recently found at north London: one was made of quartzite, the first example met with in the London gravels of this material; the other was a white implement from the "trail and warp." He also exhibited two white porcellaneous Palæolithic flakes replaced on to their original blocks: the four pieces were found by him at north London, wide distances apart, at different times during the last six years.

Mr. Smith also exhibited a large axe from New Guinea, with a keen blade of silicious schist or banded chert, 9 inches long, and weighing over 24 lbs. The axe was sent home by a sailor, and Mr. Smith purchased it of a person who was using it at north London for chopping up firewood.

The DIRECTOR read a paper, by Mr. E. H. MAN, on the Ethnology of the Andaman Islands.

The PRESIDENT then read the following paper:

ADDITIONAL OBSERVATIONS on the OSTEOLOGY of the NATIVES of the ANDAMAN ISLANDS. By WILLIAM HENRY FLOWER, LL.D., F.R.S., P.Z.S., Pres. Anthrop. Inst., Director of the Natural History Departments of the British Museum.

IN the year 1879 I read a paper before the Institute describing the osteological characters of the natives of the Andaman Islands, as derived from the examination of nineteen skeletons, and a larger number of crania. Since that time I have had the opportunity of examining ten additional skeletons, two of which are in the Museum of the University of Oxford, and eight in the Barnard Davis collection, now in the Museum of the Royal College of Surgeons. They are all from the same locality as the others, viz., the vicinity of Port Blair, all undoubtedly genuine unmixed natives, having been obtained through the kindness of Mr. E. H. Man. Five are males and five females, and all are adults.

In my former communication I gave the average measurements of the limb bones, and the principal dimensions of the cranium and pelvis, expressing my belief that the materials at my command were sufficient to draw with safety some general conclusions as to the physical characteristics of the race, and adding, "Perhaps when still larger numbers of skeletons are examined, some of the statements and average measurements and indices will have to be modified, but probably not in any essen

1 "Journ. Anthrop. Inst.," vol. ix, p. 108 (1876).

tial degree." The number of skeletons being now raised from nineteen to twenty-nine, the two sexes being represented in as nearly equal proportions as possible, we have a good opportunity of testing the truth of this supposition, and of giving still more reliable averages of the dimensions and proportions of a people of a remarkably pure and homogeneous race, which may serve as a good standard of comparison with those of other races. The measurements are all taken exactly in the same manner as in the former paper, and the comparison is instituted between the average then obtained from the smaller number of skeletons, and that now derived from the entire number now available (composed of those formerly used with the new ones added). The figures given in the present communication will therefore supersede those given before, and represent the correct averages of the race as far as our present data enable us to give them.

Limb Bones.

The average length of the fifteen clavicles of males previously given was 1169; the entire series of twenty-five gives a very slight increase of length, viz., 118.2. The average length of eighteen clavicles of females was 1071. The average length of twenty-six is 107.5. The ratio of the length of the clavicle as compared with the femur is now in the males 300 to 100, in the females, 283 to 100. Of the humerus the average length in the males is reduced from 281-3 to 276·5; in the females, on the other hand, increased from 2607 to 263-5. The radius is also reduced in the males from 228.9 to 225-2, but in the females remains practically the same, the former number given being 2100, and the present average 2101. The average length of the femur is also reduced from 398-7 to 393-4 in the males, and increased in the females from 378.2 to 3804. The tibia is slightly reduced in the males from 336-2 to 3321, and in the females scarcely undergoes any alteration, being 320.5 in the former list, and 321-0 in the present.

An inference was drawn as to the average height of both sexes from the length of the femur, upon the usual estimate derived from measurements of our own race, that the height is to the length of the femur as 1000 to 275. Whether it applies equally to the Andaman is of course not known, although the height inferred by the calculation accords remarkably with the results obtained from the measurement of living individuals. The average height of the males calculated upon the old average length of fifteen femurs was 1,448 mm., or 4 feet 9 inches. The average calculated from the new total of twenty-five femurs

The measurements are all given in millimetres.

is 1,431, or 4 feet 8.3 inches. The maximum of the old series is not exceeded, but the minimum is reduced from 1,385 or 4 feet 6.5 inches, to 1,320 or 4 feet 4 inches. The new averages in this sex are probably more to be relied upon than the former ones, which, as remarked at the time, were unduly raised by the presence of an exceptionally tall man (that presented to the British Museum by Dr. Mouatt) among the series. In the case of the females, the old height was calculated at 1,375, or 4 feet 6.1 inches; it is now placed at 1,383, or 4 feet 6-4 inches, being but very slightly above. The new numbers all fall within the old maximum and minimum.

The inter-membral index, or length of the humerus and radius added together, compared with that of the femur and tibia, the latter being taken as 100, taking both sexes together, is exactly the same as in the former list, viz., 68.3. In the males it is 690; in the females 67.5.

The femoro-humeral index (or ratio of the humerus to the femur, the latter being taken as 100) is also unchanged, being 69-8. In the males it is 70-3, in the females 69.2.

The femoro-tibial index undergoes the excessively trifling modification from 84.5 to 84:4. It is exactly the same in both

sexes.

The humero-radial index, which forms one of the most important differences between the skeleton of the Andamanese and the European, is only changed by the additional specimens from 810 to 80.6, taking both sexes together. There is some difference between this index in the male and the female skeletons, being 81.5 in the former, and 79-7 in the latter.

Cranium.

All the peculiarities of the Andamanese cranium fully described in the previous communication are strikingly illustrated by the present specimens, and therefore need not be repeated. There is not one that is in any way exceptional in any of the characters that can be judged of by the eye, although some of the measurements show slight differences, especially a greater index of breadth, due apparently to the presence of some rather peculiarly narrow skulls in the former series. It must be noted, however, that as many as four of this small series (all males) have the frontal suture persistent. In the former series of thirty-five known examples of skulls of the race, four were recorded as metopic; in forty-eight of which I have at present evidence of the condition of the frontal bone, as many as six, or 12.5 per cent., are in this condition, so rare among the dolichocephalic Melanesians and Australians.

With regard to the capacity of the cranium there appears to be little difference from the former series, although the new appear larger owing to a different method of measurement, which as far as tried gives more certain results. Some of the skulls, the capacities of which were given before, are no longer accessible to me, but those which are in the Museum of the College of Surgeons have been measured over again with very great care, using shot instead of mustard-seed as formerly, and then added to the new specimens, giving an average of 1,281 cc. in thirteen males, and 1,148 cc. in the same number of females. These numbers may be substituted for 1,244 and 1,128 of the males and females respectively given in the former communication. Among the males the maximum is 1,400, and the minimum 1,120; among the females the maximum (one of the new series, No. 1,486 of the Barnard Davis collection) attains the exceptional capacity of 1,375, giving perhaps an unduly high average to this group; the smallest is 1,040.

The average horizontal circumference in the twelve males of the former series was 480; the additional eight only change the average by 1 millimetre, it being now 481. The average circumferance in the female is absolutely unaffected, it being 462 in the original series of twelve, and the more extended one of seventeen. The average length (ophryo-occipital) of the twelve males was 167.5, that of the twenty-one 1674, showing only a difference of of a millimetre. On the other hand, the maximum parietal breadth in the new series somewhat exceeds that of the old, principally in consequence of two or three unusually narrow skulls being contained in the former. The average of the old series was 1349, that of the new 135.8, raising the index of breadth from 80.5 to 811. The average height (basi-bregmatic) in the former series was 1296, in the present series 129.5, the index being 773 instead of 77.0.

In the females the average length of the old series of twelve was 1606; that of the combined series of eighteen, 160-8. The average of breadth is changed from 132-8 to 133-2, the index. being only altered from 82-7 to 82-8, an unappreciable difference. The average height of the twelve crania was 1253; that of the eighteen 124.9, the index changing only from 77.9 to 77.7.

The average basi-nasal length of the males in the former series of twelve was 950; in the present series of twenty-one crania it is 94-6; the basi-alveolar lengths being respectively 96.3 and 94-6, reducing the gnathic index' from 1,014 to 1,000, the present series being thus somewhat less prognathous than the former one. In the females the former basi-nasal length was

1 See "Journ. Anthrop. Inst.," vol. x, p. 163.

90-7, that of the eighteen skulls now measured 90-3; the basialveolar lengths being 92.7 and 92-4, which produces practically no alteration in the gnathic index, which is given as 1,022 in the previous list, and is 1,023 in the present one.

The measurements of the nasal aperture in the males are as follows: Former series of twelve crania: height 45.8, width 234, index 511; present series of twenty-one crania, height 461, width 234, index 508; the width is therefore the same, but the increased height gives a slightly lower index. In the females: former series of twelve, height 43.2, width 22-2, index 512; present series of eighteen: height 43-3, width 22-4, index 51.5. The general average index of the male and female skulls combined is exactly the same in both the whole and the partial series, viz., 51.15, which figure may be safely accepted as the average nasal index of the race.

The orbital measurements and indices receive scarcely any alteration from the additional specimens. In the males the height, width, and index were respectively 36-3, 330, and 90-6; now they are 365, 330, and 904. In the females, the old numbers were 352, 322, and 915; now they are 351, 32.1 and 91.4.

With such close agreement in all the primary measurements of the cranium, it seems scarcely necessary to follow out in detail the numerous minor dimensions, as whenever tested they correspond equally nearly with those previously given. The main point that I wished to test has been thoroughly established, the twelve skulls of each sex previously examined furnishing a very fair average of the characters of the race.

It was stated in the former communication that "in dimensions the teeth appear equal to the average of those of Europeans, and therefore may be considered large in relation to the general size of the body." Since that time I have found it convenient to use an index to express this relation, which is obtained by comparing the length of the molar series (straight line between the anterior part of the first premolar and the posterior point of the third molar d) with the basi-nasal length (or craniod x 100

=

facial axis) BN BN ; may be called the dental index, and by it races may be divided into megadont, mesodont, and microdont. Like most of the frizzly-haired races, the Andamanese are decidedly megadont; the average index in nine males being 44.4, and in eight females 46.5. Unfortunately so many of the skulls have lost part or all of the series of teeth, that the number available is rather smaller than might be wished.

Among other points in the skeleton of primary importance, as indicating race characteristics, is the pelvic index, or the