4 BULLETIN 1015, U. S. DEPARTMENT OE AGRICULTURE. 
fruited varieties, the more distantly related small- fruited varieties, 
and the varieties recommended in the literature and in seedsmen's 
catalogues for resistance to tjomato diseases. Some of these were 
tested in 1915 and others in succeeding years. 
PREPARATION OF THE BREEDING PLATS. 
In order to have satisfactory breeding plats for making the tests. 
a piece of sandy soil and a piece of clay loam upland on the Arling- 
ton Experimental Farm, near Washington, D. C, were thoroughly 
inoculated with the wilt fungus. This was done by sterilizing 
chopped straw in a large autoclave, spraying it with spores from a 
pure culture of the fungus, and after about a month disking it into 
the surface soil. A good covering of inoculated straw was thus 
added to the soil for three successive years. Plowing under this 
straw and the dead infected vines soon thoroughly inoculated the 
soil with the fungus. 
To increase further the opportunity for infection, during the first 
two years the seed was soaked before planting in a suspension of 
Fusarium spores and sown in flats containing wilt-infested soil. In- 
fested soil was also used in the flats to which the seedlings were 
transplanted and in the holes in which they were set in the field; 
also pure cultures of the fungus were mixed with the soil in the 
flats and with the soil around the roots of the plants in the field. 
As wilted susceptible vines add large quantities of the fungus to 
the soil while resistant vines do not, the infected vines were cut into 
small pieces and distributed over the field to help maintain uniform 
infestation. This seemed desirable, because in greenhouse experi- 
ments with tomato wilt the writer found that even when all plants 
are exposed to infection the percentage of those infected varies with 
the quantity of inoculum used. Although the fungus was already 
thoroughly distributed through the soil, its unequal accumulation in 
certain parts through the plowing under of susceptible vines in one 
place and of resistant vines in another, if not corrected, would very 
likely have affected experimental results. 
METHOD OF SELECTION. 
Selections from two classes of material have been made : (1) Those 
of apparently resistant plants from wilt-infested fields, and (2) 
those of resistant strains from these selected plants. 
The original selections were made in the worst wilt-infested areas 
that could be found, and though not always free from wilt they were 
much freer from it than the plants immediately surrounding them. 
The seed from each plant was harvested separately and tested in 
progeny rows. These progeny rows or strains were graded for yield 
and quality of fruit and for wilt resistance. Those inferior to the 
