July 3 i, 1916 Life Cycles of the Bacteria 695 



motile. These filtrates were then transferred to beef agar, beef broth, 

 milk and blood serum. After incubating, the macroscopical appearance 

 was the same in all cases. On the agar slope, especially on its lower 

 moist part, a very scant, thin, slimy growth, somewhat resembling a very 

 thin layer of small droplets of dew, became visible. A stained preparate 

 from a 4-day-old slant clearly showed the gonidia germinating to minute 

 rods (fig. 41 of PI. G). The other substrates also gave no conspicuous 

 growth. 



The dark field proved more efficient in observing these almost invisible 

 forms. Figure 42 (Pi. G), made from the same 4-day-old slant as figure 41 , 

 shows clearly that the filterable gonidia also form a symplasm in the same 

 manner as the larger ones, which, in its turn, produces new small cells. 

 In order to bring out more definitely the structure of this symplasm, it 

 was necessary to make a very dark print, thereby obliterating several free 

 granules which were also visible in the field. 



These facts are in good agreement with the observations of various 

 authors concerning filterable vira. In one of the latest publications along 

 these lines Healy and Gott (9) described the filterable organism causing 

 hog cholera as small globules or rods (0.2 to 0.3/*) when growing on 

 ordinary media appearing in slimy clumps which are either well stained 

 with aqueous dyes or are not stained at all. 



As a regeneration of larger forms could in r.o case be observed on the 

 subtrates mentioned , we also transferred small quantities of filtrates of 

 cultures 31, 35, and 40 into ammonium-citrate solution. Here a quick 

 regeneration took place. After two days some sediment was already 

 formed, which after shaking caused a distinct turbidity of the solution. 

 Under the microscope in the stained preparate many pale, stained, small 

 granules and minute rods were visible, as before, and also larger dark 

 stained oval forms 0.5 to iju broad, 0.75 to 1.5/1 long. These forms still 

 differ considerably in their appearance from the normal rods of B. sub- 

 tilis or Bad. fluorescens. They may be classified as regenerative bodies. 

 That they will turn back entirely to the normal large vegetative cells is 

 not doubtful, but this still remains to be tested experimentally. So far 

 our tests have been repeated three times with Bad. fluorescens and twice 

 with B. subtilis. The results were identical. For the filtration we used 

 three different filters, which were controlled in each case by obtaining 

 sterile filtrates from i-day-old cultures of B. subtilis or B. azotobader 1. 



In carrying out experiments like these, however, another possibility 

 of- obtaining erroneous results must be kept in mind. Not only must 

 the media be carefully prepared and sterilized but all glassware must be 

 thoroughly treated with some cleaning fluid such as chromic acid, which 

 destroys entirely all bacterial forms. The fact that mere sterilization 

 is not sufficient is shown by the following test : 



Particles of symplasm containing many regenerative bodies were car- 

 ried from a mannite-nitrate solution to a similar medium and heated 



