326 FRANCIS O. HOLMES. 



plant hosts only when the insects introduce their beaks directly 

 into the latex to feed. In the cells they do not enter the cyto- 

 plasm, so far as can be seen, but inhabit the milky juice which 

 fills the long cell vacuole. There they grow and multiply 

 rapidly, as is indicated by the many dividing specimens present 

 in latex smears from infected plants stained with Wright's stain. 

 The flagellates are shut off from other cells of the plant and even 

 from other latex cells of which there may be several. Unless a 

 single plant were infected several times by insects it is unlikely 

 that all the latex ducts would become infected, even though one 

 or more might. 



LOCALIZED INFECTIONS. 



If one latex system of a plant is infected, a macroscopically 

 localized infection may result unless that system happens to 

 penetrate every leaf and flowerlet of the plant. 



Early in the season of 1924 a search was .made for such cases, 

 for the previous year all the infections had appeared systemic. 



It seemed likely that insects might repeatedly bite plants 

 during the late summer, but that during June and early July 

 any infections which might occur from insect carriage would be 

 the result of a minimum number of infective bites. 



The first example of the way in which this worked out in the 

 field was met when a plant of a group surveyed thoroughly 

 every few days gave a negative record after showing flagellates 

 on several occasions. The later examinations of this plant 

 showed that flagellates were present, but were not always to be 

 found in the single drops of latex preserved as records. For this 

 plant, then, a new system of sampling was instituted. Samples 

 were taken from ten leaves instead of from one. It was found 

 that some leaves were positive and others negative, just as one 

 might reasonably expect if only a few of the latex cells were 

 parasitized. 



The study of sections of petiole and leaf tissue gathered at this 

 time showed the even more interesting fact that in plants never 

 suspected of having localized infections only a few of the latex 

 cells were inhabited by organisms. It was easier to find negative 

 cells than positive in specimens from apparently heavily para- 

 sitized plants. Smears from these plants showed very large 



