The above table should be taken with caution. The range of variation in the species is apt to be much greater than the figures indicate. See Johnson's tests on long-leaf pine, and Lanza's on spruce, above. The weight of yellow pine in the table is much less than that given by Johnson. (W. K.)

Compressive Strengths of American Woods, when slowly and carefully seasoned.-Approximate averages, deduced from many experiments made with the U. S. Government testing-machine at Watertown, Mass., by Mr. S. P. Sharpless, for the Census of 1880. Seasoned woods resist crushing much better than green ones; in many cases, twice as well. Different specimens of the same wood vary greatly. The strengths may readily vary as much as one-third part more or less from the average.

*Specimens 1.57 ins. square × 12.6 ins. long.

+ Specimens 1.57 ins. square X 6.3 ins. long. Pressure applied at mid-length by a punch covering one-fourth of the length. The first column gives the loads producing an indentation of .01 inch, the second those producing an indentation of .1 inch. (See also page 306).

Expansion of Timber Due to the Absorption of Water.

(De Volson Wood, A. S. M. E., vol. x.)

Pieces 36 x 5 in., of pine, oak, and chestnut, were dried thoroughly, and then immersed in water for 37 days.

The mean per cent of elongation and lateral expansion were:

Expansion of Wood by Heat.-Trautwine gives for the expansion of white pine for 1 degree Fahr. 1 part in 440,530, or for 180 degrees 1 part in 2447, or about one-third of the expansion of iron,

Shearing Strength of American Woods, adapte Pins or Treenails.

J. C. Trautwine (Jour. Franklin Inst.). (Shearing across the

Pine (very resinous yellow).
Poplar.

THE STRENGTH OF BRICK, STONE, ETC. A great advance has recently been made in the manufacture of b the direction of increasing their strength. Chas. P. Chase, in Engi News, says: "Taking the tests as given in standard engineering book or ten years ago, we find in Trautwine the strength of brick given as 4200 lbs. per sq. in. Now, taking recent tests in experiments m Watertown Arsenal, the strength ran from 5000 to 22,000 lbs. per sq. the tests on Illinois paving-brick, by Prof. I. O. Baker, we find an a strength in hard paving brick of over 5000 lbs. per square inch. The a crushing strength of ten varieties of paving-brick much used in the find to be 7150 lbs. to the square inch."

A recent test of brick made by the dry-clay process at Watertown Ar according to Paving, showed an average compressive strength of 39 per sq. in. In one instance it reached 4973 lbs. per sq. in. A test was at the same place on a "fancy pressed brick." The first crack deve at a pressure of 305,000 lbs., and the brick crushed at 364,300 lbs., or lbs. per sq. in. This indicates almost as great compressive streng granite paving-blocks, which is from 12,000 to 20,000 lbs. per sq. in.

The following notes on bricks are from Trautwine's Engineer's Po book:

Strength of Brick.-40 to 300 tons per sq. ft., 622 to 4668 lbs. pers A soft brick will crush under 450 to 600 lbs. per sq. in., or 30 to 40 tons square foot, but a first-rate machine-pressed brick will stand 200 to 400 per sq. ft. (3112 to 6224 lbs. per sq. in.).

Weight of Bricks.-Per cubic foot, best pressed brick, 150 lbs.; pressed brick, 131 lbs.; common hard brick, 125 lbs.; good common b 118 lbs.; soft inferior brick, 100 lbs.

Absorption of Water.-A brick will in a few minutes absorb 1 34 lb. of water, the last being 1/7 of the weight of a hand-moulded one, of its bulk.

Tests of Bricks, full size, on flat side. (Tests made at Wa town Arsenal in 1883.)-The bricks were tested between flat steel buttres Compressed surfaces (the largest surface) ground approximately flat. bricks were all about 2 to 2.1 inches thick, 7.5 to 8.1 inches long, and 3.5 3.76 inches wide. Crushing strength per square inch: One lot ranged fr 11,056 to 16,734 lbs.; a second, 12,995 to 22,351; a third, 10,390 to 12,709. Oth tests gave results from 5960 to 10,250 lbs. per sq. in. Crushing Strength of Masonry Materials. (From How "Retaining-Walls.")

tons per sq. ft. Brick, best pressed.. 40 to 300 Chalk.... 20 to 30 Granite.... ......... 300 to 1200

tons per sq. Limestones and marbles. 250 to 1000 Sandstone................ 150 to 550 Soapstone.................................... 400 to 800

Strength of Granite.-The crushing strength of granite is common rated at 12,000 to 15,000 lbs. per sq. in. when tested in two-inch cubes, an only the hardest and toughest of the commonly used varieties reach strength above 20,000 lbs. Samples of granite from a quarry on the Cor

necticut River, tested at the Watertown Arsenal, have shown a strength of 35,965 lbs. per sq. in. (Engineering News, Jan. 12, 1893).

Strength of Avondale, Pa., Limestone-(Engineering News, Feb. 9, 1893).-Crushing strength of 2-in. cubes: light stone 12,112, gray stone 18,040, lbs. per sq. in.

Transverse test of lintels, tool-dressed, 42 in. between knife-edge bear ings, load with knife-edge brought upon the middle between bearings: Gray stone, section 6 in. wide x 10 in. high, broke under a load of 20,950 lbs. Modulus of rupture...

2,200

1,170

66

Light stone, section 814 in. wide x 10 in. high, broke under........ 14,720
Modulus of rupture.....
Absorption.-Gray stone....... .............................................................. .051 of 1%

Light stone..

.....................

Transverse Strength of Flagging.

(N. J. Steel & Iron Co.'s Book.)

EXPERIMENTS MADE BY R. G. HATFIELD AND OTHERS.

.052 of 1%

b = width of the stone in inches; d its thickness in inches; 1= distance between bearings in inches.

The breaking loads in tons of 2000 lbs., for a weight placed at the centre of the space, will be as follows:

Thus a block of Quincy granite 80 inches wide and 6 inches thick, resting on beams 36 inches in the clear, would be broken by a load resting midway 80 X 36 between the beams= 36

X.62449.92 tons.

STRENGTH OF LIME AND CEMENT MORTAR.

(Engineering, October 2, 1891.)

Tests made at the University of Illinois on the effects of adding cement to lime mortar. In all the tests a good quality of ordinary fat lime was used, slaked for two days in an earthenware jar, adding two parts by weight of water to one of lime, the loss by evaporation being made up oy fresh additions of water. The cements used were a German Portland, Black Diamond (Louisville), and Rosendale. As regards fineness of grinding, 85 per cent of the Portland passed through a No. 100 sieve, as did 72 per cent of the Rosendale. A fairly sharp sand, thoroughly washed and dried, passing through a No. 18 sieve and caught on a No. 30, was used. The mortar in all cases consisted of two volumes of sand to one of lime paste. The following results were obtained on adding various percentages of cement to the mortar: Tensile Strength, pounds per square inch.

Age.....

Days. Days. Days. Days. Days. Days. Days.

MODULI OF ELASTICITY OF VARIOUS MATE The modulus of elasticity determined from a tensile test of a b material is the quotient obtained by dividing the tensile stress in p square inch at any point of the test by the elongation per inch produced by that stress; or if P = pounds of stress applied, K= tional area, = length of the portion of the bar in which the ment is at is made, and the elongation in that length, the mo እ Pl elasticity E = The modulus is generally measured w elastic limit only, in materials that have a well-defined elastic limit iron and steel, and when not otherwise stated the modulus is under be the modulus within the elastic limit. Within this limit, for such n the modulus is practically constant for any given bar, the elongati directly proportional to the stress. In other materials, such as ca which have no well-defined elastic limit, the elongations from the be of a test increase in a greater ratio than the stresses, and the mo therefore at its maximum near the beginning of the test, and con decreases. The moduli of elasticity of various materials have alrea given above in treating of these materials, but the following tabl some additional values selected from different sources:

The maximum figures given by many writers for iron and stecl, 40,000,000 and 42,000,000, are undoubtedly erroneous. The modulus of ticity of steel (within the elastic limit) is remarkably constant, notwiths ing great variations in chemical analysis, temper, etc. It rarely is f below 29,000,000 or above 31,000,000. It is generally taken at 30,000,0 engineering calculations. Prof. J. B. Johnson, in his report on Long Pine, 1833, says: "The modulus of elasticity is the most constant and rel property of all engineering materials. The wide range of value of modulus of elasticity of the various metals found in public records mus explained by erroneous methods of testing."

In a tensile test of cast iron by the author (Van Nostrand's Science Se No. 41, page 45), in which the ultimate strength was 23,285 lbs. per sq. the measurements of elongation were made to .0001 inch, and the mod of elasticity was found to decrease from the beginning of the test follows: At 1000 lbs. per sq. in., 25,000,000; at 2000 lbs., 16,666,000; at lbs., 15,384,000; at 6000 lbs., 13,636,000; at 8000 lbs., 12,500,000; at 12,000 11,250,000; at 15,000 lbs., 10,000,000; at 20,000 lbs., 8,000,000; at 23,000 1 6.140,000.

A factor of safety is the ratio in which the load that is just sufficient overcome instantly the strength of a piece of material is greater than greatest safe ordinary working load. (Rankine.)

Rankine gives the following "examples of the values of those fact which occur in machines":

The great factor of safety, 40, is for shafts in millwork which transmit very variable efforts.

Unwin gives the following "factors of safety which have been adopted in certair cases for different materials." They "include an allowance for ordine.ry contingencies."

Unwin says says that "these numbers fairly represent practice based on experience in many actual cases, but they are not very trustworthy." Prof. Wood in his "Resistance of Materials" says: "In regard to the margin that should be left for safety, much depends upon the character of the loading. If the load is simply a dead weight, the margin may be comparatively small; but if the structure is to be subjected to percussive forces or shocks, the margin should be comparatively large on account of the indeterminate effect produced by the force. In machines which are subjected to a constant jar while in use, it is very difficult to determine the proper margin which is consistent with economy and safety. Indeed, in such cases, economy as well as safety generally consists in making them excessively strong, as a single breakage may cost much more than the extra material necessary to fully insure safety."

For discussion of the resistance of materials to repeated stresses and shocks, see pages 238 to 240.

Instead of using factors of safety it is becoming customary in designing to fix a certain number of pounds per square inch as the maximum stress which will be allowed on a piece. Thus, in designing a boiler, instead of naming a factor of safety of 6 for the plates and 10 for the stay-bolts, the ultimate tersile strength of the steel being from 50,000 to 60,000 lbs. per sq. in., an allowable working stress of 10,000 lbs. per sq. in. on the plates and 6000 lbs. per sq. in. on the stay-bolts may be specified instead. So also in Merriman's formula for columns (see page 260) the dimensions of a column are calculated after assuming a maximum allowable compressive stress per square inch on the concave side of the column.

The factors for masonry under dead load as given by Rankine and by Unwin, viz., 4 and 20, show a remarkable difference, which may possibly be explained as follows: If the actual crushing strength of a pier of masonry is known from direct experiment, then a factor of safety of 4 is sufficient for a pier of the same size and quality under a steady load; but if the crushing strength is merely assumed from figures given by the authorities (such as the crushing strength of pressed brick, quoted above from Howe's Retaining Walls, 40 to 300 tons per square foot, average 170 tons), then a factor of safety of 20 may be none too great. In this case the factor of safety is really a "factor of ignorance.

The selection of the proper factor of safety or the proper maximum unit stress for any given case is a matter to be largely determined by the judgment of the engineer and by experience. No definite rules can be given. The customary or advisable factors in many particular cases will be found where these cases are considered throughout this book. In general the following circumstances are to be taken into account in the selection of a factor:

1. When the ultimate strength of the material is known within narrow limits, as in the case of structural steel when tests of samples have been made, when the load is entirely a steady one of a known amount, and there is no reason to fear the deterioration of the metal by corrosion, the lowest factor that should be adopted is 3.

2. When the circumstances of 1 are modified by a portion of the load being variable, as in floors of warehouses, the factor should be not less than 4. 3. When the whole load, or nearly the whole, is apt to be alternately out on and taken off, as in suspension rods of floors of bridges, the factor should

be 5 or 6.

4. When the stresses are reversed in direction from tension to compression, as in some bridge diagonals and parts of machines, the factor should be not less than 6.