REPORT ON GEOPHYSICAI, LABORATORY 189 



proceed mucli more rapidly and effectively in a laboratory of fairly 

 constant temperature. Some physicists, indeed, maintain that it is 

 sufficient to attempt constant temperatures only within pieces of 

 apparatus, but in this view I cannot agree. A standard bar, for 

 example, may be measured in a case kept nearly at constant tem- 

 perature by circulating liquids or by electric resistance ; but this 

 temperature depends in part on the radiation of the case, and this 

 on the temperature of the apartment. Again, the accuracy of gal- 

 vanometers, and all similar apparatus, is greatly promoted by a 

 substantially uniform temperature. 



Special Difficulties in America. 



In the eastern United States the natural atmospheric temperature 

 varies so enormously and so rapidly as greatly to interfere both with 

 the accuracy of instruments and the capacity of observers. Heat 

 flows from piers to instruments, or in the reverse direction, so fast 

 as to be most disturbing and wholly incalculable, and this flux 

 renders the more minute measurements most uncertain. Thus, 

 even more than in Europe, it is desirable that an American labora- 

 tory of the highest class should be isothermic. 



Avoidance of Heat Flux. 



The maintenance of a temperature of 20° C. in a laboratory is at- 

 tended with other difficulties besides that of supplying cool or warm 

 air. In the laboratory of the Normal Aichungs Kommission it has 

 been found that the flow of heat downward through the piers is a 

 verj' disturbing factor at the best, and I there received the excellent 

 suggestion that the exposed portion of the piers should be made of 

 metal, in order that it should readily take on the temperature of 

 the observing room. The metal plate should stand on three points, 

 and the stratum underlying them should be of the most non-con- 

 ductive material which can be found. Hard magnesia brick is 

 almost ideal in this respect, and I suggest that the masonry of the 

 piers be faced with this material. 



Notes on Ventilation. 



Insulation, excepting in very deep subcellars, is not sufficient to 

 establish uniformity of temperature, even in Europe. In the east- 

 ern United States, the annual variation being much greater than on 

 the other side of the Atlantic, insulation is still less effective. 



