In the dam recently completed for the Mississippi River Power Company at Keokuk, Iowa, many of the conditions and methods were practically identical to the foregoing. The foundation was very favorable, being a solid blue limestone, which was found at about the same elevation all over the work and which required apparently very little preparatory excavation. It was necessary to pass large quantities of water 20,000 c.f.s. to 300,000 c.f.s. during construction.

The cofferdams were constructed of cribs exactly the same except for height, as the McCalls Ferry, and like the McCalls Ferry cofferdams, were remarkably tight. The dam itself was about four-fifths the height and twice as long. For the river section the construction bridge, in this case part of the permanent structure, was built first by a cantilever crane working from the structure itself and overhanging four bays in advance. (See Plate II, Fig. B.) Each bay consisted of a 6-ft. pier and a 30-ft. clear opening, the springing line of the arches being 11 ft. above the crest of the spillway sections between. As at McCall's Ferry the openings were left low to be later raised 5 ft. at a time by blocking off with a similar shutter handle by a derrick from the bridge.

A variation of Type 1 was the scheme employed at Spier Falls dam (1900-1903). No diversion seems to have been required for the construction of the first portion of the structure which was built upon a bench 20 ft. or so above the river level. Through this first portion were left four openings 7 ft. by 10 ft. with floor at elevation 30 designed to be closed later by wooden shutters and filled with masonry. There was also left a gap 90 ft. wide at elevation 40 in order to pass the larger floods. When it came to unwatering the main portion of the excavation immense crib cofferdams were required, the main upper one being 600 ft. long. Its average section was 150 ft. bottom width, 60 ft. high and 25 ft. top width, with a maximum section 250 ft. bottom width, 90 ft. high and 80 ft. top width. This was constructed cob-house fashion of logs drift-bolted together and filled with rock and gravel. Though portions of it were twice washed out by floods at critical times, the finally completedcofferdam was entirely sufficient and quite tight. During the entire construction period the stream flow was never less than 1000 c.f.s., and floods of 40,000 c.f.s. to 50,000 cf.s. occurred.

The total amount of excavation was some 270,000 cu. yd., the maximum depth being about 80 ft. About 4,000,000 gal. of water per twenty-four hours were pumped from the pit.

A question which naturally arises in connection with a study of the stream diversion at Spier Falls is why were the four 7 X 10 ft. openings put so high (with floor at elevation 30)? True the first section was built on a bench 30 ft. above the river, but it seems possible that they might have been placed near the river end of the section, and that an inlet and outlet channel of not excessive length might have been excavated to carry the water. If they could have been so placed, at or very near river level, that 30 ft. saved in elevation would doubtless have saved more than half of the material and labor involved in the crib work in the upper cofferdam. It would also have reduced somewhat the amount of leakage to be pumped and the danger of washouts by floods. Published accounts of the work indicate no reasons why this could not have been done, though quite possibly some existed.

Unfortunately costs are not available in connection with any of the foregoing examples.

The method employed at the Wachusett Dam, Type 2, was to carry at all times the entire flow of the river across the work through a flume at one side. It was about 250 ft. wide between rock at river level, and the loose material to be removed was gravel with a maximum depth of about 65 ft.; maximum amount of pumping was about 6,000,000 gal.per twenty-four hours at heads of 50 ft. to 70 ft.; one flume 700 ft. long, 40 ft. wide, by 15 ft. to 17 ft. high, supported across the foundations on piles and posts; grade 1 in 1000; upper temporary dam of earth with a core of sheet piling not, however, to rock; lower temporary dam of earth; another flume 7 ft. X 7 ft. X 500 ft. long carried water to the Wachusett aqueduct but was not required as far as handling the water was concerned; both flumes had head works to control the flow. (See Fig. 16, also Plate IV, Fig. C.) The cost of the temporary works was about $120,000.

The water above the upper temporary dam was held at an elevation about 20 ft. above the natural river level in order that water might be diverted during the construction period for two purposes; first through the small flume, above mentioned, for the use of the Metropolitan water district; second on the other side of the river through 3/4-mile of 24-in. cast-iron pipe for the use of a cotton mill. Of course, holding the water up for these diversions added to the expense of the temporary works, complicated the operation, increased the leakage into the pit and consequently the pumping. But for these diversions no head works would have been necessary for the large flume, and the flow could have been entirely uncontrolled.

A flood, of 9700 cf.s., the maximum known on the river, was carried by the flume a few weeks after completion. One attendant circumstance is worthy of mention. During the height of the flood, the water lost so much head in acquiring velocity after passing the head gates that its surface was some 5 ft. lower than the surface of the water outside the lower end of the flume. At this time the lower temporary dam had not been built and the water backed up to the upper temporary dam at an elevation sufficient actually to float a large length of the flume. The amount that it was lifted was afterward found to have been at least 18 in.