IteCEMBEE 24, 1915] 



SCIENCE 



913 



Without going into a detailed discussion of 

 the various conclusions arrived at, it may 

 suffice to state as one of the principal results 

 that diiferent strains of mice kept in the same 

 environment, being given the same kind of 

 food, differ very much in the frequency with 

 which cancer occurs among them. (Carcinoma 

 of the mammary gland in females is by far the 

 most common kind of cancer in mice.) In 

 some of them cancer may affect as many as YO 

 per cent, of the female mice, while in other 

 strains only 2-3 per cent, are affected. In suc- 

 ceeding generations this percentage figure is 

 fairly constant in different strains of mice. 

 Equally characteristic for different strains 

 seems to be the age at which cancer occurs; 

 while in some strains it occurs at a relatively 

 early age, in others it appears later in life. 

 Such differences in the cancer age may exist 

 even in strains in which the cancer incidence 

 is similar. 



These data are a prerequisite for further 

 studies of factors responsible for the sponta- 

 neous development of cancer as well as for at- 

 tempts to diminish or increase the cancer in- 

 cidence at will. From a practical point of view 

 such studies may ultimately lead to the dis- 

 covery of means enabling us to lower the can- 

 cer rate or to prevent cancer. As a first con- 

 tribution to this problem we undertook, on the 

 basis of our previous results concerning the 

 significance of heredity, an analysis of the in- 

 fluence of the ovary and especially of the cor- 

 pus luteum in the spontaneous development 

 of cancer of the mammary gland in mice. We 

 had shown previously that a combination of a 

 mechanical stimulus and the influence of a 

 substance secreted by the corpus luteum at a 

 definite period of its existence led to the pro- 

 duction of rapidly growing tumor-like new 

 formations with the structure of the deeidua 

 in all parts of the uterus. After a period of 

 at first rapid and then declining growth they 

 became later necrotic. We designated those 

 tumor-like formations as " Deciduomata " or 

 " Placentomata " and placed them among a 



used by Tyzzer (which was subsequently also 

 used by T. A. Murray) could, however, not be de- 

 pended upon to give decisive results. 



category of new formations which we char- 

 acterized as "transitory tumors."^ Other in- 

 vestigators showed later that the corpus luteum 

 was furthermore of significance for the pe- 

 riodic growth of the maromary gland, espe- 

 cially also during pregnancy. 



These facts suggested a possible importance 

 of the corpus luteum for the spontaneous devel- 

 opment of cancer in mice. Accordingly, we 

 undertook in conjunction with Miss A. E. 0. 

 Lathrop in Granby, experiments in this direc- 

 tion. Several lots of female mice of various 

 ages belonging to strains rich in tumors were 

 castrated, while other normal mice of the same 

 strains were kept as controls. In other experi- 

 ments a certain number of mice belonging to 

 strains of known cancer incidence were pre- 

 vented from breeding and thus the influence 

 of the non-occurrence of pregnancies was 

 tested. Under the latter condition the possible 

 influence of a corpus luteum was not entirely 

 abolished, but merely diminished. We wish to 

 state briefly the results so far obtained: Cas- 

 tration of mice at or below the age of 6 

 months (corresponding to a period of life when 

 the animals are already sexually mature) 

 diminished the cancer incidence in a very pro- 

 nounced way. The cancer rate fell from 60 

 per cent, to 70 per cent, in normal mice to 9 

 per cent, in castrated mice. Castration above 

 the age of 6 months has so far been without 

 any noticeable effect. 



Ifon-breeding mice develop cancer in a 

 somewhat smaller percentage of cases and at 

 a somewhat higher age than normal breeding 

 mice. The influence of prevention of breeding 

 is therefore much less marked than the effect 

 of castration, but considering the large num- 

 ber of mice we used it is not probable that the 

 differences which we actually did observe were 

 accidental. However, we are continuing our 

 experiments in various directions and if acci- 

 dental factors should complicate some of our 

 results, this will become apparent during the 

 further development of our work. We estab- 

 lished thus two sets of factors in the etiology 



2Loeb, L., CentralU. f. allg. Fathol., XVIII,, 

 1907, 563, and a series of papers in the Archiv f. 

 EntwicMungsmechanik. 



