Metic.

He found In going up the grades the greatest oscillation of in crease anc! decrease in the number of failures and apparently in the power of numerical analysis.

But his figures showed that the rate of decrease in error is most rapid at thirteen and fifteen, although at four teen errors are nearly as numerous as at twelve.

Girls improved till they were ten and then lost, but after a year improved again- From this work it would appear that children can not reason much beyond their experience and environment until about the age of puberty.

These conclusions are based on returns studied for the purpose of 500 or more children of each age from seven to fifteen.