impact when a locomotive first passes on a bridge, the grinding wrench when brakes are applied while it is on the structure, the vibration of any steel construction caused by the moving load, the deflection and reverse bending on continuous-span steel-bridges, the changes in volume incident to temperature variations—all these place very severe but indefinite stresses on the water-proofing blanket. At the same time, an engineer, knowing these conditions, can select a material, the physical properties of which will generally meet them. Materials which are brittle at low temperatures should receive scant consideration, because of the probability of their being fractured by vibration in cold weather. Materials which are very soft at high temperatures should likewise be regarded with suspicion, because they cannot be held on vertical or steeply-inclined surfaces. Finally, a material should be selected all parts of which are flexible and elastic, so that they will yield rather than break under the conditions mentioned. To put it briefly, the water-proofing, in so far as physical requirements are concerned, should be selected on the basis of its plasticity or flexibility at all temperatures and its. small factor of susceptibility to temperature changes. As to the deterioration of water-proofing, that is almost entirely a chemical matter. The exclusion of all materials that are affected by water, whether or not the water carries acids or alkalies, such as the ligno-cellulose compounds (jute or burlap) and the felts in which tapioca is the binding medium (asbestos felt), the non-employment of those that are physically unstable, such as most of the artificially compounded asphalts and asphalts containing organic matter, and the selection of a material which has to a large degree been pre-aged or pre-oxidized will assure long life for the protection. Indeed, water-proofing materials (asphaltic) placed in 2500 B.C. have lately been found to be still in good serviceable condition. A discussion of the various materials sold for bridge water-proofing is a complicated one and very technical. It is relevant to the subject of this chapter only in so far as it may point out the answer to the questions that have been raised. It is sufficient to say, however, that if a selection of materials is made with regard to the known conditions which must be met, rather than on the basis of initial cost, a life of at least 25 years may reasonably be anticipated, with the probability that this figure will be greatly increased.

It is obvious that, if water-proofing a bridge floor is economically wise, the resultant value of the protection afforded to the structure thereby must exceed the cost thereof. That the said value does exceed the cost may be a difficult proposition to establish by definite figures. As far as is known, no data exist concerning the amount of the damage caused by lack of water-proofing. All that can be done with this question is to indicate the agents which attack bridge floors and the effect of water-proofing in warding off such attacks, leaving it to the judgment of the individual to decide whether the probabilities in the case justify the expense which water-proofing involves.

For convenience in treatment, the effect of water on steel alone will be