406 CARNEGIE INSTITUTION OF WASHIGTON. 



In another group of experiments, methods were tested as to the 

 degree with which they might suffice to determine the acceleration 

 of gravity when a fixed installation is not assured. In the first set, 

 this constant was referred to the torsion of a steel wire, the readings 

 being made by the interferometer with an accuracy of 1 part in 

 100,000. It was found that the viscosity of the wire could be ade- 

 quately allowed for, but that the thermal coefficients, both of rigidity 

 and of viscosity, are so large, relatively, as to make the reductions 

 untrustworthy unless metals of small thermo-elastic changes like the 

 Guillaume alloys are used. The experiments are still in progress. 

 In the second set of experiments (also in progress) the acceleration 

 of gravity was expressed in terms of the pressure of the air imprisoned 

 in a Cartesian diver when floating at a given level. The limit of 

 accuracy here should be about 1 part in 10,000. Again, however, 

 changes of temperature produce a relatively enormous error, inasmuch 

 as the air is dissolved or released by the v/ater envelope in a way that 

 admits of no adequate control. Only by keeping such an instrument 

 in a thermostat could interpretable results be obtained. The experi- 

 ments have been continued thi'oughout several months, both under 

 constant and under variable temperature conditions; for it seemed 

 that an apparatus for the purpose in question, which admits of rough 

 handUng, of use on board ship, for instance, should not be hastily 

 abandoned. 



Hayford, John F., Northwestern University, Evanston, Illinois. Investiga- 

 tion of the laws of evaporation and stream flow. (For previous reports see 

 Year Books Nos. 12-16.) 



As stated in the original application for the grant under which this 

 investigation is being made, "the ultimate object is to obtain a much 

 better formulation than the engineering profession now has of the laws 

 governing the amount of the stream flow." It appeared probable 

 that progress toward this end is conditioned upon first securing a 

 better knowledge than is now available of the laws of evaporation from 

 large water surfaces and land surfaces. Hence, the immediate object 

 of the investigation was stated to be "to determine the laws of evap- 

 oration from a large water surface with much greater accuracy than they 

 are now known." For this purpose it was proposed to consider each 

 of the Great Lakes in turn as an evaporation pan and to evaluate the 

 change of content, the income, and the outgo, including evaporation. 

 As the investigation progressed, it soon became evident that it is 

 necessary to determine the effects of barometric changes, and of winds, 

 in disturbing the daily level of the lake at each recording gage which 

 is used to determine the elevation of the lake surface for the day. 

 The total content is known if the elevation of the mean surface of 

 the lake is known. Still later, it became apparent that it is necessary 



