QUANTITATIVE STUDIES ON MITOCHONDRIA IN NERVE-CELLS. 41 



the cranial nerves, where the mitochondria are, for the most part, granular, as seen by 

 the illustrations. Working particularly with the- anterior-horn cells and the large 

 ?ells of the reticular formation, he describes mitochondria which arc filamentous, 

 the filaments varying in length, some being so short as to be almost identical with 

 the granular forms, as one proceeds from the periphery to the nucleus. 



Naturally the question would arise as to whether, under these circumstances, 

 quantitative variations such as those recorded in table 1 have any real significance. 

 It is obvious that the method which I have adopted of count ing the mitochondria 

 would have to be accompanied by very accurate measurements in order to yield 

 reliable information of the relative as well as actual amount of mitochondria in 

 cells in which their size and shape differ to any appreciable extent. Filamentous 

 mitochondria, though occurring in the cell processes, are rarely found in the cell- 

 bodies of the cranial nerves and, since my observations are confined to the cell- 

 bodies, these filamentous mitochondria do not constitute a source of error. A few 

 of them are homogeneous throughout, but most can be resolved by careful focus- 

 sing into rows of discrete granules, fairly uniform in size. For the work in hand 

 this was a great advantage, for by counting the granules a more accurate index 

 of the amount of mitochondria was obtained than would have been possible by 

 counting the filament as a unit. Though the method of counting here used is not 

 ideal in the case of cells characterized by great dimorphism on the part of their 

 mitochondria, it does yield reliable results when care is taken to restrict its use 

 to cells in which the size and shape of the contained mitochondria are practically 

 uniform, as in the cells of the nuclei of the cranial nerves. 



Despite the fact that the granules were of the same size, if there were lack 

 of uniformity in their distribution, an accurate estimation of the mitochondria 

 for any one cell would be impossible unless all the mitochondria in that cell were 

 counted. It is because of their practically uniform distribution that the amount 

 in any one field can be used as typical for the whole cell. Counts were made of 

 different fields of a cell in one section, also of fields selected from the same cell in 

 successive serial sections; and the numbers were practically identical. It is true 

 that there is usually a slight crowding of mitochondria in the axon hillock, with a 

 tendency for them to be arranged in filaments or rows of granules along the long 

 axis of the cell-process (figs. 3, 4, and 5). In the canalicular system, which in 

 these preparations shows white, there are no mitochondria; where they do seem 

 to occur in the canals the appearance is due to their presence in the thin layer of 

 cytoplasm surrounding the canal. Especially in figures 1 and 5 there appear to 

 be large areas free from mitochondria. This is explained by the fact that all the 

 drawings were made in one optical plane; on a different focus mitochondria would 

 have appeared in the cytoplasm, which is now free from them. No variations 

 were noted in the density of distribution of mitochondria other than those just 

 mentioned. Any minor unevenness in the distribution of mitochondria would be 

 obviated by the fact that the square used in counting is relatively large, taking 

 in an expanse of cytoplasm extending in most instances from the nucleus to the 

 periphery of the cell. 



