and some specimens collected by trawling. All the eviscerate specimens showed clearly by the state of the mesentery edges that autotomy had just occurred, and can be presumed to be due to the severe mechanical effect of wave action in one case and of trawling in the other. Autovisceration could, however, be artificially induced in all specimens tested, which included a size range from 8 gm. to 300 gm. Although there is ample evidence to show the widespread ability of holothurians to autotomise certain organs under special conditions. there are fewer recorded cases of the occurrence of this process under natural conditions. Each stimulus varies greatly in effectiveness from species to species, e.g., the methods used by Kille (1931 and 1935) for inducing autotomy in Thyone were found quite ineffective for Holothuria parvula Kille (1937). The use of dilute ammonia was found unsatisfactory for S. mollis, whereas the injection of distilled water was effective in all specimens of this species. In view of the special conditions sometimes needed to induce autotomy in holothurians, it is surprising to find the property apparently present in all specimens when suitable stimuli are used.

The organs expelled by different holothurians vary among genera but are very constant within members of the genus. In Thyone the entire length of the alimentary canal, and its haemal vessels, the introvert and the lantern with its associated structures are autotomised, Kille (1935). The respiratory trees and the gonads are retained. Phyllophorus magnus under some conditions autotomises the whole anterior part together with the tentacles, mouth, calcareous ring and other portions “snap off from the body,” Domantay (1931). Synapta appears to have less clearly defined points of autotomy and responds to inducing stimuli by constriction of the body into fragments. Species of Holothuria lose the part of the alimentary canal between the oesophagus and the cloaca, the rete mirabile and the left respiratory tree (Semper, 1861; Kille, 1936). This differs from Stichopus mainly in the retention of the right respiratory tree. In Stichopus mollis branches of gonads are also frequently expelled. Although there is a considerable variation in the organs expelled by different genera of holothurians, there is very great constancy in the points of rupture and in the organs expelled in each genus. There appear to be constant breaking points present within each genus. Lukas (1905), in a work not available to the writer, is stated by Pearse (1909) to interpret the different results as indicating that the place where self-mutilation takes place in an animal is usually determined by certain structural characteristics of the animal concerned. There is, however, insufficient data on the morphology of the regions at which rupture occurs, to show the reason for separation at these points. Histological examination of the regions of rupture in S. mollis did not show any distinct discontinuity or constriction of any of the layers in the alimentary canal. It is understandable that the point of junction of respiratory trees and rectum with the cloaca might be structurally weaker than other regions of the alimentary canal, but the reason for the constancy in position of the rupture in the oesophagus is not clear. All specimens of Stichopus mollis examined showed identical points of rupture.

It seems rather surprising, in view of the widespread occurrence of autotomy in Holothurians and in the constancy of breaking points in each genus, that possible utility of the process in suggested only in certain species. Those which undergo binary fission probably use the process as a method of asexual reproduction. Species which show regeneration in natural conditions must be capable of responding to some inducing stimulus in the environment, and it is possible that this stimulus is predator attack. The process has been suggested as protective by Minchin (1892) for Holothuria and by Domantay (1931) for the genera Holothuria, Stichopus, Thelenota and Actinopyga, which are stated to react to irritation or strong stimulation by ejecting their viscera. A series of experiments was performed on Stichopus mollis to determine its rate of response to the type of mechanical stimuli likely to be involved in predator attack. It was found that violent agitation accompanied by extensive pricking with tacks projecting from plates which held the specimens was effective only after ten or more minutes. Crushing, although effective after prolonged action, was also very slow in inducing autotomy. The treatments used were much more severe than would be expected from a predator, yet the response was far too slow for autotomy to be of any value in diverting the attention of a possible predator to the expelled viscera. There is also no evidence of any fish or other organism which preys on S. mollis. It is difficult to visualise a protective function for the process in this species.

There is much evidence, as previously mentioned, to show that autotomy can be induced casily in many holothurians by certain chemical and physical stimuli. It is difficult to relate the response to these stimuli with any conditions which would be encountered by holothurians in their natural environment. A number of holothurians, including S. mollis, undergo autotomy when the sea-water becomes, sufficiently soul. Experiments on S. mollis were carried out to determine the relative effects of oxygen deficiency and that of the accumulation of excretory products in inducing autotomy. It was sound that specimens in previously boiled and cooled seawater would not undergo autotomy any sooner than those kept in a similar volume of fresh sea-water. Of eight specimens kept separately in sea-water in which the stirred-up faeces of Stichopus mollis was added, six eviscerated within three hours, while the eight control specimens took 15–20 hours. Two specimens died without undergoing autotomy. This shows that in S. mollis the accumulation of excretory products in the main factor in inducing autotomy in specimens contained in a small volume of sea-water. It has been suggested (Pearse, 1909; Domantay, 1931) that the ejection of the visceral organs usually decreases the total amount of metabolism in holothurians, and thus they are able to survive until better environmental conditions again arise. S. mollis was found to be even more sensitive to foulness of the water after autotomy than before the process. Specimens which had eviscerated and were not immediately transferred to fresh water showed a high mortality rate. Specimens which remained intact under the same conditions showed no after-effects. Those which eviscerated and were quickly transferred to fresh sea-water showed a higher survival rate than