PHILOSOPHICAL SOCIETY OP WASHINGTON. 



35 



There seems nothing in these figures to require a discontinu- 

 ance of the use of wooden rods. 



The Kanab base was divided into 44 sub-equal sections, and 

 two independent measurements were made of each section. From 

 the 44 resulting discrepancies the probable error of the whole 

 measurement was computed and found to be .09 ft. or 1-250,000 

 of the whole length of the base. When a similar test has been 

 applied to the steel tape, the relative value of the two apparatus 

 -can be better judged. 



Unfortunately the fraction 1-250,000 does not represent the 

 total error of the base line as determined, but only that portion 

 •of the error which depends on the manipulation of the wooden 

 rods; another portion introduced in the comparison of the rods 

 with the steel standards is much greater. The latter affects the 

 value of the Kanab base line, but is independent of the value of 

 the apparatus. 



2rf. A proposed New Apparatus and Method. 



In some methods of base-measurement a single unit of length 

 is appliea repeatedly. A record is made each time of the position 

 of the advance end, to which record the rear end is afterward 

 applied. This is the usual practice when a chain or tape is used. 

 In other methods two unit measures are used in alternation. One 

 remains stationary while the other is carried forward, and the 

 rear end of the moving measure is applied to the advance end of 

 the stationary. This is the usual practice when wooden or metal- 

 lic rods are employed. It is now proposed to combine the 

 method of the single unit and record with the use of the rod. 



Where two rods have been used a simple contact of their ends 

 has been found impracticable, and an observation has usually 

 been substituted. Where a single unit has been used the record 

 has been of the nature of an observation and subject to personal 

 equation. It is proposed to replace the observations by an auto- 

 matic record. 



In the use of two rods it is necessary, for each added unit, to 

 «arry forward two tripods and adjust them in height. It is pro- 



* U. S. Geol. and Geog. Survey of Colorado, 1876, p. 280. 



