a low conical elevation of the dorsal half of the body-wall. Fig. 15 gives the position and spread of the growth in transverse section as seen under a low magnification (12 diam.). The middle line is marked by the dorsal and anal fins, and the figure shows clearly the great deformity caused by the growth. The centre of the new formation, which has completely

FIG. 15.-Gasterosteus: Squamous-cell carcinoma of skin. Shows infiltration or myotomes, connection with the skin, central degeneration, and dense lamellated stroma. X 1 ·

12

invaded the myotomes of the affected side down to the vertebral column, is occupied by an irregular cavity whose walls are formed of necrotic tissue with ragged projections. municates with the exterior, and at the apparently continuously into the skin as

[To face p. 54.

FIG. 16.-Gasterosteus: Squamous cell carcinoma of body-wall: natural size.

X

higher magnification. The tumour parenchyma is arranged in small solid alveoli, the peripheral cells of which are cubical or low-columnar in form, with scanty protoplasm and densely staining nuclei. Passing towards the centre of the alveoli, cells resembling those at the periphery are almost immediately succeeded by larger vacuolated cells with faintly stained nuclei, and in the most central parts even such indications of structure may be absent, and a granular mass of degenerated protoplasm takes their place. Here and there the peripheral cells of the more healthy alveoli are connected with each other by delicate fibrils as in the normal skin (fig. 17). Having regard to the features of the skin in fish, the growth can be pronounced to be a product of the covering squamous epithelium, and the wide invasion of the muscles as shown in fig. 15 leaves no doubt as to its malignancy. The connective tissue stroma of this growth is very characteristic and peculiar. It consists of very distinct fibrillar lamine with long spindle-cells between, arranged concentrically around the alveoli. In structure it closely resembles the dense lamellated fibrous layer underlying the normal epidermis elsewhere in this fish. It is impossible to decide whether this character is due to growth of the fibrous layer of the skin pari passu with the extension of the tumour, or is of the nature of a specific stroma reaction to the presence of the tumour on the part of the subcutaneous and intermuscular connective tissues of the animal, as has been shown by Bashford and Murray for transplantable mouse tumours.

The ulcerated surface of the growth has already become the nidus for many bacteria and parasitic protozoa, but their presence is obviously an added circumstance, as they are entirely confined to the surface and do not penetrate beyond the most superficial layers.

An interesting tumour of the jaw, having relations in its structure to odontoma on the one hand and squamous-celled carcinoma on the other, has been observed by Dr. Haaland in a specimen of Sebastes marinus var. viviparus caught by himself off the coast of Norway, and will be described and figured separately at a future date.

With the accumulation of fresh evidence, the conclusions tentatively advanced in the First and Second Scientific Reports as to the ubiquity of malignant new growths in the Vertebrata gain in certainty. The Reptilia, from which no malignant tumour has yet been recorded, still form an exception, due with great probability to the long span of

life of members of this class. The condition for discovering a considerable number of malignant new growths in animals is still that a sufficiently large number of aged individuals should be carefully examined. This conclusion, first clearly enunciated in a paper to the Royal Society and in the First Scientific Report, has been endorsed by Dr. M. Plehn in her studies on cancer in lower vertebrates (Zeitschrift für Krebsforschung, 1906, p. 525) and by v. Hansemann (Berliner Klinische Wochenschrift, 1907).

Thus, although cancer is universal in vertebrate animals escaping the effects of many of the chronic forms of irritation affecting man referred to above (pp. 19–23), its occurrence is frequently associated with other external irritants, and it may not be merely the ease with which lesions on the surface are observed, which has led to the accumulation of our extensive series of skin cancers. As in man, so in animals, no one form of external agency is constantly associated with cancer. The fundamental common factor is the peculiarity of the living cell to exhibit malignant growth under the action of most diverse agencies in divergent forms of life. The wide zoological distribution of malignant new growths, while affording the completest answer to the myriad speculations on the etiological association of conditions peculiar to mankind with the incidence of cancer, cannot be expected to furnish other than the most general indications of the essential, as contrasted with the incidental or subsidiary factors in its development. Nevertheless the surprising homogeneity of the malignant new growths in their general histological characters, points insistently to the cellular aspects and general biological importance of the problems of cancer.

At the time when the First Scientific Report was written (March 1904), it appeared as if the nature of the cellular transformation had been discovered. Farmer, Moore and Walker had described appearances in cancer-cells in man during mitosis, and had interpreted them to mean that the cells of the body had acquired the characteristic forms of celldivision until then only observed in reproductive tissues. We were able to demonstrate in cancer in animals the same morphological appearances. described by these observers, and to extend them down the vertebrate scale to the trout. We dissociated ourselves from the other conclusions. of these authors, and particularly from their statement that malignant could be distinguished from benign new growths by this means.

In following up this line of investigation discrepancies were met with in the preparations of malignant new growths, when the attempt was made to bridge over the lacunæ between different stages of the