INFECTION THROUGH STOMATA. 59 
enlarging unaffected parts of the green fruits. The only conclusion I could come to, there- 
fore, from extensive field observation made during 1902 and 1903, and from a careful 
study of serial sections made through many young spots, was that infection invariably 
takes place through the ordinary stomata, favored by the presence of rain drops or dew. 
Opportunity to test the validity of these conclusions by actual inoculation experiments 
did not occur for some time. During the first two years no trees were available. In 1903 
trees were obtained, but through neglect to make transfers at the right time (many other 
lines of work being conducted simultaneously) the cultures were allowed to die, the only 
organism remaining alive when the cultures were needed being a greenish-yellow one, 
which was applied to the trees freely, but which proved to be destitute of pathogenic 
properties. One other set of experiments miscarried by reason of leaf-miners. 
In the summer of 1903 cultures of the right organism were obtained once more from 
the Michigan orchard where the disease prevails every year, and in the summer of 1904 a 
thorough test was made, great numbers of stomatal infections being obtained both on leaves 
and green fruits by simply placing cultures of the organism in sterile water and atomizing 
this upon the tree. A Japanese plum tree about 5 years old and of the variety known as 
Abundance was selected for this experiment. The tree was about 12 feet high, with a 
corresponding spread of branches. It was very leafy and full of green fruits about one- 
third grown. The leaves and fruits were very smooth and free from gnawings of insects 
or fungous injuries, the only injury on the fruits being an occasional healed (corked out) 
curculio puncture. The tree stood in a garden at Takoma Park, a suburb of Washington, 
where no such disease was ever known before. No wounds were made, but the organism 
was scraped from several young slant-agar cultures into several hundred cubic centimeters 
of sterile water, and a portion of this was sprayed upon the tree by means of the apparatus 
shown in vol. I, figs. 92 and 93. The sprayings were made on the afternoon and evening 
of June 1, during a light drizzling rain, moist cloudy weather being selected for the experi- 
ment so that conditions might be as near as possible like those which frequently occur in 
Michigan during the period when most of the natural infections take place. The weather 
continued cloudy and moist for many hours, after which it was clear and favorable for the 
trees, with drouth at no time. For a week after the spraying there were no visible results. 
On June 8, on a few plums, the writer thought he detected some incipient spots, but the 
signs were very obscure even under the hand-lens and serial sections through the particular 
suspected tissues showed them to be free from bacteria. Distinct spots, yielding bacteria 
in abundance on sectioning, were visible, however, on many of the leaves and green fruits 
on June 14. It is probable, therefore, that they might have been detected in some cases 
as early as thetenth or twelfth day, but not much earlier, since the largest spots were still 
quite small. From this time on, the bacterial spots became numerous on both leaves and 
fruits and passed through their customary changes in an entirely typical way, so that when 
spotted plums were sent on from the Michigan orchard in July and were compared with 
those obtained by the spraying, they could not be distinguished (fig. 13). The organism was 
plated out of a dozen or more spots by several of my assistants, the poured-plates yielding 
in great abundance colonies of what was sprayed upon the tree. When made from young 
spots these plates generally contained pure cultures, but sometimes there were a few con- 
taminating colonies of various sorts. When made from old cracked-open spots, the number 
and variety of contaminating organisms was greater and then often included fungi. More- 
over, serial sections made from many spots yielded the same results as the similar sections 
already described as made from plums obtained from the orchard in Michigan, to wit: In 
early stages constant presence of the bacteria and absence of injuries due to fungi or insects. 
The writer has slides of sections cut from these sprayed plums showing all stages of the 
disease, from simple occupation of the substomatic chamber, to the formation of extensive 
closed bacterial cavities with large destruction of tissues. These experiments show also 
