HYPOTHESIS OF WELCH. 335 



by bacterial complement, by the circulating com- 

 plement of the body or by endocomplements of 

 the tissue cells. One could with equal reasonable- 

 ness assume that they may be complete toxins, 

 receptors of the second order, with a haptophorous 

 and a toxophorous structure. 



A well-known statement of Metchnikoff is to 

 the effect that a particular bacterium when viru- 

 lent is not so readily taken up by leucocytes as is 

 an avirulent strain. This fact has been noted re- 

 peatedly in recent times in the study of phagocy- 

 tosis in the test tube. This may be because the 

 organism, in its virulent parasitic state, secretes 

 substances which repel the phagocytes, neutralize 

 the opsonins, or because of the formation of actual 

 leucocytic toxins. 



One of the most widely known phenomena in 

 relation to the virulence of some organisms is that 

 their pathogenicity may be increased by passing 

 them through suitable animals repeatedly. The 

 best results are obtained when intermediate arti- 

 ficial cultivation is avoided and the inoculations 

 are made directly from the dead into the living 

 animal. It may, with all reason, be assumed that 

 by continued residence in the host the bacterium 

 has been trained to produce a greater quantity of 

 toxic substances which are inimical to the host, 

 and that the increased virulence of the parasite 

 depends on this condition. 



Although up to the present time systematic 

 attempts to place the hypothesis of Welch on a 

 firm experimental basis appear not to have been 

 made, the observations cited, as well as others 



