474 journal of Cojnparative Neurology and Psychology. 



These three conditions satisfy Pearson's skew frequency curve 

 of Type I, while for the nuclei we have 



Ki > o; /?!> o; /?2 > 3; and /Cj > o and < i, 



which calls for Pearson's curve of Type 4. 



The frequency distributions and their fitted curves are shown 

 graphically in Figs, i and 2. The equations for the curves are: 



For the cell-bodies (Type i)'^ 



, / X \3-6io7 / X \35-6365 



,= 140.0657 (i+g^^J (^.-g-— j 



origin at mode. 



For the nuclei (Type 4) 



(\ 24.0118 

 Cosd 1 £11.20380 



origin at 7.4341 micra, 



X = 11.6002 tan d 



Examining Figs, i and 2 we see at once that the theory and 

 observation do not agree at all well. The theoretical curve in 

 both cases considerably underestimates the observed ordinates 

 for the smaller values of variates x, and overestimates the same 

 for the larger values of x. The degree of deviation between the 

 observed and theoretical curves is most pronounced at or in the 

 neighborhood of the mode. This unexpected results forced the 

 writer to reinvestigate the following points : 



I. Since the spinal ganglion contains various sized cells it 

 may be possible that these cells are not uniformly distributed from 



' Original formula of Type i is given by 



y=yo{ ^+- \ I I — -I ,whrre;yo=- 



(-0"(-0 



' r{m,+ l)r{m,+ j) 

 and for Type 4 



y = yo {Cos 0) £ , where ^o = - 



6 / , \ tn, + mo 



rVTT 



r 



■Sin d e . di 



