DEVELOPMENTAL DYNAMICS 204:391406 (1995)

Formation of New Myotubes Occurs Exclusively at the

Multiple Innervation Zones of an Embryonic Large Muscle
M.J. DUXSON AND P.W. SHEARD
Department of Anatomy and Structural Biology (M.J.D.) and Physiology (P.W.S.), University of Otago Medical School,
Dunedin, New Zealand

ABSTRACT
This work examines the general
principle of whether production of embryonic
muscle fibres is invariably linked to sites of innervation, as we have previously reported in small
rodent muscles (Duxson et al. 119891 Development
1OE74.3-750). The experimental strategy has been
to make a detailed electron microscopic analysis
of the formation of new myotubes in a large muscle having multiple, discrete innervation zones.
The particular model system is the guinea pig sternomastoid muscle, a strap-like, parallel-fibred
muscle with four distinct endplate bands, both in
the embryo and the adult. Primary myotubes in
the developing muscle extended from tendon to
tendon of the muscle and were innervated at each
of the multiple endplate zones. Each point of innervation of the primary myotubes was a focus
around which many new secondary myotubes
formed, and each secondary myotube was approximately centred on one of the innervation
sites of its supporting primary myotube. This confirms our previous report, in rat IVth lumbrical
muscle, of an invariable association between sites
of formation of new secondary myotubes and sites
of innervation. We suggest that, in vivo, nerve terminals either directly induce the initial myoblast
fusions which give rise to new secondary myotubes, or induce some precondition for fusion. An
alternative hypothesis is that a common patterning influence in the muscle localizes both innervation and secondary myotube formation to the
same zone. The pattern of secondary myotube
production in the embryo has important implications for the size and final architecture of muscles
in larger animals, and some of these are discussed. o 1995 Wiley-Liss, Inc.

ples which guide early muscle development must be

capable of producing muscles with this great final diversity. In the past, our studies have examined the
general processes and mechanisms of control of muscle
fibre production in embryos of small laboratory animals. A central interest has been the role of the nervous system in initiating new muscle fibre formation.
Recently we have shifted our attention to larger muscles because of some interesting and potentially illuminating differences in their structure. Firstly, the
pattern of innervation of larger muscles is much more
complex than that of most small muscles: whereas most
rat muscles are innervated at a single endplate zone in
approximately the mid-belly of the muscle, many muscles in large animals are innervated at multiple bands
of endplates distributed along the whole length of the
muscle (e.g., Coers and Woolf, 1959; Wilson et al.,
1992). Secondly, whereas all muscle fibres in small
mammalian muscles extend from end to end of the
muscle, in larger muscles (where fascicle length is
greater than 2-3 cm) this is rarely true. Most fibres are
substantially shorter than the fascicles in which they
run, and end intrafascicularly (e.g., Huber, 1916; Barrett, 1962; Loeb et al., 1987; for review see Gans and
Gaunt, 1992). Thirdly, muscles of larger animals (especially domestic meat animals such as sheep and cattle) have much larger numbers of muscle fibres in the
adult than do rats or mice, although the initial muscle
rudiments may not be so very different in size.
We wished to understand how muscle fibre generation in larger muscles occurs, with the general aim of
understanding how the form of a large muscle arises
during embryogenesis. In particular, we wanted to
know if general rules governing the production of new
muscle fibres in embryos of small mammals also applied in the development of larger animals. As background, we first summarize what we know of muscle
Key words: Muscle development, Innervation, morphogenesis in the much-studied small animal sysSecondary myotube, Muscle architec- tems.
The development of muscles in small laboratory anture, Electron microscopy, Light miimals, such as the rat, mouse, chicken, and frog, follows
croscopy
a general pattern. First, a small number of primary
INTRODUCTION
Skeletal muscles come in an enormous range of size
and form, and fulfill an equally large range of physioReceived February 27, 1995; accepted July 28, 1995.
logical requirements. As a result, no two muscles apAddress reprint requests/correspondence to Dr. M.J. Duxson, Depear to have quite the same architecture. It is clear, partment of Anatomy and Structural Biology, University of Otago
however, that the basic processes and general princi- Medical School, P.O. Box 913, Dunedin, New Zealand.
0 1995 WILEY-LISS, INC.

392

DUXSON AND SHEARD

myotubes form; these extend the full length of the muscle anlage and define its shape, origins, and insertions.
Slightly later a second generation of myotubes appears,
and it is these secondary myotubes which eventually
form the main bulk of the muscle. Secondary myotubes
appear asynchronously over a period of some days to
weeks, are initially binucleate cells which lie under the
basal lamina of the primary myotube, and use the primary myotube as a framework which guides their progressive growth towards the tendons (Kelly and Zacks,
1969a; McLennan, 1983; Ross et al., 1987a). However,
new secondary myotubes do not appear at just any site
within the muscle. Our studies of rat IVth lumbrical
muscle have shown that the only place where secondary myotubes are normally produced is in the region of
innervation of the supporting primary myotube (Duxson et al., 1989). The absolute nature of this result is
somewhat surprising: although it is well known that
denervation or nerve blockade severely reduces the
number of secondary myotubes produced in embryonic
muscles (Harris, 1981; McLennan, 1983; Ross et al.,
1987b),it is also true that there are several experimental situations where myotubes do form in the absence of
nerve. For example, formation of secondary myotubes
has been reported in mutant mouse muscles which
were never innervated (Ashby et al., 1993), and myotube formation routinely occurs in tissue culture without nervous input.
Muscle development in larger animals has been less
studied than that in small laboratory rodents. Several
electron microscopic studies (e.g., lamb, Ashmore et al.,
1972; pig, Ashmore et al., 1973) have described, in single section analyses, the general cellular characteristics of the developmental processes. Formation of primary and secondary myotubes appears to occur broadly
as in small animals, although there has been no demonstration that primary myotubes necessarily form a
continuous, full-length framework for the muscle. At
later times, an additional, tertiary, generation of myotubes is reported in the largest muscles (sheep, Wilson
et al., 1992; human, Draeger et al., 1987). These tertiary myotubes use secondary myotubes as their physical support during the growth process, allowing a
faster rate of formation of new myotubes than if only
the (less numerous) primary myotubes were available
as a growth substratum. However, little is known
about the three-dimensional geometry of myotube formation within large muscles, and in particular about
the relationship between the multiple innervation
zones and sites of formation of new myotubes or about
the longitudinal extent of primary and secondary myotubes.
Here, we present a detailed analysis of the formation
of new myotubes in an embryonic larger muscle which
addresses general questions of large muscle morphogenesis, and also tests the general principle, first propounded in the small animal systems, that formation of
new secondary myotubes is restricted solely to sites of
innervation. As a model system we have chosen the

sternomastoid muscle of the embryonic guinea pig, because it has multiple innervation zones, a simple strap
form, and is not too large to study in its entirety in the
electron microscope. Our results demonstrate that in
this muscle, as in small rat muscles, all secondary myotube production is restricted to the close vicinity of endplates on primary myotubes. However, many more secondary myotubes are produced, because each of the
multiple innervation sites along a primary myotube
acts as a generative centre.

RESULTS
Time Course of Development of Guinea Pig
Sternomastoid Muscle
The time course of formation of primary and secondary myotubes in sternomastoid muscle was briefly assessed by electron microscopic examination of at least
two muscles at each of embryonic day 29 (E291, E31,
E33, E36, E40, and E50, in order to determine the most
appropriate time t o examine early secondary myotube
formation. In our colony, guinea pigs are born at 62-64
days of gestation.
At the earliest age studied, E29, well-formed primary myotubes were clustered in groups of 2-5 cells
(Fig. l a ) and already present in their full numbers (Table 1).No secondary myotubes were present. Primary
myotubes had mostly separated from their original
groups by E31 (Fig. lb), and occasional secondary myotubes were seen in close association with the primary
myotubes. Secondary myotubes were still rare a t E33,
with many primary myotubes having no associated secondary myotubes, but by E36 vigorous secondary myogenesis was underway, with every primary myotube
now having 1-6 associated secondary myotubes of
varying maturity. Primary and secondary myotubes
were easily distinguishable on the basis of the smaller
size and peripheral position of the secondary myotubes
in the star-like clusters (Fig. lc). By E40 most secondary myotubes had separated from the primary myotubes and had a mature organization with nuclei positioned peripherally; primary myotubes were still the
largest of the myogenic cells, and maintained their central nuclei. Formation of new secondary myotubes continued, but apparently at a slower rate, as most myotube clusters contained only one recently formed
secondary myotube, with sparse myofilament and central nuclei, lying under the basal lamina of the primary
myotube. The muscle had an essentially adult form by
E50 with all myogenic cells having the angular profile
and peripheral nuclei of mature muscle fibres, although there was great variation in fibre diameter
(Fig. Id). The fascicular arrangement of the muscle
was established a t this time, although development of
the perimysial sheaths was still rudimentary.
As a result of this preliminary work, embryonic day
36 was selected as the best time point for examination
of early secondary myotube formation in the guinea pig
sternomastoid muscle, as secondary myotubes were

393

MORPHOGENESIS OF STERNOMASTOID MUSCLE

Fig. 1. Electron micrographs showing stages in the embryonic development of guinea pig sternomastoid muscle. a: Embryonic day 29. A
group of four tightly clustered primary myotubes is shown. No secondary
myotubes are present. b: Embryonic day 31. Primary myotubes have
largely separated from their initial groups, and are larger and more mature. Secondary myotubes are rare, and none are present in this view.

c: Embryonic day 36. A single primary myotube is shown, surrounded by

a number of associated secondary myotubes
of varying maturity. This
is a period of rapid secondary myogenesis. d: Embryonic day 50. Myotubes have matured to become muscle fibres, and their origins as primary

r)

or secondary myotubes can no longer be distinguished. Calibration bar

5 pm in all cases.

394

DUXSON AND SHEARD

TABLE 1. Number of Primary Myotubes in the

Sternomastoid Muscle at Various Embryonic Ages"
Embryonic age
E29
E31
E33
E36

No. of primary myotubes

256
236
235
242

"Counts are from a single muscle at each stage. Primary myotubes could not be unambiguously identified at stages later
than E36. Average of all ages examined = 242.

forming rapidly, but were still morphologically distinguishable from primary myotubes.
Electron microscopic examination revealed no evidence of the formation of a tertiary generation of myotubes in the guinea pig sternomastoid muscle. Newly
formed myotubes were always attached to the primary
myotube, and never seen in close association with mature secondary myotubes, as has been described in the
sheep tibialis cranialis muscle (Wilson et al., 1992).
This point is further addressed later in this paper.

Pattern of Innervation of Embryonic and Adult

Sternomastoid Muscle
Acetylcholinesterase (AChE) histochemistry demonstrated that the muscle contained four endplate bands
at all embryonic stages studied, and in the adult. A
muscle from an E50 embryo is illustrated in Figure 2.
Thus, the broad patterning of the innervation zones is
established from the outset of muscle formation.
Semi-serial Electron Microscopic Analysis of
Myotube Extent, Position, and Innervation
The three-dimensional arrangement of primary and
secondary myotubes, and their relation to the innervating nerve were examined in a semi-serial electron
microscopic study of a single sternomastoid muscle,
taken from an E36 embryo. Sections were collected a t
10 pm intervals over a region extending from one tendon through three of the four innervation zones of the
muscle; this covered 4,500 pm in length, which was
approximately two-thirJs of the muscle belly. A detailed analysis was made of five myotube "clusters" in
the muscle (where a cluster is defined as the primary

myotube and its associated secondary myotubes), and

the results are presented in Figure 3.

Longitudinal Extent and Pattern of Innervation

of Primary Myotubes
All primary myotubes extended the full length of the
examined portion of the muscle; that is, they originated
from one tendon and passed through more than twothirds of the muscle belly, with an essentially unchanging cross-sectional area (Figs. 3 and 4). It thus seems
reasonable to propose that the primary myotubes of
guinea pig sternomastoid form a full-length (i.e., tendon-to-tendon) framework for the muscle, as do primary myotubes in the smaller animal models.
Each primary myotube was innervated at all three
innervation zones through which the study passed (Fig.
3). As an example, the three endplate regions of the
primary myotube of cluster 2 are shown in Figure 4a-f.
Several points may be noted, which applied generally
to all primary myotubes examined in the study.
Firstly, all three distinct innervation sites were of approximately equal size and maturity; there was no
structural evidence to suggest dominance of one site
over the others. Secondly, a t each endplate site, several
t o many nerve terminal profiles were in synaptic apposition to the primary myotube, suggesting that each
site is polyneuronally innervated (Korneliussen and
Jansen, 1976; Duxson, 1982). Thirdly, by reference to
Figure 3 it can be seen that the separation between
adjacent endplates on a single primary myotube
ranged between 1 and 1.5 mm. This is similar to the
minimum separation seen between original and experimentally induced ectopic endplates in adult muscles
(Kuffler et al., 1977, 19801, suggesting some competitive mechanism may dictate the minimum spacing of
the multiple endplates on primary myotubes. Finally,
the endplates on different clusters were not necessarily
well aligned. For example, within the third innervation zone, there was a definite staggering of the innervation sites of clusters 1 and 5 (which were immediate
neighbours) vs. those of clusters 2-4 (which were somewhat separate). Staggering of endplate zones between
adjacent fascicles is usual in adult sternomastoid muscles, and it appears that this lack of alignment is
present from the earliest stages of innervation, rather
than being a result of uneven growth of fibres.

Fig. 2. Embryonic day 50 muscle, stained for acetylcholinesterase to show endplate regions. The full
length of the muscle (approximately 15 mm) is shown. The four zones of innervation are indicated by t h e
arrowheads.

395

MORPHOGENESIS OF STERNOMASTOID MUSCLE

114
I13
I 12b

8-

7
3

Cluster 1
i
11

L
n

**

I-*
6
52-

10

*
I

A.

19

12a

Cluster 2

6a7 a w
2

. .

..

1-*
5

-*

..

6b

-7b

I-&-I~

from cluster 5
-.

**

1*-

2
3

4
4

. .

-*

..

15

-6

7a H

Cluster 4

12

Cluster 3

10

2
3

1-*
4-d

*\lo

-*

m..

16

I11

"

48

-9

*10

to cluster 3
4-

3-

16

."

..

1
I

80

160

>7

c/

I-*

Cluster 5

-9

240

Fig. 3. Position, longitudinal extent, and innervation of primary and

secondary myotubes, in five clusters of myotubes in an E36 muscle, as
determined from a semi-serial section study. Primary myotubes are represented by the bold lines numbered 1; all other lines represent secondary rnyotubes: sites of innervation are indicated by the asterisks. Secondary myotubes within each cluster are numbered in the order in which
they were encountered. The scale at the bottom represents the longitudinal axis of the muscle, and the numbers are the section level. The 460

320

400

460

section levels represent 4,600 Wm of muscle length. The manner of termination of each myotube is represented as follows: arrows = myotendinous junction: vertical line = termination on another myotube; no specific indication (at left of figure) = termination of the myotube was not
seen (myotubewas already present at start of study). The positions of the
three innervation bands are indicated by the arrows at the bottom, labelled 1-3. The fourth innervation zone of the muscle (not sectioned) lay
to the left of the origin.

396

DUXSON AND SHEARD

Fig. 4 a-d.

MORPHOGENESIS OF STERNOMASTOID MUSCLE

Fig. 4. Electron micrographs of the three regions of innervation of

cluster 2. Numbers on secondary myotubes correspond to cell numbering
used for cluster 2 in Figure 3. a: Section level 42. Cluster 2 in its entirety,
showing primary myotube and five associated secondary myotubes. The
arrowheads indicate regions of nerve terminals. The two cells with prominent heterochromatic nuclei, at bottom left, are myoblasts. b: Section
level 42. An enlargement of part of the innervation zone indicated by
arrowheads in a, and at the same orientation. The innervation zone of the
primary myotube (at right), and of a binucleate secondary myotube (cell
7a, at left) are shown. The nerve terminal marked with an asterisk is in
synaptic apposition to both primary and secondary myotubes. Although
the secondary myotube shows little sign of post-synaptic specialization,
there is a regular synaptic cleft with distinct basal lamina. c: Section level
175. Cluster 2 in its entirety, with primary myotube and four secondary

Relation of Secondary Myotube Position to Sites

of Primary Myotube Innervation
The major point arising here is that the position of

397

myotubes. A large number of nerve terminals are present in the regions

between the arrowheads. d: Section level 175. Enlargement of part of the
innervation zone indicated by the arrowheads in c, but note that the
image is rotated 90" clockwise compared to c. The innervation zones of
the primary myotube, and of secondary myotubes 5, 7b, and 8 are seen.
e: Section level 342. Cluster 2 in its entirety. The region of innervation lies
between the arrowheads. The open arrow indicates a pair of secondary
myotubes (from an adjacent cluster) which lie in close association, sharing a common basal lamina. (Note: Neither the origin of cell 12 or the
innervation of cell 9 are seen in this figure; they were both at an adjacent
section level.) f: Section level 342. Enlargement of the innervation region
indicated by arrowheads in e, at the same orientation. The innervation
zones of the primary myotube (centre bottom), and of secondary myotubes 5 and 10 are seen. Calibration bars: a, c, e = 5 pm; b, d, f = 1 pm.

tending up to cells with more than 30 nuclei (where on

average there was one nucleus per 100 pm of cell
length). The clearest association between innervation
new secondary myotubes is absolutely linked to sites of sites and secondary myotube position is seen with the
innervation of the muscle (Fig. 3). There were 56 sec- smallest cells: for example, in Cluster 2 of Figure 3,
ondary myotubes in the five clusters analysed; of these, there are five very small myotubes: cells 6a, 7a,and 12
38 were sectioned from end to end. These included a are binucleates, cells 8 and 11 each have 4 nuclei.
continuous series of myotube size, commencing with 3 These very small cells are precisely positioned over the
binucleate myotubes (size range 50-150 pm) and ex- innervation sites of their own cluster. This longitudi-

398

DUXSON AND SHEARD

nal association is even more striking when one considers that the total muscle length analysed is 4,500 pm,
the three innervation zones total only about 40 p m or
less than 1%of this, and yet new myotubes appear only
in overlap with these precise sites. The same point can
be demonstrated for any of the five clusters analysed.
The only myotubes in the study with no apparent
relation to an innervation zone are the small terminal
portions of several myotubes (e.g., myotubes 5 and 2 of
Cluster 1) which were present at the point where sectioning commenced (at left of Fig. 3). The fourth endplate zone of this muscle lay in an unsectioned region of
the muscle, to the left of the origin in Figure 3, and
these small myotube portions are presumed to arise
from myotubes associated with this fourth zone of innervation.
Longer secondary myotubes were always approximately centered on one of the primary myotube innervation sites, presumably reflecting balanced growth
about their site of origin. The secondaries in Figure 3
have been schematically represented alternately above
or below the line of the primary myotube, according to
their centering on a particular endplate zone. This allows estimation of the number of secondary myotubes
formed at each innervation site by E36. In the five
clusters presented, the range is from 2 to 5 myotubes
initiated per site.
Longitudinal Extent and Insertion of
Secondary Myotubes
The smallest secondary myotubes lay closely applied
to the primary myotube and under the primary basal
lamina along their whole length. They were attached
to the primary myotube by complex interdigitating
junctions at some points (e.g., cell 6a of Fig. 4a). Identical attachments have previously been described between primary myotubes and young secondary myotubes in rat lumbrical muscle (Duxson and Usson,
1989), and are believed to form the physical insertion
by which the young secondary myotubes maintain
their longitudinal orientation and exert tension.
As secondary myotubes lengthened and matured,
they separated from the primary myotube in their midregions, and transferred their insertion in one of two
ways. Those secondary myotubes which extended into
the region of the tendon developed a classical myotendinous junction, similar to the primary myotube. However, most large secondary myotubes did not end
within the tendon, and in these cases, the terminal
regions of secondary myotubes developed close associations with other secondaries, often displaying interdigitating secondary-to-secondary junctions, These
junctions were always formed between secondary myotubes which had different points of origin, so that the
termination of one cell generally overlapped the origin
of the other.
No secondary myotube extended the full length of
the region analysed. The longest secondary myotube
examined was cell 6 of cluster 1 (3,560 pm) (Fig. 31,

which was present in the first section of the series (i.e,,

at the origin of Fig. 3), and passed through the three
innervation zones before terminating in association
with both the primary myotube and a large secondary
myotube (cell 11) of its cluster. The longest secondary
myotube examined in its entirety was cell 10 of cluster
1 (3,530 pm) (see Fig, 3). The latter cell was seen to
insert into the tendon a t one extremity (represented by
the arrow head on cell 10 at the right hand side of Fig.
3): a t the other extremity it tapered and formed a close
association with cell 3 of the same cluster over a long
distance, finally terminating with no apparent attachment to the connective tissue framework of the muscle.
A pair of secondary myotubes in a similar close association can be seen a t the left of Figure 4e (open arrow).
This mode of termination of secondary myotube upon
secondary myotube may be the precursor of the myomyonal junctions which have been described in adult
muscles with long fascicle lengths and multiple innervation zones (e.g., Barrett, 1962; Torigoe and Nakamura, 1987; Zenker and Scheidegger, 1991). We are
currently studying the evolution of these junctions
from the embryonic to the adult form, and a fuller description of the embryonic junctions will appear in that
context. Finally, it should be noted that such pairs of
secondary myotubes lying close together have probably
been misinterpreted in single section studies as tertiary myotubes developing alongside secondary myotubes (Wilson et al., 1992). Their true nature can only
be seen after serial tracing of the cells.

Innervation of Secondary Myotubes

Synapses were observed on about half of the secondary myotubes associated with the three innervation
bands. However, no conclusion should be drawn as to
the state of innervation of the other myotubes, because
the study was designed to reliably detect only large
innervation sites on primary myotubes, and the 10 km
sampling interval almost certainly missed many small
synapses on secondary myotubes. We have previously
reported a serial section study in the rat lumbrical
which showed that about 90% of secondary myotubes
receive direct nerve contact in that muscle (Duxson,
1992).
For those secondary myotubes of guinea pig sternomastoid where innervation sites were detected, the following comments may be made. In general, secondary
myotubes had a single endplate region, close to their
mid-point. However, there were two cases where multiple, spatially separate synaptic sites were observed
(cell 5, cluster 2; cell 5, cluster 5). Cell 5 of cluster 2 is
an interesting example. In Figure 4c the cell is seen to
be innervated near its mid-point: in Figure 4e and f, a
distinct synapse is present only a few micrometers from
the end of the cell.
Synapses on secondary myotubes varied in size and
state of maturity; however, they usually displayed
multiple nerve terminal profiles apposed to a thickened
post-synaptic membrane and a regular synaptic cleft

MORPHOGENESIS OF STERNOMASTOID MUSCLE

399

Fig. 5. Synapse on a rnyoblast. The synaptic nerve terminal (arrow) shows a small pre-synaptic density
apposing the myobiast. The post-synaptic regions of the myoblast membrane also show synaptic thickening,
and there is a regular synaptic cleft with distinctive synaptic basal lamina. rnb = rnyoblast, 1 = prirnaty
rnyotube. Calibration bar = 1 prn.

with thickened basal lamina (Fig. 4b,d,f). Synaptic terminals were very commonly positioned between the
primary and a secondary myotube, with a synaptic apposition to both cells (e.g., the terminal marked * in
Fig. 4b). We have previously described such sharing
of nerve terminals in the rat (Duxson et al., 1988), and
suggest it is an intermediate stage in the transfer of
differentiated nerve terminals from an initial site on
the primary myotube, to a site on a newly formed secondary myotube.

Innervation of Myoblasts
Clear cases of synapses on myoblasts were also noted
in the course of this study (Fig. 5 ) . Our definition of
myoblasts is that used by Franzini-Armstrong and
Fischman (1994) of postmitotic, mononucleated cells
capable of fusion and of synthesizing (muscle specific)
contractile proteins. The cells were identified as myoblasts by their presence beneath the basal lamina of
the primary myotube, their spindle shape, and the
presence of only a single nucleus in the semi-serial
series. Innervated rnyoblasts were always placed close
to an innervation zone of a primary andlor a secondary
myotube. Synapses on myoblasts had an immature
form, but not notably more so than those on newly
formed secondary myotubes (compare Fig. 5 to figures
3, 6, or 8 of Kelly and Zacks, 196933; see also Duxson,
1992, for a serial section analysis of innervation of

young myotubes and myoblasts). In several cases the

myoblasts were so elongated as to be mistaken for the
extremities of young secondary myotubes (where myofibrils are usually absent), until sectioning revealed
they had only a single nucleus. A very few of these
well-formed rnyoblasts also exhibited a small amount
of organized myofilament, something rarely observed
in the myoblasts of limb muscles, and never observed
by us in our previous serial section studies of developing rat muscles. These observations make it apparent
that myogenic cells of developing limb muscles may
receive innervation and, in rare cases, differentiate to
the point of forming myofibrils, whilst still in the
mononucleated state.

Relationship of Embryonic Myotube Type to

Adult Muscle Fibre Type
Immunohistochemistry using antibodies to fast or
slow myosin heavy chain (MHC) showed that the adult
guinea pig sternomastoid is a predominantly fast muscle, with approximately 90% of cells reacting positively
with antibody MY32 (Sigma, St. Louis, MO) (anti-fast
MHC). Reaction with the antibody NOQ7.1.1A (Draeger et al., 1987) (anti-slow MHC) gave a pattern of
scattered darkly reacting fibres with a surround of nonreactive cells (Fig. 61, but with one small region entirely lacking reactive fibres (at left, Fig. 6). Adjacent
sections stained with antibody MY32 gave the inverse

Fig. 6. Distributionof slow myosin-positive fibres (dark-stained) within a cross-sectionof adult sternomastoid muscle. Note the region at left which displays no slow myosin-positive fibres. (Reacted with antibody
NOQ7.1.1A. anti-slow MHC.) Calibration bar = 500 krn.

MORPHOGENESIS OF STERNOMASTOID MUSCLE

401

Fig. 7. A small region of adult sternomastoid muscle, showing complementary distribution of fast rnyosin-positive and slow myosin-positive
fibres. a: Reacted with antibody MY32 (anti-fast myosin). Non-reacting

fibres are marked with asterisks. b: Reacted with antibody NOQ7.1.1A

(anti-slow myosin). The positively reacting fibres correspond to those
which did not react in a. Calibration bar = 100 pm.

pattern. In the mixed region of the muscle, each slow

muscle fibre was surrounded by a ring of 10-12 fast
fibres, the whole making up a small sub-fascicle (Fig.
7a,b). This pattern of staining is typical of muscles
where fibres developing from the primary myotubes
remain slow, whilst secondary myotubes form the fast
fibres (Rubinstein and Kelly, 1981).
The possibility that all slow fibres in the adult sternomastoid are derived from primary myotubes was examined by comparing the number of primary myotubes
in the embryonic muscle (Table 1)with the number of
slow fibres observed in the adult muscle. A single adult
sternomastoid was sectioned at four evenly spaced
points along the length of the muscle belly, sections
labelled with the anti-slow MHC, and the number of
reactive fibres determined in each cross section. The
muscle contained an average of 201 slow fibres per
cross section, as compared to an average of 242 primary
myotubes in the embryonic muscle. When it is considered that primary myotubes in the purely fast region of
the muscle (15-20% of total muscle area) must become
fast fibres in the adult, then this result is consistent
with the idea that adult slow muscle fibres in sternomastoid muscle have a primary myotube origin.
An apparent confounding factor is that, as may be
seen in Figure 6, a small number of slow fibres in the
adult muscle occur in doublets and triplets. However,
other studies in association with our laboratory have
shown this is not true in the first 6 months of postnatal
life, but evolves slowly as the animal ages (Dr. Wade
Grow, personal communication). The doublets and triplets may arise when sprouts from preterminal axons on
the (primary myotube-derived) slow fibres colonise
neighbouring initially fast-MHC expressing (secondary myotube-derived) fibres, and convert them to slow
fibres. Progressive expansion of slow motor units with
age has been previously reported in other systems (e.g.,
Gordon and Van Essen, 1985).

DISCUSSION
Secondary Myotubes Form Only at Sites
of Innervation
This study confirms our previous report (Duxson et
al., 1989) that, in vivo, the fusion event by which new
secondary myotubes arise from myoblasts occurs only
in intimate relation with sites of innervation. The earlier study was made on the fourth lumbrical muscle of
the rat hindlimb, a very small muscle with only a single innervation zone. In this muscle, the possibility existed that the restriction of new myotube formation to
the innervation zone was coincidental, perhaps occurring only because the central muscle region was developmentally the most advanced region. The present
study of the guinea pig sternomastoid muscle removes
this possibility, and demonstrates unequivocally that
secondary myotube formation is absolutely linked to
sites of innervation, wherever they occur along the longitudinal axis of an embryonic muscle.
It is not clear what links myotube formation to innervation zones. One possibility is that some common
patterning influence, perhaps expressed in the extracellular matrix, might determine both the sites of innervation and the sites of myoblast fusion. If a patterning influence is important it must act in concert with
activity-based mechanisms, as it is well known that
suppression of neural activity markedly depresses production of secondary myotubes (Harris, 1981; McLennan, 1983; Ross et al., 1987b; Ashby et al., 1993). Perhaps a simpler explanation is that the presence of
nerve terminals leads directly or indirectly to induction of fusion competence in the local myoblast population, or to an acceleration of processes leading to fusion. Growing nerve terminals continuously secrete
acetylcholine (Cohen, 1980) and contain other substances which may exert a trophic influence on myogenic cells. For example, calcitonin gene related peptide (CGRP) is reported to be present in the cell bodies

402

DUXSON AND SHEARD

and terminals of embryonic motor neurons, and to induce a specific increase in synthesis of acetylcholine
receptors (AChR) in primary cultures of embryonic
muscle (Fontaine et al., 1987; New and Mudge, 1986).
If CGRP induced expression of acetylcholine receptors
on myoblasts in the innervation zone of normally developing muscle, then acetylcholine released from
nerve terminals could subsequently cause depolarization and an influx of calcium ions into these myoblasts.
It has been recently reported that calcium influx into
myogenic cells is a potent inducer of changes in gene
and protein expression (Abu-Shakra et al., 1993), and
myoblasts receiving an effective innervation may be
altered in such a way as to enable them to participate
in the initial myoblast-myoblast fusion which results
in formation of a new myotube. Alternatively, the influx of calcium ions through AChR may lead directly to
myoblast fusion (e.g., Entwhistle et al., 1988). In relation to either alternative, it is notable that (1) direct
synapses on myoblasts do occur in both the rat fourth
lumbrical and the guinea pig sternomastoid muscles
(Duxson, 1992; Fig. 5) and (2) in myogenin null mutant
mice, where virtually no myoblast fusion occurs and
musclesfconsist of a mass of mononucleated cells with
only occasional myotubes, most AChR sub-units are
not expressed and there are no functional AChR (Hasty
et al., 1993; Nabeshima et al., 1993).
How can these arguments be reconciled with observations that some secondary myotube formation is seen
in mutant mouse muscles which are never innervated
(Ashby et al., 1993), and that fusion occurs freely in
culture systems where nerve is not present? If the key
to myoblast fusion or to induction of the preconditions
for fusion is calcium ion influx, this can be induced in
a number of ways other than by neural activation of
AChR. Embryonic muscles are highly susceptible to
stretch-induced depolarization (Dennis et al., 1981),
and denervated muscles to fasciculation resulting from
spontaneous depolarization, both of which result in calcium influx into the affected cells. Such spontaneous
depolarizations of aneural embryonic muscles would be
expected to affect the primary myotubes, and also myoblasts which were postmitotic, extended along the longitudinal muscle axis, and in the early stages of differentiation. In culture systems, too, spontaneous
depolarization of myogenic cells is normally found, and
again makes a link between the occurrence of fusion
and conditions permitting calcium influx, During normal development, the link between nerve terminals
and sites of secondary myotube fusion might occur because calcium influx is precociously induced in the population of myoblasts which is in contact with nerve
terminals. These myoblasts would then reach fusion
competence in advance of the remaining myoblast population, forming an expanding nucleus of secondary
myotubes with which all remaining myoblasts are able
to fuse.
If neural induction of fusion or fusion competence
occurs, it almost certainly works alongside an inhibi-

tory inf hence which delays myoblast differentiation

and fusion in other muscle regions. Two observations
make it clear that myoblast fusion in vivo is subject to
controls which do not exist in vitro. First, whilst formation of new myotubes in vivo normally occurs only
at sites of innervation (this paper; Duxson et al., 1989),
myoblasts in primary culture fuse freely in the total
absence of nerve (e.g., Entwhistle et al., 1988). Second,
in the myogenin null mutant mouse where virtually no
myoblast fusion occurs in vivo, the explanted myoblasts fuse freely in primary culture (Nabeshima et al.,
1993). This inhibitory influence controlling myotube
formation in vivo most likely resides within the connective tissue matrix of the muscle. For example, both
acidic- and basic-fibroblast growth factors (FGF1 and
FGF2, respectively) are known to stimulate proliferation and inhibit differentiation of myoblasts (Clegg et
al., 1987) and FGF2 is expressed strongly in both myotubes and the endomysium of El8 rat fetal muscle
(Gonzalez et al., 1990). Transforming growth factor
(TGF)-P also strongly inhibits differentiation of fetal
myoblasts (those giving rise to secondary myotubes),
but is interestingly without effect on the embryonic
myoblasts which give rise, in a nerve-independent
fashion, to primary myotubes (Massague et al., 1986;
Cusella-De Angelis et al., 1994). Both TGF-P and FGF
repress expression of the myogenic determining genes
myoDl and myogenin in fetal myoblasts (Brunetti and
Goldfine, 1990; Vaidya et al., 19891, and are down-regulated in the later fetal period (Cusella-De Angelis et
al., 1994; Munaim et al., 1988; Seed and Hauschka,
1988; Seed et al., 1988). Thus the presence of FGF or
TGF-P in the muscle extracellular matrix might inhibit or delay terminal differentiation of fetal myoblasts a t all points except at the point of innervation,
where a specific stimulus acts to initiate differentiation
and fusion, resulting in the formation of new secondary
myotubes. As the level of these inhibitory agents falls,
then rnyoblasts outside the innervation zone might begin to differentiate, but fuse with existing secondary
myotubes rather than with each other.
Whatever the sequence of causality, the restriction of
secondary myotube formation to sites of innervation
during normal development has two important functional consequences. First, new myotubes form only at
sites where innervation is immediately available, and
are therefore rarely uninnervated; second, growing
myotubes are centred on their innervation site. These
consequences are crucial to the survival and efficient
functioning of growing myotubes.

Relation of Embryonic to Adult Muscle Form

The relationship between the number of primary
myotubes (approximately 240, Table 1) and the number of muscle fibres in the adult sternomastoid muscle
(approximately 2,600 in a single, mid-belly, cross section; Duxson, personal observations) shows that a t
least 90% of adult muscle fibres in guinea pig sternomastoid muscle must be derived from secondary myo-

MORPHOGENESIS OF STERNOMASTOID MUSCLE

tubes. From the embryonic pattern of secondary myotube generation we would predict that, in the adult
muscle, most muscle fibres will have a single innervation zone, will be arranged in a staggered manner in
the muscle with each fibre centred on the innervation
site which was its point of origin, and that most fibres
will not extend the whole length of the muscle, but end
intrafascicularly. Preliminary studies with confocal
and electron microscopy support the last point, showing that a large proportion of fibres in the adult guinea
pig sternomastoid are significantly shorter than the
whole muscle length and end (at one or both ends) in
various forms of intrafascicular termination (Duxson
and Young, unpublished observations). Thus it seems
that at least a large proportion of secondary myotubes
never extend from tendon to tendon.
These conclusions are consistent with, and explain
the origins of, the fibre arrangements which have been
extensively described in muscles with long fascicle
lengths. For example, Huber (1916) made a survey of
the form and arrangements of muscle fibres in a number of muscles in the hindlimb of adult rabbits, and
concluded that it was only within fascicles shorter than
2 cm that muscle fibres stretched from tendon to tendon of the muscle. As fascicle length increased, an increasing proportion of fibres ended intrafascicularly by
tapering and attaching to a neighbouring fibre. Many
fibres had one tendinous insertion and one intrafascicular termination, whilst others both commenced and
terminated at points distant to the tendon. Similar reports exist for adult muscles of the human (Barrett,
19621, goat (Gans et al., 19891,cat (Adrian, 1925; Richmond et al., 1985; Trotter, 19901, horse (Zenker et al.,
1990), and for some long muscles of the rat and mouse
(Torigoe and Nakamura, 1987; Zenker et al., 1990).
The work reported in this paper now explains the origins of muscle fibre positioning and arrangement in
these larger muscles, in terms of their embryogenesis.
The fact that many muscle fibres do not stretch the
entire length of larger muscles arises from the physiological constraints imposed by action potential conduction speed in the fibres and the time to peak contraction, as has been discussed in detail elsewhere (Zenker
et al., 1990; Zenker and Scheidegger, 19911, and from
the requirement that fibres be approximately centred
on their innervation site, in order to contract efficiently. A final point is that the proportion of muscle
fibres formed from secondary myotubes is probably
much greater than the 90% stated above, as a single
mid-belly section will miss large numbers of intrafascicularly terminating fibres and therefore greatly underestimate the total fibre number in the adult muscle.
The destiny of primary myotubes in the adult sternomastoid is less clear than that of secondary myotubes. In the embryo, primary myotubes stretch the
whole length of the muscle and are innervated at multiple sites. It has been suggested that this form is not
compatible with survival in a fast contracting muscle
with long fascicles, and that primary myotubes might

403

be progressively destroyed by damage during muscle

maturation (Prof. Michel Fardeau, personal communication). However, we have demonstrated above that, in
guinea pig sternomastoid muscle, both the numbers
and distribution of slow-MHC containing fibres in the
adult are consistent with their originating predominantly as primary myotubes. If this is correct, then
most slow fibres of the adult muscle are likely to extend
the full length of the muscle, to be innervated a t multiple discrete sites, and to be collected into a small
number of discrete slow motor units whose fibres are
derived entirely from primary myotubes. We are currently testing this conclusion, by evaluating the patterns of activation and contraction of the slow motor
units in the adult guinea pig sternomastoid muscle.
The evolution of the pattern of innervation of fibres
derived from primary myotubes is a final point of great
interest. The multiply innervated fibres might either
receive nerve terminals from the same neuron a t each
distinct endplate, or be innervated by different neurons
at different sites along their length. If the former is
true this has exciting repercussions for understanding
the mechanisms by which nervous connections are selected during development. In the latter case, the interest would lie in understanding the function of muscle fibres which were asymmetrically activated, and
contributed to multiple motor units.

EXPERIMENTAL PROCEDURES
Dated pregnancies were produced by overnight matings of males with young adult females in full estrus.
The morning after the mating was designated as embryonic day 0 (EO). A total of 23 embryonic guinea pigs
from 10 litters were used a t a range of gestational ages
from E29-E60. Birth normally occurred after 62 days
in our colony. Three adult animals were also used to
obtain data on the structure of the adult sternomastoid
muscle. Anaesthesia of either pregnant dams or nonpregnant adult animals was via intramuscular injection of the tranquillizer Rompan (6 mg/kg) followed by
100 mg/kg Ketamine. After removal of embryos or of
tissues, anaesthetized animals were killed by decapitation.
Localization of Endplate Bands
Endplate zones were localized by staining whole embryonic muscles or longitudinal frozen sections of adult
muscles for acetylcholinesterase (AChE) using the
method of Karnovsky and Roots (1964).

Myosin Histochemistry
Adult sternomastoid muscles were fixed by immersion in 3% neutral-buffered paraformaldehyde on ice
for 2 hr, washed in 0.1 M glycine in phosphate buffer to
quench aldehyde radicals, then cut into four full-width
segments before inclusion in embedding medium and
freezing in isopentane cooled with liquid nitrogen. Frozen cross-sections of the muscles were cut at 8 pm
thickness and collected on gelatin-coated glass slides.

404

DUXSON AND SHEARD

Immunohistochemistry used monoclonal antibody

NOQ7.1.1A (an antibody to slow myosin heavy chain)
(Harris et al., 1989) at a dilution of 1:500, or monoclonal antibody MY32 (Sigma, St. Louis, MO) (an antibody to fast myosin heavy chain) at 1:1,000,in 0.1 M
phosphate buffer with 1%added bovine serum albumin. Binding of antibodies was revealed by sequential
reaction with biotinylated goat anti-mouse second antibody (Amersham, Arlington Heights, IL) and biotinstreptavidin-HRP complex (Vector Laboratories, Burlingame, CA). The peroxidase development procedure
used the nickel/cobalt intensification technique of Adams (1981).

ACKNOWLEDGMENTS
This work was supported by a project grant from the
Wellcome Trust, and by equipment grants from the
Lottery Board of New Zealand and the Health Research Council of New Zealand. We thank Associate
Professor John Harris for useful discussion and comments on the manuscript. The authors dedicate this
article to their friend and mentor, Professor Gerta Vrbova, on the occasion of her retirement.

REFERENCES

Abu-Shakra, S.R., Cole, A.J. and Drachman, D.B. (1993) Nerve stimulation and denervation induce differential patterns of immediate
early gene mRNA expression in skeletal muscle. Mol. Brain Res.
Electron Microscopy
18:216-220.
Adams, J.C. (1981) Heavy metal intensification of DAB-based HRP
Sternomastoid muscles of 2-3 embryos at each of
reaction product. J. Histochem. Cytochem. 29:775.
E29, E31, E33, E36, E40, and E50 were fixed by per- Adrian, E.D. (1925) The spread of activity in the tenuissimus muscle
fusion through the heart with a solution containing 1% of the cat and in other complex muscles. J. Physiol. (Lond.) 60:301315.
glutaraldehyde, 1%paraformaldehyde, and 0.05 M glucose in 0.1 M phosphate buffer. The head of the embryo Ashby, P.R., Wilson, S.J., and Harris, A.J. (1993) Formation of primary and secondary myotubes in aneural muscles in the mouse
was fixed in place to maintain the sternomastoid musmutant peroneal muscular-atrophy. Dev. Biol. 1 5 6 5 1 9 4 2 8 .
cle at approximately resting length during perfusion. Ashmore, C.R., Robinson, D.W., Rattray, P., and Doerr, L. (1972) Biphasic development of muscle fibres in the fetal lamb. Exp. Neurol.
Muscles were then excised and immersed in fixative for
37241-255,
a further 2 hr on ice. Further processing included postC.R., Addis, P.B., and Doerr, L. (1973) Development of musfixation in 2% osmium tetroxide, block staining with Ashmore,
cle fibres in the fetal pig. J. Anim. Sci. 361088-1093.
2% aqueous uranyl acetate, dehydration in alcohols Barrett, B. (1962) The length and mode of termination of individual
and propylene oxide, and embedment in TAAB epoxy
muscle fibres in the human sartorius and posterior femoral muscles. Acta Anat. 48:242-257.
resin. Embryonic muscles were divided, after processing, into three full width segments of roughly equal Brunetti, A,, and Goldfhe, I.D. (1990) Role of myogenin in myoblast
differentiation and its regulation by fibroblast growth factor. J.
length, and embedded with known orientation. For the
Biol. Chem. 2655960-5963.
description of the stages of development of the sterno- Clegg, C.H., Linkhart, T.A., Olwin, B.B., and Hauschka, S.D. (1987)
Growth factor control of skeletal muscle differentiation: Cornmittmastoid muscle, silver-gold sections were cut from apment to terminal differentiation occurs in G1 phase and is repressed
proximately the middle of each muscle.
by fibroblast growth factor. J. Cell Biol. 105:949-956.
For the systematic study of sites of secondary myo- Coers,
C., and Woolf, A.L. (1959) The Innervation of Muscle: A Bitube production, a single E36 muscle was semi-serially
opsy Study. Oxford: Blackwell Scientific Publications.
sectioned over more than two-thirds of its length (ap- Cohen, S.A. (1980) Early nerve-muscle synapses in vitro release
transmitter over postsynaptic membrane having low acetylcholine
proximately 5 mm); sectioning ended at the sternal
sensitivity. Roc. Nat. Acad. Sci. U S A . 77:644-648.
tendon. A group of ten ultrathin (90-100 nm) sections
Cusella-De Angelis, M.G., Molinari, S., Le Donne, A., Coletta, M.,
was collected onto Formvar-coated slot grids every 9
Vivarelli, E., Bouche, M., Molinaro, M., Ferrari, S., and Cossu, G.
pm. The interval of 10 pm was selected as the largest
(1994) Differential response of embryonic and fetal myoblasts to
interval likely to give reliable detection of all aligned
TGF-beta: A possible regulatory mechanism of skeletal muscle histogenesis. Development 120:925-933.
myogenic nuclei (these are 12-15 pm in length) and of
M.J., Ziskind-Conhaim, L., and Harris, A.J. (1981) Developall innervation zones on the primary myotubes. Sec- Dennis,
ment of neuromuscularjunctions in rat embryos. Dev. Biol. 81:266tions were stained with lead acetate and uranyl aee279.
tate, using a Reichert Ultrastainer. Detailed analysis Draeger, A., Weeds, A.G., and Fitzsimons, R.B. (1987) Primary, secondary and tertiary myotubes in developing skeletal muscle: A new
was made of a group of five, initially neighbouring,
approach to the analysis of human myogenesis. J . Neurol. Sci. 81:
myotube clusters a t the edge of the muscle. These were
19-43.
systematically photographed at each interval and the Duxson, M.J. (1982) The effect of postsynaptic block on development
serial photomontages used to construct a longitudinal
of the neuromuscular junction in postnatal rats. J. Neurocytol. 11:
395-408.
map indicating extent, origin, and insertion of every
myotube within the clusters. All points of innervation Duxson, M.J. (1992) The relationship of nerve to myoblasts and
newly-formed secondary myotubes in the fourth lumbrical muscle
of myotube clusters were also noted from the photoof the rat foetus. J. Neurocytol. 21574-588.
montages and later rephotographed a t higher magni- Duxson, M.J., and Usson, Y. (1989) Cellular insertion of primary and
fication so that exact details of the individual innervasecondary myotubes in embryonic rat muscles. Development 107:
243-251.
tion of both primary and secondary myotubes could be
established. We believe that this protocol located all Duxson, M.J., ROSS,J.J.,and Harris, A.J. (1988) Transfer of differentiated synaptic terminals from primary myotubes to new-formed
endplates on primary myotubes (which are generally of
muscle cells during embryonic development in the rat. Neurosci.
substantial size) but probably missed many of the
Lett. 71:147-152.
smaller synaptic contacts on secondary myotubes.
Duxson, M.J., Usson, Y., and Harris A.J. (1989) The origin of second-

MORPHOGENESIS OF STERNOMASTOID MUSCLE

ary myotubes in mammalian skeletal muscles: Ultrastructural
studies. Development 107:743-750.
Entwhistle, A., Zalin, R.J., Warner, A.F., and Bevan, S. (1988) A role
for acetylcholine receptors in the fusion of chick myoblasts. J. Cell
Biol. 106:1703-1712.
Fontaine, B., Klarsfield, A., and Changeux, J.-P. (1987) Calcitonin
gene-related peptide and muscle activity regulate acetylcholine receptor alpha-subunit mRNA levels by distinct intracellular pathways. J. Cell Biol. 105:1337-1342.
Franzini-Armstrong, C., and Fischman, D.A. (1994) Morphogenesis of
skeletal muscle fibres. In: Myology,Vol. 1,Engel, A.G., and Franzini-Armstrong, c. (eds). New York: McGraw-Hill, Inc., pp. 74-96.
Gans, C . , and Gaunt, A S . (1992) Muscle architecture and control
demands. Brain Behav. Evol. 40:70-81.
Gans, C., Loeb, G.E.and DeVree, F. (1989) Architecture and consequent physiological properties of the semitendinosus muscle in domestic goats. J . Morphol. 199287-297.
Gonzalez, A.-M., Buscaglia, M., Ong, M. and Baird, A. (1990) Distribution of basic fibroblast growth factor in the 18-day rat fetus:
Localization in the basement membranes of diverse tissues. J. Cell
Biol. 110:753-765.
Gordon, H., and Van Essen, D.C. (1985) Specific innervation o f muscle
fibre types in a developmentally polyinnervated muscle. Dev. Biol.
111:42-50.
Harris, A.J. (1981) Embryonic growth and innervation of rat skeletal
muscles. I. Neural regulation of muscle fibre numbers. Philos.
Trans. R. Sac. Lond. IBioll 293:257-277.
Harris, A.J., Fitzsimons, R.B., and McEwan, J.C. (1989) Neural control of the sequence of expression of myosin heavy chain isoforms in
foetal mammalian muscles. Development 107:751-769.
Hasty, P., Bradley, A,, Morris, J.H., Edmondson, D.G., Venuti, J.M.,
Olsen, E.N., and Klein, W.H. (1993) Muscle deficiency and neonatal
death in mice with a targeted mutation in the myogenin gene. Nature 364501-506.
Huber, G.C. (1916) On the form and arrangement in fasciculi of striated voluntary muscle fibers: A preliminary report. Anat. Rec. 11:
149-168.
Karnovsky, M.J., and Roots, L. (1964) A direct coloring thiocholine
method for cholinesterases. J. Histochem. Cytochem. 12219-221.
Kelly, A.M., and Zacks, S.I.(1969a) The histogenesis of rat intercostal
muscle. J. Cell Biol. 42:135-153.
Kelly, A.M., and Zacks, S.I. (1969b) The fine structure of motor endplate morphogenesis. J. Cell Biol. 42154-169.
Korneliussen, H., and Jansen, J.K.S. (1976) Morphological aspects of
the elimination of polyneuronal innervation of skeletal muscle fibres in newborn rats. J. Neurocytol. 5591-604.
Kuffler, D.P., Thompson, W., and Jansen, J.K.S. (1977) The elimination of synapses in multiply-innervated skeletal muscle fibres of the
rat: Dependence on distance between end-plates. Brain Res. 138:
353-358.
Kuffler, D.P., Thompson, W., and Jansen, J.K.S. (1980) The fate of
foreign endplates in cross innervated rat soleus muscle. Proc. R.
SOC.Lond. [Bioll 177:189-222.
Loeb, G.E., Pratt, C.A., Chanaud, C.M., and Richmond, F.J.R. (1987)
Distribution and innervation of short, interdigitated muscle fibres

405

in parallel-fibered muscles of the cat hindlimb. J . Morphol. 19l:l15.

Massaguk, J., Cheifetz, S., Endo, T., and Nadal-Ginard, B. (1986) Type
beta transforming growth factor is a n inhibitor of myogenic differentiation. Proc. Natl. Acad. Sci. U.S.A. 83:8206-8210.
McLennan, I.S. (1983) Neural dependence and independence of myotube production in chicken hindlimb muscles. Dev. Biol. 98:287294.
Munaim, S.I., Klagsbrun, M., and Toole, B.P. (1988) Developmental
changes in fibroblast growth factor in the chicken embryo limb bud.
Proc. Natl. Acad. Sci. U S A . 85:8091-8093.
Nabeshima, Y., Hanaoka, K., Hayasaka, M., Esumi, E., Li, S., Nonaka, I., and Nabeshima, Y. (1993) Myogenin gene disruption results
in perinatal lethality because of severe muscle defect. Nature 364:
532-535.
New, H., and Mudge, A. (1986) Calcitonin gene-related peptide regulates muscle acetylcholine receptor synthesis. Nature 323:809811.

Richmond, F.J.R., MacGillis, D.R.R., and Scott, D.A. (1985) Musclefibre compartmentalization in cat splenius muscle. J . Neurophysiol.
532368-885.
Ross, J.J., Duxson, M.J., and Harris, A.J. (1987a) Formation of primary and secondary myotubes in rat lumbrical muscles. Development 100:383-394.
ROSE,J.J., Duxson, M.J., and Harris, A.J. (1987b3 Neural determination of muscle fibre numbers in embryonic rat lumbrical muscles.
Development 100:395-409.
Rubinstein, N.A., and Kelly, A.M. (1981) Development of muscle fiber
specialization in the rat hindlimb. J. Cell Biol. 90:128-144.
Seed, J., and Hauschka, S.D. (1988) Clonal analysis of vertebrate
myogenesis. VIII. Fibroblast growth factor (FGFI-dependent and
FGF-independent muscle colony types during chick wing development. Dev. Biol. 128:40-49.
Seed, J., Olwin, B.B., and Hauschka, S.D. (1988)Fibroblast growth
factor levels in the whole embryo and limb bud during chick development. Dev. Biol. 128:50-57.
Torigoe, K., and Nakamura, T. (1987) Fine structure of myomyous
junctions in the mouse skeletal muscles. Tissue Cell 19243-250.
Trotter, J.A. (1990) Interfiber tension transmission in series-fibred
muscles of the cat hindlimb. J. Morphol. 206:351-361.
Vaidya, T.B., Rhodes, S.J.,Taparowsky, E.J., and Konieczny, S.F.
(1989) Fibroblast growth factor and transforming growth factorbeta repress transcription of the myogenic regulatory gene MyoD1.
Mol. Cell. Biol. 9:3576-3579.
Wilson, S.J., McEwan, J.C., Sheard, P.W., and Harris, A.J. (1992)
Early stages of myogenesis in a large mammal: Formation of successive generations of myotubes in sheep tibialis cranialis muscle.
J . Muscle Res. Cell Motil. 13:534-550.
Zenker, W., and Scheidegger, E.P. (1991) Quantitative aspects of muscle innervation. In: Plasticity of Motoneuronal Connections, A.
Wernig (ed). Amsterdam, The Netherlands: Elsevier Science Publishers, pp. 37-48.
Zenker, W., Snobl, D., and Boetschi, R. (1990) Multifocal innervation
and muscle length: A morphological study on the role of myo-myonal junctions, fiber branching and multiple innervation in muscles
of different size and shape. Anat. Embryol. 182273-283.