320 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 



hence must be by the products and not by the living sub- 

 stance of the meristem and its daughter-cells. The direction 

 of growth of these infection threads cannot be determined 

 by the oxygen or nitrogen (or both) of the air, for if this 

 were the case, we should find strands of bacteria running 

 from the central cells in all directions toward the periphery 

 of the tubercle. This is not the case. The strands run 

 toward the growing-point of the tubercle. In consequence, 

 the daughter-cells successively formed by the repeated divi- 

 sion of the cells of the meristem become infected. 



Not only do the infection threads run definitely toward the 

 growing-point of the tubercle ; they also grow toward the 

 nucleus of each cell which they enter. This statement has 

 been repeatedly made and denied in papers on the subject 

 of root-tubercles. In hand sections, especially if the mate- 

 rial were not carefully fixed and differentially stained, it 

 would be easy to find evidence in support of the affirmation 

 and of the denial. Microtome sections, differentially stained 

 as before described, of carefully fixed growing tubercles of 

 the species of leguminous plants which I have especially 

 studied, show that in most cases the infection threads run 

 definitely toward the nuclei of the tubercle cells. This is 

 evident in fig. 10. Figures 11 and 12 also show this. In 

 fig. 13 are shown two tubercle cells in which the main in- 

 fection thread is not directed toward the nuclei, but the 

 lower of these two cells shows that the infection thread 

 bends toward the nucleus. In the next section of the series 

 (not figured) a branch runs from the main infection thread 

 to the nucleus. In the cell shown in fig. 11 the infection 

 thread is divided, one part running beneath, the other 

 above the nucleus. For the sake of clearness this upper 

 part of the thread was omitted in drawing. The nucleus 

 of the upper cell in fig. 13 was not in the plane of the sec- 

 tion. An adjacent section (not figured) contains this and 

 has a branch of the main infection thread running to it. 

 There must be some reason for this definite growth of the 

 strands of bacteria toward the nuclei of the cells which 



