322 ON GROWTH AND OVERGROWTH 



entiated are the epiblast and hypoblast : their differentiation, 

 indeed, is one of the earliest events in developmental history ; 

 and this being the case we should expect — and we find — that 

 tissues, whether normal or neoplastic, derived in direct line 

 from these two layers, are singularly tenacious of their properties 

 as lining membranes, and so it is that Epilepidomata and Hypo- 

 lepidomata always show evidences of their lepidic nature. We 

 should not expect — and we do not find — that where this direct 

 line has been departed from, where, for example, hypoblast has 

 given off masses of cells to form mesoblast, and the epiblast 

 similar masses to form the neuroblast, that, in reversion, tissues 

 derived from mesoblast and neuroblast respectively should again 

 enter upon the lining membrane stage. Where in the process 

 of development an organ or part is formed by the cells of tissue 

 of a higher order assuming a less differentiated condition, and 

 from this lower state proceeding to develop along special lines, 

 we do not find that in reversion and degeneration that tissue 

 passes beyond the less differentiated stage, and then proceeds 

 to show characters of a primary more differentiated condition. 

 Thus it is that the mesenchymatous tumours and sarcomata in 

 general show no tendency to assume fining membrane or lepidic 

 characters, even though, without exception, all these tissues have 

 primarily arisen as derivatives from either epiblastic or hypo- 

 blastic lining membrane. 



If occasionally in gliomata we find cysts lined with columnar 

 epithelium, this is not an example of such reversion to the more 

 primitive epiblastic characters of the glial tissue. Those who 

 have studied cases of this order have, without exception, ascribed 

 such conditions to inclusions of rests of embryonic tissue con- 

 taining portions of the Anlage of the central nervous canal. 1 



But now, coming to the mesoblast, we know that it is at a 

 later date in the history of the embryo that this becomes 

 differentiated into mesothelium and mesenchyme, and the 

 development of the mesothelium is in the direction of increased 

 specialization. Thus we should expect — and we find — that in 

 processes of a reversionary type these more lately acquired 

 characters are more capable of being lost, so that growths formed 



1 [I have since pointed out (Principles of Pathology, vol. i. 1st edit. p. 564) 

 that in gliomas areas of necrosis with autolysis of the dead tissue come to form 

 cysts lined by a false epithelium. The lining cells become modified as a function 

 of their position, but present no basement membrane.] 



