402 
Journal of Agricultural Research 
Vol. XXIII, No. 6 
They are inclined to place it therefore among the Actinomycetes, but 
a comparison of their findings with those discussed in Part I of our 
Studies (25) leaves no doubt that Bacterium influenzae f too, follows 
exactly the same general lines which are recognizable in the life history 
of all bacteria. The general occurrence of filterable gonidia is further 
indicated by observations made by Heymans (13) with regard to anthrax 
and tuberculosis. It was found that the bacteria pass the filter in an 
ultramicroscopic form. Hort (14) published additional details concern¬ 
ing the various modes of bacterial reproduction; Howe and Hatch (13) 
confirmed Noguchi's finding that B. bifidus can grow as an aerobic, 
sporulating bacillus; and Orla-Jensen {38) pointed out that certain 
aerobic plectridia can be easily transformed so that they are no more 
distinguishable from the typical Proteus bacteria. Studies upon the 
conjunction of bacterial cells were made by Potthoff (59) with Chroma- 
tium and sulfur spirilla, and by Enderlein (8) with Vibrio cholerae. The 
results obtained are in good agreement with earlier findings (25, p. J95, 
202) . The same holds true in regard to more recent observations made 
by Lutz (30a) upon the different types of globoid bodies produced by 
B. anthracis. Almquist {2d) and Mellon (33a) have again emphasized 
that a thorough knowledge of the various developmental phases in the 
life histories of pathogenic bacteria is essentia! for gaining a correct view 
of epidemiological problems. 
Our observation upon the life cycle of Azotobacter, as recorded in 
our preliminary paper ( 28 ), were partially confirmed and partially con¬ 
tested by D. H. Jones {16). Confirmed were the data on pleomorphism 
and symplastic stage, contested those on conjunction and on the ability 
of Azotobacter to assume endospore formation. The discussion of these 
differences of opinion will be taken up at the end of this report after our 
experimental results, obtained since 1916, have been presented. 
METHODS OF INVESTIGATION 
To obtain complete and accurate information upon the life cycles of 
the bacteria is no more difficult but much more time-consuming than to 
make an ordinary bacteriological diagnosis. Single-cell cultures and 
continuous microscopic observation of the living organisms are by no 
means so absolutely indispensable as is sometimes asserted (25, p . 59, 
203) . The generally used methods of isolating, cultivating, and studying 
the bacteria are, as a rule, quite sufficient to collect complete information, 
provided they are applied judiciously and the investigator himself is not*" 
too preoccupied by the wide-spread prejudices concerning “normal” and 
“abnormal” growth, “involution forms,” and “contamination”. 
A sufficient number of parallel tests, the frequently repeated microscopic 
control of the cultures held for a sufficient length of time (not less than 
a month, preferably longer), and the regular repetition of all experiments 
are three points of major importance. Accordingly, apparent changes 
of growth can be correctly accepted as such, only if they have been as¬ 
certained repeatedly in parallel tests and if they have been closely followed 
under the microscope. Furthermore, an apparently new type of growth 
must always be tested in regard to its inclination to return to its origin, 
or to pass over into another well-known developmental stage of that 
particular organism. Sometimes the patience and persistence of the 
investigator is taxed very much by such experiments; it may take years 
before positive results are secured. A few examples in this respect were 
