band: nervous system of lumbeicid^. 107 



preserved by the sublimate. In the regenerating brains and ganglia, the 

 mitotic figures more often occur in cells the remainiug contents of which 

 appear to have become entirely fluid. Such figures, unsupported by a 

 dense cytoplasm and completely exposed to the action of the fixing fluid, 

 were very seriously distorted in the sublimate preparations. In all cases 

 spindle fibres were very poorly shown in the sublimate preparations. 



As for the superiority of Flemming's fluid for the demonstration of the 

 centrosome, my experience agrees with that of Ballowitz ('97*, p. 358) 

 who demonstrated centrosomes in epithelial cells of Salpa without the 

 use of stains. He succeeded in doiug this with sublimate, but Flem- 

 ming's fluid gave by far the best results. He says, " Ich behaupte auf 

 Grund meiner Erfahrungen an meinem Untersuchungeasobject, dass es 

 mit grosserer Sicherheit und mehr Constanz gelingt, die Centrosoraen 

 an dem mit Flemming'scher Losung fixirten, ungefarbten Material zu 

 erkennen, als durch specifische Tinction an den mit Sublimat behan- 

 delten Objecten sichtbar zu machen." 



My Flemming preparations were all stained with iron-hfematoxylin, 

 but they were decolorized to such an extent that, examined by a low 

 power, the cell bodies appeared scarcely darker than in the unstained 

 preparation. But examination with high power revealed the fact that 

 the cell granules had retained the stain. It was in such preparations 

 that central granules were most clearly seen, because of their deep stain 

 in contrast to the unstained ground substance of the cytoplasm. Her- 

 mann's fluid gives substantially the same results in this respect. I 

 have examined unstained Hermann's fluid preparations and have been 

 able to distinguish fairly well the granules and fibrillje of the cytoplasm. 

 In stained sublimate preparations there is apt to be present in the cyto- 

 plasm a large amount of deeply staining material scattered about in 

 irregular masses, obscuring finer details of structure. If the decolor- 

 izing is carried so far as to clear out these masses, the granules and 

 fibrillse are left much less distinct than in the decolorized osmic acid 

 preparations. 



All the material was cleared in cedar oil and imbedded in paraffin. 

 Sections were cut generally 6| fx thick. A few series were cut at 3^ /z. 



4. Stains. 



Heidenhain's iron-hsematoxylin proved to be by far the most useful 

 stain. Sections were treated with a two per cent solution of iron-alum 

 from three-quarters of an hour to three hours and stained about the 



