4 io 
Journal of Agricultural Research voi. xxm, no. 6 
regeneration of larger, in this case of coccoid cells. (See also figure 5 on 
Plate A of our preliminary paper, (28) where both processes became 
visible simultaneously, and figures 3 and 4 on Plate 1 of the present paper.) 
In figure 77 on Plate 7 the two granulated threads in the lower left part 
of the field deserve special attention, because the regular arrangement 
of the unstained gonidia contained therein is very characteristic of 
Azotobacter, as is the stainability of the same bodies as soon as they 
have left their mother cell. (See especially the dark bodies budding from 
several threads in figure 77 and analogous occurrences photographed 
in figure 60 on Plate 5.) Other types of budding, as well as the transforma¬ 
tion of regenerative bodies into larger cells, are pictured in figures 
13 to 16, 22 to 24 on Plate 2. Figure 65 on Plate 6 demonstrates the 
formation of gonidia in the small sporulating cells, and also their develop¬ 
ment to regenerative bodies and to endospores. 
Typical gonidangia of Azotobacter may be seen in figures 5 and 11 
of Plate 1 and in figures 88 to 92 on Plate 8, those of the sporulating cell 
type in figures 79, 81, and 82 on Plate 7; in the latter case the swollen 
ovals are replacing the sporulating cells, while in the two former instances 
the globular, branched, or spindle-shaped cells are acting as true sporan¬ 
gia. If small cells swell up to gonidangia they look much like regular 
Azotobacter cells (fig. 20 on PI. 2 and 57, on PI. 5), and as the latter 
are always able to produce numerous gonidia, it seems as if the true 
physiological reason for the occurrence of this developmental stage 
lies in the ability of these cells to act as gonidangia. The peculiarity 
of all bacterial gonidangia to assume either globular, oval, club-shaped, 
or thread-like forms is also characteristic of the large spore-free cells 
of Azotobacter. That they frequently multiply as such without acting 
as reproductive organs places them parallel to the dwarfed growth (of 
the gonidia) and to the coccoid growth (of the regenerative bodies). 
Instead of gonidia or endospores new cells of normal size may be directly 
formed within these large receptacles. Figures 6 on Plate A and 21 
on Plate D of our preliminary paper (28) illustrated such cases; similar 
preparates are shown in figures 90 to 92 and 94 on Plate 8 of the present 
paper. 
The fungoid type of growth was shown to be the result of a consolida¬ 
tion of slime threads containing numerous gonidial bodies. Accordingly 
it is easy to understand why in this case the formation of arthrosoores 
is very common. The gonidia simply increase somewhat in size at the 
cost of the cell and surround themselves separately with cell walls. 
Occasionally, but much more rarely, normal rods or threads act in an 
analogous manner. Figures 51, 54, and 55 on Plate 5, and figure 66 
on Plate 6 illustrate the arthrospore formation of the small non-sporulat- 
ing and sporulating growth types. An upgrowth of other globular cells 
from the encapsulated symplasm also visible in figure 66 will be discussed 
later. The fragmentation of the large threads shown in figure 78 on 
Plate 7 could also be accepted as another instance of arthrospore forma¬ 
tion, but as the globular cells produced in this case are able to multiply 
as such, this occurrence is to be considered more similar to that pictured 
in figure 21 on Plate D of our preliminary paper (28) and mentioned 
above. 
Terminal swellings—that is, regenerative bodies in a polar position, 
but still inclosed in the mother cell—are the first step toward the forma¬ 
tion of endospores, as was fully discussed in Part I (25) and frequently con¬ 
firmed in the course of these investigations. Figures 45 and 46 on Plate 4, 
