again on August 18 and finally died on August 20. There is no external indication whatever that the regenerated structure, Fig. 30 (n), is segmented. The basal portion (b) represents merely a part of the connection between the regenerated structure and the eye on the opposite side. There is little ground for supposing that, had the individual lived longer, the appendage would have become segmented; for more than three months had elapsed between the removal of the eye and the death of the crayfish and two moults had intervened. This is nearly twice as long as the period occupied in the development of the appendage described under Type II (Fig. 33, Pl. IV), and in the latter case fourteen segments were regenerated.

The horn-like processes of Types III and IV agree in the fact that in all instances the process is directed somewhat forward and is usually curved upward and at the same time slightly inward toward the median line. They further agree in the presence of a hard chitinous cuticle and in the absence of setae.

Type V: Indefinite structure regenerated. Finally there remain to be considered instances in which no definite regeneration takes place. The number of individuals, which may be considered as belonging to this negative type, are numerous. In the majority of such instances, however, the crayfish died before the occurrence of a moult, or the moult occurred in a very short time after removal of the eye. It is, therefore, impossible to say what the final outcome might have been. In most cases in which the entire eye had been removed, there were no external indications whatever of any regeneration until after a moult, but this is not true of every instance. Fig. 28, Pl. IV, shows the normal left eye and the stump from which the right eye was removed four months before. The animal died without having moulted after the operation. The new tissue that had been formed is represented by the shaded portion distal to the line

225]

REGENERATION OF CRAYFISH APPENDAGES

37

C. There are no external signs of differentiation, as constrictions or folds in the tissue, and the cuticle is not even hardened. Apparently nothing is present but a knob of indifferent tissue. Sections of this knob show nothing but a cuticle consisting of numerous lamellae and a small amount of connective tissue. In Fig. 29, Pl. IV, the uninjured left eye and the stump of the right eye of another individual are indicated. In this case, a moult took place about three months after removal of the entire eye, yet the regeneration here is apparently not further advanced than in the preceding case. The regenerated portion shows no resemblance to an eye or any other appendage. The optic nerve still persists, for it can be seen passing into the mass of tissue on the right (op). No more new tissue has been formed than would equal in volume the stalk of the normal eye. It does not seem probable in this and in the preceding instance that any definite structure would have been formed, if the crayfish had lived for a much longer period.

(c.) Theoretical Considerations. In considering these cases of heteromorphosis, it is noteworthy that in no instance is there anything like a basal protopodite present. In every case the main structure arises from a slightly rounded eminence which is in no way similar to a protopodite. This is again different from the majority of cases cited by Herbst. In all of his figures, except one, there is shown some sort of a base from which the principal portion of the appendage arises. The base is not altogether similar to a typical protopodite, yet there is sufficient resemblance to suggest it.

Undoubtedly, as Herbst suggests, the regenerated heteromorphic structure receives a plausible explanation on the ground of reversion to an ancestral condition. The question as to whether or not the eyes are modified appendages is difficult of solution, and the study of embryology has not yielded a defi

38

UNIVERSITY OF MISSOURI STUDIES

[226 nite answer. Yet, from an examination of the regeneration of the eyes of crayfish, it may at least be said that there are certain reasons for regarding the eyes as modified segmental appendages, and the heteromorphic structures which arise when the entire eye is removed indicate strongly that this interpretation is the correct one.

One of the most significant points in this connection seems to me to lie in the fact that the structure which is regenerated is not always an antenna-like organ. For instance, the structure represented in Fig. 30, Pl. IV, shows no segments whatever but has very much the appearance of the bud of an ordinary appendage before constrictions appear which mark the future joints. And furthermore, the fact that the regenerated structures sometimes appear in an antenna-like form, sometimes with one, sometimes with a pair of flagella, may point to a gradual transformation of the typical appendage, the appendage in its phylogony, having passed through an antenna-stage before reaching the condition of an eye.

There may be some objection to regarding this heteromorphosis as atavism on the ground that the regenerated structure lacks a basal protopodite. This, however, might be explained on the hypothesis of partial reversion. And, if it were possible to keep the same individual alive long enough to remove successively the regenerated structures, as can be done in other regenerated appendages, there would probably finally be regenerated an appendage more nearly approaching the typical form. However, Herbst, in some of his experiments, removed the heteromorphic structure, both wholly and in part, and found that a similar structure was always regenerated. In the shrimp, Herbst found an indication of a segmented basal portion, but the segments possessed no movable articulations. Although here there was a nearer approach to an ordinary appendage than I

have found in the crayfish, still it was far from being typical. In spite of the differences in the character of the regeneration of the shrimp and crayfish, there are sufficient similarities to point to a close relationship; and both instances may be regarded as arguments in favor of considering the Decapod eye a modified segmental appendage. Morgan's experiments upon the eyes of the hermit crab should be considered as belonging to the same series. But as he gives no details, I am unable to say whether the regenerated structures in the crab are more nearly like those regenerated by the crayfish or those regenerated by the shrimp.

It is possible that, if the conditions for experiment could be rendered more natural and hence more favorable, many additional facts could be learned which would be decisive in favor of reversion or against it, but at present the question must remain an open one.

IV. GENERAL CONSIDERATIONS.

It is not my intention to enter upon a detailed, theoretical discussion of regeneration. The phenomena of regeneration are not more explicable in the crayfish than in other animals, and at present it is not theory but fact that will prove useful in the ultimate solution of the problem if a solution is ever to be reached.

The points of interests to be mentioned here are to be found in the similarities and differences between my own experiments and those of others. For instance, in Herrick's experiments upon the lobster, he comes to the conclusion that the power of regeneration is regulated according to the uses of the parts injured. He says (loc. cit., p. 107), "The power of regenerating a lost part varies in both vertebrates and invertebrates in direct proportion to the physiological importance of the part, as Weismann has clearly shown."

Later investigation has not substantiated this view. Morgan, from his experiments upon the hermit crab, which have been mentioned in an earlier part of this paper, was led to take a precisely opposite ground in regard to the relation of regenerating capacity to the physiological importance of the part concerned and its liability to injury. He found that the hermit crab was capable of regenerating any appendage that was removed, even though under natural conditions, the possibility of injury to the appendage, as in the case of the abdominal appendages which are not exposed, was practically precluded.

He also found that appendages frequently injured regenerated not only at the usual breaking joint, but were capable of regenerating from a point either above or below that level. From these observed facts he concludes that for the hermit-crab at least, there is no relation whatever between power of regeneration and liability to injury. Not only for the hermit crab, but for all other animals that are capable of regeneration, he denies the existence of any such relation.

My own observations upon the crayfish tend to substantiate Morgan's view. I have found that appendages do regenerate which under normal conditions are not usually injured, and that they are capable of regenerating from levels other than that of the usual breaking joint. In my paper on The Crayfish of Missouri (:02) I mentioned the extreme uniformity of parts in C. gracilis, a species which spends nearly its entire life in burrows and is therefore protected from injury.

This uniformity shows that the appendages are seldom or never injured under normal conditions. In only one instance, and that has been since the above paper was written, have I ever found an adult of C. gracilis that had lost an appendage. Yet I found that when young individuals of this species were used for experiment, new appendages were restored as fre