304 CARNEGIE INSTITUTION OF WASHINGTON. 



Some experimental work was done also on cloud-and-haze horizons at var- 

 ious altitudes, but the difficulty with such observations is to determine the 

 altitude of the horizontal plane. A dip-measurer was used very successfully 

 to determine this altitude, the results giving an error of ±3' to ±5' for a 

 single determination. 



During a flight from Langley Field to Washington and return, observations 

 were made with the preliminary artificial horizon, using a small pocket- 

 sextant. On the trip to Washington, which occupied 2 hours, 7 position-lines 

 were determined, each based on 10 observations, and all work of computation 

 and plotting of this line was done in the air without previous preparation, 

 using Aquino's tables. The average time for each position-line, including 

 observations, computations, and plotting, was 9.5 minutes. The average error 

 of each fine was =*= 13' of altitude. On the return trip, which occupied 1 hour 

 and 45 minutes, 9 position-lines were determined and plotted with an average 

 error of ± 17' of altitude. This increased error was due to the irregularity of 

 motion due to "bumps," the airplane falling 50 feet in a single "bump" quite 

 frequently during observations. The results in a set of 10 observations ranged 

 over 3° at times. 



If such results as those just described can be obtained with preliminary 

 apparatus and on small airplanes, it is quite certain thait the errors can be 

 materially decreased with more refined instruments and larger airplanes. 



As previously mentioned, if only one celestial object is available, such as the 

 Sun, then to determine completely the position the azimuth, as well as the 

 altitude, must be measured. The experimental work along this line was inter- 

 rupted, but the preliminary results were very encouraging. During the flight 

 when azimuths were first measured, 80 observations were taken and the error 

 of groups of 10 was =*=0?6, the mean error of all being ±0?3. The mean 

 difference between deviations as determined on the ground and those deter- 

 mined in the air was only 0?1. These observations were made with an azimuth- 

 card the least graduation of which was 5°, the single degrees being estimated. 

 Some further observations were made by J. A. Fleming of the Department; 

 during his first flight he made 110 observations with the above-described 

 instrument. The average time for 10 observations was 1| minutes and the 

 probable error of a single determination was ±0?4. 



As to instruments, a light sextant is desirable, but no difficulty was experi- 

 enced in using the ordinary sextant. A special protractor was designed to 

 facilitate the rapid plotting of the line of position. A chart-holder and navi- 

 gator's case was designed and constructed by the Department. 



Several flights were made at night to determine the practicability of obser- 

 vations on the stars. The results showed that observations could be made at 

 night with the same ease and accuracy as during the day. The advantage of 

 night observations is the possibiUty of always having two objects on which to 

 observe. 



As to results as far as the experimental work was carried out, if two celestial 

 bodies are available for observation, a position should be determined within 20 

 minutes of time and to an accuracy certainly of =»= 15 miles. Where only a 

 single celestial body is available and where both altitude and azimuth are 

 determined, the resulting position may be in error from 30 to 60 miles. These 

 figures should be very materially reduced with refined instruments and large, 

 stable airplanes. 



Report on the 1918-1919 experiments in magnet-photography. S. J. Barnett. 



The experiments of R. Colson^ and W. J. RusselP have shown that objects 

 of zinc, brass, wood, rosin, and many other substances when placed on or near 



' Comptes Rendus (Jan. 1897). 



Troc. Roy. Soc, vol. 61, p. 424 (1897); vol. 63, p. 102 (1898); vol. 64, p. 409 (1899); Phil. 

 Trans. Roy. Soc, B, vol. 197 (1904). 



