426 
Journal of Agricultural Research 
Vol. XXIII, No .6 
bacteria, as created by Ferdinand Cohn and his successors, have become 
invalid because, in accordance with earlier observations discussed in 
Part I (25), our experiments have shown that Azotobacter, like certain 
other natural groups of bacteria, is able to grow in all forms which were 
accepted as a basis for establishing the so-called genera Micrococcus, 
Bacterium, Pseudomonas, Bacillus, and Mycobacterium. Contrary to a 
very widespread, although quite illogical, assumption the pleomorphism 
of Azotobacter, as of all other bacteria, does not obscure but clarifies and 
strengthens the character of the genus as well as of its species. While a 
study of the large nonsporulating cells alone made it probable that 
Azotobacter Beijerinckii J. G. Lipman might be a variety of A . chroococcum 
Beij., this conclusion has been made certain by the discovery that all 
other types of growth are identical in both cases. A . agile , on the other 
hand, exhibits in its large nonsporulating cell form certain features which 
might be accepted as proof that it should be classed as another variety 
of A. chroococcum , as was done, for example, by Prazmowski (40); yet its 
small nonsporulating rod form and its fungoid growth are so different 
that no doubt remains about its being a true species. That A . Vinelandii 
J. G. Lipman is identical with A. agile Beij. could be said on account of the 
identity of their large nonsporulating cell forms (30) , and all other 
developmental stages have now shown themselves equally identical. 
A. vitreum Lohnis, in its large nonsporulating cell type apparently quite 
different from the other species, has displayed in all other directions so 
much similarity with A. agile that its being a variety of this species is 
practically certain, although this conclusion had to be drawn from the 
results obtained with the only strain available. 
Three other Azotobacter species have been described as Azotobacter 
Woodstownii J. G. Lipman (22) , A. Hilgardi C. B. Lipman (79), and 
A. Smyrnii C. B. Lipman et Burgess (20) ; but they all should be canceled. 
The first one was never shown to be able to fix nitrogen, and its charac¬ 
terization is not distinct enough. The second one was only very 
incompletely described. According to information kindly furnished by 
its author it was “very similar” to A. Smyrnii. This like the other 
is no longer alive, but its illustrated description {20) leaves no doubt 
that it was identical with the sporulating large cell form of A. chroococcum. 
It is true that the original description says motility and spores were 
lacking; but the characterization of what was accepted as vegetative 
cells makes it certain that only spores were seen and photographed. 
Every mark ascribed to A. Smyrnii tallies exactly with those of quickly 
sporulating large cells of A. chroococcum (Bacillus petasites). 
The ability of Azotobacter to produce in certain stages of growth genu¬ 
ine heat-resistant endospores was accepted by the senior author (26) as 
proof that it should be classed as Bacillus azotobacter among the sporu¬ 
lating bacilli. But after our more recent studies have shown conclusively 
that all old form genera, including that of Bacillus F. Cohn emend. 
Hueppe, will have to be replaced gradually by natural genera, based on 
complete investigation of their life histories, it appeared preferable to 
retain and to emend the genus Azotobacter Beij. by adding to the large 
nonsporulating type of growth the six other growth types described on 
the preceding pages. 
Contrary to our observations upon endosporulation in Azotobacter 
cells it was asserted by D. H. Jones (id), as well as by the Committee of the 
Society of American Bacteriologists on Characterization and Classifi- 
