/. Embryo/, exp. Morph. Vol. 54, pp. 141-154, 1979
Printed in Great Britain © Company of Biologists Limited 1977
Segmental innervation of the chick forelimb
following embryonic manipulation
By M. R. BENNETT, 1 R. LINDEMAN 1 AND A. G. PETT1GREW 1
From the Neurobiology Laboratory, Department of Physiology,
University of Sydney
SUMMARY
A number of studies have shown that the segmental innervation of some muscles in the
developing limb undergoes some modification during the earliest stages of ontogeny. These
observations can be interpreted in support of the hypothesis that the motor axons and
muscles are matched during this period of development. As a further test of this suggestion
we have made a quantitative examination of the motor innervation of the chick forelimb
under conditions of controlled abnormal development.
Embryos were surgically manipulated at stages before the motor axons invade the limb.
The operations were controlled such that forelimbs were induced with segments deleted or
reduplicated or simply that a segment of the spinal cord had been deleted. In preparations
with abnormal limbs the motor innervation of the muscles present was the same as for
those muscles in the normal limb. Where a spinal segment had been deleted the limbs
developed normally and their innervation was completed by the remaining brachial segments.
These results suggest that any particular matching property of a developing muscle does
not develop as a consequence of its position in the limb relative to those segments of the
limb proximal to it. Furthermore, that some muscles which are normally innervated by two
spinal segments can be completely innervated by one of those spinal segments, in the absence
of the other, suggests that any matching between growing axons and developing muscle cells
is hierarchical rather than strictly all-or-nothing.
INTRODUCTION
The segmental distribution of motoneurones which project to particular
muscles in amphibian and avian limbs changes during the earliest stages of
ontogeny (Lamb, 1976, 1977; McGrath & Bennett, 1979; Pettigrew, Lindeman
& Bennett, 1979; see, however, Landmesser & Morris, 1975). Furthermore,
Harris & Dennis (1977) and McGrath & Bennett (1979) have shown that
functional synapses formed by some axon terminals in a muscle regress in
favour of those synapses formed by the axons which are destined to form the
mature innervation of the muscle. These results are not consistent with the
idea that axons are selectively guided to a muscle such that the initial projection
is the same as that observed in mature animals of the same species (see Sperry,
1963; Landmesser & Morris, 1975).
1
Authors' address: The Neurobiology Laboratory, Department of Physiology, University
of Sydney, Sydney, N.S.W., Australia 2006.
142
M. R. BENNETT, R. LINDEMAN AND A. G. PETTIGREW
The observations above on developing nerve-muscle systems have been
likened to those made on adult axolotls, where muscles can be competitively
reinnervated by axon terminals of both the original and a foreign nerve. In such
preparations both types of terminal compete for synaptic sites and this is
followed by the elimination of the terminals formed by the foreign nerve (Cass &
Mark, 1975; Bennett & Raftos, 1977; Dennis & Yip, 1978; Bennett, McGrath &
Davey, 1979). The suggestion has been made that motor nerve terminals
are matched to particular muscles of the limb (see Bennett & Pettigrew, 1976)
and that when well-matched and poorly matched terminals compete for the
same synaptic site, the poorly matched terminal will be eliminated (Bennett &
Raftos, 1977; Dennis & Yip, 1978; Bennett et al. 1979). Thus, during development it is possible that the changes in the segmental innervation of a muscle
reflects the elimination of poorly matched terminals from the initial projection
to that muscle.
In the present work we have investigated this suggestion further by attempting
to force newly formed muscles to accept a stable innervation from nerve
terminals which are normally eliminated. These experiments have involved the
surgical manipulation of embryos at very early stages of development to
produce either abnormal limbs, or embryos with abnormal spinal cords.
In an earlier study, Stirling & Summerbell (1977) made a largely histological
analysis of the pattern of nerves in forelimbs with truncations or deletions
along the proximo-distal axis. These authors found that the pattern in those
parts of the limb which were present was essentially normal, and that the
innervation of some of the major muscles was qualitatively similar to that in
the normal limb. The present study both confirms and extends the work of
Stirling & Summerbell (1977) by providing a quantitative analysis of muscle
innervation in such limbs with truncations and deletions and also in limbs
with reduplications of various segments or normal limbs innervated by an
abnormal spinal cord.
METHODS
Experiments were performed on White Leghorn embryos, staged according
to Hamburger & Hamilton (1951).
Operations
All operations were performed on embryos between stages 17 and 25.
Truncation of the limb at various levels was achieved by removing the entire
apical ectodermal ridge (AER) (Summerbell, 1974#). For trunction at the wrist,
the AER was removed at stage 22 and for truncation at the elbow, the operation
was performed at stage 20. To produce limbs with mirror image reduplications
(Summerbell, 19746) about the proximo-distal axis at the level of the elbow, a
small piece of tissue, which contained the zone of polarizing activity (ZPA),
was taken from the posterior edge of a donor limb at stage 21 and then placed
Segmental innervation of chick forelimb
143
in a similar sized niche on the anterior edge of a host limb at the same stage.
The graft was held in place by a fine tungsten pin. To produce limbs reduplicated
at the wrist, a similar operation was performed at stages 23-24. To delete
various portions of the proximo-distal axis of the forelimb, the wing tip,
containing the AER and some mesenchyme, of older stage embryos, was
attached to wing stumps of younger embryos (Summerbell & Lewis, 1975).
For preparations where digits were connected to the scapula, most of a stage-18
limb bud was removed and replaced by the AER of a stage-23 limb bud. For
preparations where the humerus was deleted, the donor AER was taken from
a stage-21 embryo and attached to a stage-18 embryo.
Extirpation of a spinal segment was carried out on embryos at stages 18-20.
Following identification of the segment to be removed, the spinal cord was cut
on either side of the somite using iris forceps and the segment was removed
by suction.
After each of these operations surviving embryos were incubated for about
a further week.
Estimation of muscle innervation
The pattern of innervation of various muscles in both control and operated
forelimbs was determined by recording either the electrical activity in the
muscle nerves or the contraction of each muscle in response to stimulation of
spinal nerves 14, 15 and 16. All embryos were examined at 10-14 days' incubation. Individual embryos were removed from the egg and decapitated. The
operated limb, together with the brachial spinal cord, was dissected free and
placed in a perspex organ bath containing a circulating modified Ringer
solution (Pettigrew et al. 1979). The limb was skinned and the spinal roots were
exposed and cut close to the spinal cord. The whole preparation was then
pinned at the joints so that movement artifacts would not complicate the
recording of muscle contraction. The spinal nerves were stimulated via a
suction glass capillary electrode using pulses of 0-01-0-05 msec and 1-10 V.
Muscle nerve activity was recorded in the intrinsic hand muscles (extensor
indicis brevis, eib; abductor medius, am; flexor digiti quarti, fdq) using another
suction glass capillary electrode placed over the point of entry of the individual
nerve to each muscle. The tetanic contraction of the major limb muscles
(biceps, b; triceps, t; flexor digitorum profundus,/d/>; flexor carpi ulnaris, feu;
extensor metacarpi radialis, emr) was recorded using a Grass tension transducer.
Individual muscles and tendons were freed from the distal insertion (the
proximal insertion and point of nerve entry were left intact) and connected at
rest length to the transducer using suture cotton tied securely around the tendon.
The output of the transducer was monitored on a chart recorder. Repetitive
stimulation (40 Hz) of each spinal nerve was halted as soon as the contraction
reached its maximum amplitude and this amplitude was compared to that
recorded with simultaneous stimulation of all three spinal nerves. The amplitude
144
M. R. BENNETT, R. LINDEMAN AND A. G. PETTIGREW
of the contraction produced by stimulation of all spinal nerves has been used
to provide a measure of the total number of muscle fibres which are innervated.
The comparison of the response to stimulation of one spinal nerve with the
response to combined stimulation provides, therefore, a measure of the proportion of muscle fibres in the muscle which are innervated by that particular
spinal nerve. Thus in this study, the percentage innervation of a muscle by a
spinal nerve has been determined by expressing the size of the tetanic contraction with stimulation of that nerve, as a percentage of the contraction with
simultaneous stimulation of all three spinal nerves. At the end of an experiment,
the limb was stained for cartilage using the Lundvall technique.
RESULTS
Innervation of the normal forelimb
The forelimb of the chick receives the major part of its innervation via spinal
nerves 14, 15 and 16 (Roncali, 1970). Nerve 13 also innervates the limb in some
embryos but in each case its contribution is small and has not been included
in the present study. The normal innervation of the major muscles of the limb
and the muscles in the hand is shown in the control distribution in Figs. 1 and
2. The proximal part of the limb (biceps and triceps) is innervated mainly
by nerves 14 and 15; nerve 16 contributes only about 10% of the innervation
of the triceps muscle. The more distal muscles in the limb are, in general,
innervated by more caudal spinal segments. Nerve 14 is found only in extensor
metacarpi radialis. The remaining distal forelimb muscles (flexor carpi ulnaris
and flexor digitorum profundus) are innervated to some extent by nerve 15
but mainly by nerve 16. The intrinsic muscles of the hand are innervated almost
completely by nerve 16. None of the muscles listed above is innervated by
more than two of the three spinal segments studied.
In a number of cases the summed percentage innervation of a muscle by two
spinal nerves exceeds 100 %. This observation has been noted previously in the
development of muscle innervation in the chick wing (Pettigrew et al. 1979).
The most likely explanation for this observation is that some muscle fibres are
dually innervated at their synaptic site (Pettigrew et al. 1979) by separate axons
from each of the spinal nerves. These muscle fibres will develop a maximal
tension in response to activity in either nerve terminal, and simultaneous
stimulation of both terminals will have no additional effect on the tension
developed. Thus, while the tension produced in response to simultaneous
stimulation of spinal nerves provides an accurate measure of the total number
of muscle fibres which are innervated, it also probably reflects an underestimate
of the total number of active synapses.
Segmental innervation of chick forelimb
100
145
F
S 14
S15
S 16
Muscle
Fig. 1. Innervation of the major muscles of the chick forelimb. The percentage
innervation (see Methods) of forelimb muscles by segmental nerves S14, S15 and
SI6 is shown for control limbs (filled circles) and for limbs with reduplications
below the elbow (open circles, anterior reduplication; open triangles, posterior
reduplication). Preparations were examined at 12 days' incubation. Each point
represents the mean ± S.E.M. of four determinations for each muscle in each of eight
preparations. Note that the total percentage innervation of some muscles exceeds
100%. This is believed to be the result of dual synaptic contact of some muscle
fibres by separate axons from the two spinal nerves which innervate each muscle
(see text). Abbreviations are b, biceps; emr, extensor metacarpi radialis; /, triceps;
feu flexor carpi ulnaris ;fdp, flexor digitorum profundus. Note that the innervation
of all parts of the reduplicated limbs is substantially the same as that of the control
limb.
Innervation offorelimbs with reduplications about the proximo-distal axis
Two types of reduplicated or 'twinned' limbs were examined. Embryos were
used only when the gross anatomy of the reduplicated parts appeared quite
normal and all the muscles of interest could be easily identified.
The first type of preparation had been reduplicated at, or just proximal to,
the elbow and had two distal forelimbs and hands (Fig. 3B). We have studied
only those preparations where the radii were completely separated. In all such
preparations, the segmental innervation of corresponding muscles identified in
both distal forelimbs was almost identical. Furthermore, the segmental innervation of the muscles studied, in both the proximal and reduplicated parts of
the abnormal limb, was essentially the same as that for those muscles in the
146
M. R. BENNETT, R. LINDEMAN AND A. G. PETTIGREW
100
16
St'gmental nerve
Fig. 2. Innervation of the intrinsic hand muscles in the chick forelimb. The percentage innervation of the extensor indicis brevis (A), abductor medius (B) and
flexor quarti (C) by segmental nerves S14, S15 and SI6 are shown for control limbs
(filled colums), limbs with reduplications below the wrist (open columns, anterior
reduplication on the left, posterior reduplication on the right) and limbs where the
digits were attached directly to the scapula (hatched columns). Error bars show
± S.E.M. for three determinations for each muscle in at least two preparations. In all
cases the intrinsic hand muscles are innervated almost solely by segmental nerve 16.
normal limb (Fig. 1). The only exception was a small contribution (about 20 %)
by nerve 16 to extensor metacarpi radialis in the reduplicated distal forelimbs.
In those preparations where the distal end of the humerus was also reduplicated,
the biceps and triceps muscles appeared normal, except that their insertions
near the elbow were slightly misplaced. There was no additional musculature
in the elbow region.
The second type of preparation had been reduplicated at the wrist and had
two hands (Fig. 3C). In these preparations the forearm muscles appeared
normal, though some of their insertions at the wrist were misplaced. In all
cases the limbs were innervated normally proximal to the wrist and symmetrically distal to the wrist (Fig. 2). Thus, all the hand muscles studied were almost
solely innervated by nerve 16.
Segmental innervation of chick forelimb
Fig. 3. Photomicrographs of the bones of the right forelimb (dorsal view) at 11-14
days' incubation. (A) Normal limb; (B) limb with mirror image reduplication distal
to the elbow; (C) limb with mirror image reduplication distal to the wrist, h,
humerus; r, radius; u, ulna; m, metacarpals.
147
148
M. R. BENNETT, R. LINDEMAN AND A. G. PETTIGREW
w^
D
Fig. 4. For legend see opposite.
Segmental innervation of chick forelimb
149
S 14
1
50
0
—
i
o'
100 -
S 15
1
S 16
M usclc
Fig. 5. Percentage innervation of major limb muscles in preparations where digits
were deleted (open circles), parts distal to the elbow were deleted (open triangles)
and where the humerus was deleted (open squares). The normal innervation is shown
with filled circles. Recordings were made at 14 days'incubation and each point represents the mean of three determinations for each muscle from at least two preparations; error bars show ±S.E.M. b, biceps; emr, extensor metacarpi radialis; /, triceps;
feu, flexor carpi ulnaris; fdp, flexor digitorum profundus. Note that the segmental
innervation of all muscles in the abnormal limbs is very similar to the innervation of
those muscles in the normal limb.
Innervation offorelimbs with deletions along the proximo-distal axis
In forelimbs which had been truncated so that the digits failed to develop
(Fig. 4B), the proximal and distal forelimb muscles were innervated normally
(Fig. 5). In forelimbs which had failed to develop parts distal to the elbow
(Fig. 4A), the biceps and triceps muscles were innervated normally by nerves
14 and 15 (Fig. 5). Nerve 16 was absent from these limbs but it still normally
innervated muscles adjacent to the limb (e.g. deltoids).
FIGURE 4
Photomicrographs of the bones of the right forelimb (dorsal view) following induced
deletion of segments along the proximo-distal axis. (A) Deletion of parts distal to the
elbow; (B) deletion of parts distal to the wrist; (C) deletion of humerus; (D) deletion
of the humerus, radius and ulna (hand attached to scapula), s, scapula; h, humerus;
r, radius; //, ulna; m, metacarpals.
150
M. R. BENNETT, R. LINDEMAN AND A. G. P E T T I G R E W
Fig. 6. Photomicrographs of the right brachial plexus (ventral, rostral uppermost)
in normal preparation (A) and a preparation where segment 14 had been removed
at stage 20 (B). Calibration is 240/*m for (A) and 150 /*m for (B). sc, Supracoracoideus nerve. The inset in (B) shows a low power magnification of the brachial
region. Even though spinal nerve 14 is missing the basic geometry of the plexus has
been retained.
In preparations where the proximal ends of the radius and ulna were attached
to the scapula (humerus deleted) (Fig. 4C), the distal forelimb had developed
normally and the three muscles examined received their normal pattern of
innervation (Fig. 5). Similarly, where the hand was attached directly to the
scapula (Fig. 4D), the three hand muscles examined were innervated almost
solely by nerve 16 (Fig. 3).
Thus in all the experimental situations described, where growing axons have
invaded a grossly abnormal limb, the innervation of each of the muscles studied
at later stages was the same as for those muscles in the normal embryo.
Innervation of the forelimb following removal of a spinal cord segment
In embryos where a segment of the spinal cord was removed at stages before
the axons have left the cord, the forelimb developed normally, with normal
size muscles, and appeared to show correct movement and coordination.
Similarly, the brachial plexus in these embryos retained the basic geometry of
Segmental innervation of chick forelimb
151
100 F—
S 16
50
0 b_
100
S
S 15
50
o
~
0
100 F
S 14
50
0 -.
feu
Muscle
Fig. 7. Innervation of the forelimb in control preparations (filled circles) and
preparations where segment 15 had been removed at stage 18 (open circles).
Recordings were made at 12 and 14 days' incubation respectively and each point
represents the mean of three determinations for each muscle in each of two
embryos; error bars shown ±S.E.M. b, biceps; emr, extensor metacarpi radialis;
t, triceps; feu, flexor carpi ulnaris; fdp, flexor digitorum profundus. Note that
muscles which are normally partially innervated by SI 5 are now solely innervated
by the remaining segmental nerve and not at all by the 'foreign' segmental nerve.
normal embryos, except that one complete spinal nerve was missing (Fig. 6).
In general, the innervation of the forelimb muscles was complete and the
remaining spinal segments innervated those parts of the limb musculature
which would normally have been innervated by the missing segment (Fig. 7).
It is interesting to note, however, that the innervation of muscles which normally
receive axons from two adjacent spinal nerves was completed by the remaining
spinal nerve of the pair, and not by a spinal nerve whose axons are never found
in that muscle in the mature animal. For example, the biceps and extensor
metacarpi radialis muscles are normally innervated by nerves 14 and 15. When
segment 15 is missing, both muscles are innervated solely by nerve 14 and not
at all by nerve 16.
In this study individual spinal segments were removed from embryos at
stage 20. The full complement of motoneurones is not achieved, however,
until about stage 27 (Hollyday & Hamburger, 1977). As yet, we have not
examined the fine structure of the spinal cord in these preparations. It is possible
that neurones were established after the operation and that they migrated into
the space remaining to reform the missing segment. If this were the case, it
might be expected that such neurones would send axons to the limb via adjacent
152
M. R. BENNETT, R. LINDEMAN AND A. G. PETTIGREW
spinal nerves, because of the abnormal bone structure in this region. Thus, in
preparations where segment 15 had been removed we would have expected to
find the biceps muscle innervated in part by spinal nerve 16. In all such preparations, however, the biceps was innervated solely by nerve 14.
DISCUSSION
Innervation of abnormal limbs
The initial operations on the embryos used in this study were performed at
stages before the axons which are to innervate the limb have reached the
shoulder region. Nevertheless, in the series of experiments where the morphology
of the limb was abnormal, the motor innervation of all but one muscle examined
at 10-14 days' incubation was the same as that found in the control forelimb.
This observation was made in limbs with either deletions along the proximodistal axis or with reduplications about this axis. Thus, wherever a muscle was
placed in the limb in relation to the parts proximal to it, it was almost always
innervated normally. These results both confirm and extend the work of
Stirling & Summerbell (1977).
The interpretation of our data relies on the observation that the motor axons
of each spinal nerve at the brachial level have had prior access to muscles of
the forelimb. All the spinal nerves which form the brachial plexus during
normal development are known to supply axon terminals to muscles of the
forelimb (Pettigrew et al. 1979), although it has been claimed that the spinal
nerves which form the lumbar plexus only supply axon terminals to the muscles
of the hindlimb identical to those found in the mature muscles (Landmesser &
Morris, 1975; for commentary on this paper see the Discussion in Pettigrew
et al. 1979). In the present work operations which produced abnormal limbs
were performed at stages before the motor axons have invaded the limbs. It is
reasonable to assume, therefore, that at least some of the muscles in abnormal
limbs receive an initial widespread segmental innervation, just as they would
during normal development (Pettigrew et al. 1979). That these muscles have
the same more restricted segmental innervation as in the normal limb by later
stages of development, suggests that the processes which result in the loss of
part of the early innervation of a muscle are operating normally, even though
the muscle is in an abnormal position.
The development of limb innervation
The observation by Lamb (1976, 1977), McGrath & Bennett (1979) and
Pettigrew et al. (1979) that the segmental projection to at least some major
limb muscles undergoes considerable modification at early stages is more easily
explained in terms of matching between growing axons and muscle cells (see
Hughes, 1968; McGrath & Bennett, 1979; Pettigrew et al 1979), rather than
in terms of strict fibre guidance (see Weiss, 1934; Sperry, 1963; Landmesser &
Segmental innervation of chick forelimb
153
Morris, 1975). From the present experiments on abnormal limbs, it seems
likely that any selective matching properties of embryonic muscle cells may not
be determined by the position of the muscle in the limb relative to the parts
proximal to it.
As a further test of the possible matching between embryonic nerve and
muscle cells, it was of interest to determine to what extent a normal limb would
be innervated by a spinal cord which was deficient in a group of motoneurones
of a particular spinal segment. In such preparations the limbs and musculature
were fully developed and the motor innervation of each limb was completed by
the segments which remained. Previous studies, using only anatomical techniques, have described the pattern of nerves in developing chick limbs following
early removal of two or three of the brachial (Castro, 1963) or lumbar spinal
segments (Kieny & Fouvet, 1974). While Castro describes hyperplasia in the
one or two remaining nerves, it was nevertheless insufficient to reconstitute the
normal size of some of the limb nerves. The results of the present physiological
study imply, however, that if only one spinal segment is removed, rather than
two or three, hyperplasia in the remaining segmental nerves is sufficient to
compensate for the missing segmental nerve. The observations of Kieny and
Fouvet, in which the anatomical pattern of nerves in the hindlimb was determined after removal of part of the lumbar spinal cord and associated somites,
are difficult to interpret. In this study destruction of the somites produced gross
abnormalities in hindlimb structure.
It is likely that early during the development of preparations with a single
segment deleted, the remaining segments initially supplied axons to most muscles
of the limb, as they do during normal ontogeny and that some axons were
subsequently lost from individual muscles during later development (Pettigrew
et al. 1979). In some cases, however, muscles were left completely innervated
by axons which only partially innervate that muscle in the normal embryo. This
observation suggests that any matching between growing axons and muscles
may be hierarchical rather than strictly all-or-nothing, as is the matching
between axons and muscles in the axolotl (Bennett et al. 1979). Furthermore,
such a hierarchy may be related to the general rostro-caudal spinal gradient
along the proximo-distal and antero-posterior gradients of innervation in the
normal limb. For example, the biceps muscle (anterior) is normally innervated
by nerves 14 and 15. When segment 15 has been removed the biceps is innervated
solely by nerve 14 (rostral) and not at all by nerve 16 (caudal). In this context,
it will be of interest to examine preparations where the spinal segment which
has been removed is normally solely responsible for a muscle's innervation;
for example, examination of the intrinsic hand muscles after removal of segment
16.
We are very grateful to Ms J. Stratford for her excellent technical assistance. This work
was supported by the Australian Research Grants Committee.
154
M. R. BENNETT, R. LINDEMAN AND A. G. P E T T I G R E W
REFERENCES
M. R., MCGRATH, P. A. & DAVEY, D. F. (1979). The regression of synapses formed
by a foreign nerve in a mature axolotl striated muscle. Brain Research. (In the Press.)
BENNETT, M. R. & PETTIGREW, A. G. (1976). The formation of neuromuscular synapses.
Cold Spring Harb. Symp. quant. Biol. 40, 409-424.
BENNETT, M. R. & RAFTOS, J. (1977). The formation and regression of synapses during
reinnervation of axolotl striated muscles. /. Physiol. 265, 261-295.
CASS, D. T. & MARK, R. F. (1975). Reinnervation of axolotl limbs. I. motor nerves. Proc.
R. Soc. B 190, 45-58.
CASTRO, G. (1963). Effects of reduction of nerve centers on the development of residual
ganglia and on nerve patterns in the wing of the chick embryo. /. exp. Zool. 152, 279-295.
DENNIS, M. J. & YIP, J. W. (1978). Formation and elimination of foreign synapses on adult
salamander muscle. /. Physiol. 274, 299-310.
HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the development of
the chick embryo. /. Morph. 88, 49-92.
HARRIS, A. J. & DENNIS, M. J. (1977). Deletion of 'mistakes' in nerve-musle connectivity
during development of rat embryos. Soc. for Neurosciences, 7th Meeting, Anaheim,
California. Abstracts, vol. in, p. 107.
HOLLYDAY, M. & HAMBURGER, V. (1977). An autoradiographic study of the formation of
the lateral motor column in the chick embryo. Brain Res. 132, 197-208.
HUGHES, A. (1968). Development of limb innervation. In Growth of the Nervous System,
pp. 110-117. London: Churchill.
KIENY, M. & FOUVET, B. (1974). Innervation et morphogenese de la patte chez l'embryon
de poulet. Archs Anat. microsc. Morph. exp. 63, 281-298.
LAMB, A. H. (1976). The projection patterns of the ventral horn to the hind limb during
development. Devi Biol. 54, 82-99.
LAMB, A. H. (1977). Neuronal death in the development of the somatotopic projections of
the ventral horn in Xenopus. Brain Res. 134, 145-150.
LANDMESSER, L. & MORRIS, D. G. (1975). The development of functional innervation in the
hind limb of the chick embryo. J. Physiol. 249, 301-326.
MCGRATH, P. & BENNETT, M. R. (1979). The development of synaptic connections between
different segmental motoneurones and and striated muscles in an axolotl limb. Devi Biol.
69, 133-145.
PETTIGREW, A. G., LINDEMAN, R. & BENNETT, M. R. (1979). Development of the segmental
innervation of the chick forelimb. J. Embryol. exp. Morph. 49, 115-137.
RONCALI, L. (1970). The brachial plexus and the wing nerve pattern during early developmental phases in chick embryo. Monitore Zool. Ital. 4, 81-98.
SAUNDERS, J. W. & GASSELING, M. T. (1968). Ectodermal-mesenchymal interactions in the
origin of limb symmetry. In Epithelial-Mesenchymal Interactions (ed. R. Fleishmajer &
R. E. Billingham), pp. 78-97. Baltimore: Williams & Wilkins.
SPERRY, R. W. (1963). Chemoafnnity in the orderly growth of nerve fibre patterns and
connections. Proc. natn. Acad. Sci., U.S.A. 50, 703-710.
STIRLING, R. V. & SUMMERBELL, D. (1977). The development of functional innervation in
the chick wing-bud following truncations and deletions of the proximo-distal axis. J.
Embryol. exp. Morph. 41, 189-207.
SUMMERBELL, D. (1974a). Interaction between the proximo-distal and antero-posterior coordinates of positional value during the specification of positional information in the
early development of the chick limb-bud. J. Embryol. exp. Morph. 32, 227-237.
SUMMERBELL, D. (19746). A quantitative analysis of the effect of excision of the AER from
the chick limb-bud. /. Embryol. exp. Morph. 32, 651-660.
SUMMERBELL, D. & LEWIS, J. H. (1975). Time, place and positional value in the chick limbbud. /. Embryol. exp. Morph. 33, 621-643.
WEISS, P. (1934). In vitro experiments on the factors determining the course of the outgrowing
nerve fibre. /. exp. Zool. 68, 393-448.
BENNETT,
(Received 5 February 1979, revised 8 April 1979)