DEPARTMENT OF TERRESTRIAL MAGNETISM. 303 



The experimental work carried on by the Department of Terrestrial Mag- 

 netism at Langley Field, Virginia, in the development of methods and instru- 

 ments for determination of geographic position of airplanes by astronomical 

 observations, consisted in making sextant observations in the air, to try out 

 the feasibility of different artificial horizons, and in the study and use of different 

 methods of rapid calculation of the position hne. The first apparatus used was 

 a preUminary instrument consisting of a mirror mounted on small gimbal 

 rings with a counter-weight suspended in oil to damp the vibrations. The 

 results obtained with this instrument, using different types of sextants, gave 

 an average error for a single observation of ±25'; the error of a group of 10 

 observations was ± 12'. During the second flight 59 observations were made, 

 giving an average error of ±12', rejecting only 3 which were obviously in 

 error; all the others were less than 36' in error. 



A more accurate instrmnent was next constructed by the Department and 

 used in the experimental work at Langley Field. The mirror was made of 

 speculum metal and the gimbals were mounted on steel-ball bearings. The 

 results obtained with this instrument gave an error for a single observation 

 of ±15' to ±29' and an error for a group of six observations of ±7' to ±12'. 



Through efforts of the Department a sextant with an artificial level-bubble 

 attachment was secured from Professor R. W. Willson, of Harvard University. 

 With this instrument Dr. H. N. Russell obtained results which gave an error 

 for a single observation of ±12' to ±21' and an error for a group of five 

 observations of ±6'. The experience with this sextant showed material 

 improvement over the mirror-and-gimbal horizon, both in ease and conven- 

 ience of handling as well as in rapidity and accuracy. 



After the altitude is measured, the next step is to make the calculations and 

 to draw the position-hnes on the chart, or by some method to determine the 

 position of the observations. Several methods were investigated. First, the 

 tables devised by Radler de Aquino, of Brazil, were used, the computation 

 with these tables requiring about 3 minutes and the plotting of the position- 

 line about 2 minutes. These tables are published by the United States Hydro- 

 graphic Office, PubUcation No. 200. Second, differentmethods of precalcula- 

 tion were studied. The best of these precalculation methods seems to be that 

 outlined by G. W. Littlehales, where some central position on a chart is taken 

 as the assumed position of the observer and tables are precalculated on this 

 basis. If Lambert's conformal conic projection map is used, the arcs of great 

 circles appear as straight lines and the altitude-intercept may be very large 

 without appreciable error, so that one assumed position may be made to cover 

 a wide extent of territory. Third, an instnmient called the "hne of position 

 computer," designed and loaned to the Department by Professor C. L. Poor, 

 of Columbia University, was used. This instrument is probably the best that 

 has been devised up to date for calculating the position-line in the air. It is 

 made on the principle of a circular slide-rule; both the altitude and the azimuth 

 can be calculated in less that 1.5 minutes of time and to an accuracy of 2 

 minutes of arc. 



Most of the experimental work in computing and plotting positions in the 

 air was done by using Aquino's methods. With his tables, if both the altitude 

 and the azimuth are observed, a previous knowledge of the dead-reckoned 

 position is not necessary, except to determine the magnetic declination of the 

 place of observation. With the natural horizon an observation was made, 

 computed, and the position-lines plotted in 4.1 minutes of time; the mean 

 error of 4 positions thus determined was ±1'. This gives some idea of the 

 accuracy which can be obtained in making sextant observations where the 

 uncertainty of the horizon is eliminated. 



