362 
Miscellanea 
For the number of pocks corresponding to Classes II and III I have only very rough 
evidence. In two large scale photographs of cases in these two classes I estimated the number 
of pocks at about 5000 and 2000 respectively. The pocks were so densely crowded that it was 
impossible to count them accurately. All I could do was to count those in one square inch of 
each photograph and multiply by the approximate area of the photograph. These photographs 
were of individual cases only and were not used as divisions between classes, nor as types. 
Consequently I have not used these figures in the table. 
Tabulating the above results we get : 
Class 
Class 
No. of pocks 
No. of cases in class 
No. of cases per 
range of 100 pocks 
VI 
0— 100 
2851 
2851 
V 
100— 500 
1385 
346 
IV 
500—1300 
1141 
142 
III 
1015~| 
II 
1300—00 1 
291 [1572 
? 
I 
266 J 
Total 
6949 
which are exhibited alongside in chart form. 
ClassV 
Class IV 
wo 200 300 400 500 600 700 800 900 WOO 1100 1200 1300 
No of ^ 
Pocks." 
The skewness of the diagram is of a high order. 
Classes I to M 
cases exceeding' 
1300 pocks. 
Further Remarks on the Distribution of Severity in Cases of Smallpox. 
By KARL PEARSON, F.R.S. 
In his paper in Biotnetrika, Vol. iv. pp. 483-504, Dr Turner obtained a series of values for 
the relation lietween severity of disease and vaccination, and in a discussion on these suggested 
that a normal distribution ought rather to be assumed for the whole population e.\posed to risk 
of infection than for the population actually attacked hy the disease. He suggested that the 
attacked population is really a " curtailed " normal distribution and considered formulae for such 
" curtailed " distributions. 
In a note on Dr Turner's memoir I took the only test of smallpox severity which was at my 
disposal, namel}' the distribution of intervals which the physicians at Glasgow allow to elapse 
between (i) onset and (ii) eruption and the first bath ; this I have been assured is a rough but 
fair measure of the severit}' of the attack. I showed that in these cases the maximum severity 
