Knowles Patent Steam Pumps THE STANDARD. Every Variety of Steam Pumping Machinery, viz.: Knowles Steam Pump Works, 86 LIBERTY STREET, NEW YORK. BLAKE'S IMPROVED STEAM PUMPS. MORE THAN 13,000 IN USE. SIMPLE, POSITIVE, COMPACT, DURABLE. Combined Pumps and Boilers for Railroad Water Stations a Specialty. Improved Compound Duplex Pumping Engines for Water-Works. GEO. F. BLAKE MANUFACTURING CO., june '81 ly 88 LIBERTY STREET, NEW YORK. THE Franklin Institute is not responsible for the statements and opinions advanced by contributors to the JOURNAL. ON THE EFFECT OF PROLONGED STRESS UPON THE STRENGTH AND ELASTICITY OF PINE TIMBER. BY PROF. R. H. THURSTON. Presented to the American Association for Advancement of Science, In papers read before the American Society of Civil Engineers at various dates,* the writer has given the results of investigations made to determine the behavior of metals under loads of varying magnitude and under intermitted stresses, and to ascertain in what cases and under what conditions the variation, with period of stress, of the normal line of elastic limits, discovered and announced by him in the year 1873, occurs in practice. Experiments made by Mr. Herman Haupt,† forty years ago, revealed a fact not even now generally understood and appreciated—that timber may be injured by a prolonged stress far within that which leaves the material uninjured when the test is made in the usual way and occupies a few minutes only. Thus, using pieces 60×3×1 inches (152.4×7·62×2.54 cm.) set as *Trans. Am. Soc. C. E., 1873-80; Jour. Franklin Inst., etc., etc. + Haupt on Bridge Construction. WHOLE NO. VOL. CXII. THIRD SERIES, Vol. lxxxii.) 11 = cantilevers with a breaking moment, due the load, of P 48P inch-pounds (122P kilog.-metres) he obtained for the value of 6 w l m R = the following figures: b d2 All samples tested were considered good selected timber. An extended series of experiments made intermittently in the mechanical laboratory of the Stevens Institute of Technology, Department of Engineering, during some years past,* had included an examination of this subject and the result has confirmed Haupt's earlier work and has given a tolerably good idea of the effect of prolonged stress in modifying the primitive relation of stress and strain where the wood is good Southern yellow pine. A selected yellow pine plank was obtained for test, the history of which was known. The stick was cut at Jacksonville, Florida, in October, 1879, was received early in the following year and was piled in the yard, air-seasoning, until taken for test in the spring of 1880. The plank measured 4" × 12" × 24′ (10·16 × 30-48 × 731-52 cm.). When tested, it had been seasoning six months, the latter part of the time indoors. From the middle of this plank a stick was first cut 3" x 3" × 24' (7·62 × 7·62 × 731·5 cm.) and from this was cut a set of ten pieces from 40" to 54" long (101-6 to 137-2 cm.) and from 11" to 3" square in cross-section (3·16 to 7·62 cm.) square. These latter pieces were *Trans. Am. Assoc. for Advancement of Science, 1879-1880; Journal Franklin Inst., Oct., 1879, Sept., 1880. tested on various conditions, as then reported,* to determine the values of their moduli of elasticity and of rupture. The moduli of rupture were usually 11,000 to 12,000 for the expression R = PI bd2 (in metric measure, 773-3 to 843-6) and the moduli of elasticity ranged from two to two and a quarter millions (in metric measure, 106 X 1406 to 158175 X 10'). In specific gravity the wood ranged from 0.75 to 100, usually about 0-85. When kiln-dried to a moderate extent, the density was but little altered, if at all, but the modulus of elasticity rose to two and a half millions (17375 × 103) and the modulus of rupture was increased about 20 per cent. From the previously unused part of the plank a set of three test pieces was cut about 1 inch (2.54 cm.) square in section and tested on supports 40 inches (101.6 cm.) apart, to determine their breaking loads. The result is shown in detail in the appended table. In these specimens the annual rings were in the cross-section of each piece, indicated by lines making angles of 45° with the edges. These pieces broke at 345, 380 and 410 pounds respectively. The weakest piece broke by splintering, and had it been as sound as the others would probably also have sustained a somewhat heavier load. As will be seen by comparison with the other and with subsequent tests, the deflection of the strongest piece in the set is exceptionally small and the piece probably exceptionally strong and stiff. We may therefore take 375 pounds (170 kilog.), or a trifle over, as a good average for loads breaking pieces of this size. Nine other pieces were cut and dressed to the same size and were mounted on supports 40" apart, in a frame arranged for the purpose in the workshop of the Institute, in three sets of three each. 350 pounds (158.1 kilogs.); Table 4. 3d set, Or to about 60, 80 and 95 per cent. of their probable maximum strength, as indicated by ordinary test of the companion lot above described. Their deflections were measured when set, and at intervals subsequently, by means of an accurate micrometer reading to tenthousandths of an inch. *Trans. Am. Assoc., 1880; Journal Franklin Inst., 1880. The whole set of bars, loaded most heavily as above, broke within two days; one bar yielding, as shown in Table 2, at the end of a period included between observations taken at 4 and 133 hours from the beginning, the second breaking at some time between 27 and 301 hours and the third giving way at the end of 43 hours. A load of 87 per cent., the maximum obtained by usual methods of test, is thus shown to be capable of breaking the piece under the conditions here described, and an apparent "factor of safety" of 14th is evidently not a factor of safety at all when time is given for the piece to yield. The second set, loaded with 0.75 the maximum momentary weight, all broke, as is shown by Table 3, one at the end of about 3 days, another after 5 days, and the third at the end of a little more than a month. It is probable that these differences of time are due to differences of strength more than to variations of the effect of time of stress. A "factor of safety" of 13 is evidently not a real factor of safety for wood in such cases as this. The behavior of the third and last set of test pieces is shown in Table 4. These pieces were loaded with 60 per cent. of the average breaking weight under ordinary test. Left under this load, the deflection, in every instance, slowly and steadily increased from about one inch (2·54 cm.) to some considerably larger amount at the end of the period of investigation. Fortunately, as is indicated by a comparison of these initial deflections with those observed under the same weights when testing the first set, and by their close accordance with each other, these pieces were all good samples of a good quality of yellow pine. The increase of deflection was almost precisely the same for all for several months, a fact which is of importance, as showing not only the gradual progress and the steadiness of yielding, but also that no accident produced final rupture. Finally, after several months (about 6000 hours; the exact time is uncertain), the piece which had at the beginning shown most pliability broke completely down. The next piece to break was that which was intermediate in stiffness between the two others; it broke at the end of about 9000 hours-precisely one year from the date on which the load was imposed. The last of the three pieces of this set still carried its load of 60 per cent. of the maximum under ordinary test at the last date, but it was still very slowly but unmistakeably yielding, its deflection having increased nearly 04 inch (1016 cm.) during the preceding five |