April 28, 1905.1 



SCIENCE. 



673 



A CONNECTION BY PRECISE LEVELING BETWEEN 

 THE ATLANTIC AND PACIFIC OCEANS. 



On October 4, 1904, a Coast and Geodetic 

 Survey party, running eastward from Seattle, 

 Wash., met a similar party, running westward, 

 at Hunts Junction, in the southeastern part 

 of Washington. The party running from east 

 to west had started in the beginning of the 

 season from a bench mark of which the eleva- 

 tion had been fixed by a long line run during 

 several seasons and extending westward from 

 the precise level net, composed of many cir- 

 cuits, which covers the eastern half of the 

 United States. As far west as Norfolk, Nebr., 

 the elevations in this net had been checked 

 by completed circuits of precise leveling of 

 the highest grade of accuracy. The joining 

 of the two lines at Hunts Junction completed 

 the first connection by precise leveling be- 

 tween the Atlantic and the Pacific. 



The discrepancy developed at the junction 

 was .615 ft. (=187.5 millimeters), the Pacific 

 being apparently higher than the Gulf of 

 Mexico and Atlantic. 



The old question at once arises: is the 

 Pacific at a different elevation from the At- 

 lantic? The discrepancy of .615 ft. must be 

 due either to errors in the tidal observations 

 which furnished the connection with mean 

 sea level, or errors in the leveling, or to a real 

 difference in the elevation of the mean sea 

 surface at the points at which the tidal ob- 

 servations were made. 



The three principal connections with sea 

 level concerned are at Sandy Hook, near New 

 York City, at Biloxi, Miss., and at Seattle, 

 Wash. Six years of tidal observations were 

 taken at Sandy Hook and five years at each 

 of the other points. The range in the six 

 annual means at Sandy Hook was .322 ft., 

 and of the five annual means at the stations 

 at Biloxi and Seattle was, respectively, .100 

 and .204 ft. These ranges are not sufficient 

 to account for the discrepancy of .615 ft. 



The shortest line of leveling of the highest 

 grade of accuracy from Seattle to Sandy Hook 

 is 4,600 miles (7,400 kilometers) ; to Biloxi, 

 3.500 miles (5,700 kilometers) ; and to Norfolk, 

 the point at which the line to the westward 

 leaves the thoroughly checked portion of the 



precise level net, is 2,000 miles (3,300 kilo- 

 meters). 



If it is assumed that the discrepancy (.615 

 ft.) is simply an acctunulated error in leveling 

 and that the rate of accumulation is uniform 

 between Seattle and Biloxi, it is at the rate 

 of one foot in 5,700 miles (.033 millimeters 

 per kilometer). Even if it is assumed that 

 the accumulation all occurred between Seattle 

 and Norfolk, it is at the rate of one foot in 

 3,300 miles (.057 millimeters per kilometer). 

 This is an extremely small error of leveling. 



Another test to determine whether the dis- 

 crepancy is a possible error of the leveling 

 may be applied. The probable error of the 

 elevations at Seattle, as carried westward from 

 the Gulf and Atlantic, the computation being 

 based upon the discrepancy developed in the 

 circuits in the eastern part of the United 

 States, was ± 76 millimeters. The actual dis- 

 crepancy is two and one half times this. Ac- 

 cording to the doctrine of chances, such a 

 discrepancy, two and one half times the prob- 

 able error, should occur about once in ten 

 times. 



Therefore, it is not safe to make the state- 

 ment that the Pacific is higher than the Gulf 

 and Atlantic ; the extremely small discrepancy 

 being well within the possible limits of error 

 of the precise leveling alone, even though it 

 be assumed that the leveling in question is of 

 as high a grade of accuracy as any yet done 

 anywhere in the world. 



One is apt to associate observations of such 

 extreme accuracy as this precise leveling with 

 slow progress and high cost. It is interesting, 

 therefore, to note that three thousand miles 

 out of the thirty-five hundred between Seattle 

 and Biloxi have been leveled since the begin- 

 ning of the field season of 1899, and that the 

 average rate of progress, during the period 

 1900-1904 (a total of 3,900 miles), with the 

 new type of precise level now in use in the 

 Coast and Geodetic Survey, was 64 miles of 

 completed line per month, for each observer, 

 and that the average cost, per completed mile, 

 was $10, including salaries, transportation 

 and bench marks. Each completed mile was 

 leveled over at least twice, and, in some cases, 

 four or more times. This rate of progress is 



