132 



SCIENCE 



IN. S. Vol. XXVII. No. 682 



more diverse organs. Many, althougli not- 

 all, of the tumors are transplantable to 

 new individuals of the same species and 

 race, never to animals of another species, 

 and less often to those of another race of a 

 given species. 



There is something extremely subtile in 

 the conditions underlying successful trans- 

 plantation within one race since it may be 

 determined by such minor factors as en- 

 vironment and mere quality of food. A 

 tumor used to growing in Danish white 

 mice may fail utterly to gTOw in Berlin 

 white mice, and one used to growing in 

 Berlin mice and unable to grow in Nor- 

 wegian mice, may lose its capacity to grow 

 in Berlin mice transported to Christiania 

 and kept there for a period before inocula- 

 tion. 



Artificial selection of fast-gi-owing 

 strains can modify a slowly-growing into a 

 rapidly-growing tumor and increase the 

 percentage of successful implantations. 

 Once a high degree of power of growth is 

 secured, it can be maintained. The method 

 of selection for virulence is analogous to 

 that for securing virulent strains of bac- 

 teria. But the analogy does not go much 

 farther. It may be set down as a rule to 

 which at present the exceptions are in- 

 significant, that the more virulent a tumor 

 is, the less it tends to produce secondary 

 growths at a distance from the primary 

 nodule. Tumors which grow slowly cause, 

 not very infrequently, large secondary 

 growths especially in the lungs; but those 

 which 'grow rapidly never. Moreover, 

 once the original graft of tumor has begun 

 to grow vigorously, it is almost impossible 

 to implant successfully a second graft. 

 The metastases originate from tiunor cells 

 which have entered the circulation and 

 been deposited in the capillaries of the 

 lung. Now, highly virulent tumors, as is 

 to be expected, invade the blood-vessels 

 just as the less virulent tumors do, but in 



one case the cells at a distance develop into 

 new tumors and in the other they lie 

 dormant. If such a distinction exists in 

 human beings, the subjects of malignant 

 tumors, it has not been made out ; but there 

 are great variations in metastases in human 

 cases of tumors, which have never been 

 brought under any law governing tumor 

 gi'owth. My purpose, therefore, in speak- 

 ing of this peculiar feature of tumor 

 growth in mice is to bring out the fact of 

 the existence of a form of immunity to 

 tumor cells which may be restricted to one 

 part of the body, or be general to the whole 

 body. This immunity Ehrlich called 

 atrepsy and he conceives it to be an ex- 

 pression of deprivation of the peculiar 

 foodstuff required for tumor growth. In 

 any body this peculiar nutriment is limited 

 in quantity and hence if cells very highly 

 avid with reference to it are growing ac- 

 tively in any part of the body they may 

 draw all that is available to them and leave 

 none for cells, even of the same nature, 

 which are less avid or away from the focus 

 of its accumulation. The immunities of 

 species, possibly of races also, and the more 

 subtile forms of immunity alluded to, Ehr- 

 lich thinks may be atreptie; and for this 

 view there is more or less foundation in 

 facts developed by experiment. 



There remain many features of the ex- 

 perimental study of tumors of absorbing 

 interest to the student of tumors in human 

 beings, but they can not be discussed here 

 without extending widely the length of this 

 address. I may, perhaps, allude, before 

 passing to another and quite dift'erent 

 topic, to the interesting fact of the occur- 

 rence of grades of acquired active im- 

 munity in rats and mice that have been in- 

 oculated with tumor grafts which have 

 grown but slightly, or which, having grown 

 to a considerable size, later undergo com- 

 plete absorption. Mice and rats which 

 have recovered spontaneously from tumors 



