Page:Proceedings of the Royal Society of London Vol 69.djvu/352

336

Over 6"'0o.

Under 6"'05

Totals.

Honours

284

240

524

J?

a

Pass

250

237

487

x>

<J

Totals

534

477

1011

The correlation therefore between ability and broad heads is r = 0'0450 O0322. Were these results sensible we should have to conclude that ability was directly correlated with increased length and breadth, and further with a dolichocephalic tendency. But if we compare the numbers with their probable errors, we see that the corre- lation is non-significant in the cases of both cephalic index and breadth ; Avhile in the case of length of head, the correlation is onty two to three times the probable error, and, even if this be taken as significant, which is doubtful, the value of the correlation is so small that it can have no practical importance, when we attempt to judge ability from physical measurements of the head.*

So far then as our Cambridge results go, they thoroughly confirm Dr. Lee's investigation as to the capacity of the skull. There is no marked correlation between ability and the shape or size of the head.

(4.) It seemed possible although on less safe data to consider the same problems from our school-measurement series. f I say on less safe data for two reasons : (a) Because the measurements being made in schools of all types up and down the country really involve a mixture of classes with very different forms of nurture, and (b) Because we have to allow rather roughly for growth.

There is little doubt that the cephalic index remains sensibly con- stant during growth, and therefore in this character we can compare children of different ages. J On the other hand, the length and breadth

would be first and second class honours men together and third class honours men with poll men. Accordingly, Miss Lewenz undertook the calculations for this divi- sion. She found the correlation between ability and short heads 0'10660'0343, and between ability and broad heads 0'03470'0345. The latter correlation was thus practically unaltered, and the former increased only within the limits of the- probable error. Compare the results in the summary for honours men only.]
 * [It was suggested to me when this paper was read that a better grouping

f This series was carried out by assistance from the Government Grant Com- mittee.

J I find for Cambridge men, cephalic index = 79 '5 ; for school children (all classes) 78'4, and Dr. Macdonell for crimimal classes 78'5. Compare also the valuabb