156 FISCHER: LENGTH STANDARDS AND MEASUREMENTS 



sive. The other method designed by Professor Woodward con- 

 sists in measuring the base with 100-meter steel tapes, equipped 

 with special stretching devices and standardized on the kilometer 

 measured in terms of the 5-meter iced bar. Very rapid progress 

 could be made in measuring with these tapes and the prepa- 

 ration of the ground was reduced to a minimum. 



The tapes used were ordinary steel tapes and measurements 

 had to be made at night to avoid the direct radiation from the 

 sun; but the results were nevertheless so satisfactory that the 

 method at once attracted attention, the greatest objection to 

 the method was the necessity of carrying the iced bar in the 

 field for the purpose of laying out a hundred meter standard 

 with which the length of the tapes could be frequently checked. 

 While even then the tape method possessed certain obvious 

 advantages the base bar method still had its advocates. After 

 the measurement of the Holton, Md., and St. Albans, W. Va., 

 bases two duplex base bars designed by Wm. Eimbeck were 

 constructed by the Coast and Geodetic Survey and a base line 

 was measured with them near Salt Lake City, Utah, in 1897, 

 with a probable error of 1 part in 800,000.^ A duplex bar con- 

 sisted of two concentric brass tubes in the inner of which were 

 mounted a steel and a brass measuring bar. Holes were pro- 

 vided in both the outer and inner tubes for the insertion of ther- 

 mometers, and the apparatus was so constructed that the inner 

 tube could be rotated 180° and thus equalize the temperature 

 of the brass and steel components if one side of the apparatus 

 should be more exposed to the direct radiation of the sun. Mr. 

 Eimbeck's idea was that if the bars were standardized under a 

 sufficiently wide range of temperature the length of either the 

 brass or steel component could be determined as a function of 

 the difference in the lengths of the bars, and consequently the 

 apparatus itself could serve as a thermometer. Subsequently 

 in measuring a base line the difference in the length of the base 

 as given by the two components would furnish data from which 

 the length of the base could be computed. As a precautionary 

 measure the temperature was observed when standardizing 



' Appendix 12, Coast and Geodetic Survey Report for 1897. 



