Feb. io, 1923 Life History of Azotobacter 
granulated slime threads of the dwarfed cell type shown in figures 28 and 
29 on Plate 3, the development proceeds over the steps illustrated by 
figures 37 to 39 on Plate 4 further to a very stable type of irregular, 
fungoid growth, which ultimately acquires all the qualities character¬ 
istic of a Mycobacterium. Wedge-shaped cells, budding and branching, 
club formation, etc. are all noticeable, and the same holds true concern¬ 
ing the cultural features of this “genus”. The clubs again are no “in¬ 
volution forms,” but steps in the development of rods with terminal 
spores (fig. 45). Dwarfed cells, coccoid forms, rods of various size, as 
well as large non-sporulating globules and ovals were all evolved from this 
type of growth, as illustrated by figures 40 to 48 (PI. 4). Cultivation 
on potato was found useful for stimulating the tendency to produce 
clubs and rudimentary endospores. Potato agar favored again the 
development of coccoid forms. In water, broth, and milk branching 
was very pronounced, but continued cultivation in milk established a 
growth of regular spore-free, slime-producing rods. Alkaline mannite- 
nitrate solution and agar (P H 7.5) proved helpful for reestablishing the 
typical Azotobacter growth. It is especially noteworthy that this was 
secured on these substrates with one of our cultures of Azotobacter 
Beijerinckii (No. 15) which had grown for six years in 145 transfers 
on practically every substrate as a highly pleomorphous mycobacterium. 
It had been plated repeatedly, strains representing the dwarfed growth, 
coccoid growth, sporulating and non-sporulating rods had been branched 
off, until ultimately after a passage on potato agar (fig. 44) the restora¬ 
tion of the original Azotobacter type was effected (fig. 47). 
The typical form of the small spore-free rods of Azotobacter chroococcum 
and A. Beijerinckii is shown in figures 49 and 50 on Plate 5, while those 
of A. agile and A. vitreum are usually of more slender shape, similar to 
that type of growth of A. chroococcum visible in figure 51; fairly typical 
rods of A. agile are also to be found in the left upper comer of figure 60 
on Plate 5. All these rods are Gram-negative, those of A. chroococcum 
and A. Beijerinckii immotile, those of A. agile and A. vitreum motile by 
usually three polar flagella. It should be noted that to start the 
cultures reproduced in figures 49 and 50 the same inoculation material 
was used; the difference in the substrates was the only cause of the very 
marked difference in appearance. A comparison of figure 49 with figure 
9 and figure 24 will demonstrate the gradual changes from the large to 
the small non-sporulating cells, while the small coccobacilli in figure 50 
should be compared on the one side with the dwarfed growth shown in 
figures 25 and 26 on Plate 3 and on the other hand with the short rod 
forms which always accompany the mycobacterium type, as demonstrated 
in figures 40 to 45 on Plate 4. The larger rod form producing a short 
branch, which is visible in the lower right corner of figure 50, is also of 
interest in this connection. A direct transformation from the original 
cultures of A. chroococcum and A. Beijerinckii to small non-sporulating 
rods was recorded only three times, while they were more frequently 
derived from the fungoid growth (nine cases observed). A. agile and 
A. vitreum behaved differently; several changes from the large cells to 
small rods, but no direct transformation from the fungoid growth could 
be ascertained. Figures 51 to 53 on Plate 5 demonstrate the trans 
formation from the short non-sporulating rods to coccoid and fun 
goid growths, the latter identical with that of figures 43 and 45 on’ 
Plate 4. That sporulating bacilli which have lost their ability to 
