mann. The latter contrast is somewhat unsatisfactory owing to the usual reason-the small number of skulls available for examination. The latter shows that the Hottentots, still more therefore the Bushmen, are distinctly separated in all dimensions from the Kaffirs, and in most dimensions from the peoples of Nyasaland, but although they probably differ equally from the Herrero tribes of German territory, the differences are not demonstrably unaffected by chance.

A figure above 20 indicates considerable odds against the explanation of the differences between the respective average being random sampling, while a figure above 30 indicates the practical exclusion of chance.

The apparent relationships between the dimensions of the crania from the several groups having been determined, it seemed advisable to endeavour to determine if any of these groups were more mixed than the others. The groups showing the widest range of variation in dimensions might be expected to be more probably a mixture of individuals of different stocks. For this purpose the coefficient of variation of each group was calculated.

The coefficient of variation is one hundred times the quotient of the standard deviation divided by the actual value of the Standard deviation × 100 Average

average:

=

Coefficient. Thus the standard

deviation of head length in the Strandloopers is 5.6276, the average

DIMENSION.

head length is 181.07 millimetres, and the coefficient of variation 5.6267 × 100

181.07

= 3.10.

The following table indicates the coefficients in the groups previously contrasted :

From this it would appear that the Strandloopers are a less variable group than the skulls collected under the general title of Bushmen-a conclusion amply confirmed by inspection. The Hottentots show a lesser range of variation in most dimensions than do the Bushmen, though somewhat greater than either the Strandloopers or the Kaffirs.

This scarcely agrees with the idea that the Hottentots are a recent racial admixture recruited irregularly from both Bushmen and Kaffirs. It is doubtful whether this view has ever rested on a firm historical basis. On the other hand, if an equal number of Strandlooper and Kaffir skulls be taken and treated as a simple group, the resulting average dimensions and coefficients of variation agree closely with the numbers obtained from the Hottentots. Yet in the case of such a mixture it is comparatively easy by the eye to pick out the majority of the component skulls and place them under their respective categories. While in dealing with the Hottentot series, although many could be at once selected as indistinguishable from Bushmen, fewer would be taken for Kaffirs.

On the other hand, it is far less easy to separate the skulls of a mixed group of Hottentots and natives of Nyasaland. The latter. are in many respects a more variable group, and, it may be, contain

Hottentots.

among them the racial elements from which the Hottentots are derived. The skulls from the northern part of German South-West Africa and the south of Portuguese territory in Angola are more easily distinguished from the Bushmen-Hottentots, and resemble those of the Eastern Bantu group.

It has been suggested by von Luschan and others on linguistic grounds that the Hottentots are related to the Hamitic peoples of North-Eastern Africa. To inquire into this is a matter of some importance, but of great difficulty. In the first place it has to be determined who are the Hamites-a question as yet unanswered. The tribes of Abyssinia and Somaliland are possibly largely Hamitic, but no collections of skulls from these areas appear to be available. The Egyptians, ancient and modern, and the Soudanese of the Nile Valley, no doubt present a large Hamitic element. Of the osteological features of the latter little is known. Of the former there are large collections from all periods, even as far back as predynastic times, which can only be approximately dated. Are these to be looked upon as representative of the Hamites? This is quite uncertain; they undoubtedly contain a large mixture of the Mediterranean man. The relationships of the early, yet still surviving, inhabitants of the South-Eastern Mediterranean to the Hamitic stock are quite unknown, so that it may be a gross error to select them as examples of Hamites. Arthur Thomson and Randall MacIver, in a recent publication, claim to have disentangled a Negroid element from among the early inhabitants of the Thebaid. It is disputed whether the type in question represented a separate race element or only one extreme of the variation of a fairly uniform population, and if the type does represent a separate element it is. not clear whether it should be looked upon as Hamitic or Negro. The only other people differing from the surrounding Negro tribes, and who may in their distant affinities be Hamitic, are the Masai. From this group only six male skulls are available, so that they can be but sparingly employed for comparison.

The Egyptian skulls chosen for comparison were from the predynastic period, and are compared in two series-one containing all the skulls, the other only those of the Negroid type. The average and deviations have been calculated from the tables in the appendix to Thomson and MacIver's work. The following table gives the ratio between the differences in the averages of the contrasted groups and

the value derived from the formula- M,

M.

:

+

N,

In this table a number above 2 shows a probability of a real difference existing between the contrasted groups in respect to the given dimension, while a number above 3 makes this difference a practical certainty. It will be seen that the Hottentots are quite distinct from the general average of the predynastic Egyptians, but show fewer differences from the Negroid type. While the Negroid type is in other respects separated from the Bantu-speaking Negroes from Nyasaland. The latter group was selected for comparison as appearing to present more points of similarity to Thomson's Negroid. type of Egyptian than any other of the Negro groups investigated. Although in certain respects the Hottentots and the Negroids are clearly separated, it is interesting to note that in any large collection of early Egyptian skulls a certain number can be picked out by the eye as distinctly resembling the former. The Bushmen, and still more the Strandloopers, are further removed from the Negroid type than are the Hottentots. The Masai skulls differ from the Negroid type in presenting longer faces and longer and narrower noses, otherwise there is a considerable similarity in their dimensions. The small number of Masai crania leads to so large a probable error that direct contrast, as in the previous table, is of much less value. Yet even with this the Negroid type are clearly distinct from the Masai, who more closely resemble the general average of predynastic Egyptian, and especially in some dimensions the so-called non-Negroid type. The important basi-nasal and basi-alveolar dimensions are unfortunately not available. The chief interest of

Negroid Type and

Masai.

All Egyptians and Masai.

this comparison is that it would tend to show that the Negroid type is not the Hamitic element, and the non-Negroid type is widely different from the Bushmen-Hottentot group.

The features of resemblance of the Negroid type with the Hottentots may be due to some former northern extension of the Bushmen and Hottentots, as has been suggested, or to the influence of the Central African Pigmies.

A comparison of the Bushman and Strandloopers with the Central African Pigmies is best made by a list of actual dimensions and indices, since only three examples of the latter skulls are available. The measurements of a Sicilian Pigmy described by Kollmann are also appended.

→ From measurements of Flower, Journal Anthr. Inst., vol. xviii.

The resemblance is not great, but individual Bushman and Strandlooper skulls approach the others closely in dimensions, and with such small numbers doubts must arise as to the representative nature of the examples available.

The great feature distinguishing the Central African Pigmies from the Bushmen is the prognathism. This character suffices to say that at present they are not the same race, it does not show that they may not have a common ancestry. The Pigmies of the forest zone, so far as it is at present known, seem to be largely mixed with Negroes. This has given rise to the colour types among them, one with a dirty yellow-coloured skin, the other black. Johnston regards