34 



this was not necessary. The soil was from Washington where this dis- 

 ease does not occur, and melon ])lants grown in quantity in the same 

 hothouses, in the sauie soils, and watered with the same water, subject, 

 in a word, to the same conditions except as to inoculation, did not con- 

 tract the disease. These experiments were in i^rogress parts of two 

 seasons. In many pots of the inoculated soil every plant contracted 

 the disease within a few weeks (PI. IX, 1). With two exceptions', not 

 one case of the disease appeared in the uninoculated soils, although 

 several hundred check plants were grown in the latter. 



During the course of these inoculation experiments, which were per- 

 formed in 1894 and 1895, it was found to be as easy to produce the 

 melon wilt with mycelium derived from the external couidia (PI. X) 

 as with that derived from the internal couidia (PL YIII). The usual 

 method of infection was by putting a little fragment of a pure culture 

 of the fungus into the soil at some distance from the plant and at no 

 great depth, so as to avoid breaking the roots. The wilt usually 

 appeared suddenly in three to six weeks, whichever spore form was 

 used for infection. The internal fungus in the vessels of the plant was 

 the only one to be found at the time of the wilt, but later when the 

 plant was dead the external couidia beds often appeared, and always 

 if the air was not too dry. These infection experiments were repeated 

 several times on a large scale with uniform results. Those with the 

 two sorts of conidia were conducted in different greenhouses, with 

 very great care, and, exclusive of the 4 cases already mentioned as 

 occurring late in the course of two of the experiments, the numerous 

 control plants in all cases remained healthy. 



NUMBER OF SUCCESSFUL INFECTIONS. 



Altogether more than 500 successful infections have been obtained 

 with watermelon plants by simply inoculating the soil with the melon 

 fungus, the above-mentioned 175 cases not including those obtained in 

 189G and 1897 and described in the following section. All varieties 

 of the watermelon appear to be susceptible. No attempts have been 

 made to induce the disease by inserting the fungus directly into wounds, 

 except a few futile ones made in the fields in South Carolina, when 

 the disease was first discovered. 



1(1) One case whicli appeared late iu the course of the experiment (internal conidia, 

 1894) and the only plant out of many checks to contract the disease, although all 

 were on the same bench and cockroaches ran about in the house. 



(2) Three cases whicli appeared late in the course of the experiment (external 

 conidia, 1895). These were iu 2 pots only out of many checks which had been 

 standing on the experimental bench for several months within 20 inches of numer- 

 ous pots of inoculated soil, iu each one of whicli several plants had wilted. Under 

 the circumstances there can be no doubt that these were accidental infectious 

 derived from the inoculated soil or the wilted plants. 



