MITOCHONDRIA VARIATIONS IN NERVE CELLS 337 



normal states and in pathological conditions. Two important 

 interpretations have been advanced. 



Rubaschkin ('10, p. 428) found, in the study of guinea pig 

 embryos, that the mitochondria were granular in the primordial 

 germ cells and filamentous in other more specialized epithelial 

 cells. He arrived at the general conclusion that the primitive 

 granular form of mitochondria is peculiar to undifferentiated 

 cells and that the process of differentiation shows itself, by a 

 change of the primitive granular type into chain-like and fila- 

 mentous forms. This view has been much criticised. It is in- 

 consistent with the investigations of Swift ('14, p. 495) who 

 found that in the primordial germ cells of the chick the mito- 

 chondria are rodlike and do not differ from those in somatic 

 cells. It is sufficient merely to state that my observations, that 

 granular mitochondria occur constantly in some types of nerve 

 cells and filamentous ones in others, are also at variance with 

 Rubaschkin's hypothesis, because both the types of nerve cells 

 in question (large cells of the Gasserian ganglion (fig. 7) and 

 anterior horn cells (fig. 1) ), are undoubtedly highly differen- 

 tiated. 



Dubreuil ('13, p. 137), on the other hand, is of the opinion that 

 granular mitochondria are in a state of rapid multiplication by 

 division and are characteristic of active stages in the life of the 

 cell and that filamentous ones are indicative of rest. He bases 

 this belief upon his study of the changes which the mitochon- 

 dria undergo in the development of fat cells from fixed connect- 

 ive tissue cells. He found that when the cells are most active 

 the mitochondria are most numerous and are granular; when 

 the cells are less active the mitochondria are filamentous and less 

 abundant. He adds to this the observation that when inflam- 

 mation sets up, the mitochondria immediately increase greatly 

 in number and are granular. The observations recorded in this 

 paper would, at first sight, seem to support this view, for an in- 

 spection of the plates reveals at once that where the mitochon- 

 dria are most abundant, that is to say in cells of the mesencepha- 

 lic nucleus of the fifth nerve (fig. 5) and in the large cells of the 

 Gasserian ganglion (fig. 7) they are also granular. The conten- 



