1903.] of Malignant Growths in Man. 50o 



latter. Furthermore, it is clear that the loops and rings characteristic 

 of this stage of the cellular development of the malignant growth are 

 arranged lengthwise on the spindle, and so are ultimately divided 

 transversely, exactly as in the corresponding heterotype mitosis of the 

 reproductive elements. 



Subsequent divisions that occur behind this zone appear to resemble 

 the somatic form, but retain a reduced number of chromosomes, just 

 as do the cells that arise from a parent cell that has once exhibited the 

 heterotype character. But irregularities of various kinds usually 

 supervene amitosis is of frequent occurrence, and the number of the 

 chromosomes in those nuclei that may continue to divide mitotically 

 often exhibits irregularity. These facts do not, however, seriously 

 affect our position, for in many plants similar irregularities occur in 

 post-heterotype cells that are not destined to give rise to actual sexual 

 cells. 



The above-described series of cellular and nuclear changes are not 

 confined to epitheliomata, but recur in an essentially similar manner in 

 other carcinomata and sarcomata. For example, in a rapidly developing 

 growth of a sarcomatous type from the cervix uteri we were able to 

 distinguish near the growing edge a well-marked zone of cells, 

 characterised by the somatic (and amitotic ?) types of mitosis, whilst 

 this was succeeded towards the interior by a band of heterotypically 

 dividing cells, and within this again the cells showed the somatic type 

 with reduced number of chromosomes, together with other cells in 

 which a mitosis was going on. 



In the case of slow-growing tumours which obviously tend to produce 

 a considerable amount of normal somatic tissue, such as the fibrous 

 tissue in scirrhus of the breast, cells showing these phases are, as 

 would naturally be expected, far more difficult to find than in rapidly 

 growing tumours. In such growths, cells showing the figures of 

 ordinary somatic division are numerous in comparison with those 

 showing heterotype figures. This would seem to indicate that the 

 cells which are destined to form fibrous tissue never divide hetero- 

 typically. 



It thus becomes evident that in a most important respect the various 

 types of malignant growths present certain features which are common 

 to all and that these features are similar to those to be observed in 

 the process of differentiation of reproductive cells from the preceding 

 somatic tissue. We feel that the evidence justifies us m deliberately 

 correlating the appearance of these " gametoid " neoplasms with the 

 result of a stimulus which has changed the normal somatic course o 

 cell development into that characteristic of reproductive (not embryor 



tissue. ,. ,, 



We look, then, upon this transformation as representing the 

 immediate cause of the development of the malignant growth but 



