66 BOTANICAL GAZETTE [JULY 
found in the infecting threads which are found in the apical areas of 
tubercles and in the cells just within the phellogenic layers. Additions 
to the growth of the tubercles are also made at these points. The 
starch, which was deposited just outside of these meristematic areas in 
the fall of the year, is now again assimilated by the host plant. 
It is evident that perhaps about one-half, or somewhat less, of the 
thizobia existing within the infected areas of tubercles are killed 
during the unfavorable winter conditions. Freezing alone does 
not kill them; it is rather a combination of conditions, the lack of 
food supply perhaps being the most important. The tubercles and 
soil examined were taken from near the surface of the frozen ground 
when the temperature was from —10° to — 20°F. Cultures were 
made from the tubercles as well as from the soil by the usual plate isola- 
tion methods. The growths showed the presence of rhizobia and other 
soil bacteria. Streak and stab cultures were also made from the infected 
area of tubercles. “A careful examination of growths and culture 
media at the point of inoculation showed a number of impoverished 
tubercle organisms which had evidently lost the power of dividing. 
. These stained very feebly and the cell-wall was partially destroyed, 
having a roughened perforated appearance. Soil cultures showed the 
presence of rhizobia, besides numerous soil bacteria. 
Examining tubercles which were more deeply situated, about one 
foot below the surface of the soil, showed that the destruction of rhi- 
zobia had been less and the number of dead but not destroyed rhizobia 
was also less, which would seem to indicate that cold was also a factor 
to be considered in the killing of rhizobia. It seems probable that 
rhizobia of tubercles below the freezing depth develcp and multiply to 
some extent, though the tubercles do not increase in size, as is indi- 
cated by the tensely filled glistening appearance of such tubercles. A 
careful examination of a number of such tense, brittle tubercles showed 
that they contained numerous rhizobia imbedded in a large amount of 
a mucilaginous substance. The cells of the infected area were loosely 
united and almost spherical in form. The contents of these tubercles 
require further study. 
The observations were made chiefly upon tubercles of Meltlotus 
alba, Trifolium pratense, and T. repens, and the conclusions with refer- 
ence to these plants are that root tubercles are mostly biennial, the 
tubercles attaining their full growth during the first year and gradually 
dying and decaying toward the close of the second year.*. With the 
* New tubercles are, of course, added each season along with the development of 
new rootlets. 
