20 Journal of A gricultural Research \o\. xxii. No. i 



inverted pot was placed about 3 inches below the surface instead of at 

 the bottom in the hope of maintaining a more uniform and rapid dis- 

 tribution of moisture. This appears to have been an unfortunate 

 change in method, since at the higher temperatures roots tended to col- 

 lect around these pots, where they apparently developed more ex- 

 tensively and produced more nodules than they had in the previous 

 series. Inasmuch as no accumulation of roots took place at lower tem- 

 peratures, it is not easy to explain this fact. 



Attention should here be drawn to the fact that two distinctly dif- 

 ferent methods of securing data have been used. In the preliminary ex- 

 periments, the plants were first grown in the cans at ordinary greenhouse 

 temperature for about two weeks before inoculation was made, with 

 the suitable strain of Bacillus radicicola Beyr., by pouring a water sus- 

 pension of the organism around the base of the plants, and the cans were 

 placed in the tanks adjusted at the predetermined temperatures. In the 

 later series the plants were grown from seed in inoculated soil held at the 

 required temperatures from the beginning. Several reasons led to the 

 change of method. In the first place, one could not be certain that the 

 bacteria poured around the plant in the first instance would become 

 rapidly distributed through the soil at all temperatures. This inequality 

 in rate of distribution might affect the number of infections and hence the 

 number of nodules formed. At least it might tend to limit the formation 

 of nodules to the roots near the surface of the ground where temperature 

 is less exactly controlled. A second objection to this method appeared 

 when the marked effect of temperature upon the morphology of the roots 

 themselves was observed. The number of root hairs through which 

 infection has been found to take place is much greater at lower tempera- 

 tures than at higher. In view of the possible effect of this difference 

 it appeared preferable to grow the plants from the beginning in inoculated 

 soil at the designated temperatures, even though the plants thus produced 

 would necessarily vary considerably in size. Data obtained by each of 

 these methods will be presented. 



MEASUREMENT OF EFFECT OF SOIL TEMPERATURE UPON NODULE 



FORMATION 



When the experiments were begim it was assumed that different tem- 

 peratures, if they are at all potent, would produce such a marked effect 

 upon the number of nodules that count alone would give a significant 

 expression of results. This expectation was fostered by the fact that 

 Wilson 1 and nearly all previous investigators have used numbers to 

 express similar experimental results. It will be seen from data given 

 later that this hope was early disappointed. Different temperatures 

 usually seemed to affect number not nearly so much as rate of develop- 



» Wilson, J. K. op cit. 



