[1527] Brit. Barrows, p. 568.

[1528] See A. de Quatrefages and E. T. Hamy, Crania Ethnica, 1876-82, pl. lv, and Mem. Anthr. Soc., i, 1865, p. 150.

[1529] Brit. Barrows, p. 621. For some valuable remarks on the permanence of cranial types notwithstanding changes of environment, see Mr. J. L. Myres’s paper in Geogr. Journal, xxviii, 1906, p. 559, with which cf. pp. 555-6.

[1530] See my Caesar’s Conquest of Gaul, 1899, p. 246, n. 1.

[1531] A. H. Pitt-Rivers, Excavations in Cranborne Chase, ii, 1888, pp. 205-6. See also Dr. Beddoe’s remarks in Journ. Anthr. Inst., xvii, 1888, pp. 202-9.

[1532] I say ‘for our purpose’ because of many of the skeletons with which we are here concerned the only relevant measurement that exists is that of the thigh-bone. Dr. Garson (Journ. Anthr. Inst., xxii, 1893, p. 9) thinks that the ‘most reliable estimate of stature is obtained from the length of the femur and tibia added together’ according to the formula (Femur + Tibia x 100) / 49.4

[1533] Ib., xvii, 1888, p. 204. Dr. Beddoe’s arguments are very strong; and his method appears to me better, on the whole, than that of Dr. Topinard, expressed by the formula (Femur x 100) / 27·1 (ib., xxii, 1893, p. 9), or than that of M. Rollet (ib., p. 19, note), expressed by the formula (Femur x 100) / 27·3. So far as I know, the most exhaustive discussion of the question is that of M. L. Manouvrier (Mém. de la Soc. d’anthr., 2e sér., iv, 1892, pp. 347-402), who points out defects in the methods of Dr. Beddoe and MM. Topinard and Rollet; but although he has perhaps shown how greater accuracy can be achieved, the more or less approximate results that have been already obtained are sufficient for our purpose: we should not be in a better position for solving the problems of the ethnology of Ancient Britain even if the Britons whose skeletons have been preserved had been measured in their lifetime, and the measurements recorded.

[Since the foregoing note was written I have read a most interesting paper by Dr. Beddoe (Journ. Roy. Inst. Cornwall, xv, 1902, pp. 161-78), which confirms my conviction that his is the best method of measurement, although he confesses (p. 165) that it ‘probably errs by excess in the higher statures’. Remarking (p. 163) that prehistoric bones ‘have lost much of their original substance, and are probably from 1 to 3 millimetres short of their original length’, he says, ‘Manouvrier does not seem to have made any provision for this reduction; and I apprehend that his computed statures must on an average be a little too low’. See Addenda.]

[1534] See pp. 25-30, supra.

[1535] See p. 35, supra.