A NEW EVAPORATION FORMULA 5 
Part I of this publication deals with the investigation of the laws of evapora- 
tion from observations on Lakes Michigan-Huron and Superior. It shows how the 
investigation was made; how the final evaporation formula was derived. It 
necessarily overlaps, to a certain extent, Publication No. 317, and represents the 
outcome of one application of the results of that part of this investigation contained 
in that publication. 
Using the evaporation formula derived from observations on the Great Lakes, 
the attack on formulating the laws of stream-flow was begun. This part of the 
investigation is presented in Part II of this publication. 
For the reader, therefore, who is interested in the theory of wind and barometric 
effects and the method and results of applying those effects to correct gage readings 
on Lakes Erie and Michigan-Huron, Publication No. 317 should be consulted. If 
the reader is interested only in the method of deriving the evaporation formula, 
assuming the corrections for winds and barometric pressures to be known, he is 
referred to Part I of this publication. If the reader is interested only in this new 
method of estimating stream-flow, assuming the evaporation formula to be known, 
Part II of this publication should be consulted. 
DATA USED AND ACKNOWLEDGMENTS 
No new observations were made in this investigation. The observations used 
covered the period July to October inclusive of 1909 and May to October, inclusive, 
of each of the years of 1910, 1911, 1912 and 1913. The observations used covered 
the 28 months named on each of Lakes Michigan-Huron and Superior. They were 
necessarily confined to the warmer months of each year. The observations con- 
sisted of the following items : 
(1) The daily mean elevations of the water surface at each of the gages Milwaukee, Harbor 
Beach and Mackinaw on Lake Michigan-Huron, and at Marquette on Lake Superior. 
Each daily mean was the mean of 24 hourly values for that day obtained from the 
records of the automatic gages which had operated at the stations named. 
(2) The observed barometric pressures at six points near Lakes Michigan-Huron and Su- 
perior were read from the isobars (lines of equal barometric pressure) as shown on the 
daily forecast maps as prepared at the U. S. Weather Bureau at Chicago for the use 
of the forecaster at that station. Two such maps are made for each day, one showing 
the facts at 8 a. m., 75th meridian time, and the other the facts at 8 p. m., 75th meridian 
time. The barometric pressures used in this investigation were for the two times named 
on each day at each of six stations for the 28 months named. 
(3) The observed mean daily wind velocity which was used had been obtained from record- 
ing anemometers as operated at the U. S. Weather Bureau stations at Chicago (111.), 
Milwaukee (Wis.), Grand Haven (Mich.), Green Bay (Wis.), Escanaba (Mich.), Sault 
Ste. Marie (Mich.), Alpena (Mich.), Port Huron (Mich.), and at the Canadian Meteor- 
ological stations of Parry Sound and Saugeen, all near Lake Michigan-Huron. On 
Lake Superior the observed mean daily wind velocities which were used had been 
similarly obtained at the U. S. Weather Bureau Stations at Sault Ste. Marie (Mich), 
Marquette (Mich.), Houghton (Mich.), Duluth (Minn.), and at the Canadian Meteor- 
ological station at Port Arthur. The mean daily wind velocity for each lake was taken 
as the arithmetic mean of the velocities observed at each of the stations named on each 
lake for each day of the 28-month period mentioned. 
(4) The observed mean daily air temperature had been obtained at the U. S. Weather 
Bureau stations and the Canadain Meteorological stations mentioned in item (3). The 
mean daily temperature at each station was taken as the mean of the maximum and 
minimum at that station. The mean temperature for the lake was taken as the arith- 
metic mean of that at the several stations, 10 near Lake Michigan-Huron and 5 near 
Lake Superior. 
