DEPARTMENT OF TERRESTRIAL MAGNETISM. 267 



was also taken on the Galilee. The error in the height caused by an inclina- 

 tion of 10° is 0.3 foot for a height of 24 feet on even keel, and would cause 

 an error less than O'.l in the deduced refraction. An error of this size is not 

 likely to occur as one can estimate even-keel position closer than 10°, but 

 errors caused by the rolling and pitching of the vessel will always be in the 

 same direction and can not be entirely eliminated by the mean of many 

 observations. 



As the ship rides the waves and drops into the troughs of the sea, the 

 vertical height of the instrument above mean sea-level changes. The visible 

 horizon is delineated as a straight line by the crests of the distant waves, if 

 the sea is not too high and irregular. Accordingly, observations are made 

 only luhen the vessel rides the waves, on the supposition that the waves are all 

 of the same height and that the instrument will be raised a like amount. This 

 precaution, however, does not eliminate the whole error, for when the vessel 

 rides the waves, her load water-line is not tangent to the crests of the waves, 

 but always below them. The adopted height, therefore, is toohigh by the 

 vertical height of the waves above the load water-line. 



Observations on the Galilee were made bj^ one observer, using dip-measurer 

 No. 4048. Two readings were always made, sometimes 4, 6, and 8, one-half 

 with instrument erect, the other with inverted position. Those made on 

 the first three cruises of the Carnegie were obtained with the same instrument 

 (No. 4048) by different observers. Another instrument, No. 4031, was 

 courteously loaned by the United States Coast and Geodetic Survey during 

 the first two passages of the fourth cruise, after which it was replaced by 

 instrument No. 5490. Every determination of the dip-of-horizon on the 

 fourth cruise of the Carnegie, therefore, is the result of 2 instruments, and 

 observers were changed every passage or every two passages. 



The data recorded include the date, latitude, longitude, approximate local 

 time, height of instrument above sea, aspect or definition of the horizon, 

 wind direction and force, air and water temperature, barometer reading, the 

 directions sighted, cloud notes, and direction of the Sun. 



In the adjustments that have been made only those observations were 

 used that were obtained under good conditions of the horizon. In the last 

 few years German investigators of the refraction of the horizon have intro- 

 duced a temperature coefficient in the adjustment of their observations. 

 The temperature of the air can be measured on the ship, but there is no 

 means of measuring the temperature of the air in contact with the sea at the 

 horizon, so it is assumed that the temperature is the same as the surface of 

 he water at the ship. Both temperatures as ordinarily observed may be 

 subject to errors resulting from the methods used. The water temperature 

 is obtained by immersing a thermometer in a bucket of sea water umnediately 

 upon hauling it aboard. This is a rather crude procedure, but more precise 

 methods are not likely to give better results in the adjustment until larger 

 systematic errors are eliminated. 



The Galilee observations, and also those of the fourth cruise of the Carnegie, 

 were adjusted to Koss's equation,* which for a constant height of eye, may 

 be written 



D-D' = x+y ita-tr,) 



The results are given in the original paper. 



The adjustments of the observations of first, second, and third cruises of 

 the Carnegie were deferred, pending the result of the adjustment of the fourth 

 cruise, since the observations on this cruise having been made with two 



*Meyer, H. : Kimmbeobachtungen, Ann. Hydrogr, Berlin, v. 34, Heft 9, 1908, p. 438. 



