THE WILT DISEASE OF PIGEON-PEA. 55 
not been of sufficient duration to indicate how long this period may 
be, but it certainly exceeds a year. Most of the conidia found in 
cultures nearly a year old were still capable of germination, and 
then usually formed secondary spores very rapidly. The chlamy- 
dospores are likely to retain their vitality for a still longer period. 
The prevalence of the disease and the extent of damage caused by 
it are largely due to soil infection from the preceding crop. It is 
on this account that sufficient rotation is of such paramount im- 
portance in pigeon-pea cultivation im India. Spread within a field 
takes place chiefly through the soil. It is, however, dependent 
largely on the space between the plants and the size of the latter. 
It is noticeable that extension occurs much more rapidly where the 
plants are crowded in pots than in the field. Similarly the exten- 
sion is more rapid the larger the plants are. It cannot be doubted 
that this is due in both cases to extension occurring more rapidly 
along the roots than across the intervening soil. Hence the nearer 
together the roots of different plants are, whether as a result of 
overcrowding or of extensive lateral growth in the roots of large 
plants, the more rapid the spread of the disease. 
Aerial spread by wind-borne spores probably occurs, though 
to a lesser extent and much more slowly than through soil. The 
following observations bear on this pomt. Up to the present no 
direct attempt to isolate the parasite from the bark of the stem of 
wilted plants has been made, though it is proposed to do so shortly. 
Fusarium beds and microconidial colonies are common on such 
places, but they have not been separated and their pa ra sitism tested. 
Some at least belong to other species. The control pots growing 
in the neighbourhood of infected pots have in several cases ultimate- 
ly developed the disease. This has not occurred until the plants 
were a good size, from three to five months alter sowing. The 
control plot of Series XVI remained free from wilt for eight months, 
though only separated from the inoculated plot by a narrow trench 
a foot deep. Aerial infection is not likely to be responsible for 
much loss since it occurs rather late in the season, and though 
isolated plants may be killed before harvest, centrifugal spread in 
