243
Variations in histology of abdominal stretch receptors of
saturniid moths during development
By L. H. FINLAYSON and D. J. MOWAT
(From the Department of Zoology and Comparative Physiology, University of Birmingham)
With 2 plates (figs, i and 2)
Summary
In the stretch receptors of Lepidoptera there are normally two giant nuclei, but supernumerary giant nuclei are frequently present. In a study on Samia cynthia supernumerary giant nuclei have been found to occur most frequently in the first, second,
third, and fourth abdominal segments. They are present in the larva, pupa, and adult.
Small nuclei in the muscular component of the receptors are infrequent in the larva
and in the fourth, fifth, and sixth abdominal segments of the adult; but in the first,
second, and third abdominal segments there is an increase in small nuclei at metamorphosis. Small nuclei also multiply in the receptors of the eighth and ninth segments
but these receptors degenerate. The muscular component of the receptors that metamorphose at first becomes reduced and loses its striations and then differentiates anew.
In those receptors that disappear (eighth and ninth) dedifferentiation is succeeded by
histolysis. A third category of nucleus, larger than the small nuclei but smaller than
the giant nuclei, was found in a few receptors of larva, pupa, and adult. Such nuclei
appear to be absent from the fifth and sixth segments, but they were too rare for
reliable statistics to be obtained.
Introduction
I N the Lepidoptera the stretch receptors of thorax and abdomen are more
complex in structure than those of any other group of insects that have been so
far studied (Finlayson and Lowenstein, 1958; Osborne and Finlayson, 1962).
Each consists of a sensory neurone whose dendrites run into a tube of connective tissue that lies along a special muscle-fibre. In the central region of the
receptor are two giant nuclei. The larger nucleus lies in a mass of cytoplasm in
the middle of the muscle and the smaller one lies in the tube containing the
dendrites, the 'fibre tract' (Finlayson and Lowenstein, 1958). The receptors
are slung between the intersegmental folds in the dorsal region of the body, one
pair in each segment in which they occur. They run in parallel with the dorsal
longitudinal muscles, but they do not degenerate and disappear during metamorphosis like the ordinary longitudinal muscles (Finlayson, 1956, i960).
In the third, fourth, fifth, and sixth abdominal segments of the pupa certain
longitudinal muscles of the larva persist, but in the adult these muscles (the
imaginicaducous muscles, Finlayson, 1956) degenerate and disappear in the
first few days of adult life. The stretch receptors do not disappear but are conspicuous in old adults. Clearly the muscular component of the stretch
receptors is of peculiar interest in the light of previous work on the longitudinal
muscles, which showed that their degeneration during pupation was influenced
[Quart. J. micr. Sci., Vol. 104, pt. 2, pp. 243-51, 1963.]
244
Finlayson and Mowat
by an anterior and a posterior gradient and that denervation in the larval
stage could, in certain species, bring about the degeneration, during pupation,
of longitudinal muscles that would normally have persisted (Finlayson, 1956,
i960). Preliminary experiments indicated that the muscles of the stretch
receptors, even in the most anterior segments of the abdomen, do not degenerate when their motor innervation is severed (Finlayson, 1959). These
experiments also revealed that the stretch receptor undergoes a metamorphosis.
Further observations have revealed profound differences in the degree of
metamorphosis of the stretch receptors in different segments of the body. The
present paper gives details of certain aspects of the morphology of the stretch
receptors and of their metamorphosis, with particular reference to the variation in numbers and sizes of nuclei.
Material and methods
Most of the work was carried out on a non-diapausing strain of Samia
cynthia, probably ricini (Finlayson, i960), but some observations were made
on Antheraea mylitta. Samia was reared in the laboratory: Antheraea pupae
were kindly provided by Dr. A. D. Blest from supplies he received from
India. Details of the structure of the receptors were studied in whole mounts
and sections using chiefly Ehrlich's haematoxylin and the Feulgen technique.
Phase-contrast microscopy was used extensively.
The giant nuclei
There are normally two giant nuclei lying near the middle of the stretch
receptor (Finlayson and Lowenstein, 1958). The larger of the two, the major
giant nucleus, lies in the core of the receptor (figs. 1, A, c; 2, B), and the other,
the minor giant nucleus, lies in the 'fibre tract'(fig. i, A, D, E, F), a tube of connective-tissue fibres which is continuous with the sheath surrounding the
dendrites of the sensory neurone and inside which the dendrites run (figs. 1,
B; 2, A, c).
The major giant nucleus varies considerably in size and shape. In the larva
it tends to be plump and relatively short. In the pupa it is similar in shape and
size but in the adult it tends to be thinner and longer. The range of size
variation is about the same in larva and adult, about 170 to 680/z in length and
10 to 60 fj, in width. An average value for 12 receptors, 2 from each of the
first to the sixth abdominal segments taken at random from preparations made
from 8 adults and 10 larvae, is 410 X 33 JJ. for larva and 428 X 21 fj. for adult.
The adult receptors are much shorter than those of the larva so the major
giant nucleus of the adult, although it is about the same length as that of the
larva, is much longer in relation to the length of the receptor (table 1).
During metamorphosis the major giant nucleus undergoes a shrinking process
which reduces its thickness but not its length. In all stages the nuclear membrane is thrown into a series of longitudinal, sinuous furrows. Both ends of
the nucleus may be clearly demarcated by the nuclear membrane but often
Abdominal stretch receptors of moths
245
one or both ends are indistinct because they become narrower and less
granular and cannot then be seen in the central mass of cytoplasm.
The minor giant nucleus is similar in appearance to the major giant nucleus
but is always smaller. It ranges in size from 60 to 180 /u. in length and 5 to
12-5 n in width. There is no obvious difference in the size of the minor giant
nucleus between larva and adult. It is situated in the fibre tract (fig. 1, A, D, F)
and is therefore intimately associated with the dendrites of the sensory neurone.
In a few preparations this nucleus has been seen partially inserted into one of
the dendritic branches of the neurone (fig. 1, E). The minor giant nucleus, like
the major, is often rounded at the ends and the nuclear membrane can then be
clearly seen. In many cases, however, it also fades away at one or both ends as
though its contents have been lost leaving only the folded and furrowed
nuclear membrane as a 'tail' which is difficult to distinguish among the cytoplasm of the core.
Variation in numbers of giant nuclei
In the majority of stretch receptors at all stages in the development of the
insect there is one major and one minor giant nucleus in each receptor. In the
others there are more than the basic pair. The numbers of giant nuclei found
in all stages (larva, prepupa, developing adult, and adult) are shown in fig. 3.
It is clear from the counts that there is a tendency for supernumerary giant
nuclei to occur in the first four segments of the abdomen. Most often there is
only one extra nucleus but there may be several. It is not always easy to see
whether supernumerary giant nuclei are majors or minors, especially in whole
mounts, and so the two categories have been combined in the totals on which
the histograms are based. Supernumerary giant nuclei have been found in all
stages of development from final instar larvae to adults. There is no strong
evidence of a change in their numbers during metamorphosis, although the
supernumerary giant nuclei of adults appear not only to be more evenly distributed throughout all the segments of the abdomen but to occur with the
greatest frequency in the first abdominal segment, not in the second segment as
in the other stages. However, the sample is small and the material is not easy
to prepare and examine, so that errors of observation are probable.
The small nuclei
In addition to the giant nuclei, which are unusual objects, there are nuclei
of a normal size range. These are the nuclei of the tracheae and tracheal end
cells, the motor-endings, the fibre tract, and of the muscle-tissue. The greatest
number occurs in association with nerve- and tracheal-endings in the central
region near the major giant nucleus, forming a 'cap' of nuclei (fig. 1, A). In the
larva each of these categories of small nucleus can be identified but the muscle
nuclei are few in number and in many preparations it is impossible to identify
any with certainty. At best their identification in the larval receptor is based on
the distance at which they are located from other structures such as motorendings, to which they may or may not belong. This is unsatisfactory but the
246
Finlayson and Mowat
difficulty of finding even a few muscle nuclei serves to emphasize the contrast
between the larval receptor and those metamorphosed, adult receptors in
which there are many muscle nuclei.
In the adult of Samia the receptors of the fifth and sixth abdominal segments are always 'larval' in appearance (fig. 1, B) but those of the first, second,
third, and seventh segments are metamorphosed (fig. 2, D). The receptors of
the fourth segment may be altered to a lesser degree. The principal difference
is the presence in the metamorphosed receptor of very large numbers of small
nuclei arranged in rows throughout the muscular part of the receptor (fig. 2, D).
In the 7-day developing adult each row of nuclei is associated with a musclefibril. The increase in numbers of these nuclei between larva and adult is
striking.
Attempts have been made to follow the multiplication of these small nuclei,
but it is surprisingly difficult to obtain a clear picture of the sequence of events
that lead to their increase in number. The earliest record of mitosis comes
from an early ('green') prepupa in which one mitotic figure was seen near the
central region of the receptor occupied by the major nucleus (fig. 1, G). This
and other prepupae and early developing adults have shown a group of small
nuclei in the central region beginning to spread out along the receptor. A
question arises as to whether these really are the muscle nuclei, because in the
central region of the larval receptor there do not appear to be any musclefibrils (Finlayson and Lowenstein, 1958). In the early developing adult
pycnotic and mitotic figures (fig. 2, B, c) are frequent and the multiplication of
the small nuclei proceeds rapidly until the final number is reached after about
a week of adult development. Mitoses and pycnoses are common in the fibre
tract and dendritic branches of the neurone during the same period.
In the developing adult the small nuclei in rows vary in size from about
10x6/1 to 3 x 3 JJL. In the adult these nuclei are shrivelled and misshapen
(fig. 2, D). The number of small nuclei in adult receptors shows some variation
within those segments in which the small nuclei multiply. In the seventh
segment there are fewer nuclei than in the first, second, or third. In the fourth
segment there may or may not be an increase in numbers of nuclei. It is
possible that there are also differences in the density of small nuclei between
FIG. 1 (plate). All whole mounts of receptors from S. cynthia stained with Ehrlich's
haematoxylin.
A, central region of receptor from first abdominal segment of full-grown larva showing major
giant nucleus (maj), minor giant nucleus (min), and 'cap' nuclei (c).
B, part of receptor from second abdominal segment of full-grown larva showing part of
major and minor giant nuclei, fibre tract (/), vacuolated cytoplasm of core.
c, D, receptor from third abdominal segment of 1-day-old pupa (developing adult) showing
two major giant nuclei (maj) and two minor giant nuclei (min),
E, F, receptor from third abdominal segment of full-grown larva showing two minor giant
nuclei in fibre tract. One is partially inserted into a dendrite (E). The receptor is broken in the
region shown in E and the cytoplasmic core is bulging out. d, dendrites.
G, receptor from first abdominal segment of early (green) prepupa showing group of small
nuclei below major giant nucleus (maj).
H, large nuclei in receptor from fourth abdominal segment of adult.
Abdominal stretch receptors of moths
247
the receptors of the first, second, and third abdominal segments, but as the
nuclei are very numerous in all of these segments and the receptors vary in
length it would be difficult to make an accurate comparison. There is no
doubt, on the other hand, that in the fourth and seventh segments the
tendency for the small nuclei to multiply is not so pronounced, or the rate of
multiplication is slower during the period when it takes place. In the receptors
of the eighth segment of the abdomen the small nuclei have increased considerably in numbers by the second day of adult development. In the receptors
of the seventh segment at the same time there is only a slight increase in the
numbers of small nuclei. We have never found receptors in the eighth or
ninth abdominal segments of the adult, so presumably they degenerate during
development. The multiplication of the small nuclei is a prelude to histolysis
in these receptors.
The large nuclei
Among the hundreds of receptors of Samia that have been examined in this
study about a dozen were found which had nuclei that would not fit into the
above categories. They were larger than the small nuclei and smaller than the
giant nuclei. They were classed as large nuclei. In size they overlap with small
minor giant nuclei but in the few preparations where their position can be
clearly established they are seen to occupy the core of cytoplasm and not the
fibre tract where the minor giant nuclei are situated (fig. 1, H).
Although the records of large nuclei are few they have been found in larva,
prepupa, developing adult, and adult. Although too few examples have been
found to permit accurate comparisons, the large nuclei of larvae appear to be
larger than those of adults. For example, the receptor from the second abdominal segment of a full grown larva had large nuclei measuring 90 X 7 n,
6 7 x 8 ^ , 48X8/X, 4 0 x 7 ^ , whilst a receptor from the fourth abdominal
segment of an adult had large nuclei measuring 40X4/X, 38x7/^, 25x8/A,
23 X 7 /x. In other adults the large nuclei were more restricted in size range and
were smaller. A receptor from the seventh abdominal segment which had
been mounted in a relaxed condition has a dozen spherical large nuclei ranging
from 13-6 to 6-8 [x in diameter. The receptor from the other side of the same
segment also has large nuclei, but these are elongated and measure about 16 fx
in length and 6 /x in width. At one time we thought it possible that some of the
small nuclei of the larva grew rapidly and formed large nuclei, which then
FIG. 2 (plate). All stained with Ehrlich's haematoxylin.
A, whole mount of receptor from seventh abdominal segment of diapausing pupa of A.
mylitta showing large nuclei in core, n, motor-nerve ending;/, fibre tract.
B, c, longitudinal section of receptor from 5-day developing adult to 5. cynthia showing
major giant nucleus, pycnotic nuclei, a mitoticfigure(marked by arrow) and muscle striations.
/ , fibre tract.
D, whole mount of receptor from second abdominal segment of adult »S. cynthia showing
typical rows of small nuclei.
248
Finlayson and Mowat
divided to form the rows of small nuclei found in the adult. It now seems more
likely that these large nuclei are produced by the same 'nuclear-promoting'
influence that causes the production of supernumerary giant nuclei. Large
nuclei are found in the first, second, third, fourth, and seventh abdominal
segments, just as the majority of supernumerary giant nuclei. Moreover,
as already mentioned, large nuclei occur in the cytoplasmic core in which
the major giant nuclei are situated. Similar large round nuclei occupying
precisely the same position have been seen in the diapausing pupa of
A. mylitta (fig. 2, A). We may perhaps conclude from this evidence that large
nuclei are in the same category as major giant nuclei, to be simply regarded
as smaller versions.
An objection to this hypothesis lies in the difference in structure between
large nuclei and major giant nuclei. The large nuclei are vacuolated and
irregularly granular and usually have prominent nucleoli (fig. 1, H). The
major giant nuclei are more densely granular with granules of a more uniform
size, distributed in a more regular fashion. At present the origin and significance of both large and major giant nuclei are obscure.
The muscle-fibrils
The muscle-fibrils of the receptor have been studied in the first, second, and
third abdominal segments of developing adults of Samia. During development
the fibrils lose their striations. Loss of striations begins in the receptors of the
first segment, and is followed by those of the second segment and then by those
of the third segment. For example, in an animal 4 days from the larval-pupal
moult there are no signs of striations in the receptors of the first or second
abdominal segments but those of the third abdominal segment are still clearly
striated. By 7 days the striations of the receptors in the third abdominal segment have also disappeared. By this time there is a string of small nuclei in
each fibril.
In the adult receptor of the first abdominal segment the arrangement of the
muscular component differs from that of the larva. In the larva there is a prominent swollen central region occupied by the major giant nucleus. The
muscle-fibrils, as far as we can see, do not pass through this region. In the
adult the major giant nucleus is much narrower, there is no central swelling
and the muscle-fibrils clearly pass from one end of the receptor to the other
without interruption at the centre.
It is possible that the metamorphosis of the muscular component of the
stretch receptor is correlated with the decrease in length that takes place
during the metamorphosis of the animal. The greatest decrease in length
takes place in the first, second, seventh, and eighth segments of the abdomen
(table 1) but during pupation the other segments also decrease by almost half
their length and it is difficult to see why the slightly greater decrease in length
of the anterior and most posterior segments should lead to such a profound
difference in the receptors.
Abdominal stretch receptors of moths
249
In the third segment the receptors undergo a profound metamorphosis but
this segment does not decrease in length as much as do the fifth, sixth, and
seventh. Furthermore, the receptors of the fourth segment frequently have a
greatly increased number of small nuclei, although that segment shrinks least
of all. There does not appear to be a simple correlation between decrease in
segment length and metamorphosis of the receptor.
TABLE I
Comparison of lengths of abdominal segments in larva and pupa of S. cynthia
Segment number:
Length in mm. (larva
(mean of 12) jpupa
Decrease in mm
% decrease
1
2
3
4
5
6
7
8
49
2-1
S'4
26
28
5-5
29
26
63
63
5'3
3-7
26
3'2
3'i
61
31
30
2-4
2-9
4-3
1-9
2-4
52
47
41
49
49
SS
56
28
57
Discussion
The muscular component of the lepidopteran stretch receptor must at one
time in the course of evolution have been an ordinary muscle-fibre or fibres.
It belongs to the series of longitudinal, intersegmental muscles which behave
in different ways at metamorphosis according to their position in the body
(Finlayson, 1956, i960). They degenerate in all segments of the body except
the third, fourth, fifth, and sixth abdominal segments. Previous work has shown
that degeneration of these muscles takes place along two gradients, one passing
backwards and the other forwards. Experiments involving denervation of
these muscles revealed that the waves of histolysis can be extended into the
segments in which the longitudinal muscles normally persist. The degree of
extension varies considerably between species of moth (Finlayson, i960). It is
possible in all species to induce the muscles of the third abdominal segment to
degenerate by severing their innervation, but only in certain species is it possible to extend this effect to the other segments. The muscle of the stretch
receptor does not degenerate, except in the eighth and ninth abdominal segments, and each of the recognizable abdominal segments of the adult (segments 1 to 7) has receptors. Stretch receptors in the fifth and sixth segments
remain larval in appearance, the ones in the first, second, third, and seventh
segments undergo a striking metamorphosis and those in the fourth segment
may metamorphose to a lesser degree or may remain larval in appearance. The
onset of dedifferentiation of the muscle-fibrils, marked by loss of striation and
multiplication of nuclei, begins at the anterior and posterior ends of the abdomen and then spreads to the second and third segments and to the seventh
segment. This is another manifestation of the anterior and posterior gradients
that influence the course of muscle degeneration (Finlayson, 1956, 1961) and
in this case dedifferentiation and metamorphosis. The most striking feature of
the metamorphosis is the multiplication of the small, muscle nuclei. Dedifferentiation may be followed by metamorphosis, as in the anterior segments, or
250
Finlayson and Mowat
by histolysis, as in the posterior segments. Multiplication of muscle nuclei as
a prelude to metamorphosis or histolysis is described by Hufnagel (1918) who
reviews earlier literature on muscle metamorphosis.
The metamorphosis of the stretch receptors may in part be correlated with
their decrease in length in the adult. It is true that the most extreme changes
take place in those segments where the decrease in length of the receptor is
greatest, but the receptors of the fourth abdominal segment often metamorphose although that segment decreases in length least of all. It is also sur-
1—1
S 20%
r-
—
n n n
FIG. 3. Histogram showing incidence of supernumerary giant nuclei in receptors taken
from all stages of S. cynthia. The following figures in italics represent the number of
preparations on which the counts were based (Ai = ist abdominal segment, etc.): Ai, 49\
Az, 77; A3, So; A4, 78; As, 72; A6, 78; A7, 23.
prising that the considerable reduction in length that takes place in the other
segments does not also necessitate a similar drastic reorganization of the receptor muscle. There is, therefore, no necessary correlation between decrease in
length and degree of metamorphosis of the stretch receptor. Another possible
explanation is that the receptor muscles react according to their position in the
body. In those segments of the body in which all or most of the larval muscles
break down the receptor muscle may be 'driven' to dedifferentiate and has then
to regrow its muscle-fibrils. In the other segments the receptor muscle behaves
like the other muscles of the nearby longitudinal band.
The giant nuclei are unusual structures and with the possible exception of
the Trichoptera they do not occur in the stretch receptors of any other insects
(Finlayson and Lowenstein, 1958; Osborne and Finlayson, 1962). The minor
giant nucleus is closely associated with the dendrites of the neurone and it is
probably an extra large glial or Schwann cell nucleus. The major giant is an
enormous nucleus by any standard of comparison, but its origin is obscure.
It may be pertinent that the small muscle nuclei multiply during metamorphosis whilst the major giant nucleus is shrinking. The major giant nuclei of
the anterior stretch receptors appear to degenerate more than those in the
middle segments of the abdomen. It is possible, therefore, that the major
giant nucleus is a hypertrophied 'larval' muscle nucleus and that its presence in
Abdominal stretch receptors of moths
251
the larva inhibits the multiplication of the normal small nuclei. Hufnagel
(1918) describes large 'larval' nuclei and small 'adult' nuclei in the muscles of
the larva of the moth Hyponomeuta.
The occurrence of supernumerary giant nuclei in some stretch receptors is
unlikely to have an adaptive significance. Only in a minority of receptors are
additional giant or large nuclei present, and even within the same animal the
receptor on one side of the body may have one or more additional giant
nuclei whilst the receptor on the other side in the same segment has only the
normal complement. The frequency of occurrence of supernumerary giant
nuclei is not constant throughout the abdomen. They are much more
frequent in the anterior segments, with a probable maximum in the second
abdominal segment. This distribution of additional giant nuclei is already
present in the larva, so it has been established long before the multiplication
of the small nuclei. However, there is the same sort of distribution of additional giant nuclei as of additional small nuclei and it seems conclusive that the
giant nuclei or their precursors were influenced by a similar factor during
embryonic or larval growth. This factor is strongest in its influence in the
anterior segments of the abdomen, but even there it only produces a minority
of receptors with additional giant nuclei. In the development of the majority
of receptors the tendency for production of extra giant nuclei does not upset
the normal equilibrium which results in the production of two giant nuclei
only. A similar situation in relation to the small nuclei is seen in the seventh
segment of the abdomen, where there may be fewer small nuclei than in the
first, second, third, and eighth segments. In the fourth segment there is a series
of possibilities ranging from no apparent multiplication to an obvious increase
in numbers of small nuclei, although never so great an increase as in the preceding segments.
The large nuclei have been seen in so few receptors that at present they
must be regarded as abnormal. They may be supernumerary 'giant-type'
nuclei or enlarged 'small-type' nuclei. Their position in the central core of
cytoplasm supports the idea that they are major giant nuclei, but their
structure is different.
References
FINLAYSON, L. H., 1956. Quart. J. micr. Sci., 97, 215.
i960. J. Ins. Physiol., 5, 108.
1960. In The ontogeny of insects (Act. Symp. Evol. Insect. Prague 1959).
and LOWENSTEIN, O., 1958. Proc. roy. Soc, 148, 433.
HUFNAGEL, A., 1918. Arch. Zool. exp. g£n., 57, 47.
OSBORNE, M. P., and FINLAYSON, L. H., 1962. Quart. J. micr. Sci., 103, 227.