Atmospheric-Electric Observations, 1915-16 379 



consequently insulation is a much more important factor than is the case with low-resistance 

 batteries. Thus, for example, if the terminals of such a Kriiger battery are joined through 

 a resistance as high as 10^ ohms, the potential is lowered by 0.1 volt. While it should be 

 possible to construct batteries of this type to give a very high degree of constancy, those at 

 present supplied are apt to be defective. For many purposes, e.g., maintaining the potential 

 of the needle in a Dolezalek electrometer, they are excellent, but in an instrument which 

 responds to momentary fluctuations of small amount they do not appear to be very satis- 

 factory, except under the very best conditions of internal and external insulation. For 

 this reason small batteries of dry cells of low resistance have been used. These can be 

 bought conveniently in 100-volt units, but those in use on the Carnegie have all been made 

 up specially, and are free from fluctuations of the kind cited. Of course, the essential 

 requirement in the use of the Einthoven electroscope is that if the potentials of the plates 

 fluctuate they shall do so by equal and opposite amounts. The employment of a battery of 

 reasonably low resistance enables us still further to meet this requirement by connecting 

 the battery and the two plates to the two ends of a megohm, the mid-point of which is 

 earthed. In this way it has been found possible in the laboratory to work with the instru- 

 ment sensitive to 1,000 divisions per volt. 



Since good insulation is essential to such measurements as are here discussed, it is 

 desirable that those parts of the apparatus which contain the insulating materials shall be 

 well protected. To this end a small observing-house has been erected on the Carnegie. 

 In this observatory the greater part of the apparatus is mounted in such a way that the 

 electroscopes, batteries, and other parts which requu-e good insulation are permanently 

 inside, and as an additional precaution the essential parts are protected by drying-bulbs. 

 Only those portions of the apparatus which must be exposed to the open are so exposed; 

 these pass through holes in the roof of the observing-house, and are protected by suitable 

 covers when the apparatus is not in use. One of the cliief advantages attending this 

 arrangement Ues in the fact that, in so far as the temperature of the mside of the house 

 is always above the dew-point of the air outside, condensation on the insulating parts is less 

 likely to occur than if the apparatus were exposed to the open air. This point is of con- 

 siderable importance, especially in observations taken towards evening, for, if the apparatus 

 is exposed on the open deck, it frequently becomes so wet that observations are impossible. 



Another point in favor of the observing-house lies in the fact that it results in a great 

 economy in time. The instruments and the various appurtenances incidental to their use 

 are left permanently connected up, so that when the observer desires to start observations, 

 he has only to remove the covers and conmnence work. The principle of reducing to a 

 minimum the time wasted in setting up apparatus before each set of measurements has been 

 adhered to throughout. In this way it has become possible for one observer, with a little 

 assistance from another, to obtain within a period of 2^ hours the meteorological data and 

 observations of all the quantities from 1 to 5, page 377, including the determination of leakage 

 corrections and the standardization of the various electroscope systems. As will be seen 

 from a glance at the tabulated results in Tables 79-83, it has been possible to obtain complete 

 sets of observations on almost every day, some of the sets being obtained when the ship was 

 rolling through an angle as large as 30°. Further, it is particularly gratifying to notice 

 that, even in the case of the sub-Antarctic cruise, Table 83, where precipitation of some 

 kind or other was recorded on 100 out of 115 days, the table of observations is almost 

 complete. Pictures of the atmospheric-electric house are shown in Plate 22, Figures 1 and 2. 



