211 
found the cells of the infected area entirely filled with curved 
Rhizobia, resembling the comma bacillus of Koch. — Staining 
with eosin or fuchsine brought into view a spore at either end. 
_ Their form was so constant and even that I could not bring 
myself to imagine that they were identical with the R/dzodia of 
Melilotus alba, for example. Furthermore they retained their con- 
stant form throughout the entire year. The Rhizobia of Phaseolus 
vulgaris were still different—they were straight, in form constant, 
in size quite variable. There were spherical spores, one at either 
end, rarely only one. The spores of the curved Rhizobia were 
hemispherical. The Rhizobia of Pisum sativum corresponded 
exactly with those of Phaseolus vulgaris, only that they were 
smaller. Dalea alopecuroides, Robinia pseudacacia and Cassia 
Chamecrista contained a Rhizobium differing from any of the 
ones described above. It presented a knotted, irregular outline; 
generally three spores, sometimes one, rarely two. The RAzzobia 
of Amphicarpea comosa differed from the one just described in 
that the spores were placed some distance from the ends. They 
were also somewhat smaller, as will be seen from the table of 
Measurements. From these observations I felt satisfied that there 
Was more than one species of Rhizobium, and that they did not 
all take part in that change of form so well marked in the first 
described species. The question might be raised whether the - 
Protoplasmic contents of the RAzzobia are in all cases reabsorbed 
by the plant or whether they, at least in some cases, simply die 
and disintegrate for want of food supply. It is known that in all 
cases the tubercles are destroyed towards the close of vegetation, 
thus liberating the remaining Rhzzobza. 
As already stated, Schroeter took the Infektionsfaden to be 
true plasmodia of his Phytomyxa leguminosarum,=¥ rank's Rhi- 
Z0bium. Now the Infektionsfiden are very prominent in Trifo- 
‘ium pratense but its Rhizobia contain no spore whatever so far 
as I have been able to make out, so the Infektionsfaden cannot 
be the plasmodium of the Rhizobium of white clover. Further- 
more the Infektionsfaden are absent in many species infected with 
the Rhisobia or Phytomyxa. So instead of placing the Rhizobia 
with Myxomycetes, according to Schroeter, it is better to — 
them with the Schizomycetes, according to Frank; because their 
