362 P. R. BILLINGSLEY AND S. W. RANSON 
width of the ruled square, using some well-isolated fiber as a 
landmark. This was continued until a column of fibers was 
counted extending anteroposteriorly clear through the section. 
Then by means of the lateral movement of the stage, the section 
was moved the full width of the ruled square and a second column 
of fibers counted, and so on until the field was covered. 
The number of cells in the ganglion was determined by count- 
ing the nucleoli in every fourth section and multiplying the result 
by four. The method of using the square ruled ocular and me- 
chanical stage was thesameasin counting the fibers. Here especial 
care had to be taken to avoid overlooking small nucleoli which 
fell behind the ruled lines as well as those which might be out of 
focus. 
There are several possible sources of error in counting the cells 
by this method. Since only every fourth section was counted 
and the result multiplied by four to find the total number of 
cells, an inaccuracy is introduced, which, however, is made 
negligible by the large number of sections counted. A second 
source of error may be found in the fact that some few ceiis 
contain two nucleoli and the knife may pass between them and 
they will then lie in adjacent sections and may each be counted as 
representing a cell. This possibility would represent an error 
so small as to be negligible. <A third and real source of error is 
found in the fact that a certain percentage of all nucleoli are 
cut and the parts come to lie in adjacent sections. Parts of 
nucleoli would then be counted as whole ones. 
Measurements showed that the diameter of the average nucleo- 
lus is 2.25u, and since the sections of the ganglion were 10u in 
thickness we must assume that 224 per cent of all nucleoli were 
cut at some point in their diameters. If the knife passes through 
the nucleolus at any point in the middle one-half of its diameter, 
each of the resulting parts will probably be thick enough to per- 
mit of its being seen and counted as if it were an entire nucleolus. 
If the cut passes through either of the outer one-fourths the major 
part will be counted but the minor part will be so thin as to be 
overlooked. We may therefore assume that one-half of the 223 
per cent of cut nucleoli will be so cut as to be seen in two sections 
