13 



It will be seen that some of the best kinds are among those most 

 injured by this disease; but there were one or more plants in each 

 race that entirely withstood the disease, and the seed from these has 

 been saved with the intention of securing valuable resistant strains 

 by cross-breeding them. 



The ability of certain cotton plants to grow on infected land is due 

 to the fact that the wilt fungus is unable to enter their principal root 

 system and not to any lack of infection. This has been determined 

 by microscopic examination. That infection of these plants has 

 really taken place may be demonstrated by an examination of their 

 roots for the little tufts of rootlets which mark the location of infec- 

 tions (see p. 8). The roots of plants taken from the row of Janno-. ' 

 vitch cotton shown in PI. II, fig. 1, were attacked by the fungus in 

 over a hundred places, as found by actual count, yet in no case did 

 the parasite penetrate as far as the main stem, while plants of King 

 cotton in the adjoining row were completely overcome. A part of 

 tlie root system of one of these resistant plants is represented in PI. 

 Ill, fig. 1. As determined by numerous microscopic examinations 

 each little tuft of roots marks a point attacked by the fungus, so that 

 there can be no doubt of the thoroughness of the infection and, fur- 

 thermore, no doubt that such plants are actually resistant to the 

 fungus. 



It is evident that such an effect as the fungus has produced here 

 must injure the plant considerably and -this was found to be the case. 

 The average height of plants grown on the infected land was 23 inches, 

 while plants on adjoining land very slightly infected grew 42 inches 

 high. Plate III, fig. 2, shows the difference between these plants. 

 Such injury as this would of course greatly shorten the crop, but the 

 indications are that seed selected from the most vigorous plants will 

 be more resistant than the average. The best plants in our experi- 

 mental plots on the infected land were nearly equal to those grown 

 on healthy land and also showed a smaller number of root tufts. 



In this connection the most important question is whether this 

 quality of resistance to disease is transmissible through the seed to 

 succeeding generations. An experiment designed to settle this point 

 proved a remarkable success. It was carried out by Mr. Elias L. 

 Rivers, of James Island, S. C, who selected a healthy plant of sea 

 island cotton that gi-ew in a badly blighted field in 1800. The seed 

 from this i-esistant plant was saved and planted in a single row through 

 a field, that luul been infected with the wilt disease for several years. 

 The adjoining i-ows were planted with seed from his main crop, grown 

 on noninfected land. The result is indicated in the photograph (PI. 

 IV) taken September, 1000. The wilt disease made almost a clean 

 sweep through the oi'dinary cotton, 95 per cent of the plants l)eing 

 killed, Avhile in the row planted with seed from the resistant plant not 

 a single plant teas killed by IJie wilt disease. 



These plants were vigorous and productive. The dwarfing noted in 



