actual measurement being about two miles north of the town. The base 

 was 2400 feet long and was measured twice by means of the Invar Tape ; 

 the probable error being 1/50,000. The longitude of one end of the 

 base was assumed, and the latitude calculated from astronomical 

 observations. This was subsequently corrected by the telegraphic 

 longitude reduced to the South East Gate of the City. 



(b) On our arrival at Yen-an Fu, the surveyor found himself at 

 fault, and it was considered advisable to measure a new topographical 

 base. This base was 8 miles long and connected with the longitude by 

 occultation. 



Datum Level. 



From the T'ai-yiian base-line, five elevated points were fixed by 

 triangulation, their vertical angles were taken with the theodolite, and 

 their altitudes determined trigonometrically. The altitude of the end of 

 the base from which they were determined was assumed to be 2700 feet. 

 As there is only a very gradual rise to this point from T'ai-yiian Fu (alt. 

 2600 feet as found by the French Railway Survey) we think that the 

 height assumed may be taken as correct to within 50 feet. 



Plane-table Survey. 



The plane-table work was executed by Hazrat Ali on the J-inch 

 scale. The country traversed was, generally speaking, of the very 

 hardest that could be found for the plane-table. With the exceptions of 

 the mountain country extending 80 miles west of T'ai-yiian, the range 

 immediately to the west of Ku-yiian Chou in Kansu, and the mountains 

 in the vicinity of Lan-chou Fu, the whole terrain was covered with loess. 

 Such a surface with its bare, rounded hill-tops — all usually of the same 

 level and rarely varying more than 100 feet — affords no salient features 

 suitable for triangulation ; whilst the deep gullies cutting it in every 

 direction make access difficult. In addition the valleys of even the 

 larger water- courses offer but very little better fields of view, being never 

 more than a mile, and but seldom more than a few hundred feet, wide. 

 In every way, a loess country taxes the abilities and patience of a surveyor 

 to the utmost, and I do not believe that too much praise can be given to 

 Hazrat Ali for the consistent and conscientious accuracy of his work. 

 Never was he seriously out but once. At Yii-lin Fu the work checked 

 out to within 5 ' of Long, and 4' of Lat. ; but our last three marches before 



131 



