IRON COMPOUNDS IN ANIMAL AND VEGETABLE CELLS, 221 
A similar conversion of a compound in which the iron is 
easily attacked by ammonium hydrogen sulphide and by acid 
alcohols into one from which the liberation of the iron is more 
difficult, obtains in the dim bands of muscle-fibre in Inverte- 
brates (Oniscus, Chironomus, Musca), but in this case the 
transformation does not proceed as far as the production of 
hemoglobin or myo-hematin, if one may judge from the 
absence of pigment and from the fact that the liberation of 
the iron, though difficult, is possible, while in the case of hemo- 
globin the use of acid alcohols and of ammonium hydrogen 
sulphide is ineffective for that purpose.! 
In the development of the blood-corpuscles in the larve of 
Amblystoma there is, as I have pointed out,? a conversion of 
the chromatin of the hematoblast into hemoglobin, a change 
that is analogous to that described above as occurring in muscle- 
fibre. In hematoblasts, however, the chromatin so trans- 
formed is not directly derived from that of the yolk-spherules, 
as is the case in muscle-fibre, but from that compound after it 
is transformed into nuclear chromatin. ‘This is very distinctly 
seen in sections through the aortic arches of the larve, which 
have been treated with acid alcohol to liberate the iron. In 
the concave side of the arches are seen hematoblasts in all 
stages of division, and in these one may, by the iron reaction, 
differentiate between hzmatoblasts in which there is no cyto- 
plasmic chromatin, and those in which the cytoplasm between 
the chromatin loops of the mitotic figure contains dissolved 
chromatin to an extent varying with the example of hemato- 
blast noted. This cytoplasmic chromatin does not act in the 
same way as ordinary nuclear chromatin does towards staining 
reagents, as, for example, hematoxylin, eosin, and safranin, 
1 The fact that ammonium hydrogen sulphide will liberate the iron from 
hematin in solution, while it does not attack the iron in the compound called 
myo-hematin by MacMunn, indicates that the latter cannot belong to the 
hematin class. Its property in this respect shows that it is related to hamo- 
globin. ‘lhe name given to it by MacMunn certainly appears to be a mis- 
leading one. 
Loc. cit. 
