Peripheral Nerve and Regeneration in Crustacea

The limbs of the 7th thoracic segment were chosen for experiment ; they are not specialized in any way and thus may be regarded as typical of the thoracic limbs in general. One of the pair served as a control, regenerating under normal conditions after amputation, under light ether anaesthesia, at the normal autotomy plane. The other limb was similarly amputated at the same time, and the nerve to the stump was severed where it passes from the body to the coxa, close to the ventral body wall, about mm. proximal to the level of amputation. The damage to the superficial tissues was not great under these conditions ; visible haemorrhage ceased within half a minute and the wound cicatrized readily. A trace of proflavine was used in the water as an antiseptic both during operation and subsequently. Right and left limbs were used equally for experiment, to allow for any possible intrinsic difference between the two in their rate of regeneration (Needham, 1943, p. 57). In a parallel series of operations, serving as additional control, a wound equal in magnitude to that caused by denervation, was inflicted on the base of one limb from the dorsal surface. In neither series was the blood supply to the regeneration blastema directly affected by the operation, so that the second series may be regarded as a Suitable control for the effect of mere tissue destruction in the limb base. Microscopical sections indicated that, in general, the dorsal wound caused more widespread damage than the ventral denervation, since muscle tissue was largely affected; the muscles concerned subsequently degenerated along their length, often remote from the initial lesion.

The regenerating limbs were measured, as in a previous paper (Needham, 1943), on their first appearance externally, usually at the first post-operational ecdysis (scale 56 units =1 mm.). The significance of the mean control-experimental (C.-E.) difference in total length of the regenerate was tested statistically (Tables 1, 2). In both series of experiments data were obtained for twenty-one individuals, representing a survival of about 37 % in the denervation series and 66 % in the other. In three individuals of each series the experimental limb had not appeared externally when the control was measured. These zero figures increase the significance of the mean c.-E. difference in the type of test used, more than is perhaps justified (Table 3), and they were therefore omitted in making.the crucial tests. The test adopted was that the mean c.-E. difference should be more than 3 × its probable error.

Denervation under the conditions of this experiment caused a significant retardation of regeneration (Table 1). The significance is not masked by individual variations in rate of regeneration and in the time interval between operation and measurement of the regenerates (p. 146). An equivalent wound not affecting the peripheral nerve may also have a retarding effect but was scarcely significant (Table 2), even when including the three individuals in which there was no external regeneration of the experimental limb (Table 3).

No evidence was obtained concerning the nature of the particular functional component of the peripheral nerve which affects regeneration of these limbs. The effect may be purely trophic, since no marked abnormalities in the form of the regenerates were obtained.

The results suggest not only a difference between the mean length of experimental and of control limb in each of the two series, but also a difference in the mean length of the control limb between the two series. The various mean lengths were as follows:

The shorter experimental limb of the denervation series is correlated with the longer control limb, suggesting the possibility that regeneration of the control limb may be accelerated in proportion to the degree of retardation of its experimental partner. The difference between -the means of the two controls is, however, not statistically significant either when using actual limb lengths or relative limb lengths (limb length/body length) as the basis for the tests. The difference is more nearly significant when estimated by the first method, which is no doubt partly explained by the slightly greater mean body length in the denervation series (8·1 mm. compared with 7·9 mm.). Nevertheless, some support for the above hypothesis of competition between the two limbs of a pair of regenerates was provided by the great length of the control limb in four of the six individuals with a zero measurement for the experimental limb (see also p. 146 and Needham, 1943, P-59)

The absence of qualitative abnormalities in the present experiments is not complete proof that the nerve exerts only a trophic effect, as in Vertebrata (p. 144). The limbs were always reinnervated by their original, appropriate nerve, which, if it determines the nature of the regenerate, will no doubt usually restore the status quo. The classical experiments of Herbst (1896) strongly indicated that the nerve ganglion determined the nature of the regenerate: that this effect may be indirect (Child, 1941, pp. 342–3) does not alter the main fact. Previous tentative suggestions that the nervous system may determine morphology (Needham, 1941, pp. 79–81; 1943, p. 69) merit further investigation.

It may be necessary to revise Millot’s rule (p. 144) ventually. The nervous system would seem to be important for regeneration in two groups, the Crustacea and the Amphibia, with relatively extensive powers of regeneration, and the correlation would seem to be a direct one, between the importance of the nervous system for regeneration and the complexity of organization of the animal rather than an inverse correlation between the former and relative power of regeneration.

The results of the dorsal wound series suggest that any tissue damage proximal to the level of amputation may retard regeneration to some extent. This might be due to competition for materials essential to regeneration (cf. p. 145), since neither nerve nor blood supply was affected. In the denervation series, however, the factor of competition probably did not play any significant part ; microscopical sections indicated that no appreciable regeneration occurred until the blastema was reinnervated, by which time repair of the damage due to the operation of denervation was practically complete.

It is important that statistically significant differences can be demonstrated in these experiments in spite of the difficulty of obtaining strictly comparable measurements in the different individuals. No accurate measurements can be made until after the ecdysis at which the regenerate appears externally, and this occurs a very variable time after amputation, even in individuals of the same size and when amputation is performed at a fixed interval after the previous ecdysis ; thus the measured length of the control limb shows considerable individual variation. It is possible to-offset this variation to some extent by using the percentage C.-E. differences in the tests for significance rather than the actual differences of Tables 1 and 2, and such tests do show an increased significance. Since the curve of regeneration is sigmoid and not linear, however, the percentage C.-E. difference is not a strictly comparable estimate, at all stages of regeneration, of the difference between the two regenerates. In view of the unavoidable variables it was decided not to restrict the experiments to individuals of one particular body size. It seems probable that, with precautions to minimize all variables, statistically significant results might generally be expected in experimentation regeneration in the Crustacea.

In the denervation series, but not in the other, the difference between the mean experimental limb length and the mean control limb length is found to be less significant than the mean c.-E. difference. This probably indicates that the variance of C.-E. is less than that of control or experimental themselves, and that denervation arrests regeneration completely for a certain time after which it proceeds at about the same rate as in the control limb. This is supported by microscopical examination of denervated blastemas (see above). If, on the contrary, the effect of denervation were to reduce the rate throughout the period of regeneration then the C.-E. difference might be expected to vary as much as control and experimental themselves.

1. The peripheral nerve appears to have an important trophic influence on the regeneration of thoracic limbs in the crustacean Asellus aquaticus, and section of the nerve proximal to the level of amputation causes a significant retardation of regeneration.

2. The retardation is statistically demonstrable in spite of the number of variables affecting the rate of regeneration.

3. Regeneration is probably completely inhibited until the blastema becomes reinnervated by the regenerating nerve.

4. Comparable damage to the limb base, not affecting the nerve, may also cause some retardation of regeneration, although, in these experiments, this was not clearly significant.

I am greatly indebted to Prof. H. S. Raper for the facilities I have enjoyed in carrying out this work, to Prof. J. Gray for his helpful advice on the manuscript, to Dr G. L. Camm for assistance with the statistical tests, and to my wife for her continued help throughout.