408 
Journal of Agricultural Research voi. xxin, no. 6 
produce regular endospores may display a very similar morphology may 
be derived from a comparative examination of figure 54 with figures 51 
and 53. Bacillus pumilus , shown in figure 54, is identical with the sporu- 
lating rod form of A. chroococcum , as will be discussed presently. The 
return from the small rods to large globular cells is illustrated by figures 
55 to 57; figure 58 pictures the transformation into the typical fungoid 
growth, similar to that of figures 37 and 38 on Plate 4, and in figures 59 
and 60 the simultaneous upgrowth of small and large rods from the 
symplasm is demonstrated, figure 59 making a counterpart to figures 34 
and 35 on Plate 3. 
The small sporulating rods are of special interest because, when the 
endosporulation first becomes noticeable in an Azotobacter culture, 
nearly always these small rods, not the large sporulating cells, are the 
first ones to appear. Referring to what was said in Part I (25, p. 137,139) 
about the transformation of terminal regenerative bodies into polar 
exospores and endospores, it is evident that this observation is in full 
agreement with all earlier findings. On the other hand, it must be 
pointed out that these rather fragile, Gram-negative, slender rods with 
large polar spores are by no means a frequent occurrence in our labora¬ 
tories. To discard them lightly as “contaminations” would be entirely 
unjustified. Identical strains were isolated from various Azotobacter 
cultures in Leipzig, Washington, D. C., Urbana, Ill.; and a culture 
recently received from Vienna gave again the same small sporulating 
rod. But never in our experience, and especially not in any one of 
approximately 5,000 transfers made in the course of these experiments, 
were such sporulating small rods found as contaminations. On account 
of their weak growth we even lost two of them by replating; but in view 
of this experience it is especially interesting that in old Azotobacter 
cultures kept undisturbed for years on mannite-nitrate agar or in man- 
nite-nitrate solution just these small sporulating rods were the only 
survivors. Figure 62 on Plate 6 shows such a strain, which should be 
compared with the typical picture reproduced as figure 61. Figures 
63, 71, and 72, on the other hand, illustrate the tendency to pass over 
into large sporulating rods. That transformations from small to large 
sporulating cells were seen occasionally in our preliminary experiments 
was mentioned before ( 28 , p. 681 ). But as the first appearance of these 
small sporulating rods caused considerable surprise in 1912 ( 26 ), so it 
was again the case when one of these cultures (No. 6), after having dis¬ 
played a very constant behavior, ultimately in the spring of 1921, when 
cultivated on alkaline mannite-nitrate agar, reverted to the large-cell 
type and assumed gradually all the characters of the original strain 
(No. 3), from which it had been branched off nine years before. A 
tendency to change to the fungoid growth is noticeable with some of the 
cells in figures 64, Figures 65 to 70 illustrate certain reproductive 
phases to be discussed on the following pages. 
The large sporulating cells are shown in figures 73, 74, 76, and 77 on 
Plate 7 in what may be accepted as their typical appearance. They are 
motile by peritrichous flagella and are Gram-positive. The tendency of 
the threads, very common in the cultures of this type, to make short, 
oval, Gram-negative cells is very pronounced. Figure 78 illustrates the 
analogous behavior of a thread in a spore-free Azotobacter culture. In 
figure 75 many branched rods and threads are visible, representing the 
fungoid type of the large rod form, which, however, could not be sta¬ 
bilized, although with Bacillus anthracis , B. subtilis , and B. mycoides 
