bg: THE MAPS [Chap. III
scarcely necessary to point out that the use of these distinctions can serve only for approximate guidance.
Here I may conveniently mention that for quasi-technical reasons, names, usually end-
ing in jilya( Turki ), indiscriminately applied both to streams and the Lettering of temporary valleys containing them, have always been printed in blue without re-
gard to whether the streams generally carry water or not. For the same reasons the lettering of all temporary water-forms, such as flood-beds, seasonal marshes, etc., has been shown in blue.
Marshy areas, being subject to considerable seasonal changes in arid regions, such as those represented by most of the maps, introduce a special element of
Representation of uncertainty into cartographical representation. I have tried to restrict
the use of marsh symbols in blue to ground which is likely to be covered with water for the greater portion of the year. But it should be remembered that where drainage beds, ordinarily dry, terminate in areas shown as sandy tracts with living desert vegetation ( reeds or scrub ), extensive portions of the ground are likely to be found boggy during the season of spring floods from the melting snows of the mountains. S
Salt-encrusted areas, marking dried-up lake-beds and marshes or ground subject to inundation or percolation at intervals followed by evaporation, form
Salt-encrusted areas. very important features of the surface in the lower portions of the Tàrim
basin and of the adjoining drainageless regions. They date from different periods and present considerable geographical or geological interest. I have accordingly thought it desirable in the course of our later surveys to distinguish areas showing marked differences in the salt formation covering their surface.
Three varieties of new symbols have been introduced for this purpose, all derived from the conventional symbol in use for water of lakes and marshes, but all
Varieties of salt shown in black. Thin lines with hook-like ends turned upwards on the
right indicate a crust of hard salt, practically pure, crumpled up into hummocks or small ridges, such as covers the greater portion of the ancient Lop sea bed. The crust of salt-permeated clay, usually formed into big lumps quite as hard as the first variety, and usually found near the edges of ground still receiving some water, has been shown by similar lines with hooks turned downwards on the left. Finally, ground where a soft crust of salt of varying depth covers tiie underlying soil is indicated by lines with downward hooks on the right. 3 Ordinary salt efflorescence ( known as in India by the term shbr) is present to a greater or less extent almost everywhere in the Tarim basin where subsoil drainage comes to the surface or temporarily inundated areas are near; it has hence not been specially marked. Among other permanent surface features drift-sand is the most important in the regions represented by the majority of the map sheets; for it covers
Representation of the portion greater ortion of the Tarim basin. For drift-sand bare of vege-
tation, or nearly so, brown stipple has been used. The indication of dunes within areas of bare drift-sand by small clusters of thicker stipple facing to the southwest is meant to be purely conventional. It has been chosen with regard to the prevailing wind direction, though the bearing of individual dunes varies greatly in different desert areas. But where large accumulations of dunes in the shape of high ridges or dawûvs are met with, usually near river courses or parallel to them, they have been shown with the true bearing of their axis as actually surveyed. 4
Two surface formations of desert ground often associated with drift-sand are of sufficient importance to claim representation by special symbols. The
Representation of one adopted for ardati s or wind-eroded clay ridges and-trenches so
ydrdangs and •Mesas' p ✓ ✓ y r
characteristic of different parts of the Lop depression, shows a form meant to represent the usual bearing of the ridges from N.E. to S.W. and their tail-like end tapering to S.W. For the high clay terraces or ' Mesas' also due to erosion but of earlier