132 Journal of Agricultural Research voi. xix. no. 3 



1902 that he succeeded in inoculating healthy plants with the disease by injecting 

 sap from diseased plants. We have repeated these inoculation experiments as far as 

 we have been able to obtain data about them, but without success. Neither have we 

 been able to find any other indication that the disease is contagious. 



They conclude with Kobus and van der Stok that the mosaic is an 

 expression of bud variation. No reference is made to successful inocu- 

 lation experiments in the numerous papers on mosaic in the Hawaiian 

 Sugar Planters' Record for 1911-1919. Stevenson (5) reports hundreds 

 of inoculations of many cane varieties by various methods during 191 7 

 and 1918, all with negative results. Prof. F. S. Earle, in an unpublished 

 paper, calls attention to a method of inoculating with juice expressed 

 under oil to prevent oxidation. Some of the plants he inoculated be- 

 came diseased, but the experiment was inconclusive and open to the 

 criticism that it was carried on v/ithout control plants in a field where 

 cases of the disease were appearing naturally. 



Various writers have called attention to the possibility of insect 

 carriers of the mosaic disease, but no published proof has appeared, 

 and the statements have been based on analogy with other apparently 

 similar diseases and on field observations. The failure of all efforts to 

 obtain uniform or dependable results with either artificial methods of 

 inoculation or with insects has been one of the conspicuous peculiarities 

 in the behavior of sugar-cane mosaic. In all inoculation work in plant 

 pathology it is necessary to secure a very high percentage of infection 

 in inoculated plants where control plants are not absolutely protected 

 from extraneous infection. In diseases like cane mosaic, where, for 

 reasons which we are not in a position to discuss at present, the percent- 

 age of infection resulting from experimental inoculation is not high, it 

 is not only necessary that all experimental plants be apparently healthy 

 but also that they be of known healthy parentage for at least one gen- 

 eration back and preferably more. Further than this, the experiments 

 should be performed under absolutely controlled conditions. The pre- 

 vention of contamination of experimental plants with diseased material 

 by direct or indirect contact must be absolute. Special precautions 

 must be taken to prevent the admittance to treated plants of insects or 

 any other animals other than the ones being experimented with. 



The writer became convinced, after observations and experiments 

 with the mosaic disease dating from, the summer of 191 6, that more 

 reliance can be placed on the results of experiments performed in some 

 region far removed from any chance of accidental infection. It was 

 owing to these considerations that the experiments recorded here were 

 performed at a distance from the seat of any natural infection, because 

 the required conditions would be practically impossible to obtain where 

 the disease is prevalent. 



The first experiments were conducted in a quarantine greenhouse 

 near Garrett Park, Md. Later experiments were made in several green- 



