There are four fibres of giant dimensions in each cord of the abdominal ventral nerve-chain of the locust. These fibres are the axons of large neurons situated in the extreme posterior region of the last abdominal ganglion, and run uninterruptedly through the abdominal nerve-chain to the metathoracic ganglion. Posteriorly these neurons synapse with preganglionic fibres which appear to be cereal nerve-fibres and anteriorly with motor pathways to the metathoracic legs.

Oscillographic observations have been correlated with those from the histological study. The possible function of these pathways has been discussed and comparison made with the cockroach, where a similar organization has been described in detail.

The method developed by Pumphrey and Rawdon-Smith (1937) to record electrical activity from the ventral nerve-cord of a cockroach offers a simple preparation for the study of a central synapse in an invertebrate. It is the synapse in the last abdominal ganglion between the primary vibration receptor afferents from the cereal nerve and ascending axons of giant neurons in the abdominal cord. This preparation involves a single synapse in each of a relatively few ascending pathways, thereby simplifiying analysis of the records and approximating to a single fibre preparation. Lowenstein (1942) used the time of abolition of the response to ‘air-puff’ stimulation in this preparation as a method of bio-assay of pyrethrum extracts. Roeder et al. (1947) used the preparation for the study of the action of various drugs on the cockroach synapse and Roeder (1948) extended the study of the organization of the giant fibre system both electrically and histologically.

A similar preparation is at present being used by the author for the study of the mode of action of certain neurotoxic insecticides on Locusta migratoria migratorioides (R. & F.). Since the literature contains little or no detailed description of this region of the nervous system in the locust, this histological study was undertaken, partly as a necessary background for the physiological work, and partly to provide anatomical information not as yet available in the case of the locust. This information is important for comparisons of the functions of the giant fibre systems in Periplaneta and Locusta.

Adult locusts were dissected, and the nervous system was exposed by the removal of the alimentary canal and genital organs with the nervous system intact in situ. The dissection was then fixed in Bouin’s fluid for 24 hours. After fixation, short regions of the abdominal cord, including the last two ganglia and cereal nerves, were dissected out and embedded in paraffin wax. In other cases the ventral nervous system, excluding the cephalic, suboesophageal, and prothoracic ganglia, was embedded in ester wax (Steedman, 1947). Serial sections were cut at 10μ and stained regressively in Heidenhain’s haematoxylin.

The method used in recording the electrical activity from the exposed ventral nervous system is essentially that described by Pumphrey and RawdonSmith (1937), Lowenstein (1942), and Roeder (1947). It will be described in detail elsewhere.

The acridian central nervous system (C.N.S.) as described by Snodgrass (1935) consists anteriorly of a dorsal cephalic ganglion united by a pair of connectives round the oesophagus with a sub-oesophageal ganglion. This is followed by three thoracic and five abdominal ganglia. The present study is concerned only with the posterior region of the C.N.S. of Locusta migratoria as far forward as the metathoracic ganglion (fig. 1). This does not differ in any essential details from that of Dissosteira (Snodgrass, 1935; Nesbitt, 1941) or Rhomalea (Nesbitt, 1941). In Locusta adipose tissue forms a complete sheath around each abdominal cord, leaving only the ganglia exposed. Anteriorly, paired segmental nerves, whose ganglia are situated far forward, are closely applied to the abdominal cords and are consequently included in the fatty sheath. The various nerves entering the posterior region of the last abdominal ganglion on either side are similarly ensheathed in adipose tissue and may thus appear as a single pair. Each member of this pair contains four separate bundles of nerve-fibres. One is an afferent sensory nerve from the cercus of that side, one innervates the genitalia, and the other two the musculature in this region.

FIG. 1.

Dissection of an adult male Locusta migratoria migratorioides to show the posterior region of the central nervous system. The locust was dissected from the dorsal side; a median dorsal strip of body-wall, the alimentary canal, and most of the gentalia were removed. A sheet of tissue, which separates off a median ventral portion of the haemocoel in which lies the nervous system, was carefully dissected away. The adipose tissue deposits around the nerve-cords were also removed.

FIG. 1.

Dissection of an adult male Locusta migratoria migratorioides to show the posterior region of the central nervous system. The locust was dissected from the dorsal side; a median dorsal strip of body-wall, the alimentary canal, and most of the gentalia were removed. A sheet of tissue, which separates off a median ventral portion of the haemocoel in which lies the nervous system, was carefully dissected away. The adipose tissue deposits around the nerve-cords were also removed.

A transverse section of these nerves shows them to be composed of nervefibres all of a diameter less than 5 μ. The general appearance is one of homogeneity, with little variation in fibre size. A transverse section of the abdominal cord between the ultimate and penultimate ganglia has a markedly different appearance. Approximately two-thirds of the area of cross-section of each cord is taken up by closely packed fibres running parallel to the longitudinal axis of the cord, whilst the remaining third, ventro-lateral in position, contains only a few scattered distinct fibres, the remainder of the tissue showing up rather indistinctly in sections (fig. 2A). While the majority of fibres are less than 5 μ in diameter there is a definite variation in fibre size with four fibres of diameter greater than 5μ and one of 5μ diameter in each cord.

FIG. 2

Transverse sections at various levels in the abdominal region of the central nervous system of one locust. A. The left cord between abdominal ganglia 5 and 4. c. The left cord between abdominal ganglia 4 and 3. E. The left cord between abdominal ganglion 1 and the metathoracic ganglion. B. Posterior region of abdominal ganglion 4. Low magnification. D. Left dorsal region of abdominal ganglion 4 as indicated in B. F. Posterior region of abdominal ganglion 1.

I, single median giant fibre; 2, 3, 4, peripheral group of three giant fibres; dotted line encloses indistinct region containing few axons..

FIG. 2

Transverse sections at various levels in the abdominal region of the central nervous system of one locust. A. The left cord between abdominal ganglia 5 and 4. c. The left cord between abdominal ganglia 4 and 3. E. The left cord between abdominal ganglion 1 and the metathoracic ganglion. B. Posterior region of abdominal ganglion 4. Low magnification. D. Left dorsal region of abdominal ganglion 4 as indicated in B. F. Posterior region of abdominal ganglion 1.

I, single median giant fibre; 2, 3, 4, peripheral group of three giant fibres; dotted line encloses indistinct region containing few axons..

These four large fibres can be followed from the anterior region of the fifth abdominal ganglion right up to the posterior region of the metathoracic ganglion. They pass through all the intervening ganglia without interruption (fig. 2, B, D, E). These four fibres are arranged in two groups. In each cord there is a single median fibre of an approximate average diameter of 13μ, and a peripheral latero-dorsal group of three fibres, two of which have an average diameter of 12μ and the third an average diameter of 8·5μ. This arrangement is maintained throughout the length of the abdominal cord. In each ganglion the dorsal group of three remains on the periphery following the dorsal contours of the C.N.S. while the median fibre takes a straighter course, being median in the cord but more dorsal in the ganglia. These fibres do not retain a constant diameter throughout their length but show a general tendency to taper towards their anterior ends. In addition they are constricted as they pass through each ganglion. Table I shows average measurements for these fibres at different levels in the ventral nerve-cord made on one series of sections. All the measurements given are taken from fixed material; there does not, however, appear to be any gross shrinkage, since in vivo measurements for the large fibres are 12—16μ. These in vivo measurements were not made from transverse sections, but from whole cords dissected in saline (Belar, 1929) and completely freed from fat tissue. The maximum diameter of the largest fibres visible was then measured. Attempts to prepare transverse sections of fresh material have not proved successful.

TABLE I.

Measurements, in microns, of Giant Fibres concerned in Cereal Response on Left Side of C.N.S.

All measurements are of fixed material. The fibre measurements are the means of ‘horizontal’ and ‘vertical’ dimensions.

Measurements, in microns, of Giant Fibres concerned in Cereal Response on Left Side of C.N.S.
Measurements, in microns, of Giant Fibres concerned in Cereal Response on Left Side of C.N.S.

The single median fibre in each cord is the axon of a large neuron, the nearly spherical cell body of which is situated in the extreme posterior region of the last abdominal ganglion. It has a diameter of about 10μ. Processes from this giant neuron are in close contact with the most median group of preganglionic fibres entering the fifth abdominal ganglion (fig. 3).

FIG. 3.

A. Transverse section through extreme posterior region of abdominal ganglion 5 showing the cell bodies of the two median giant fibres, one in each cord (i.e. fibre 1). B. Transverse section slightly anterior to that shown in A (30 μ). The axon can be seen emerging from the right cell body. This section has, unfortunately, been damaged in the region of the left axon.

FIG. 3.

A. Transverse section through extreme posterior region of abdominal ganglion 5 showing the cell bodies of the two median giant fibres, one in each cord (i.e. fibre 1). B. Transverse section slightly anterior to that shown in A (30 μ). The axon can be seen emerging from the right cell body. This section has, unfortunately, been damaged in the region of the left axon.

The peripheral group of three axons could not be traced back to their connexions with individual. cells. These three fibres could be followed quite readily through the anterior region of the fifth abdominal ganglion; behind this the three fibres become very closely applied to the dorsal boundary of the ganglion, and finally disappear in the posterior third of the ganglion where a number of fairly large cells is situated. Some of these cells also have processes in close contact with the most median group of preganglionic fibres, which would therefore appear to be those derived from the cercus.

Anteriorly the giant fibres cannot be traced farther forward than the posterior region of the metathoracic ganglion. The cord between the metathoracic and the mesothoracic ganglia carries many fibres of diameter comparable with that of the abdominal giants. Such fibres are also found in considerable numbers in the nerves leaving the metathoracic ganglion.

In the abdominal cord the four fibres described are not the only ones of giant dimensions; sections anterior to the fourth abdominal ganglion show other such fibres. In the cords between ganglia 4 and 3, 3 and 2, and 2 and 1 are found a number of fibres resembling the smaller giants, i.e. approximately 9μ diameter; their origin is unknown (fig. 2C). One such fibre resembling in size the larger giants is found in the cord between the metathoracic and the first abdominal ganglia (fig. 2E).

Oscillographic recording of the impulses from the cereal nerve shows a burst of impulses coinciding with a puff of air in the region of the intact cercus (fig. 4A). Recording electrodes placed on the abdominal cord pick up a similar burst of impulses on stimulation of the cereal receptors, except that two much larger sizes of spikes are present (fig. 4B). These spikes are also considerably faster than those in the cereal nerve and are therefore considered to belong to fibres of greater diameter than those in the cereal nerve. Electrodes placed on the large nerve from the metathoracic ganglion innervating the metathoracic leg pick up a burst of larger slower impulses every time there is a puff of air directed at the cerci. Removal of the cerci causes complete disappearance of such impulses in all positions on ‘air-puff’ stimulation. Electrical stimulation of the cereal nerve provokes a response similar to that of natural stimulation.

FIG. 4.

Oscillographic recordings from (A) cereal nerve, and (B) abdominal cord, between ganglia 4 and 3 on natural cereal stimulation. Upper time marker in each case represents 50 per second; lower action signal in each case represents duration of stimulation.

FIG. 4.

Oscillographic recordings from (A) cereal nerve, and (B) abdominal cord, between ganglia 4 and 3 on natural cereal stimulation. Upper time marker in each case represents 50 per second; lower action signal in each case represents duration of stimulation.

A few comparative behaviour experiments on Periplaneta and Locusta were carried out in a hot room (30° C.). A puff of air was directed on to the cerci of both these insects by means of a glass jet attached to a length of rubber tubing. The jet was moved about near the insects before the experiments in order to accustom them to its presence, and to ascertain that it did not, itself, produce a reaction. When a puff of air was then directed at the cerci of Periplaneta it caused the insect to move away rapidly from the region of the glass jet. The distance covered in this ‘evasion response’ was usually only a few inches, but the ‘locomotor activity which may carry the insect several feet’ (Roeder, 1948) was not observed. The locusts, on the other hand, behaved very differently. When the glass jet was placed near the cerci without disturbing the insect, a puff of air did not produce any response resembling the ‘evasion response’ in the cockroach. Locusts, blinded by painting the compound eyes with black paint, were much more quiescent and showed no response at all to movements in the region of the cerci, nor did a decapitated locust show any response to this stimulus, despite the elimination of possible inhibitory centres in the cephalic ganglion.

It is, perhaps, useful at this stage to compare the giant fibre system of the locust with that of the cockroach. Periplaneta americana has been most commonly used in insect neurophysiological investigations, due mainly to its large size and availability. Thus it is the only species amongst the Orthoptera in which the giant fibre system has hitherto been studied histologically. Roeder (1948) describes six giant fibres in each abdominal cord ranging from 20 to 45 μ in diameter and 10 to 12 fibres of 5 to 20μ in diameter. The fibres are arranged in two main groups, a ventral group formed by the three largest fibres and a smaller fibre and a dorsal, similarly arranged, quartet of fibres which are all slightly smaller than their ventral counterparts. The largest fibres in Periplaneta are thus considerably bigger than the four large fibres in Locusta, which only reach a diameter of 1 2 to 15 μ. Both in Periplaneta and in Locusta the giant fibres pass uninterruptedly through the abdominal ganglia and are considerably reduced in diameter while passingthroughthe abdominal ganglia. Roeder was not able to trace connexions of these fibres with individual cell bodies, but thought that they are multicellular, and arise from groups of cell bodies situated in the posterior peripheral region of the last abdominal ganglion. This region is close to the point of entrance of the cereal nerve. While it was not possible in the cockroach to pin-point the synaptic region between giants and efferents, there was no morphological evidence for the presence of interneurons.

Not only are the giant fibres of Locusta thinner than those of Periplaneta: there are also fewer of them, viz. only five fibres of a diameter of 5 μ or over. The most medianly situated of these fibres could be traced to its connexion with a cell body and is, in all probability, unicellular. The other three large fibres were not traced to individual cell bodies, although they disappear in the region of large cell bodies in the peripheral part of the posterior region of the fifth abdominal ganglion. The possibility cannot be excluded that the cell bodies of these fibres lie in the metathoracic ganglion. This, however, does not seem likely since these fibres, like the median fibre, taper towards their anterior ends.

Roeder (1948) considered it probable that one of the two quartets of fibres in Periplaneta served as a pathway for ascending impulses, the other quartet probably being concerned with descending impulses. In Locusta, both natural and electrical cereal stimulation evoke spikes in the abdominal cord which are both larger and faster than those recorded in the cereal nerve. These large spikes are not uniform in size, so that it would seem probable that more than one large fibre is involved in the response.

Prosser (1950) states that giant nerve-fibres have evolved many times, and that whereas they are present in some species, they can be absent in closely related forms. Wherever giant fibres are present they seem to function in rapid escape reactions, their efficiency depending on the fact that a single nervous unit brings about a complex fast response. Roeder (1948), in summarizing the observations of Pumphrey and Rawdon-Smith and himself, states that ‘Characteristic of cockroach behaviour is the burst of evasive locomotion which follows a sound or gentle air movement near the insects’. This reaction is called the ‘evasion response’. It is considered to be initiated by the stimulation of fine hair sensillae on the cerci. Impulses are carried into the sixth abdominal ganglion by the preganglionic cereal fibres, spatially summated at synapses with the giant fibres, and carried forwards to the metathoracic ganglion. Here two other types of synapse are found, one with shorter giants running forward, the other with fibres innervating the metathoracic legs which are the most powerful legs. This initiates the ‘evasive locomotion’ which is then sustained by descending impulses from the head region. Decapitated cockroaches show only a brief twitch of the legs, or at the most a few rapid steps, after stimulation in the same manner.

The ‘evasion response’ as such is not strikingly evident in the locust. According to Ewing (1904) a grasshopper with the cephalic ganglion removed reacts to all external stimuli with exaggerated movements, so that if an ‘evasion response’ similar to that in the cockroach did exist in the locust, decapitation would be expected to exaggerate the initial reaction in this response. No such increase could be observed in Locusta. The fact that the cereal ‘evasion response’ is absent in the locust is not altogether surprising. The locust’s cerci are smaller and less prominent than in the cockroach and are to some extent shielded by the wings. On the other hand, it is highly probable that the compound eyes and aerodynamic sense organs (Weis-Fogh, 1949) on the head, and auditory organs play a more important part in locust behaviour. The habitats of the two genera are also very different. Whereas the cereal ‘evasion response’ in the cockroach would seem to be of definite survival value, its value for the migratory locust is not so evident.

The functional significance of the giant fibre system in the locust is not thus very obvious although, oscillographically, the systems in Periplaneta and Locusta behave very similarly. The extra large multicellular giant fibres of the cockroach may conceivably represent a specialization, perhaps an aggregation of several neurons, tending to give an increase in fibre size with a corresponding increase in rate of conduction.

I wish to express my sincere thanks to Dr. O. Lowenstein for his very kind help, interest, and criticism in this work, also to Dr. H. F. Steedman for his help and advice over the choice of an embedding medium and with the photomicrographs. This work was carried out with the aid of a grant from the AntiLocust Research Centre (under the direction of Dr. B. P. Uvarov, F.R.S.).

Belar
,
K.
,
1929
.
Arch. EntwMech
.,
118
,
359
.
Ewing
,
H. E.
,
1904
.
Kansas Univ. Sc1. Bull
.,
2
,
305
.
Lowenstein
,
O.
,
1942
.
Nature
,
150
,
760
.
Nesbitt
,
H. H. J.
,
1941
.
Ann. ent. Soc. Amer
.,
34
,
51
.
Prosser
,
C. L.
,
1950
.
Comparative Animal Physiology. Philadelphia (W. B. Saunders Company)
..
Pumphrey
,
R. J.
, and
Rawdon-Smith
,
F.
,
1937
.
Proc. Roy. Soc. B
,
122
,
106
.
Roeder
,
K. D.
,
Kennedy
,
N. K.
, and
Samson
,
E. A.
,
1947
.
J. Neurophysiol
.,
10
,
1
.
Roeder
,
K. D.
1948
.
J. exp. Zool
.,
108
,
243
.
Steedman
,
H. F.
,
1947
.
Quart. J. mici. Sc1
.,
88
,
123
.
Snodgrass
,
R. E.
,
1935
.
Principles of Insect Morphology. London
.
Weis-Fogh
,
T.
,
1949
.
Nature
,
164
,
873
.