanson, 1985; Landmesser, 1981). The role of basal lamina has been suggested to have both positive and negative aspects in neural differentiation. Clearly, the basal lamina must be absent for neural crest cells to leave the neural tube, as noted by Sternberg and Kimber (1986) and Martins-Green and Erickson (1986). Dell and Munger (1986) and Moore and Munger (1989) have noted areas of deficient basal lamina at the junction of growth cones and Merkel cells. Small bundles of filaments are occasionally seen that may be the substate for neural growth. The role of the myotomal basal lamina may well be important, a s suggested by Loring and Erickson (1987). The findings of the present study would confirm those of Loring and Erickson (1987), and we would also agree with Tosney and Landmesser (1985) that the most important association is interneuronal. As the segmental spinal nerve is established, the somatotopic map is also established (Dell and Munger, 1986; Munger and Jones, 1987). The motor and sensory axons of each segment must know their relative segmental position. We will address this problem again when we describe the development of the spinal cutaneous nerves. In terms of trophic interactions, Cajal stated the problem clearly in 1919 (see also Cajal, 1960, for a restatement of the concept). In fact, Cajal would probably be impressed by how long the list of neurotrophic and neuronal growth factors is a s summarized by Berg (1984) Barde et al. (1983), and Thoenen and Edgar (1985).As noted in the classic studies of Lehman (1927) and Detweiler (1934,1937) and more recently by Lewis et al. (1981) and Teillet and LeDourain (1983), the somites have a critical role in the differentiation of dorsal root ganglia. Because the periphery can control the differentiation of motor as well a s sensory systems, a s noted in transplantation of limb buds, the trophic factors of both somites and skin must have a powerful role in guiding the course of axonal growth. We can extend these conclusions on the interrelationship of neural and cutaneous differentiation to include CNS development a s well. Dating from the classic studies of Hamburger and Levi-Montalcini ( 1949), the concept that DRGs differentiate over several successive days is considered axiomatic. Munger and Rice

R.L.MUNGER

(1986)applied this concept in the r a t mystacial pad and were able to correlate peripheral waves of neural development with the sequential changes of vibrissal barrels in SMI cortex noted previously by Rice (1985). Thus, the results of the present study document that the earliest wave of sensory and motor development in the chick embryo involves the myotome first and skin shortly thereafter. These conclusions lead us to consider the more sweeping generalization that afferent neural development occurs in concert with efferent neural development. The 2-day chick embryo has the somite directly apposed to the neural tube. As noted by Filogamo and Sisto Daneo (1977) and reviewed by Filogamo (19811, cellular projections extend from the somite to abut cells of the neural tube, implying extensive interaction of somites and neural tube by HH stage 14. The metamerism of the segmental plate mesoderm is well under way by this stage in the chick based on studies in the turtle of Packard and Meier (1984) and Meier and Packard (1984). This period of cellular contact between somite and neural tube is followed by the invasion of neural crest cells into the anterior half of this space between neural tube and somite (Keynes and Stern, 1984; Rickman et al., 1985; Stern et al., 1986; Loring and Erickson, 1987; Stern and Keynes, 1987). In the 3-day chick, this process is still under way in the caudal regions of the embryo. Coincident with this migration, differentiation into neuroblasts begins in some cells immediately, and axonal processes can be identified in suitable silver impregnations a s well a s electron micrographs in HH stage 16 embryos. Axons extend not only from the dorsal root ganglia, but also the motor root toward the somite, and thus sensory differentiation is occurring coincident with motor differentiation. We would concur with Filogamo and Sisto Daneo (1981) t h a t axons can be identified leaving the area of the myotome in 52hour chick embryos. They noted that axons could be identified close to the heart a t that stage. Although this conclusion is at variance with standard textbooks (Jacobson, 19781, it is consistent with experimental studies of Moody and Heaton (1983a-c) that implicate the trigeminal ganglia in the differentiation and migration of trigeminal motor neurons. Motor axons do not leave the brainstem if sensory neurons are not present. We have been impressed with the precocity of differentiation of the trigeminal complex in our studies on chick embryos by light microscopy. Inhibition of motor differentiation in the absence of sensory differentiation would cast significant doubt that motor differentiation in the spinal cord could precede sensory differentiation by days. How can we account for this discrepancy in the literature‘? Cajal published his original findings on neural development before 1900 and, due to the obscurity of the original journals, began to compile a republication a s his Studies in Neurogenesis. He compiled these previously published papers with comments and deletions to “bring them up to date” a s the “French Edition” of his Studies in Neurogensis in the 1920s. Few people knew the French edition existed when Prof. F. DeCastro made a copy available to L. Guth, who translated the French edition with a publication date of 1960. The footnotes to the French edition provided by Cajal to bring the book up to date are clearly identified in the

MYOTOME A N D EARLY N E U R O G E N E S I S

Fig. 7a. (legend follows)

205

Fig. 7b.c.

MYOTOME AND EARLY NEUROGENESIS

207

Fig. 7d.

Fig. 7. These electron micrographs are all taken from the junction of the spinal nerve and myotome of a stage 19 embryo. The survey micrograph (a)is also depicted a t higher magnification (d). The upper arrow in both micrographs illustrates the same growth cone. a: The ventral ramus ( V )of the spinal nerve is present to the lower right and consists of presumptive Schwann cells surrounding bundles of axons and growth cones. b: The axons leave their cellular investment and course toward the myotome. Many relatively pale and homogeneous cellular processes in a can be identified as axons and growth cones, and several examples have been indicated by the arrows. b d : Axonal profiles and growth cones can be seen to better advantage a t higher magnification. b The presumptive Schwann cells (XIhave prominent free ribosomes, polyribosomes, and granular ER. Arrowheads indicate profiles consistent with the typical contents of growth cones including one that is relatively empty. The circle indicates a bundle of presump-

tive collagen fibrils in the extracellular matrix. c: Taken from the lower portion of the myotome, where the spinal nerve is closely applied to the inner surface of the myotome as in Figure 5c. The cellular elements separating the spinal nerve from the myotome as depicted in a are not present at this level. The axons and growth cones in c are identical to those seen in b. When the ventral rami tighly abut the myotome, a basal lamina does not separate the axons and myotomal cells (empty arrows), which is in contrast to the situation in d. The myotomal cells contain portions of sarcomeres (S). Other bundles of thick and thin filaments in this micrograph are also portions of sarcomeres. d: Field of the upper arrow in a, included to permit identification of growth cones that are scattered among the cells separating the spinal nerve from the myotome and also directly abutting the basal lamina of the myotome. a ~ 3 , 1 0 0M, ; x 11,500.

208

E.1). KING AND R.1,. MUNGER

Fig. 8.This micrograph represents a fortuitous section in which an axon (arrows) can be traced into the mytotome crossing the basal lamina (clear arrows). The numerous axonal and growth cone profiles (arrowheads)are scattered throughout the matrix and lack any sign of cellular investment. x 12,425.

such as Visintini and Levi-Montalcini (1939) did not have access to this restatement of Cajal’s findings. Cajal added a n addendum to the paper on the development of the chick spinal cord noting that he finally “obtained impregnations of the chick spinal cord a t the third day of incubation,” and illustrated this statement with a drawing clearly showing dorsal a s well a s ventral roots and axons of the dorsal root course in close proximity to commissural neurons. The opening paragraphs of that paper state that the results were incomplete due to the lack of a good reaction product prior to day 5. This statement has been frequently quoted out of context in that the failure is due not to a lack of neuronal differentiation, but rather to a technical problem of silver staining. Axonal growth, as noted in the addendum, is definitely present by 3 days, and our findings thus confirm Cajal’s addendum. We presume that other studies have all had technical problems with the silver methods and, as we note here, by electron microscopy we can clearly identify a n abundance of axons not visualized by any of the silver methods. In the absence of confirmatory electron microscopy, we would not be comfortable contradicting Cajal’s major statements even with a n addendum in which he contradicts himself. Tell0 (1922) also had difficulty impregnating the 3-day and younger chick and is also in part the source of the compounded error, as he notes that the separation of cutaneous and motor branches begins a t day 5 to 6. As we note in the present study, this time table is off by a significant factor. Our conclusion that differentiation of afferent neural elements occurs much earlier t h a n previously suspected has a profound implication on the relationship between neural differentiation and differentiation of peripheral elements such as cutaneous appendages. Dell and Munger (1986) hypothesized that the overlapping dermatotopic map might have a role in the differentiation of the papillary ridges of glabrous digital skin and thus the fingerprint. That suggestion is strengthened on the basis of the present study and preliminary experimental studies published to date (Munger et al, 1985; Jones and Munger, 1986, Munger and Jones, 1987). These studies describe abnormal cutaneous differentiation in opossum pups with surgically produced deletions of the spinal cord and DRGs or trigeminal ganglia. We can thus confirm Mauger (1972), who noted that in the absence of the neural tube feathers do not differentiate. Similar conclusions have been reached by Zelena (1957) with respect to muscle spindles. Deafferentation blocks the differentiation of muscle spindles but not de-efferentation. These conclusions in no way contradict the critical role of the mesoderm in the differentiation of the ectoderm as reviewed by Sengel (1983, 1986). Instead, we propose that the role of the nervous system has not been precisely defined, but the available data suggest t h a t a tight link exists between the development of afferent nerves and their targets in both muscle and skin. ACKNOWLEDGMENTS

translation. Thus, these papers were generally not known prior to 1960 and to our knowledge rarely, if ever, were cited in works prior to that publication date. Thus, studies on dorsal root ganglia differentiation

The authors wish to thank Debbie Hinton for invaluable assistance in the light microscopic preparations and Brian Tucker and Roland Myers for help in electron microscopy. Rebecca Ulrich and Therese Segneri assisted us in drafting the numerous revisions of the

209

MYOTOME AN11 EARLY NEUROGENESIS

manuscript. This research was supported in part by USPHS NS19462 and BRSG SO7 RR05680. LITERATURE CITED Al-Ghaith, L.K., and J . H . Lewis 1982 Pioneer growth cones in virgin mesenchyme: An electron-microscope study in the developing chick wing. J. Embryol. Exp. Morphol., 681149-160. Barde, Y.A., D. Edgar, and H. Thoenen 1983 New neurotrophic factors. Annu. Rev. Physiol., 45t601-612. Bartlett, W.P., and G.A. Banker 1984a An electron microscopic study of the development of axons and dendrites by hippocampal neurons in culture. I. Cells which develop without intercellular contacts. J. Neurosci., 4:1944-1953. Bartlett, W.P., and G.A. Banker 1984b An electron microscopic study of the development of axons and dendrites by hippocampal neurons in culture. 11. Synaptic relationships. J. Neurosci., 4:19541965. Bastiani, M.J., S. duLac, and C.S. Goodman 1985 The first neuronal growth cones i n insect embryos. Model system for studying the development of neuronal specificity. In: Model Neural Networks and Behavior. A.I. Selverston, ed. Plenum, New York, pp. 149174. Berg, D.K. 1984 New neuronal growth factors. Annu. Rev. Neurosci. 7~149-170. Brauer, P.R., D.L. Bolender, and R.H. Markwald 1983 Localization of 'H-fucosylated substances in the pathway of migrating cephalic neural crest cells. In: Developing and Regenerating Vertebrate Nervous System. P.W. Coates, R.W. Markham, and A.D. Kinney, ed. New York: Alan R. Liss, pp. 151-157. Bressler, M.F., and B.L. Munger 1983 Embryonic maturation of sensory terminals of primate facial hairs. J . Invest. Dermatol., 80; 245-260. Bunge, M.B. 1973 Fine structure of nerve fibers and growth cones of isolated sympathetic neurons in culture. J. Cell Biol., 56t713735. Bunge, M.B. 1977 Initial endocytosis of peroxidase or ferritin by growth cones of cultured nerve cells. J. Neurocytol., 6t407-439. Bunge, M.B., M.I. Johnson, and V.J. Argiro 1983 Studies of regenerating nerve fibers and growth cones. In: Spinal Cord Reconstruction. C.C. Kao, R.P. Bunge, and P.J.Reier ed., Raven Press, New York, pp. 99-120. Caial. " . S.R. 1919 Accion neurotrouica de 10s enithelios. Trab. Lab. Invest. Biol., Madrid 17:181-2!28. Cajal, S.R. 1960 Studies on Vertebrate Neurogenesis. L. Guth, translator. Springfield, Illinois: Charles C. Thomas Publishing. Carry, M.R., M. Morita, and H.O. Nornes 1983 Morphogenesis of motor endplates along the proximodistal axis of the mouse hindlimb. Anat. Rec., 207t473-485. Davis, M.R., and M. Constantine-Paton 1983a Hyperplasia in the spinal sensory system of t h e frog. I. Plasticity in the most caudal dorsal root ganglion. J. Comp. Neurol., 221t444-452. Davis, M.R., and M. Constantine-Paton 198313 Hyperplasia in the spinal sensory system of the frog. 11. Central and peripheral connectivity patterns. J. Comp. Neurol., 221:453-465. Dell, D.A., and B.L. Munger 1986 The early embryogenesis of papillary (sweat duct) ridges in primate glabrous skin: The dermatotopic map of cutaneous mechanoreceptors and dermatoglyphics. J. Comp. Neurol. 244t511-532. Detweiler, S.R. 1934 An experimental study of spinal nerve segmentation in amblystoma with reference to the plurisegmental contributory to t h e brachial plexus. J. Exp. Zool., 67t395-441. Detweiler, S.R. 1937. Observations upon the migration of neural crest cells and upon the development of the spinal ganglia and vertebral arches in amblysoma. Am. J. Anat. 61t63-94. Filogamo. G. 1981 The first stage in myoblast development. In: Studies in Developmental Neurobiology. W.M. Cowan, ed. Oxford University Press, New York, pp. 171-187. Filogamo, G., and L. Sisto Daneo 1977 Nervous projections and myotome maturation. J. Submicrosc. Cytol., 9t307-310. Fishman, D.A. 1986 Myofibrillogenesis and the morphogenesis of skeletal muscle. In: Myology: Basic and Clinical. A.G. Engel and B.Q. Banker, eds. McGraw-Hill, New York, pp. 5-38. Gamble, H.J. 1966 Further electron microscope studies of human fetal peripheral nerves. J. Anat., 100:487-502. Gasser, R.F. 1979 Evidence that scleromal cells do not migrate medially during normal embryonic development of the rat. Am. J. Anat., 254t509-524. Hamburger, V., and H.L. Hamilton 1951 A series of normal stages in the development of the chick embryo. J. Moruhol., 88t49-92. Hamburger, V:, and R. Levi-Montalcini 1949 Proliferation, differen-

tiation and degeneration in the spinal ganglia of t h e chick embryo under normal and experimental conditions. J. Exp. Zool., lllt457-501. Harrison, R.G. 1910 The outfrowth of the nerve fiber as a model of protoplasmic movement. J. Exp. Zool., 9t787-848. Holtzer, H., J.M. Marshall, and H. Finck. 1957. An analysis of myogenesis by the use of fluorescent antimyosin. J. Biochem. Biophysic. Cytol. 3t705-724. Jacobson, M. 1978 Developmental Neurobiology, 2nd. ed. Plenum Press, New York. Johnson, R.N., and N.K. Wessels 1980 Regulation of the elongating nerve fiber. Curr. Top. Dev. Biol., 16t165-206. Jones, T.E., and B.L. Munger 1985 Early differentiation of the afferent nervous system in the glabrous snout skin of the opossum (Monodelphis domestica). Somatosens. Res., 3t169-184. Jones, T.E., and B.L. Munger 1986 Intraepidermal axonal invasion and epidermal hyperplasia secondary to partial neuralectomy in postnatal opossum pups. Anat. Rec., 213:62A3. Kawana, E., C. Sandri, and K. Akert 1971 Ultrastructure of growth conesin the cerebellar cortex of the neonatal r a t and cat. Z. Zellforsch. Mikrosk. Anat., 115t284-298. Kelly, A.M., and S.I. Zacks 1969 The fine structure of motor endplate morphogenesis. J. Cell Biol., 42t154-169. Keynes, R.J., and C.D. Stern (1984) Segmentation in the vertebrate nervous system. Nature, 310t786-789. Laczko, J., and G. Levai 1975 A simple differential staining method for semithin sections of ossifying cartilage and bone tissue embedded in eopxy resin. Mikroscopie, 31tl-4. Landis, S.C. 1983 Neuronal growth cones. Annu. Rev. Physiol., 45: 567-580. Landmesser, L. 1981 Pathway selection by embryonic neurons. In: Developmental Neurobiology. W.M. Cowan, ed. Oxford University Press, New York, pp. 53-73. LeDouarin, N. 1981 Migration and differentiation of neural crest cells. Curr. Top. Dev. Biol., 16t32-85. Lehman, F.E. 1927 Further studies on the morphogenic role of the somites in the development of the nervous system of amphibians. The differentiation and arrangement of the spinal ganglion in Pleurodeles uoalthi. J. Exp. Zool., 49t93-131. Letourneau, P.C. 1982 Nerve fiber growth and its regulation by extrinsic factors. Curr. Topics Neurobiol., 4t213-254. Lewis, J.,A. Chevallier, M. Kieny, and L. Wolpert 1981 Muscle nerve branches do not develop in chick wings devoid of muscle. J. Embryol. Exp. Morphol., 64r211-232. Lofberg, J., and K. Ahlfore 1978 Extracellular matrix organization and early neural crest migration in the axolotl embryo. In: Formshaping Movements in Neurogenesis. C.O. Jacobson and T.E. Almqvist, ed. Almqvist and Winksell, Stockholm, p. 87-101. Loring, J.F., and C.A. Erickson 1987 Trunk neural crest cell migratory pathways in the chick embryo. Dev. Biol., (in press). Luduena, M.D., and N.K. Wessells 1973 Cell locomotion, nerve elongation and microfilaments. Dev. Biol., 30,427-440. Martins-Green, M., and C.A. Erickson 1986 Development of neural tube basal lamina during neuralation and neural crest cell emigration in the trunk of the mouse embryo. J. Embryol. Exp. Morphol., 98t219-236. Mauger, A. 1972 A role du tube neural dans le development du plumage dorsal de l'embryo de poulet. William Roux's Arch., 170t244266. Meier, S. and D.S. Packard 1984 Morphogenesis of the cranial segments and the distribution of neural crest in the embryos of the snapping turtle chelydra serpentina. Dev. Biol., 102t309-323. Moody, S.A., and M.B. Heaton 1983a Developmental relationships between trigeminal ganglia and trigeminal motoneurons in chick embryos. I. Ganglion development is necessary for motoneuron migration. J. Comp. Neurol., 213t327-343. Moody, S.A., and M.B. Heaton 1983b Developmental relationships between trigeminal ganglia and trigeminal motoneurons in chick embryos. 11. Ganglion axon ingrowth guides motoneuron migration. J. Comp. Neurol., 223t344-349. Moody, S.A., and M.B. Heaton 1983c Developmental relationships between trigeminal ganglia and trigeminal motoneurons in chick embryos. 111. Ganglion perikarya direct motor axon growth in the periphery. J. Comp. Neurol., 213t350-364. Moore, S.J., and B.L. Munger. 1989. The early ontogeny of the afferent nerves and papillary ridges in human digital glabrous skin. Dev. Brain Res. 48t119-141. Munger, B.L. 1971 Patterns of organization of peripheral sensory receptors. In: Handbook of Sensory Physiology. W.R. Lowenstein, ed. Surineer-Verlae. New York. DD. 523-556. Munger, 'B.L:: and T.E. Jones 1987 Origin and modulation of the I

.

I

210

E.1). KING A N D R.L. M U N G E R

dermatotopic map. In: Effects of injury on spinal and trigeminal somatosensory systems. L.M. Pubols and B.J. Sessle, ed. Alan R. Liss, Inc. New York, pp. 61-68. Munger, B.L., and F.L. Rice 1986 Successive waves of differentiation of cutaneous afferents in rat mystacial skin. J. Comp. Neurol., 252t404-414. Munger, B.L., E.D. King, A. Mauger, and R., Saxod. 1990. Tissue interactions and cutaneous differentiation. I. The ontogeny of feathers and cutaneous innervation in chick embryos. In preparation. Munger, B.L., S. Terry, and M.L. Kirby 1985 Hyperplasia of dorsal root ganglia and expansion of the dermatotopic map following neural crest deletion in chick embryos. Anat. Rec., 221:134A. Nathanson, M.A. 1989 Differentiation of musculoskeletal tissues. Int. Rev. Cytol., 116:89-164. Norlander, R.H., and M. Singer 1982 Morphology and position of growth cones in the developing Xenopus spinal cord. Dev. Brain Res., 4t181-193. Packard, D.S., and S. Meier 1984 Morphological and experimental studies of the somitomeric organization of the segmental plate in snapping turtle embryos. Embryo. Exp. Morphol., 84t35-84. Panattoni, G.L., and L. Sisto Daneo 1979 Three-dimensional models of chick embryo. Somites obtained by mathematical methods applied to ultrastructural data. Part I. Thoracolumbar somite. Wilhelm Roux’s Arch. 186.273-287. Rees, R.P., M.B. Bunge, and R.P. Bunge 1976 Morphological changes in the neuritic growth cone and target neuron during sympathetic junction development in culture. J. Cell Biol., 68t240-263. Rice, F.L. 1985 Gradual changes in the structure of the barrels during maturation of the primary somatosensory cortex in the rat. J . Comp. Neurol., 236t496-503. Rickman, M., J.W. Fawcett, and R.J. Keynes 1985 The migration of neural crest cells and the growth of motor axons through the rostra1 half of the chick somite. J. Embryo. Exp. Morphol. 90: 437-455. Sanes, J.R. 1983 Roles of extracellular matrix in neural development. Annu. Rev. Physiol., 45581-600. Sengel, P. 1983. Epidermal-dermal interactions during formation of skin and cutaneous appendages. In: Biochemistry and Physiology of Skin L.E. Goldsmith, ed., pp. 102-131. Oxford, New York. Sengel, P. 1986. Epidermal-dermal interaction. In: Biology of the Integument, Vol. 11, Vertebrates J. Bereiter-Han, A.G. Matolsty, and K.S. Richards, eds., pp. 374-408. Springer, Berlin. Sevier, A., and B.L. Munger 1965 A silver method applicable to paraffin sections of formol-fixed tissue. J. Neuropathol. Exp. Neurol., 24t130-135. Skoff, R.P., and V. Hamburger 1975 Fine structure of dendritic and axonal growth cones in embryonic chick spinal cord. J. Comp. Neurol., 153t107-148. Stern, C.D., S.M. Sisodiya, and R.J. Keynes 1986 Interactions between neurites and somite cells: Inhibition and stimulation of

nerve growth in the chick embryo. J. Embryo. Exp. Morphol., 92:209-226. Stern, C.D., and R.J. Keynes. 1987. Interactions between somite cells: The formation and maintenance of segment boundaries in the chick embryo. Dev. 99:261-272. Sternberg, J . , and S.J. Kimber 1986 The relationship between emerging neural crest cells and basement membranes in the trunk of the mouse embryo: a TEM and immunocytochemical study. J. Embryol. Exp. Morphol., 98251-268. Swanson, G.J. 1985 Paths taken by sensory nerve fibers in chick wing buds. Embryo. Exp. Morphol, 86t109-124. Tam, P.P.L., S. Meier, and A.G. Jacobson 1982 Differentiation of the metameric pattern in the embryonic axis of the mouse. 11. Somitomeric organization of the presometic mesoderm. Differentiation, 22:109-122. Teillet, M.-A., and N. LeDouarin 1983 Consequences of neural tube and notochord excision on the development of the peripheral nervous systemin the chick embryo. Dev. Biol., 98:192-211. Tello, J.F. 1922 Les differenciations neuronales dans l’embryon du poulet, pendant les premiers jours de l’incubation. Trav. Lab. Rech. Biol., Univ. Madrid, 21:l-94. Tennyson, V.H. 1970 The fine structure of the axon and growth cone of the dorsal root neuroblast of the rabbit embryo. J. Cell Biol., 445-79. Thoenen, H., and D. Edgar 1985 Neurotrophil factors. Science, 229: 238-242. Tosney, K.W., and L.T. Landmesser 1985 Development of the major pathways for neurite outgrowth in the chick hindlimb. Dev. Biol., 109: 193-2 14. Visintini, F., and R. Levi-Montalcini 1939 Relazione tra differenziazione strutturale e funzionale dei centri e delle vie nervose nell’embrione di pollo. Schweiz. Neurol. Psychiatr., 43t381-393. Webster, H.deF., and S.M. Billings 1972 Myelinated nerve fibers in the Xenopus tadpoles: In vivo observations and fine structure. J . Neuropathol. Exp. Neurol., 31:102-112. Yamada, K.M. 1980 Structure and function of fibronectin in cellular and developmental events. In: Current Research Trends in Craniofacial Development. R.M. Pratt and R.L. Christensen, eds., ElsevieriNorth Holland, New York, pp. 277-295. Yamada, K.M., B.S. Spooner, and N.K. Wessells 1971 Ultrastructure andand function of growth cones and axons of cultured nerve cells. J. Cell Biol., 49:614-635. Yasuda, M. 1964 Comparative and topographical anatomy of the fowl XXXIVdistribution of cutaneous nerves of the fowl. Jpn. J . Vet. Sci., 26:241-248. Yntema, C.L., and W.S. Hammond 1947 Depletions in the thoracolumbar sympathetic system following removal of neural crest in the chick. J . Comp. Neurol., 86:237-266. Zelena, J. 1957. The morphogenetic influence of innervation on the ontogenetic development of muscle spindles. J. Embryol, Exp. Morph. 5t283-292.