CONTRIBUTIONS PROM THE WISLEY LABORATORY. 495 



uninoculated, but even here there are three out of the six rows giving 

 a smaller average yield from the inoculated seeds. In comparing the 

 total yields from the different plots it was found that four of the plots 

 gave a higher yield from the uninoculated seed (the last four). Here 

 again, therefore, the two methods of estimation to a great extent support 

 one another. 



In the twenty-four pairs of rows seven only gave a greater average 

 weight of pods from each plant in the inoculated rows than in the un- 

 inoculated, in one case the average was the same in the two rows and in 

 sixteen a diminished average yield was shown. 



Either inoculation had no effect whatever, or it had an adverse effect 

 in a majority of cases. If the former suggestion be true, then the yield 

 from comparable plots must vary normally between very wide limits ; if 

 the latter, we must account in some way for the adverse effect. 



It cannot justly be concluded from the results of this experiment 

 that the inoculation actually caused a diminution in the yield, for in the 

 case of such exact comparisons as would enable this question to be 

 definitely answered the space occupied by a given number of plants must 

 be the same in the two rows (albeit, the plants in this experiment all had 

 ample room for development), and other precautions would have to be 

 taken which are practically impossible in a field or garden trial. The 

 results detailed here, however, as well as those obtained by the method 

 of calculation adopted in the former portion of this Report (the method 

 by which comparison would be made in an ordinary garden trial), suggest 

 the possibility that certain races of the bacilli which produce the nodules 

 on the roots of peas are capable of fixing a greater amount of nitrogen 

 than are other races, and that a race more powerful in this direction 

 than that in " Nitro-Bacterine " may be actually present in the soil 

 already. 



The roots of all the plants, as previously pointed out, were well 

 supplied with nodules, and apparently the experiment resolved itself into 

 a competition between the races of bacteria already in the soil and those 

 added in the inoculation of the seed. The much more frequent associa- 

 tion of the native bacilli with the larger yield makes the question of the 

 relative virulence of the races a very pertinent one, but it can only be 

 settled after further investigation ; as suggested in the introduction to 

 the Report (p. 235), it may be a question of the gravest importance. 



Another hypothetical suggestion may be offered to account for the 

 apparent adverse influence of inoculation. Is it certain that very early 

 inoculation is as beneficial to the host plant as later inoculation ? 

 Possibly, with a race of bacteria which have recently been in active 

 growth, the power of rapid infection may be greater than that of those 

 which have lain dormant in the soil for some time, and nodules might 

 be formed earlier. We have no data showing that such earlier nodule 

 formation occurs, however ; but it is conceivable that if it were so, since 

 it is probable that the bacteria are for a time at least parasitic, harm 

 might be done at a critical period of the plant's existence. Or, again, the 

 bacteria might be manufacturing and handing on to the young plant 

 materials which it could not, at that stage of its existence, profitably 

 make use of. 



