THE MITOTIC CYCLE 



nuclear diameter rather than the cubed values for volumes, for the 

 errors involved in measurements under the microscope from serial 

 sections which have been shrunk in dehydration and embedding are 

 already formidable enough without being raised to higher powers. 

 Figure 20 shows that Jacobj's data for nuclear diameter in a mouse 

 embryo when plotted on probability paper suggests a normal distribu- 

 tion about a single mode, for the points may approximately be joined 

 by a straight line. Harding^^^ showed that a bimodal distribution when 

 plotted in this way gives a sigmoidal curve, and the abscissa of the point 

 of inflection indicates the approximate proportion of the representatives 



.1 



^9 



001 0050-10205 1 2 



5 10 20 30 V0 50 60 70 80 

 Cumu/ofi^e percentages 



90 95 98 99 998 99-9 



Figure 20 Nuclear diameter in kidney tubules of 20 mm. mouse embryo. 



Data taken from JacobjI^^ and plotted as cumulative percentages on 



probability paper. The distribution is unimodal, as the points may 



approximately be represented by a straight line. 



of each mode. Such a distribution is exemplified by the results of 

 Beams and King^^^ for the nuclear diameter of mononucleate cells in 

 the normal adult rat liver* (Figure 21). The data for the mouse liver is 

 less easily interpreted. In addition to Jacobj's observations, nuclear 

 diameter in this tissue has also been investigated by Heiberg,^"" 

 Voss^oi and Muller^os and the data of all these authors for mono- 

 nucleate cells is plotted in Figure 22 in the same manner which has 

 been described. It is clear that the points for each set of data could 

 only very approximately be represented by a straight line, and one 

 cannot point to any sigmoidal inflections with confidence; nor are 

 there any constant features in the shape of the curve for each set of 

 points. Nuclear size in the adult mouse liver is therefore too complex 

 an example to be analysed in this way from the existing data. 



• The corresponding data of Jacobj also suggests a bimodal distribution, but is curiously 

 divergent from that of Beams and King in absolute size. 



58 



