Of all structures designed and built by engineers, a masonry dam is probably the one about which there is the least amount of exact knowledge. Lest there be a disposition to question this proposition, let us examine it.

Other structures may usually be resolved into their separate parts or members, and the function of each member may be analyzed to the end that its design may and does answer the purpose. Experiments may often be conducted upon actual full-sized members, the conditions of their construction may be accurately known and controlled and thus the adverse conditions may be eliminated.

A masonry dam is usually assumed and designed as an entirely homogeneous structure; in other words, as a structure of one member. The only tests are use and failure. Use, however long continued, is a negative test which conveys very little information; failure conveys only a sharp admonition to be more careful next time and usually leaves the specific cause unsatisfactorily determined. We do know that many masonry dams have stood for years, and that some have failed.

In the case of a failure a reason is usually assigned that satisfies our intelligence, but only because it seems to conform to other, more common and more appreciable experiences, and not because correct relative values can be assigned to the various possible contributing causes. Experiments under comparable conditions and on a working scale are impossible except as we regard all actual dam failures as experiments.

Now to take up the various features one by one. We assume the rock foundation to be unyielding, and such is probably the case as far as any effect upon the masonry is concerned; but it is probably often the case that the superimposed weight of masonry produces new conditions in the rock and in the possible channels for leakage. The function of other foundations is simply to sustain weight, and the determination of their adequacy is a relatively simple matter compared with the problems which arise in connection with the foundation for a dam. In the latter case the foundation must be tight enough to prevent any significant amount of leakage under the dam, in addition to sustaining the weight. The obscurity of many of the conditions, the fact that the foundation must be prepared and accepted once for all with very limited possibilities for future inspection, repairs or reinforcement combine to render this a problem calling for the most experienced judgment.

Though we may know the strength of mortar briquettes or concrete cubes manufactured under conditions insuring a high degree of uniformity, we can obviously know very little about the behavior of masses of masonry comparable in size to the structure under consideration. Temperature conditions will vary widely during construction, and even if it were possible to make a sufficient number of observations, still the effect of the changes could not be accurately determined.

The point of least uncertainty is the load that the dam has to carry. We may know within very close limits the weight per cubic foot of the masonry that resists the weight of the water, but immediately we are in the midst of an argument as to whether to reduce that figure by 10 per cent., 20 per cent, or 30 per cent, to allow for an unknown but possible uplift pressure.

We assume that a mass of masonry may resist by a certain amount any tendency to slide, when every observation we have as to coefficient of friction was made under circumstances so widely different as to be absolutely inapplicable. We make no allowance for ability to resist shearing stress, when we know that a sliding failure is inconceivable without shear in a vertical plane.

We proportion a dam to resist overturning, and assume for the purpose no tensile strength in the masonry, though an actual overturn would be resisted by tension and also by shear.

We assume an ice thrust equal to the crushing strength of the ice, without any reason to suppose that the ice can exert any such pressure. We regard the masonry as homogeneous, when every large dam is constructed under conditions such as to produce initial temperature stresses. Under certain conditions a curved dam recommends itself to our reason and intelligence as being more stable for the same amount of masonry, or of equal stability for less masonry, yet when we try to analyze the structure we are immediately involved in a mass of assumption which may with equal justice be made to yield widely differing results.

When we can assign but one limit to the range of values possible for any factor entering into the design, we work close to that limit; when both limits are known we are cautious and work nearer the more conservative one, trusting that the margin of safety will cover the conditions of unknown value. Thus the limits for value of tension and shear are zero and some unknown finite quantity, and we assume zero. The amount of uplift pressure is between zero and full head applied to full area, hence we are cautious and work near the latter.

For ice thrust we assume a fair value as its crushing strength, and a maximum thickness. We apply it all, knowing that the truth is probably much less. Finally, we trust that the margin of safety is sufficient allowance for the factors of foundation, temperature and settlement cracks and initial stress, to whose limits we can assign no values though they may appear small. Though such procedure is usually sound, an attempt will be made to show that certain fears are groundless or much exaggerated, and that certain alternate measures are entirely adequate.

Very many dams (and this is especially true of large dams) are designed by one person and constructed by another. However each may be assisted or consulted by the other the fact remains that the two functions are in different hands, and there are occasional indi-cations that they are imperfectly correlated. This condition may manifest itself in several ways as follows: