1375 
Table Il contains the average number of mistakes, with the mean 
deviation and the central value, made by the children of different 
ages, by boys and girls collectively and by each sex separately. 
TABLE II. 
Age (years) Boys | Girls | Children 
| AM. | MD. | Cv. | AM. | MD. | CV. | AM. | MD. | CV. 
8 50.5 | 23.1 | 47 53 | 25.3| 44 51.7 | 24.1 | 45.5 
9 40 17.5 | 36 36.5 | 17.5 | 35 | 38.2| 17.2 | 36 
10 40-3) Bie) 33-5 | 33.4 |. 12.2) al 36.1 | 16.9 |- 32 
11 sorot 27 83°} TES) | SS ides) sp 
12 23.5 | -f1.8 | *'92 27 14.8-| 25 25.2 | 13.3| 23 
Also the number of mistakes, as well as the time-values decreases 
rather regularly, both for the boys and the girls collectively and 
for each sex separately, with the increase of years. The boys of 
10 years form the only exception. While the average decrease of 
the number of mistakes for the boys of 8—12 years is 46 °/,, that 
for the girls is much larger, viz. 51°/,. Since the working-speed 
in the same space of time increases for both categories only with 
30°/,, the years seem to exert a greater influence upon accuracy 
than upon speed. Boys of 8, 9 and 10 years generally make more 
mistakes than girls of the same age. When considering the arith- 
metic mean the boys of 8 years seem to form an exception to this 
rule, but the rule holds good also for them when we consider the 
central value. From the 11% year, however, the boys have the better 
of the girls; however, the differences are generally inconsiderable. 
For either sex and all lifetimes the central values are invariably 
smaller than the arithmetical means. The mean deviations, on the 
other hand, are rather considerable, as they amount to */,—’/, of 
the arithmetical mean; here also they lessen with the increase of 
age. As for accuracy the individual differences become less significant 
as compared with the typical regularities, characterizing a certain age. 
Times and mistakes per se do not throw much light upon the 
nature and the quantity of the work done. To realize both we first 
have to reduce the quantitative data of times and mistakes to one 
experimental value, in which either the time values have been 
reduced to mistake-values or the reverse. We followed the first 
method and obtained our experimental value by adding to the average 
89 
Proceedings Royal Acad. Amsterdam. Vol XXIII. 
