the thickness was about 6 feet. Below the intersection of this slope, with the surface of the ground, the face of the dam is made perpendicular, evidently depending on the earth backing as a portion of the resistance.

This dam was completed on December 1, 1909. It was said to have

contained 15,780 cubic yards of concrete, 7,925 cubic yards of excavation, and 6,360 cubic yards of embankment. The cost is given as $71,821, this account appearing in the Engineering News of March 17, 1910.

It will be seen by those familiar with dam construction that this cost was very low, and that the ratio of concrete masonry to excava

tion was extremely high. In other words, the amount of concrete was double that of excavation. Compare this with a few of the large dams of which the writer has recently had charge.

All these dams extended down into the rock a great depth. The Tub Mill dam went down to a maximum of 50 odd feet. The North Branch of the Conemaugh, about 40 feet. Indian Creek dam, to a maximum of 20 feet. Lloydell dam to a maximum of 80 feet. As these dams were solidly bedded against the down-stream rock vertical face there was no possibility of their sliding, which was, undoubtedly, the cause of the failure of the Austin dam.

The Austin dam was built of concrete, stated to be mixed in proportions of one, three, and six, and an examination would indicate that it was well built in about the proportions stated.

The sketch, Fig. 7, will show in plan and elevation about how the failure occurred.

According to the article in the Engineering News, the dam contained steel reinforcement uniformly located, but an examination of the fractures will indicate that this was not as shown. In fact, the reinforcement cannot be considered as of any value whatever, and it was unnecessary, except to prevent temperature cracks, or to anchor it to the bottom, but, as for either of these purposes, many times the amount used would be required, it is seen that this reinforcement is of no consequence whatever.

According to the same article, this dam practically failed in January, 1910, and how those responsible for the dam could have assumed that it was safe after that failure is almost inexplicable. The

case is parallel to that of a beam or girder uniformly loaded to such an extent as to cause it to crack in several places. The load being immediately relieved, the beam stands in its weakened condition, and, afterwards, when loaded up to even a greater extent than originally caused the break, final failure must inevitably take place.

It is said that at the time the dam was originally cracked, a hole was dynamited through it which caused the water level to go down, and naturally relieved the pressure. It is also stated that the reservoir had not been filled up again until just before the final failure. The information, which the author obtained second-hand, is that

several women who lived near the dam, and who frequently walked out on it, noticed a general movement down-stream. One of them ran to a telephone and notified the operator at Austin. It required three or four minutes before the final bursting and release of the water. It is also stated that the wave required about seven minutes to reach the manufacturing plant of the Bayless Pulp and Paper Mill, which was three-quarters of a mile down-stream, and four or five minutes longer to reach the town of Austin, which was a half-mile below the paper mill.

It would seem, from the conditions and the position of the fractured

sections, that the dam must have failed by sliding, because all the large sections stand on their original bases, but at a considerable distance down-stream.

With the exception of two, all the lines of fracture were in existence more than a year ago, and observation showed that the fractures were partly along planes between concrete built at different times. In fact, in nearly all cases such planes could be seen, showing that they represented the lines of least resistance. In some cases these planes were horizontal and one of such planes at the west end of the dam was 30 odd feet long and passed through the entire section. Another plane that could be seen distinctly was several hundred feet long and probably represented the line between the finishing of one period of work and commencing of the next. This could be seen plainly on the down-stream side near the natural ground level. However, none of these fractures could be considered responsible for the weakness or failure of the dam. This was due to some deeper cause, and when the strain became too great, naturally, the breaks would occur along these planes of cleavage, but it must be understood, that if the dam was in other respects perfect, these cleavage planes would not have been a source of weakness. Such cracks may, however, be criticized, because temperature cracks are almost sure to occur at these places, and, therefore, water could leak through the dam, generally, however, in very small quantities. A gravity dam is built on the principle that, if it were sawed through transversely in sections of one or more feet in length, the sections would stand just as safely as if they formed one continuous solid mass. It has been said that this dam was built of boulder concrete, often called cyclopean concrete, but this is not generally true. Most of the dam was composed of concrete, pure and simple, with some boulders near the bottom. However, this is not important.

Now, if the dam failed by sliding, what was the cause? The weight of the concrete is found to be about 135 pounds per cubic foot, and, figuring the weight of the dam and its coefficient of friction. against the rock underneath, it is found that there would not be anywhere near sufficient water pressure to slide the dam on the surface of the rock; and, judging from its appearance, the engineer or men in charge of the construction of the dam must have taken pains to see that the bottom of the dam was in intimate contact with solid rock surface. It is, therefore, entirely unlikely that sufficient water could percolate between the bottom of the dam and the rock to exert a