430 SCIENCE PROGRESS 



" elementarkorperchen " of v. Prowazek, and in another form 

 with Guarnieri's bodies {Cytoryctes). Here, again, a resemblance 

 to Rickettsia would seem to be indicated. 



No systematic search for the agent in transmitting insects 

 appears to have been made. That there is a gradation of size 

 in the infective agent in the different filterable virus diseases 

 of plants is evident from the filtration experiments, the cu- 

 curbit mosaic, for instance, being more easily removed by 

 filtration than that of tobacco. 



Various indications suggest that the parasite must often be 

 closely associated with the living protoplasm of the host cells. 

 The plant juice is not usually as highly infective as the crushed 

 tissues, and transmission by grafting is one of the most successful 

 methods and the only one known in several cases. There is 

 not the slightest evidence that the infective agent can live 

 saprophytically, and the failure to get infection through the soil 

 in very many carefully conducted tests, though the underground 

 parts of the plant are exposed to frequent wounding in many 

 ways, is conclusively against a soil life. Failure to cultivate 

 the active agent is common in the obligate plant parasites, none 

 of the rusts (of which there are several thousands) having ever 

 been grown apart from its host. 



The localisation of active symptoms to developing parts of 

 the plant is a remarkable feature of these diseases. Infection 

 is followed by the appearance of symptoms (chlorosis and 

 various other manifestations) only in the newly formed shoots 

 and leaves. Those that have already reached their full size 

 remain unchanged, though in many cases they are highly infec- 

 tive. The virus can at least live in the mature tissues, but it is 

 apparently unable to do harm to any but developing tissues, 

 except where it is able to cause necrosis, as in leaf roll disease 

 of potatoes and allied diseases. 



Though there is no evidence in many of these cases that the 

 virus passes through any necessary part of its life cycle in the 

 insect carrier, it is sometimes very noticeable that only certain 

 species of insect can transmit it. In the tobacco mosaic some 

 species of aphids will cause infection, but others fail, though 

 they feed and multiply readily on tobacco ; and in sugar-cane 

 mosaic only Aphis maydis has been proved capable of trans- 

 mitting infection, though many other sucking insects feed on 

 cane. The case is possibly similar to that of Japanese river 

 fever, where it would seem that the mite which ordinarily trans- 

 mits the disease is not a mere mechanical carrier, but its inter- 

 vention is not always necessary to secure infection, since all 

 stages of development of the parasite can take place in man. 

 Or we can imagine a condition similar to that obtaining in some 

 heteroecious rusts, which can propagate indefinitely in the uredo 



