26 TRANS. ST. LOUIS ACAD. SCIENCE. 



expansion for i° Fahr. in one tape-length is 0.025 inch. The 

 probable error in measurement for a distance of 36 tapes 

 ( = 10,767 feet) is ±0.0719 inch, or i in 1,180,000; and that 

 for the determination of a tape-length (=-299.079 feet, since the 

 standard tape is from zero (o) to the 299-foot graduation of the 

 tape) is ±0.004 ^"-i °^ ^ ^" 897,237 parts of a tape-length. 



This latter probable error combined with that before given for 

 the several bases would, of course, increase each to more than i 

 in 1,000,000. But it is believed that the standard tape may be 

 determined within i in 2,000,000 by direct comparison with the 

 primary base apparatus, and that the probable error of a base-line 

 measurement will then be less than i in 1,000,000. 



