460 MUCOUS MEMBBANES. 



large or very numerous ; on the contrar}^ the stroma of con- 

 nective tissue decidedly predominates. But the form of the 

 cell-nests is comparatively regular ; they are elongated tubuli, 

 circular in transverse section, in which the epithelial elements, 

 of small but uniform size, are disposed almost radially, though 

 an axial lumen may not be discoverable. 



Colloid cancer of the stomach is also distinguished by its 

 prevailing tendency to cause destruction, as opposed to ulcera- 

 tion ; even before giving way, it invariabl}^ becomes zonular. 

 It may thus give rise to very serious stenosis. The ulcerative 

 process is a slow disintegration, proceeding from within out- 

 wards, without haemorrhage or abundant secretion of any sort. 

 Were it not for the risk of extension to the peritoneum, colloid 

 cancer of the stomach would agree with scirrhus in the com- 

 parative innocence of its primary manifestations. 



As regards the minute structure of this form of cancer, the 

 horizontal sections recommended by Koster are so far important, 

 that they undoubtedly shift the onus of infiltrative propagation 

 to the lymph-paths of the gastric walls. This observer is inclined 

 to transfer the results which he has obtained from the investiga- 

 tion of Cylindromata (§ 173), with reference to the share taken 

 by the endothelia in the process of morbid growth, to colloid 

 cancer of the stomach. My notions concerning hard glandular 

 cancer, as laid down in § 159, make it quite obvious that I can 

 have no a priori objections to this view. I would regard these 

 appearances as indicative of an " epithelial infection " of the 

 endothelial elements of the lymphatics, starting originally from 

 the glands. But I must beg the reader to observe that my 

 attitude, especially as regards the colloid variety of cancer, is 

 one of expectation. 



§399. The spread of these three varieties of cancer 



OF THE stomach TO NEIGHBOURING ORGANS is of high cliuical 



importance and anatomical interest. The muscular coat is in- 

 volved after the submucous laj^er. The cancerous infiltration 

 follows the narrow bands of interstitial connective tissue between 

 the muscular fasciculi. These fasciculi themselves, owing per- 

 haps to the continued irritation to which they are exposed, 

 undergo a hyperplastic thickening. This is the rule at least in 

 the colloid and scirrhous forms of cancer. In vertical sections 

 through the walls of the stomach, we can trace, even with the 



