12 NATURAL SCIENCE. July, 



westerly — , and that if the total of the easterly and westerly devia- 

 tions be taken one from the other and the result divided by the total 

 number of observations, this will give the constant error of the 

 compass* on each point. This co-efficient is named A. In a ship 

 where the compass is placed exactly in the fore-and-aft line, and 

 where the iron is evenly distributed, there will be no A. This was 

 found to be practically the case in the " Challenger." By swinging 

 the ship outside Bermuda by azimuths of the sun the total error of 

 the compass, i.e., variation + deviation, was obtained on each point. 

 Taking the algebraical mean of all the observations, the constant 

 error on each point is determined. This error is the variation + the 

 A. In a ship, such as the " Challenger," with practically no A, the 

 result is the variation. 



This swinging in deep water proved so satisfactory that the system 

 was adopted at many stations throughout the voyage, thus not only 

 furnishing base-stations, independently of the land, for correcting the 

 observations made between the base-stations, but also serving to 

 ascertain whether there was any magnetic attraction on shore. Since 

 the "Challenger's" voyage all vessels instructed to obtain magnetic 

 observations have been ordered to adopt the system introduced by 

 the Expedition at Bermuda. 



2, Although, before the "Challenger" started, many ocean 

 soundings had been obtained, especially in the North Atlantic 

 between England and America, and in the Mediterranean, along the 

 routes chosen for submarine cables, most of the earlier soundings 

 were of somewhat doubtful value, as the means at the disposal of the 

 first investigators were insufficient, and the method of obtaining 

 accurate depths, when the soundings exceeded 2,000 fathoms, could 

 only be ascertained after much experience. Depths of 7,000 to 

 10,000 fathoms were asserted to have been obtained, and certain text- 

 books quoted these statements as facts. 



By the investigations made in the " Challenger," the contour 

 lines at each 1,000 fathoms of depth were for the first time drawn 

 with some degree of accuracy, and it was shown that the great depths 

 formerly reported had been much exaggerated. Soundings of up- 

 wards of 3,000 fathoms were seldom found, and the deepest cast was 

 4,475 fathoms in the neighbourhood of the Mariana Islands. The 

 investigations undertaken at the same time by the United States 

 confirmed the results obtained in the " Challenger," which the sub- 

 sequent experience of twenty years has shown to be correct ; the 

 many soundings obtained by other vessels since the expedition 

 returned to England have not altered in any material degree the 

 contour lines originally drawn by the officers in 1876, or resulted in 

 the discovery of any depth exceeding five statute miles. Conse- 

 quently, 4,500 or 4,600 fathoms may be fairly accepted as the 

 extreme oceanic depth. 



3. The laborious work of obtaining ocean temperatures at a 



