THE WILT DISEASE OF PIGEON-PEA. 7 
similar and quite agree with those described above for pigeon-pea 
wilt. Within the tissues were found small unicellular conidia 
(microconidial or Cephalosporium stage), and on the surface larger 
falcate spores with several septa (macroconidial or Fusarium stage). 
Further, bright red perithecia occurred frequently on the roots of 
wilted plants and were proved by single-spore culture to give rise to 
micro-and macroconidia indistinguishable from those found in 
nature on diseased roots, and finally, in old cultures, to chlamydo- 
spores similar to those found accompanying the microconidia on the 
bark. The perithecia differ from those of Nectria in being found 
usually on the roots (whereas the pigeon-pea Nectria occurs on the 
stem, never on the roots) and in containing characteristic unicellu- 
lar spores with a thick wrinkled brown wall. For this hitherto 
undescribed fungus Smith established the genus Neocosmospora, 
with one species N. vasinfecta (Syn. Fusarium vasinfectum, Atk.) 
on cotton and varieties tracheiphila and nivea on cow-pea and 
water-melon respectively. 
The biology of the fungus was very fully investigated by Smith, 
particularly its behaviour as a pigment-producer, in which it shows 
certain definite characters. His experiments regarding its powers 
as a parasite led to less definite results, or rather showed that 
there were points in its life as a parasite that required elucidation. 
The chief of these were the following. 
Though the fungus on cotton and cow-pea resembled that on 
melons in cultural characters and anatomical details and the 
diseases on all three were similar, he entirely failed to produce wilt 
in the first two cases by inoculating healthy plants with cultures 
of the fungus associated with the disease. In the water-melon on 
the other hand disease was easily induced by pure cultures of the 
fungus obtained from melon roots. All attempts at cross inocula- 
tions, using the fungus obtained from one host to inoculate the 
other two, failed. Hence, as stated on p. 46 of his paper, the 
parasitism of the variety on water-melon was fully established, 
while that of those on cotton and cow-pea was ‘‘not proved.’’ 
