theodolite particularly trying.
In addition to the considerations already mentioned, the total absence of local resources,
often even of water, both in the deserts and mountains, obliged us to
The rapid succession of daily marches, often over 25
Marching distances. move quickly.
miles in length, which such conditions entailed, is illustrated by the fact that during the two years and four months which on my second expedition were spent on Chinese soil and used for survey work, there were 488 shifts of camp, the aggregate of marching distances for the same period amounting in my own case to close on 8300 miles. On the third journey when the corresponding period was just under two years, the rate of progress was practically the same, the total length of my marches on Chinese soil being close on 7000 miles, and that covered by R. B. Lai Singh, my indefatigable chief surveying assistant,.probably even more. In addition it should be borne in mind that time spared for prolonged halts was absorbed mainly by exacting archæological labours, carried on generally at desert sites and involving further detail surveys.
Observations for longitude would not have been practicable under such conditions, and chronometers would not have been reliable. That our equipment
Uncertainties in on the third journey did not include an apparatus for receiving wire-
less time signals must, however, remain a matter of special regret to me. Its use would have obviated those considerable uncertainties in longitude inseparable from plane-table surveys extending for great distances mainly from west to east and checked by triangulation only for a comparatively small portion of their length.
Apart from heights measured by theodolite during the triangulation, altitudes were
obtained by Watkin mountain aneroids of the Survey of India. Those Altitude observations. used on the second and third journeys were checked at intervals with
two mercurial mountain barometers and, as judged by this test, preserved a very uniform rate of index error. None of the mercurial barometers survived the trials of the journeys. One, however, of those carried on the third journey was brought back safely as far as Kashgar and could be compared there with the instruments of the fully equipped Russian meteorological station before it, too, succumbed on its way across the Kara-koram passes. In addition hypsometrical observations were concurrently taken with boiling-point thermometers, some of which were kindly lent to me by the Royal Geographical Society. In the course of plane-table work on mountainous ground clinometers were regularly used during the second and third journeys to secure readings to prominent intersected heights. On high elevations special care was taken to obtain clinometric readings preferably from points where the mercurial barometer was available for observation of absolute heights.
With the object of covering as much geographically interesting ground as possible,
I detached my topographical assistants from my own party for independDetaching of surveyors. ent work whenever practicable routes, means of transport, the attitude
of the Chinese administration, etc., allowed it to be done with a reasonable degree of safety. During such periods of detachment which on occasions extended over several months, I carried on the plane-table work along my own routes myself. The comparison of the positions indicated by our traverses at the points where the surveyors and my own route-lines crossed or joined, provided a useful means of controlling the results.
Wherever we moved together, plane-tabling was . done under my direct supervision and with my assistance. The latter was particularly needed in order
Record of surface
features. to secure a systematic record of such geological or physiographic
surface features as belts of desert vegetation, living or dead; drift-sand formations stationary or liable to movement; tamarisk-bound sand-cones; ridges and mounds produced by wind-erosion; salt-encrusted ground of different types, etc., which are charac-