DEPARTMENT OF TERRESTRIAL MAGNETISM. 335 



Among the greatest difficulties to be contended with in work at sea is that 

 of securing good insulation. It has consequently been arranged that in all 

 measurements the results are, as far as possible, independent of this require- 

 ment. In the measurement of potential-gradient, very serious errors may 

 arise because of faulty insulation, if the ordinary method involving the use 

 of some form of collector is employed. For this and other reasons, during the 

 third cruise of the Carnegie, a form of instrument was employed of the type 

 described by the author in a previous paper. ^ This instrument has lately been 

 modified somewhat in design in order to render it more convenient for manipu- 

 lation under ship conditions, and as at present used on the fourth cruise of the 

 Carnegie, it consists of a brass tube fixed at one end to an axle so that it can 

 rotate in a plane containing the fore-and-aft line of the ship. The axle is fitted 

 to the stern rail of the ship, and the free end of the brass tube carries a gauze 

 disk made somewhat in the form of a parasol. The handle by which the 

 rotation is brought about is insulated from the axle, and the latter is itself 

 insulated from earth by causing it to work in brass tubes fixed into their 

 supports with sulphur insulation. The axle is connected by a thin wire to a 

 Wulf bifilar electroscope, the wire and the axle being in the same line. It is 

 arranged that when the brass tube is vertical and the parasol attachment is 

 downward, the electroscope system is earthed. On rotating the tube to some 

 other position fixed by a stop, a deflection is obtained in the electroscope which 

 is proportional to the potential-gradient. Insulation difficulties are entirely 

 overcome, since the leak occurring during the turning of the handle from one 

 position to the other is negligible. The sensitivity is considerable, and it is 

 easy to arrange that deflections amounting to the whole scale-length are 

 obtained for the normal value of the potential-gradient. 



In addition to the potential-gradient, the atmospheric-electric observa- 

 tions comprise measurements of the ionic content and conductivity for posi- 

 tive and negative ions, the radioactive content of the atmosphere, and the 

 number of ions produced per cubic centimeter per second in a closed vessel. 

 A small observatory has been erected on the Carnegie, abaft the after dome, 

 and it has been arranged that, as far as possible, the instruments are perma- 

 nently mounted in this observatory, only those portions projecting through 

 the roof which have to be exposed to the air. In this way, since the inside 

 of the observatory is always above the dew-point for the air outside, a great 

 deal of trouble arising from deposition of dew on the apparatus is avoided. 

 The ionic content is measured with the instrument described by the author in 

 a former paper.^ The whole instrument is supported on a gimbal, and only 

 the open end of the receiving tube projects above the observatory. The con- 

 ductivity is measured on the general principles of the method devised by 

 Gerdien, but several modifications have been made in the apparatus. The 

 concentric cylinders between which the air passes are mounted on the roof of 

 the observatory, so that they can be rotated in such a way that the open end 

 of the apparatus faces the wind. The electroscope system hangs from a 

 gimbal inside the house, and the fan is driven by a small electric motor. This 

 latter arrangement makes it possible to obtain a much greater rate of flow of 

 air, with the result that higher potentials can be applied to the central system 

 and greater sensitivity attained. It is further arranged that the air is blown 

 down into the observatory, whence it departs slowly through the windows. 

 There is thus less chance that ah' which had been robbed of its conductivity 

 by the apparatus will reenter the apparatus again. 



The Elster and Geitel stretched-wire method for measuring the radioactive 

 content of the atmosphere has several disadvantages which have been dis- 

 cussed in a former paper.^ Even when the instrumental uncertainties have 



iTerr. Mag., vol. 19, pp. 182-185, September 1914. ^Ibid.. pp. 171-175. ^Ibid., pp. 

 176-182. 



