Practically all masonry dams show leakage upon their downstream face; in fact only in the case of one curved dam is it claimed that absolutely no leakage can be detected. Some dams show only a slight sweating, which may disappear on a dry sunny day, while others may leak quite appreciable streams.

Leakage may result from temperature cracks or by percolation through the body of the masonry. In all cases, probably, it results from these two causes although in different and unknown proportions. Speaking generally and with some caution, it might be said that in dams of rubble masonry the leakage is somewhat more attributable to percolation or to the presence of numerous and minute cracks; while in the more recent dams of concrete or cyclopean masonry it seems to be due to a smaller number of wider cracks. It was probably this localization of cracks observed in the earlier cyclopean dams that led to two features of design, namely, longitudinal reinforcement near the top and expansion joints. The latter seems to have become standard practice, in theory at least, even if the details are yet subject to improvement. Actual figures regarding effect of temperature changes, and leakage, are quite scarce, and what follow are taken from a paper by Charles S. Gowen (with discussion by George C. Honness, Thaddeus Merriman, William L. Brown and Charles S. Gowen) on "The Effect of Temperature Changes on Masonry" which appeared in the Transactions of the American Society of Civil Engineers, Vol. LXI.

The Boon ton dam is 2150 ft. long, with a maximum height of 114 ft. It is constructed of cyclopean masonry approximately in the proportions of 50 per cent, stone and 50 per cent, of concrete the average mixture of which was 1-2 3/4-6 1/4.

"Across the entire top of the dam extend thirty-three cracks, seventeen of which total 2 1/2 in. in thickness, and the remaining sixteen average 1/32in. each. In addition to the above are thirty-three more which do not extend across the top but are confined to one-half or less of the crest width, and with a total thickness of 1/2 in. Thus the total width of all the cracks was 3 1/2 in. and the largest ones could be traced down the face of the dam for a distance of 60 ft. from the top. The larger cracks occur at very regular intervals of about 100 ft. except near the ends of the dam. Through the main cracks is considerable seepage, naturally greater in the winter than in the summer. Through the two principal cracks, one near each end of the dam, the seepage was measured on March 17,1908, and found to be 23,000 gal. per twenty-four hours".

From a somewhat incomplete series of thermophone observations on the temperature of the interior of the masonry Mr. Merriman deduces, subject to future verification, the following formula for range of masonry temperatures between depths of 0.5 ft. and 20 ft. from the surface. If R be the total range in temperature at any point in the mass, in degrees Fahrenheit, and D be the distance in feet to the nearest face of the dam, then R = 135/3ûD where 135 is the total atmospheric range observed during the period.

In the New Croton dam were five large cracks, one measuring 1/4 in. while the others averaged 1/8 in. each. In addition were a number of smaller cracks, but their number and size were not determined. Mr. Gowen presents some interesting curves showing the variation in width of cracks corresponding to changes of atmospheric temperature. One or two of the principal cracks were treated with grout and later caulked with lead. Since then one of them, which extends down the face of the dam for 70 ft., has shown in winter some spraying seepage at certain elevations. During warm weather this seepage reduces to a sweating or dampness too slight to create a flow.

In the Cross River dam (completed in 1907) the upper 30 ft. was reinforced with 1 1/4-in. square twisted rods or rods of equivalent area. No temperature cracks were discovered during the winter of 1906-07. Work was resumed on the dam in March, 1907, and it was completed by the end of August. Thus that part of the section of the dam which is affected by temperature cracks was constructed during warm weather. During October, 1907, four cracks were discovered, which by the following March had developed so that three of them extended down the face of the dam for 70 ft. from the top and the fourth for 42 ft. The aggregate maximum width of the cracks was 7/8-in. The dam was apparently cracked entirely through for a distance of about 43 ft. from the top at which elevation the section is 31.75 ft. wide.

"No leakage or seepage.through the dam or through the cracks was noticed until January 16, 1908. On this date the water stood at a point 53 ft. below the top of the dam, and the leakage was noticed at a distance of 70 ft. from the top of the dam." An illustration "shows the crack at Station 14 89, which extends through a header block. A slight amount of leakage continued until February 7, when it suddenly increased, indicating that there was a free passage for water through the dam section at the cracks. On February 10, a measurement showed a leakage of 6.6 gal. per min., and by February 17, this had increased to 23.9 gal. per min., which is the maximum amount of leakage measured at any time. As soon as the marked increase in leakage was noticed, steps were taken to control it. This was accomplished by caulking the crack at the upstream face with lead wool and grouting the crack. These measures were successful to the extent of reducing the leakage to 4.5 gal. per min., and this has continued to diminish since the warm weather set in".