6 A NEW METHOD OF ESTIMATING STREAM-FLOW 
(5) The observed vapor-pressure had likewise been obtained at the U. S. and Canadian 
meteorological stations mentioned in item (3), on each day at 8 a. m. and 8 p. m., 75th 
meridian time. The mean vapor-pressure at each station for each day was taken as the 
mean of that recorded at the two times mentioned. The mean vapor-pressure for the 
lake for each day was taken as the arithmetic mean of the mean daily values at the 
several stations, 10 near Lake Michigan-Huron and 5 near Lake Superior. 
(6) The daily rainfall on the drainage areas of the two lakes had been observed at the regular 
meteorological stations already mentioned in the United States and Canada, and also 
at additional stations on these areas maintained by cooperative observers. Most of 
the rainfall data which was used in this investigation appears recorded in the Monthly 
Weather Review for the 28 months stated, in District No. 4. 
(7) The daily water levels at the Harbor Beach and St. Clair Flats Canal gages from which 
the daily discharges of the St. Clair River were computed, as well as the daily water 
levels at the S. W. Pier, Sault Ste. Marie, from which the daily discharges of the St. 
Mary's River were computed, had been observed and recorded by the U. S. Lake 
Survey (Survey of Northern and Northwestern Lakes). 
Grateful acknowledgment is hereby extended to the organizations which have 
furnished the data used in this investigation ; to Professor Charles F. Marvin, Chief 
of the United States Weather Bureau; to Professor Henry J. Cox, Senior Meteor- 
ologist in charge of the Chicago office of the U. S. Weather Bureau; to Sir R. F. 
Stupart, Director of the Meteorological Service of Canada; to various officers of the 
United States Lake Survey; and to various officers of the Corps of Engineers of the 
United States Army. 
METHODS USED IN AND SCOPE OF THIS INVESTIGATION 
The methods used in this part of this investigation are the same as those used 
in the study of wind effects and of barometric effects on the Great Lakes. A full 
statement of these methods appears on pages 6 and 7 of Publication No. 317, and 
will not here be repeated. 
In the broad investigation of wind effects, barometric effects and evaporation, 
98 least-square solutions and their corresponding studies were made. These 98 
solutions include the 74 reported upon in Publication No. 317; therefore since the 
issue of that publication, 24 additional solutions were made in the investigation of 
the laws of evaporation proper and of the run-off into the lake. These 24 solutions 
each contained from 2 to 9 unknowns and each included from 30 to 800 observation 
equations, each observation equation having for its absolute term the observed rise 
of the mean surface of the whole lake, corrected for the various influences for which 
it was possible to apply reliable corrections. 
Of the 24 solutions mentioned in the preceding paragraph, only 2 were directly 
used in establishing the constants in the final evaporation formula as hereinafter 
presented. The other 22 solutions served to give general information about the 
run-off into each lake from its adjacent drainage area, and to lead the investigator 
gradually to the final form adopted for the evaporation formula. 
In addition to the 98 least-square solutions mentioned in the second preceding 
paragraph, a total of 69 solutions and their accompanying studies was made in the 
stream-flow part of this investigation. A more complete statement concerning the 
extent of these 69 solutions is given in the appropriate context in Part II (see pages 
146 and 147). 
The total expenditure of time on the complete investigation of wind effects, 
barometric effects, evaporation and stream-flow, excluding over 2,000 hours spent 
