Whipple and Frantz, 1935)
. I n others it appears evident because of number, distribution, close relationship to, or continuity with, islet cells, that ducts may form an integral part of the tumour (O' Leary and Womack, 1934 ;Laidlaw, 1938 ; and others). Indeed, as ducts may occur occasionally in normal islets, and it is known that islet tissue differentiates from duct elements, it is to be expected that such structures would also appear in islet-cell tumours. The tumour we have described shows all gradations from the more typical solid masses and trabeculae to areas consisting almost exclusively of ducts.
The intimate relationship of vessels and cells which we have described is a feature of the normal islet and of tumours derived from islet tissue. Perivascular hyalinization in mature tumours similar to that detailed above has been emphasized by Laidlaw and by O'Leary and Womack.
Attempts to demonstrate specific granulation in the tumour by the methods of Warren (1938) and Gomori (1939, 1941) were unsuccessful despite secondary treatment with selective fixatives. This may have been due to initial fixation in formol-saline. Alternatively, it is possible that, as this tumour was non-functional, satisfactory initial fixation might not have demonstrated granules. Gomori (1943) stated that 80 per cent of islet-cell tumours are nonfunctional and that the only one of these subjected to special staining up to the time of his review had shown a-type cells.
It might be suggested that the great size of this adenoma was incompatible with an islet-cell origin. Certainly the vast majority of these tumours are small, but O'Leary and Womack (1934) and Sailer and Zinniger (1946) have recorded pancreatic tumours which were approximately half the size of the present one. That of Sailer was malignant, non-functional, and probably of &cell origin.
In view of the histological evidence of capsular and blood-vessel invasion, it is of some significance