were in place. When the flash boards were in perfect condition, this would give a slight overestimate which, however, would be offset by the leakage of the flash boards at other times and by their being occasionally somewhat bent over. For the twelve years 1888-1899, it is stated that during the period when flash boards were off the flow over the dam was computed by the Mullen formula of the East Indian engineers. I doubt if this formula applied so accurately as the Francis dam formula. For that part of the year when flash boards were in use, Mr. Rafter used the Francis formula for a sharp edged weir, which is doubtless the most accurate of any that could be applied to this condition. Comparing the mean monthly flow for the years 1898 and 1899 as computed under supervision of Mr. Rafter from the East Indian formula and the weir formula aș just described, with the results of computations made for the same years under Mr. Bloss day by day with the Francis dam formula, the discrepancy in monthly means seldom exceeds 5 per cent. in any one month, and taking the whole year through, the flow as computed by the East Indian formula with flash boards, averages about 2 per cent. greater than for the computation by the Francis dam formula. 1 During the seventeen years of gaging prior to the summer of 1904, there is perhaps a little occasional uncertainty about leakage through the flash boards and probably the actual flow would therefore equal that recorded, notwithstanding the difference in formulas. Since in 1904 when the flash boards were abandoned due to the addition of the rounded concrete crest, gaging of the portion of the river discharge that flows over the dam is presumably very accurate. The uncertainties and differences due to the different formulas used in the previous years are less than the probable errors introduced by the turbine measurement. These errors may be due to turbines being not always run at full gate, and due to the leakage through those turbines that are shut down. There is doubtless at times some retardation of flow due to ice or other obstruction upon the forebay screens or in the turbines. The writer happens to have had about ten years' experience in making daily estimates of river flow under somewhat similar circumstances, and where great pains were taken to confirm the results of turbine measurement by independent methods of gaging, and from this experience, although no proof can be given as to the precise margin of error, he inclines to the opinion that because of uncertainties about the turbines and flash boards, although occasional errors would balance, the precision of measurement attained in the Mechanicville gaging as a whole, through turbines and over dam, cannot be relied on as closer than 5 per cent. for the average of, say, a three year term as used in the mass curve. For the average of a single week, errors of 10 per cent. are not improbable, particularly in the period prior to the summer of 1894, at which time the masonry crest replaced the flashboards on the dam, and for single days there is a possibility of still larger error. It is to be borne in mind that the two check measurements by current meter found the method by turbine measurement 5 per cent. in excess of the current meter. For these reasons, I have been conservative in the final estimate of Sacandaga power and made what I believe due allowance for all margin of uncertainty in calling the mean flow available from Sacandaga reservoir 1,700 cubic feet per second; for this figure in effect discounts 15 per cent. from the quantity that is most probable on the face of the returns after studying the greater rainfall on the Sacandaga. INTERPOLATIONS. No records are found published for the flow at Mechanicville in the month of January, 1902, or months of November and December, 1905. To complete our record for the mass curve computation, an interpolation based upon the Fort Edward record has been used in these places, making due allowance for the difference in drainage areas. Similarly for July, August and October, 1904, while the record at Mechanicville was interrupted during the building of a new crest on the dam, the vacancies have been supplied by the use of the Fort Edwards records. DIVERSION FOR THE CHAMPLAIN CANAL. The published record of Hudson river flows at Fort Edward and Mechanicville requires a small correction to compensate for the diversion of water for the supply of the Champlain canal, and since the records previously published in the State Engineer's reports did not make the amount of this correction as definite as desired, a special investigation was made at my request under the supervision of Mr. H. K. Barrows of the U. S. Geological Survey, from which he concludes that during the season while the canals are in use there should be added 200 cubic feet to the Mechanicville record and 180 second feet to the Fort Edward record. This accords well with the result of sundry previous but less complete gagings. These corrections of 180 and 200 cubic feet per second are equivalent to additions of .07 inches depth of monthly run-off on the watershed tributary to Fort Edward and to .05 inches depth of monthly run-off from the watershed above Mechanicville while the canals are in use. The length of the navigation season according to the table given in page 1,060 of the history of the New York canals, published by the State Engineer in 1905, during the period covered by the gagings at Mechanicville and Fort Edward, has averaged seven months, opening about May 1st and closing about December 1st. These corrections have therefore been applied during the months from May to November inclusive, and are included in the tables on pages 104 and 105 upon which the mass curves of Diagram No. 2 for computing the Sacandaga storage are based. The total diversion into the Champlain Feeder canal from the Feeder dam above Glens Falls, in recent years probably has been about 220 cubic feet per second, being now much less than in former years when as high as 383 second feet have been measured (October 8, 1895). This lessening of the diversion at the State feeder dam is due to 'repairs and to stoppage of the leaks along the course of the feeder canal through the town of Glens Falls. This feeder canal supplies the summit level of the Champlain canal, and therefore a portion of the water flows northward into Lake Champlain, and a portion flows southward, returning into the Hudson where the canal enters the river for the boats to cross it above the Saratoga dam. A second diversion from the Hudson occurs into the canal leading southward from the Saratoga dam, and a gaging made in the canal as nearly opposite as practicable to the gaging station at Mechanicville shows this to average about 110 second feet, which added to the 90 second feet diverted northward into Lake Champlain makes the 200 second feet of diversion that must be added to the Mechanicville gaging station when reckoning the yield per square mile from its watershed. Of this 220 second feet now diverted from the Hudson, about 90 second feet flows northward into Lake Champlain, and although nearly all of the remaining 130 feet ultimately reaches the Hudson, some of it by leakage above Fort Edward, the most of it does not get back into the Hudson until some miles below Fort Edward where the canal crosses the river. The addition necessary to compensate the Fort Edward gagings so they will show the entire yield of the drainage area is found by measurement to be 180 cubic feet per second. RUNOFF OF HUDSON RIVER AT MECHANICVILLE FOR THE (In inches depth over entire catchment area.) (Catchment area 4,500 square miles.) Corrected for canal diversion, May to November, inclusive. |