682 Journal of Agricultural Research vol. vi. No. is 



regular spores. This possibility will be discussed at once. But, as men- 

 tioned before, regenerative bodies may also produce forms belonging to 

 types E and F, making two of the "outside" connections between the 

 upper and the lower circles as given in text figure i. 



Type J characterizes another rather rare occurrence also studied by 

 Prazmowski. Forms belonging to type A, B, or occasionally L, increase 

 in size and inside themselves develop one or more new full-sized cells of 

 type A or B (fig. 6 of Pi. A and fig. 21 of PI. D). These new cells are the 

 result of the growth of the gonidia. 



Type K comprises all those cells of type A, B, E, F, G, I, or L which 

 produce one or more well-stained, round, oval, or rodlike buds, which, in 

 the case of the large forms, occasionally cause a close resemblance to 

 some budding yeast cell. These "buds" are gonidia developing into 

 regenerative bodies, seldom directly into full-sized forms. An added 

 a, /S, e, <p, 7, 1, or X indicates the relation to type A, B, E, F, G, I, or L, 

 respectively (fig. 3 to 6 of PI. A; fig. 14, 16, 17 of PI. C; fig. 20, 21, 24 of 

 PI. D). 



Type L is made up of the large spore-free and spore-forming rods ~L,a 

 and LjS, as well as of free spores and long threads. Germinating spores 

 of this type produce either long rods or rather sho^t ovals resembling 

 type A. The big spore-free rods and threads resulting from type A 

 (fig. 1, 2 of Pi. A) seem to be unable to develop directly the faculty to 

 form endospores. At least, we have never observed such a change, and 

 this also would be in accordance with the fact that a direct transformation 

 of a spore-free into a spore-forming bacterium has never been observed. 

 As mentioned above, type F/3, too, does not develop from type Fa, but 

 directly from the symplasm. As is also noted above, these small spore- 

 forming rods occasionally convert themselves into large spore-forming 

 bacilli. Usually, however, the regenerative bodies formed by type D 

 seem to be the normal predecessors. Under conditions, which will have 

 to be studied more closely, these round cells acquire the tendency to 

 produce endospores, which, in their turn, go back into the symplastic 

 stage. This second symplasm then produces another set of regenerative 

 bodies which stretch out to large granulated rods and threads. They 

 later form the normal endospores. 



TypeM represents another rather unusual form. It originates from type 

 L and passes over into type D (fig. 10 of PI. B and fig. 19 of PI. D). 



It is hardly necessary to point out that sometimes our separation of 

 the forms observed into different types becomes more or less arbitrary. 

 For example, there are no absolutely sharp lines separating t)^pes A 

 and B or the regenerative bodies I from the full-grown cells. Thick- 

 walled cells of types E and F, when produced in the symplasm, might 

 just as well be considered as "regenerative bodies." A germinating cell 

 of type EjS and a budding form of the type Ke resemble each other very 



