the load required to produce settlement will give the safe load, according to circumstances.

The regulations of the New York Building Code allow a safe load not exceeding two-thirds the final test load, a ratio of 1:1.5. (See p. 21.) From "American Civil Engineers' P. B.," 2nd Ed., p. 526:

"In the case of the Congressional Library, Washington, D. C., the ultimate supporting power of "yellow clay mixed with sand" was 13 1/2 short tons per square foot, and the safe load was assumed to be 2 1/2 short tons per square foot." (A ratio of 1:5.4, Auth.)

"From the experiments made in connection with the construction of the capitol at Albany, N. Y., the conclusion was drawn that the extreme supporting power of that soil was less than 6 tons per square foot and the load which might be safely imposed upon it was 2 short tons per square foot. (A ratio of 1:3, Auth.) The soil was blue clay containing from 60 to 90 percent of alumina, the remainder being fine siliceous sand. The soil contains from 27 to 43, usually about 40, percent of water, and various samples of it weighed from 81 to 101 lb. per cubic foot."

For a dense blue clay subsoil, on the site of the Michigan Central Terminal, Detroit,' a safe load of 4,000 lb. per square foot was adopted after tests showed that 5,500 lb. exceeded the safe capacity of the soil -a ratio of 1:1.38. In explanation of this small ratio, however, it should be stated that it was known that a satisfactory foundation could be designed within this loading; i.e., the tests were confirmatory rather than exploratory.

On soft material the test post will sometimes squeeze up the surrounding earth, thus invalidating the results of the test. This may be obviated by back-filling around the post so as to maintain the original soil surcharge at the level of the post bottom.

The following observations on loading a test-post and interpreting the results are extracted from an article by Mr. E. McCullough of Chicago, Ill., in Eng. News of Sept. 24, 1914, p. 648. The article is valuable, also, for a cut illustrating, in detail, a loading arrangement that has been used in a number of soil tests for building foundations.

"Elevation should be read after loading, and 6 hours later. The readings should be plotted and if the curve is a decided parabola the test load may be left until no settlement occurs. If after 18 hours the curve shows no sign of being a decided parabola, the test load should be lightened. One-fourth of the test load which the post can carry for 24 hours without settlement should be a safe load per square foot. The settlement during the loading of the platform may be several inches, due to unavoidable rocking of the test arrangement, but this should be ignored and only the settlement occurring after the full load is applied should be considered."

1 See description and reference, p. 24.

Indicating Soil Test Results by Curves.-Fig. 171 shows the results of two soil tests in graphical form.

A study of the curves (of two separate tests) will show that the heavy lines are ideal curves drawn between the points of the "actual settle

"

"

"

0 " " " % ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ 1′′ 1%
Settlement in Inches

FIG. 17.-Graphical record of soil tests.

ment curves" which are shown in light dashes. Fig. 18 shows the relative values of various strata at the location of test, and indicates very clearly that the sandy strata are by far the best in bearing value. The location of the "yield point" on the curve is instructive, but the

article does not state what was the assumed safe bearing value of the soil selected.

Alternate Methods.-The following directions for testing are abridged from Patton's "Treatise on Civil Engineering." Dig pits 2, 4, 6 and

1 Eng. News, May 7, 1914, p. 1025; article by M. W. Manz.

8 ft. deep, and test with a 12 X 12 timber in each. Observe by a level, or by a string or straight-edge across the face of the timber: (1) the least load that produces appreciable setting, (2) extent and rate of settling under this load, (3) effect of the continued application of this load. Make observations with increments of about 100 lb. For softer materials, increase size of base to 2, 3 or 4 sq. ft., for marshy soils up to 6 or 8 sq. ft. Note whether settlement produces: (1) compression, or, (2) bulging, by driving stakes in the surrounding material and noting the levels.

When the proposed structure is one subject to vibrating loads, this condition may be roughly anticipated in the test by strongly vibrating the loaded post from time to time and noting the increased settlement; the effect is illustrated in Fig. 17.1

In tests conducted for the design of the foundations of the Michigan Central Terminal at Detroit, in a blue clay formation of varying texture, four different methods were employed, so as to obtain a comprehensive knowledge of the sub-surface conditions for this important structure. The first consisted in loading a platform 11 ft. square supported on four 12 X 12 in. square posts. The second device consisted of a single 12 X 12 in. post with a loading platform on top. All of these posts were in back-filled pits 5 ft. deep. The third test was conducted in a 7-ft. diameter sheathed well 19 ft. deep; the bearing plate was 2 ft. square, surmounted by a platform (at the bottom of the well) loaded with short pieces of rail: readings were taken on top of a 1 1/4-in. iron rod 18 ft. long, extending from the bearing plate to the surface through a piece of pipe to protect it from the loading material The fourth test

was designed to approach more nearly actual load conditions, and consisted in loading a concrete pier 9 ft. 4 in. in diameter at the base reduced to 5 ft. in diameter at the top.

All the tests indicated substantially similar results, although the last showed smaller settlements than the others, giving finally a practically stable condition at a load of 5,500 lb. per square foot. A safe load of 4,000 lb. per square foot was accordingly adopted; or, rather, this proposed figure was demonstrated to be safe.

The article in question is of particular value on this subject, being well illustrated with cuts and curves.

For tests conducted in St. Paul, Minn.,3 to determine the feasibility of the soil holding up a new 12-story building, under a load of 4 tons per square foot, a block 4 ft. square was used consisting of a 1-in. thick steel plate under a three-high layer of 12 X 12 X 48 in. timbers. This was placed in a sheathed pit 15 ft. deep, and the platform therein was

1 Eng. News, May 7, 1914, p. 1024.
2 Eng. News, Oct. 1, 1914, p. 704.
3 Eng. News, Oct. 8, 1914, p. 736.

loaded with pig-iron: timber dollies, acting on vertical strips of wood, were used to prevent the load from tipping against the sides of the well. As a previous test with a 2-ft. square block had failed by reason of the material being displaced and forced up outside of the base, the remainder of the pit was covered with 2-in. planks weighted to 475 lb. per square foot to meet as nearly as possible the actual conditions that would exist were the building completed. A 1-in. pipe extended from the steel plate to the top of the pit to provide a bench on which to take levels. The soil was a mixture of coarse sand and comparatively small gravel with about 10 per cent. of clay; and, under a final load of 8 tons per square foot a total settlement of 1/4 in. was recorded.

SEC. III. STREAM GAUGING

STREAM GAUGING; IMPORTANT OBSERVATION

The record of a stream gauging on some one day, without data as to its relation to other flows, is worthless, and is liable to lead to financial wreck if acted upon. The most important record is that taken at time of lowest water.

For important installations, gaugings are taken daily for as long a period as is necessary to give a record of maximum, minimum and average flow, so that fairly exact calculations may be made as to probable flow throughout the year.

If, however, it is only possible to submit the result of a single gauging, a report should be submitted, giving, as accurately and fully as possible, the relative flow to be expected at all other times of the year.

STREAM GAUGING BY CROSS-SECTION AND VELOCITY METHOD

The following method applies more particularly to small streams or rivers; for larger rivers or for more accurate methods see any of the standard text books on surveying or hydraulic engineering. It is not as accurate as the method of measuring by a weir described on p. 26.

Select a point in the stream having as uniform a cross-section as possible in the length to be gauged, from 50 ft. in slow streams to 150 ft. in swift ones. If necessary, remove sub-surface obstructions such as snags, weeds, rocks, etc., so that the bed of the stream may be fairly uniform in cross-section and free from obstructions that would vary the velocity of the float.

Obtain a cross-section of the stream at a place in the selected stretch where it remains tolerably uniform for some distance. This may be done by stretching across the stream a cord carrying tags at measured intervals, and measuring the depth at these points with a pole or sounding line, platting the results and calculating the area.

The velocity may be obtained by planting range poles (say one on each side of the river) at the upper end of the stretch and two at the lower, measuring the distance between these stations and observing, with a seconds watch, the time required for the float to traverse this distance. It should be noted that the distance to be measured is not necessarily the distance between the poles but the distance traversed by the float as sighted by the range poles. The float should be launched in the swiftest part of the current (point of maximum surface velocity) and some distance above the first range poles, so that it may have time to attain the velocity of the stream. The float may be a small block of wood or a weighted cork, painted so as to be easily visible. Several tests should be made and the mean result taken; also the tests should be made in perfectly calm weather so that the velocity of the float may not be affected by the wind.

The velocity thus measured will be the maximum surface velocity and should be multiplied by .8 to obtain the approximate mean velocity. The mean velocity (in feet per second) multiplied by the area of cross-section of the stream (in square feet) will give the discharge of the stream (in cubic feet per second).

STREAM GAUGING BY MEANS OF A WEIR

This method of measuring the flow of water in streams, discharge from pumps, etc., when carefully executed, gives very accurate results. Refinements of calculation dependent on variations in conditions are not here considered; see text books on hydraulics or experimental engineering for more exact methods.

FIG. 19.-Stream gauging by means of a weir. FIG. 20.-Measuring the head.

The apparatus is illustrated in the accompanying figures.

The conditions affecting the accuracy of the weir are as follows:1

(1) The weir must be preceded by a straight channel of constant cross-section, with its axis passing through the middle of the weir and perpendicular to 1 Carpenter's "Experimental Engineering."