190 F. HAGUENAU AND K. H. HOLLMANN 



observed for example in hepatoma (unpublished data) or in uterine cervix 

 cancer (Hinglais-Guillaud, pers. comm.). 



In experimentally induced mammary tumours cells of the mouse and 

 of the rat ergastoplasm is again never found in its typical organized form 

 as will be developed now. 



The mammary neoplasms of these rodents have been studied in the 

 electron microscope almost uniquely from the angle of their possible viral 

 aetiology. Most of these papers, concerned with virus detection (Bittner 

 agent in the mouse), mention only briefly if at all the general histological 

 aspect of the tumour or the characteristics of the constituting cells. 



But if one concentrates on this aspect, one observes that no " organized 

 ergastoplasm" is to be found in these cells where no lactation occurs. 

 Ergastoplasm is present mostly under the form of RNP-granules, abundant 

 in some cells, sparse in others (Figs. 17, 18). Not only are these cells rapidly 

 dividing but also they are deeply involved in virus formation as evidenced 

 by the presence of virus particles both mature (in the acinar lumen) and 

 in the process of formation (at the cellular membrane) (Fig. 18). 



In the rat, likewise, mammary tumours are devoid of "organized 

 ergastoplasm" (Fig. 19). There also, it is possible that virus formation 

 occurs since elements have been discovered in the electron microscope [23] 

 which from their morphology could be virus particles. The role they play 

 in the aetiology of the tumour has not yet been established. 



From all these observations on the ergastoplasm of the mammary 

 gland cell the following conclusions may be drawn. 



(i) In the normal gland, typical "organized" ergastoplasm is charac- 

 teristically related to physiological secretion. It is an attribute of the 

 normal lactating cell. 



(2) In experimentally induced tumours, ergastoplasm is found mainly 

 under a "non-organized" form, principally RNP-granules [i, 24]. 



These thus appear to be related to rapid cell replication or to forms 

 of protein synthesis other than normal secretion, virus formation in 

 particular. Such aspects of ergastoplasm are also characteristic of 

 embryonic cells which produce protein but do not excrete it. These 

 have been termed "retaining cells" by Birbeck and Mercer in a recent 

 paper [25]. 



(3) In human mammary tumours, most cells also contain ergastoplasm 

 in the form of RNP-granules or scattered dotted membranes and this 

 ergastoplasm is not on the whole, particularly abundant (B and transitional 

 types of cells). In some less numerous cells however (A cells of Caspersson 

 and Santesson) ergastoplasm is remarkably dense and its pattern of a 

 peculiar type not related to the specific tissue of origin. The significance 

 of these cells is not yet known but they merit further intensive examination 



