28 THE NOBIANS, THE NUBA l. 2. xxxi.

depth all along it, or if one prefer there is a more direct route, that

followed by the Turks in 1821 and by most of the Arab caravans,

via the deeper desert wells of el Safia, Hobagi, and Elai. Further

west and very roughly parallel to the Wadi el Mukaddam is the

similar Wadi el Melik running from the Darfur border into the Nile

at Debba, only forty-five miles west of Korti.

At el 'Ayn on this Wadi and at Abu Sufian to the west of it are

similar traces of human occupation well known to the nomads^, and

at the southern end of it near the Darfur border they are very common

indeed.

XXXII In these parts the underground water supply has decreased

to a very striking extent of late years, and that in ancient days the

rainfall was considerably more heavy is proved, I think, by the

presence of gigantic baobabs {hiheldi: Adansonia digitata) that

are centuries old and could hardly have passed through the early

stages of growth had the country been as dry as it now is. Some hundreds of years ago the country on both sides of the two great Wadis mentioned may have been habitable all the year through^.

XXXIII Now it is certain that in past centuries Danagla from the

Nile have settled at el Haraza, and fresh colonies have joined the

older ones within recent years^. It is also believed by the people of Gebel Mi'dob in Darfur, about 140 miles west of the W^adi el Melik,

that their ancestors were Mahass and Danagla from the Nile"^, and,

as Professor Seligman has pointed out^, there are very close linguistic resemblances between a list of their numerals which I collected in 1912^ and those of the Barabra on the river. Emigrant Barabra may have reached Gebel Midob by way of

el Haraza and Kaga, or more probably by way of the Wadi el Melik'. But if Barabra from Korti and Debba, which are between the third and fourth cataracts and only a few miles from the pyramids of

Barkal, could settle at el Haraza, as we know they did, and at Midob,

as we may be fairly sure they did^, there is no reason to deny the

probability of corresponding movements along the same lines in the

^ I have not visited personally these two sites.

- During six years that I passed in northern Kordofan I never saw a young baobab growing self-sown. The Arabs also declare there is no such thing and say the "tubeldis" date from the time of Noah.

' For details see MacMichael, Tribes..., Chap. vi.

* Ibid. Chap. vil. Matrilinear descent with inheritance by the sister's son still holds good at Mid6b as it did in Christian Dongola. See p. 59.

* Loc. cit. Cf. the vocabularies of Kenuz and Nubians given by Burckhardt.

''' Published mjourn. Anthrop. Inst, xlii, 1912, p. 339.

' Marauders from Mfd6b and thereabouts follow this line in the rainy season and early winter when raiding the Arabs.

* See Part I, Chap. 4.