392 
Journal of Agricultural Research 
VdL XXIII, No. s 
on white blotting paper which before sterilizing was dipped in a rice- 
starch solution so that the color reaction of the fungus may be observed 
as soon as the growth of the fungus begins. The color produced on the 
rice-starch paper is much darker purple, more like the iodin starch 
reaction, than that produced on rice alone, but it is quite constant and 
was considered dependable in this test. 
On May 3 infected seed was planted in three short rows in soil which 
had never before been in cotton. About 1,000 seeds were planted, and the 
germination was good. On May 27 a dozen or more of the seedlings from 
this planting were found dying and thoroughly infected with wilt. On 
June 9 the plants were thinned to two or three in a hill, leaving 378 plants 
in the plot. All plants removed were examined, by sectioning, for signs 
of wilt. Of the 493 plants examined, 32 gave evidence of wilt infection, 
making about 5 per cent of the plants showing infection up to that time. 
How long the spores on the seed will remain viable in this test still 
remains to be determined, but it is very evident that it will live the usual 
period between picking time and planting. It is probable that the 
number of spores occurring on the artificially inoculated seed is con¬ 
siderably greater than is likely to occur naturally, but Edgerton 6 has 
shown that the number of spores of various fungi occurring naturally on 
cotton seed may be enormous, Under favorable conditions the number 
of Fusarium vasinfectum spores could easily be as high as reported by 
Edgerton for other fungi, and, under the ordinary bin storage of cotton 
seed, at least a considerable number should remain viable until planting 
time. 
It would seem likely that spores of the wilt fungus might be introduced 
into a field without infecting the soil, or at least that several years might 
elapse after its introduction before the infection would become noticeable. 
In the light of the results obtained in the writer’s work, it would seem 
that some such explanation must be necessary to account for Gilbert’s 7 
apparent negative results. In the planting of artificially infected seed 
mentioned above, at least 80 per cent and probably 100 per cent of the 
seed carried the wilt organism at the time of planting, yet only a very 
few plants died of wilt in the early seedling stage and those could easily 
have been overlooked if the plot had not been examined carefully. At 
the time of thinning, also, only a few of the plants having wilt showed a 
general infection, and many of those infected might not have given much 
external evidence of the disease if they had remained in the field. 
Undoubtedly those seeds which carry the wilt internally stand a poor 
chance of producing a plant that will live more than a very short time, 
and it is probable that an introduction of the wilt by such seed would 
escape notice the first year unless secondary infection of adjoining plants 
should take place. 
It would seem that under the present system of saving cotton seed and 
the planting of very large quantities of seed to insure a stand, every step 
in cotton planting favors the introduction of the wilt disease into new fields 
and its general distribution in all fields. The results of the present 
work would seem to justify the recommendation that seed for planting 
purposes shall not be saved from fields badly infected with wilt. Acid 
• Edgerton, C. W. the rots of the cotton boll. Ea. Agr. Exp. Sta. Bui. 137,113 p., 13 pi. (in text). 
191a. Bibliography, p. 81-85. 
1 Gilbert, W. W. cotton diseases and their control. U. S. Dept. Agr. Fanners’ Bui. 1187, 32 p., 
18 fig. 1921. 
