ox ClUSTISPIKA VKXEKIS XOy, SPliC. 523 



ances which I have described. It is easy to understaud why 

 the <^ranules — in reality masses of granules — appear to be of 

 different sizes in such organisms (Sg. 1), and wliy they 

 always appear at the edges of the organism, which we know 

 to be really cylindrical Avhen alive. 



AVhen no fixation previous to drying occufs, the cell under- 

 goes ])lasrnolysis and complete flattening on the slide. The 

 small granules run together in various ways, giving rise to 

 the various "nuclear" figures which have been described (cf . 

 fig. 10). It is easy to understand how the flowing of tlie 

 granules through the walls of the chambei-s, and their massing 

 together in various ways, can give rise to the appearance of 

 transverse bars, spirals, tetrads, etc., of " chromatin." It will 

 hardly be necessary to descril)e in detail the several ways in 

 which such appearances may be caused. 



The staining reactions of the substance of which the 

 gmnules are composed require a brief consideration. I have 

 already noted that the granules stain a deep blue after a long 

 exposure to osmic yapour, purple after a brief exposure, red 

 Avhen osmic fixation is omitted, and only absolute alcohol is 

 employed after previous drying. I believe these differences 

 are directly due to the action of the osmic yapour — pro- 

 lono'ed action of which so chang-es the on-anules that thev are 

 unable to take up the red-staining element in the Komanowski 

 stiiin. I have observed this action of osmic acid in the case 

 of Bacteria and many Protozoa, and I believe it must have 

 been noticed by many other workers who employ Romanowski 

 staining after osmic fixation. A short exposure to osmic 

 vapour permits the granules to stain red — as they do when 

 not acted upon by it. A longer exposure permits them to 

 stain red to a less extent, and gives rise to a purple colora- 

 tion. Still longer action of the osmic vapour renders the 

 granules incapable of taking up the red element in tlie stain, 

 and they therefore appear blue — the blue element alone being 

 capable of staining. 



There can be no doubt, I believe, that the granules are 

 composed of a substance Avhich is different from that of the 



VOL. 58, PART 3. NEW SERIES. 36 



