Another fact pointing to the same conclusion was found in the difference of their spectra. In the early part of the evening, while both the zodiacal cone and the aurora were at their brightest, an opportunity was taken to compare their spectra. The observations were made with Eaton's direct vision spectro scope. The zodiacal cone gave a faint, short, continuous spectrum, brightest near its least refrangible end. It was most sharply terminated at the same end, fading off gradually toward the more refrangible end. It was very faint throughout, and could be seen only through a wide slit. The aurora gave a spectrum much longer than that of the zodiacal cone, though of about the same pale greenish color. At or close to the less refrangible end was a sharp, bright line of a grayish-green color. The spectrum from this line toward the more refrangible end was continuous and gradually diminishing in brightness, except that a very faint brightening near the blue showed that a brighter aurora might have given a line at that place. The green line of the aurora could be seen with a very much narrower slit than that required to see the spectrum of the zodiacal cone. If the spectra of the aurora and the zodiacal cone could be superposed, it would be found that the green line of the former is close to the more refrangible end of the spectrum of the latter. The spectra show that while the light of the zodiacal cone is polychromatic, that of the aurora is nearly monochromatic, or, as it might be expressed, oligo-chromatic. The former is such as would be given by sunlight reflected on diffuse matter in space ; the latter might be given by an electrical discharge through a gas. Later in the evening, the faint zodiacal band was observed stretching completely across the sky along the ecliptic, while a round patch of light on the boundary between Virgo and Libra, lying apparently about 3° above the ecliptic, is recognized as the Gegenschein. It is perhaps worth noting that no change of consequence in the weather or temperature took place during the two or three days preceding and following this aurora. EXPERIMENTS ON THE STRENGTH OF YELLOW PINE. BY PROF. R. H. THURSTON, of Hoboken, N. J. [ABSTRACT.] In a paper read at the Saratoga meeting of the American Association for the Advancement of Science the writer presented the result of a series of experiments on the strength of timber, in which were given several unusual figures.1 To determine how far these results were due to peculiarities of the selected samples supplied from the Navy Yard, and to determine to what extent size affects the resistance, a more extended series of transverse tests were made, and the results of experiments upon yellow pine, of the ordinary market qualities and of various dimensions, are now presented below, as determined in the Mechanical Laboratory of the Stevens Institute of Technology. In the paper referred to, the modulus of elasticity was given for yellow pine as an extraordinarily high figure. It will be observed that the best wood here described gives also very high values of E, and a comparison of the pieces of the first with these test specimens shows the selected Navy Yard specimen to have been of better material than either of the latter. Samples marked F1, F2, F3, were from the same plank—a piece of yellow pine cut in Georgia, April, 1879, and tested after several months of seasoning, when it had become thoroughly dry. The three specimens were considered good material. F, was not straight-grained and broke obliquely, giving a much lower modulus of elasticity, as well as of mixture, than its companion specimens. Samples B1, B2, B3, were cut from a stick ten inches square in section, which had been lying under cover, seasoning, nine or ten years. Numbers 1 to 12, inclusive, were small sticks sawn out of the middle of a plank, originally four inches thick, one foot wide, and twenty-four feet long. A stick was first cut three inches square and twenty-four feet long, which was then cut into strips of varying smaller dimensions. The wood was selected from lumber-yard stock, and was considered to be fairly representative of average timber. It was cut in Florida, in October, 1879, and reached Hoboken in January, 1880. 1Among these were several due to printers' errors, which, owing to the illness of the writer, were not corrected in proof. The tenacity of yellow pine, for example, was made 2070.2 when it should have been given at 20,702. Specimens 1 to 8, inclusive, were too green for use in construction; Nos. 9, 10 and 11 were kiln-dried 56 hours, at a temperature of 130° Fahr., No. 12 was dried 12 hours, at 210° Fahr., at which temperature the pitch exuded from the wood quite freely. Still another specimen, not here recorded, was heated 14 hours at 420° F., and was somewhat charred. Under test it gave a modulus of rupture, R= 9000. In the table containing a résumé of results is also given the figures obtained in the earlier experiments on selected Navy Yard material, marked J and those of a sample, K, tested by another observer.2 Specimens A, A,, were picked up in a workshop, and used simply to determine specific gravity; they were probably not the best southern long-leaved pine, such as it was intended that the samples tested should be. "Van Nostrand's Magazine, Feb., 1880, p. 166. The following table contains the values of the Moduli of Rupture and of Elasticity deduced from the whole series of experiments : W 13 Calculating E= for heavy loads and considerable deflections, the following figures were obtained for this pseudo-modulus : TABLE II. Values of E at one-third to one-half total load. As a general rule, the maximum figure was given at one-third, rarely at one-half, and about as often at the beginning of the test. The latter case occurred most frequently with the unseasoned timber. From the test of B, we obtained W. E. TABLE III. W. E. 1050 2,363,000 1250 2,381,000 W. 2,394,000 Heavy loads, long applied, produced fracture of pieces, the companions to which resisted considerably more when the load was steadily increased up to the moment of fracture. The maximum permanent load was apparently something less than one-third and probably greater than one-half the maximum load which could be sustained under ordinary test. A set of four yellow pine beams of large section was tested to determine the value of the Modulus of Elasticity in large timbers, and also to ascertain whether experiments upon small sticks, such as are usually tested, may be relied upon to give correct values of this coefficient. |