(4) Structures whose floor areas are not an integral part of the foundation.

(5) Masonry dams.

The history of tunneling, to which some reference has been made, and of all braced structures, conclusively proves that buried structures are not buoyant.

It is believed that a partially buried structure can be designed to resist buoyant pressure providing it is designed with a small margin of excess weight above that shown under the theory noted in this paper. As, however, no engineer will design a bridge without a reasonable factor of safety, so no one should fail to provide against contingencies even though convinced that they may not arise. Therefore the writer concurs with those who are of the opinion that the design of partially buried structures should provide against full upward pressure.

As to those floor areas which are an integral part of the structure foundation, it is readily seen that, being part of the foundation, they must bear on the solid material and must, therefore, have some of the area exposed to upward pressure reduced.

On the other hand, those floors not forming a part of the foundation may be called upon, in certain classes of material, to resist full upward pressure.

As to the masonry dams, it is not possible to consider such structures resisting full or even a large percentage of full upward pressures, unless built on materials so porous that their being built thereon would be an absurdity.

In considering all these conditions, it must be noted that the conditions under which a structure is built, or, rather, those obtaining during its construction, are vastly different from those which eventually obtain, i. e., in connection with subaqueous tunneling it has been often noted during construction work, owing to incidental loss of air and stirring up of the material, that the material was in a "soupy" condition, whereas soundings before the beginning of work and the stability of the structure afterward tend to show that the material before being disturbed was firm and hard, as it likewise became after the disturbance ceased. It is also well to note that water through the ground is constantly flowing along minute channels, and that as it flows into and through an abnormal void, it finally fills it with the finer material obtained elsewhere. It is probable, therefore, that the abnormal voids will not long continue to exist in contact

with any structure, except in possible isolated instances, in very heavy material, such as gravel, and where protected, as in those cases noted where the foundations surround and extend below the floor area, and in allied cases.

While the question of pressures on shafts presents some unusual conditions, in dry material it can probably be considered in the same way as, or rather in direct relation to, the pressures on grain-bins; i. e., if one assumes that a grain-bin carries a centrally located pipeshaft, one must conclude that the pressures on the shaft bear some direct relation to those on the bin itself, bearing in mind that the pressures are intensified by their convergence. Another element however comes in to offset the effect of some of this pressure, i. e., the horizontal arching properties of the material; and it is not possible to determine this conclusively except by experiments on a large scale. In general, one may assume that the sides of small shafts up to five or six feet, in normally dry material, will not be subject to excessive or increased stresses due to increased depth, as the horizontal arching action of the material establishes a constant pressure beyond depths equal to four or five diameters of the shaft. The writer has supervised the sinking of numerous pits from four to six feet in diameter, for depths of from 20 to 45 feet, and has never seen in them any evidence of increased pressure due to depths. When the proportions of a shaft become greatly enlarged over the above figures, the pressures may be considered to be generally the same as those in trenches. In aqueous and semi-aqueous material over a percentage of the area the pressure of the water must be added to that of the solid material, as already noted.

The resistance of earth in its relation to the foundations of structures is a subject too broad to be considered herein. The questions relating to either piling or caissons, which are essential elements of foundations in aqueous or semi-aqueous materials, might, with difficulty, be encompassed in the limits of papers devoted exclusively to either. The writer wishes merely to emphasize the fact that in ordinary firm materials, such as sharp sand or gravel, or a mixture of both, we do not attach to the factors of resistance sufficiently high values. The fact is lost sight of that, when a reasonable depth of foundation is reached, the resisting power of firm material is increased, not so much because the material is more compact at great depths, but because the opportunity for lateral displacement is eliminated. A test has been made in which a 16-inch hollow pipe was cleaned

out to its bottom and a 14-inch piston placed therein, in which, at a depth of 77 feet below the curb, or 37 feet below ground-water, the piston supported 28 tons without further settlement, after an initial settlement of about 234 inches; while under a load of 15 tons the following observations were made, the material being ordinary sand:

While not conclusive, this test would tend to show that depth does not necessarily add to the stability of ground.

Tests have also been made on a 14-inch hollow pile in firm water bearing gravel, in which a measured circumferential area of 61⁄2 square inches resisted a measured load of 60 tons, with no initial observed settlement.

Conclusions follow that any foundation on firm ground, deep enough to be guarded against and protected from lateral displacement, can be compacted by ramming or by driving short piles into it, or, if possible, by subjecting it to excess weight, to avoid the usual initial settlement due to compacting, and that it will then, without further settlement, resist pressure greatly in excess of that usually allowed.

Before concluding, the writer desires to note a few observations and reasons for his belief that the general principles outlined in this paper are true. In the first place, it is assumed that ground pressures are not subject to the same laws as aqueous pressures. If this were not true, it would be impossible to excavate deep trenches or tunnels, even in dry ground, without air-pressure. Not only is it possible to work safely at great depths in tunnels and trenches, but any one familiar with such work must realize that the bottom or floor of a deep tunnel or trench exposed for a large area shows no evidence of pressure in normally dry ground. The fact that pressure is not transmitted directly to the exposed bottom should be conclusive proof that arching action does exist in earth. It is also true that coffer-dams can be sunk to great depths in coarse sand or gravel adjacent to deep bodies of water by means of pumping, i. e., without air-pressure, showing that the presence of water alone does not give aqueous properties to some materials. If, then, the arching action of normally dry earth exists to some degree,

as shown by the experiments, and in countless other instances, and if it exists to a large degree, as shown by the fact that deep excavations or tunnels can be safely made, as safely as those at shallow depths,why should we not accept it as a practical factor and not rather as an occasional freak of nature? The writer has never seen an instance in which the pressures in normally dry ground were greater than those accounted for in the body of this paper, and he can further give numerous observations showing conclusively that the pressures were not in excess of those allowed for herein in normally dry clays, loams, sands, gravel, or some mixtures of these. It will be necessary, however, to note but one or two here. The writer's attention has often been called to, and he has frequently examined, tunnels and large sewers in which the roof arches had cracked under pressure-one in particular, that of a cast-iron lined 15-foot tunnel, the roof plates of which were badly cracked after the passage of the shield. The writer believes that all these conditions-certainly those observed by him-can be explained by the fact that, on backfilling structures in trenches, or after the passage of the shield in tunnels, voids were left along the sides, and the normal subsidence of the ground above forced out the sides or haunches of the arch. The point of interest in all these cases is that the cracked arch sustained the ultimate loading, which had apparently been so great as to cause the initial rupture of the sound arch, and yet many engineers hold that the arching conditions may exist for a time, but eventually the full superimposed loading will come upon the structure. The writer holds that the arching properties are most effective when there is no further possibility of subsidence.

Another observation: The grade of the Joralemon Street approach to the Battery Tunnel was corrected above the water-line by cutting out sections of the bottom and lowering it from one to two feet, while in the roof 4 x 10 foot sections were cut out and jacked up 30 inches for long distances. The excavation of the bottom was accomplished by digging out, as in the ordinary trench work, after bracing the tunnel, while the roof plates were jacked up into the voids caused by displacing small quantities of sand around the exposed edges and through weep-holes. This work could not, of course, have been done had the full weight of the ground above, as is generally conceded, borne upon the full area of the roof of the tunnel.

As to deep trenches, the writer has often seen bracing crack near the top of a trench to such an extent that it had to be reinforced,

due to the fact that the trench was being deepened at that point. To use in a 30-foot sand trench bracing just strong enough at a point 15 feet down and not strengthen it on excavating the same trench to a depth of 60 feet would be suicidal, whereas at a point 10 feet above the bottom of a 60-foot trench bracing need not be any heavier than that 10 feet up from the bottom in a 30-foot trench, always assuming that the sand is normally dry. The danger of deep tunneling or trenching lies not in the normally dry, homogeneous materials, or even in firm ground when saturated, but it is rather due to the pockets of so-called quicksand, or "near quicksands," and to the treacherous soft clays, or those with well-defined seams of soft material along which they tend to slide in mass. In rare instances, even in rock, pressures may be found to be greater than in soft ground, where the stratification is vertical, or where pockets of disintegrated rock become detached from the solid mass around them. The writer desires, however, to impress upon the reader his belief that it is not depth which causes these conditions, necessarily, but that they are as likely to occur in tunnels at shallow depths as in those at very great depths; and to emphasize the original observation, that the greater the angle of repose in firm materials, the greater the pressure on a tunnel structure. Finally, the writer reiterates the plea that wherever possible the engineer will experiment on a large scale and note, wherever practicable, the results of observations, which may be of value. As he has already stated, to be of real value experiments must be made on full areas, and not on those which are a small proportion only of that affected.

Many of the experiments here noted are more fully described in the writer's paper, "Pressure, Resistance, and Stability of Earth," published in Volume lx of the "Transactions of the American Society of Civil Engineers," and much of the matter of this paper is transcribed therefrom in substance.

The writer desires to thank Mr. Frederick L. Cranford and Mr. James W. Nelson for valuable assistance and for apparatus for making the experiments noted.