ENGINEERING: H. L. ABBOT 
43 
Report of the United States Geological Survey for 1900-1901. These 
data are supplemented and discussed by the writer in papers published 
in the Monthly Weather Review of February, 1904, and of June, 1908, 
together with some later data collected by the American engineers. In 
March, 1913, Mr. Caleb M. Saville contributed a valuable paper upon 
the Hydrology of the Canal to the Transactions of the American Society 
of Civil Engineers, bringing the records up to 1910. The Annual 
Reports of the Isthmian Canal Commission, of course, cover the Ameri- 
can operations. The first point for consideration iS, how these data 
can best be grouped for study? 
The narrow limits to which the earlier observations were restricted 
suggest that the basin of the Chagres River above Bohio should be 
adopted until Gatun Lake began to fill early in 1910. Its area has been 
accurately determined by recent surveys to be 779 square miles. The 
early rain records were mostly restricted to Colon and Gamboa. For- 
tunately, Colon is situated near the Atlantic Coast line where the rain- 
fall is largest, and Gamboa well represents the Pacific limit of the water- 
shed. A careful analysis of the more ample records of recent years 
(1898-1907) has shown that the average rainfall in the basin above 
Bohio is about 89% of that at Colon, 124% of that at Gamboa, and 
52% of the aggregate of the two stations. To avoid a change in the 
standard, these ratios have been used throughout the following table 
to estimate the annual rainfall in the basin; and where a few dry months 
are missing in the early railroad records the vacancies have been sup- 
plied by the mean values of the missing months. 
The Chagres River is a torrential stream, and the first French Com- 
pany early established a fiuviograph at Gamboa to register continuously 
the heights of the water. This record has been carefully kept since 
1883; until the rise of the lake in 1910 began to affect the local water 
level. At the times of freshets the oscillations are so sudden that their 
number and duration are readily noted, and they furnish valuable 
checks upon the discharge estimates. For this study they are assumed 
to begin and end at a stage 10 feet above low water, which is rarely 
exceeded. 
Accurate measurements of discharge were inaugurated by the Liquida- 
tion and the New French Company, the continuous record dating from 
1890. The annual outflow is given in the tables under the forms of 
cubic feet per second and depth in inches upon the watershed; the 
latter to permit a direct comparison with the rainfall. Years of great 
floods are also indicated, with the same object in view. 
The first table is intended to present, for the basin above Bohio, all 
