376 CARNEGIE INSTITUTION OF WASHINGTON. 



to build up the columns bj^ cementing plates of glass together. No 

 long rods of glass tested were sufficiently homogeneous to be available. 



An earlier report on the coronas of cloudy condensation shows that 

 on using homogeneous mercury light, the maximum of green illumina- 

 tion is found alternately in the disk of the corona and in the first green 

 ring, as the particles increase in size successively. This superposition 

 of an independent periodicity contains the key to the optics of coronas; 

 experiments were therefore made at some length to utilize these obser- 

 vations for the practical case of the green coronas and white light. 



Finally, the experiments on the diffusion of gases through liquids 

 were continued, particularly in the direction of diffusions through 

 solutions of different strengths and composition. Apparatus was also 

 installed for observing the diffusions of a gas through the unequal 

 columns of a liquid, in a sealed tube, in the lapse of years. 



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

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

 Year Book No. 12.) 



The general plan of the investigation of evaporation is to consider 

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

 income, outgo (other than evaporation), and change of content of the 

 lake day by day as accurately as possible, and from these evaluations 

 to determine the daily evaporation and its relation to meteorological 

 conditions. The change of content of each lake from day to day is de- 

 termined by the change in mean level of the water-surface. The level 

 on Lake Michigan-Huron, for example, is observed by automatic gages 

 at Harbor Beach, Mackinaw City, and Milwaukee, maintained by the 

 survey of the Northern and Northwestern Lakes, which continuously 

 record the water-level at these points. To derive from any one of these 

 records at a single point the fluctuation of the mean surface of the whole 

 lake, with the degree of accuracy necessary for this investigation of 

 evaporation, one must be able to correct the record at the gage for the 

 effects of winds and of barometric changes. The winds tend each day to 

 lower or to raise the level of the water-surface at the gage independently 

 of any change which may have taken place in the total content of the 

 lake. The principal attack of the past year in the least-square solution 

 has been upon the problem of determining with sufficient accuracy the 

 local effects at the gages of wind changes and barometric changes, and 

 of correcting for these local effects in the computations from which the 

 evaporation constants are to be derived. This particular problem has 

 proved to be more difficult of satisfactory solution than it was supposed 

 it would be. Progress is being made, though but slowly. The evalua- 

 tion of wind effects and barometric effects bids fair, however, to be an 

 unexpectedly valuable by-product of the investigation. 



At this stage of the investigation, when it is little more than half 

 completed, as measured by the cash expenditure, it is premature to 



