98 LECTURES 



bar, which is supported parallel to the iron and brass bars', and which 

 is used for making the contact of the two compound bars during the 

 operation of measuring. On the opposite end of each pair of compound 

 bars a corresponding small auxiliary rod is placed, and also a sector 

 for measuring the slope at which the whole bar is placed. 



Two measuring bars being made exactly alike as above described, 

 are placed in line upon the ground and adjusted so that the small 

 rods at the end shall be at the same height, the lever end of one 

 opposite to the sector end of the other ; the rod at the sector end of 

 the foremost bar is so attached to a small spirit level that by a slight 

 pressure the bubble will be brought to the centre of the level. The 

 rod on the end of the other bar, which we have described as abutting 

 against the lever of compensation, has a sliding motion, which is 

 resisted by a delicate spiral spring ; the end of this little rod is termi- 

 nated by a smooth agate plane surface, and the end of the opposite 

 rod by an agate knife edge. The contact is made by bringing these 

 ends together, the spirit level insuring the same degree of pressure in 

 every case. 



This is the general description of the apparatus ; the sector is used 

 for measuring the slope of the ground along which the line extends, 

 the resulting angle giving the means of reducing the measurements 

 to horizontal lengths. Each of the compound bars are enclosed in 

 large tin tubes, to protect them from the air and weather and sudden 

 change of temperature, and in use these tubes are supported on trestles 

 made for the purpose. 



There are many interesting details of construction which can only 

 be explained by reference to drawings or to the apparatus itself'^ a full 

 description of which may be found in the Coast Survey Report of 1854. 



With the apparatus we have described a distance of a mile may be 

 measured in one day, and a distance of ten miles may be measured 

 with no greater error than a fraction of an inch. The length of each 

 of the compound bars is kept constantly compared during the meas- 

 urement of a base line with a standard metallic bar, the length of which 

 at a given temperature is known. 



The necessity for the greatest attainable accuracy in such opera- 

 tions is apparent, when we remember that any error in the base line 

 will be greatly multiplied in the triangulation depending upon it. 

 The true cause of the erroneous results of the measurement of the 

 first French arc was an error in the base line of Picard, which was 

 subsequently discovered and corrected as before stated. 



A base was measured on Kent island, in Chesapeake bay, to verify 

 the triangulation between a base on Fire Island (on the south side of 

 Long Island) and the Chesapeake. The shortest distance between 

 these base lines is two hundred and eight miles, but the distance 

 through the triangulation is three hundred and twenty miles. The 

 Kent Island base, ilve miles and four-tenths long, was tirst calculated 

 from the triangulation, which embraces thirty-two triangles. The 

 computed length differed from the length found by measurement only 

 four inches. 



On some parts of the coast, where it is desirable to commence the 

 work of the survey in advance of the main triangulation, preliminary 



