to weld together, they slide upon one another, the tissue is lengthened, and the fibrous condition results. Whenever a slight indication of fibres is seen in large granular bars, it will be found to occur at the junction of the different pieces forming the pile, where, owing to the interposition of scoriæ, the welding has not been perfect.

The microscope shows that while in granular iron, the degree of homogeneity is considerable, the contrary is the case with that of a fibrous nature. Moreover, it confirms the statements respecting the purity and density of granular iron, and at the same time demonstrates the cellular and imperfectly amalgamated condition of the fibrous metal. The microscope also shows the want of uniformity of texture in mixed irons, and the presence of foreign bodies in iron made with coal, after the English method. The fineness of the grain, considered absolutely, is not due to the presence of carbon, and, consequently, does not necessarily indicate a steely character. Swedish irons are examples of this; for, although of a decidedly steely nature, they frequently are very large grained. The fine grain merely insures iron of superior quality as regards its tenacity. The microscope leaves no doubt whether coal or wood has been employed. Fibrous iron made with wood fuel always presents a bright section, never the dull, black appearance belonging to that made with coal.

To prove the evil effects of imperfect welding, an experiment was made in which 33 lbs. of old iron were added to a charge, in the hope that it would become incorporated with the whole mass made from the pig. The old iron became oxidised, and formed a scoria which remained in the metal. When old iron was not added in this manner, the quality was excellent. Granular iron being more readily welded than fibrous, its density is greater in the proportion of 7.791 to 7·751. Other figures give 7.78 and 7.60. For the same reason, granular iron resists strains of tension and compression better than the other, but it is weaker with regard to flexure, although possessing greater elasticity than fibrous iron. With respect to malleability and ductility, the fibrous or non-welded iron is more malleable, but less ductile than the granular. The metalloids sulphur and phosphorus play opposite parts in their influence upon iron. The former prevents the perfect welding, and, owing to the formation of scoriæ, the fibres are short and black. On the other hand, phosphorus assists this amalgamation or welding of the particles and produces a granular iron. From this it is maintained that iron has but one normal condition of particles, the granular, which is based upon the essential property of welding or amalgamation. All other textures are simply the result of a defective and imperfect welding in the process of manufacture.

C. T.

On the Mechanical Properties of Gun-metal. By M. TRESCA,

(Annales du Conservatoire des Arts et Métiers, No. 38, pp. 324–334.)

It was found during the siege of Paris that specimens of gunmetal, when tested by tension, gave different results. This led to careful experiments on three varieties of bronzes, of which the composition was as follows:

Rectangular bars, 2 inches by 1 inch (0.05 mètre by 0.025 mètre) in section, were tested by flexion, and bars 1 inch (0.025 mètre) square, and round bars inch (0·012 mètre) in diameter by tension. The mean results of the tests are as follows, the units of length and of section being 1 mètre (3.28 feet) and 1 square mètre respectively.

The English weights are referred to 1 square inch as the unit of section. B in its fractured section had a metallic lustre and

numerous grains of tin; P had a dull appearance, grained surface, and uniform texture; L had a metallic lustre, grained surface, and very uniform texture. It appears from the above table that the moduli of elasticity of B, P, and L are in the ratios of 1.00, 1.09, and 1.20, and that the limit of elasticity of L exceeds that of B and P by one-fourth, the elongation at the limit being the same for all.

The mechanical work expended is represented by the products of the weights into the corresponding elongations; and in the case of rupture the following proportions exist:

so that 7 times more mechanical work must be expended to break a bar of L than is required for B. The superiority of the Laveissière bronze in every respect over the other two is evident; and the phosphoric bronze is shown to be better than the ordinary bronze of Bourges.

L. V. H.

Experimental and Geometrical Investigation of Internal
Ballistics. By GENERAL MORIN.

(Annales du Conservatoire des Arts et Métiers, No. 38, pp. 304-323, 1 pl.) While acknowledging the efforts of Hutton, and while giving due credit to the elegant experiments of Rumford, it must be admitted that Piobert, following the wise precept of Bacon to base his reasoning on experiment, was the first to establish a mathematical theory of the expansive forces of explosives in fire-arms, and of the laws of motion communicated to the projectiles, which laws are called internal ballistics. In 1846-47 the relative merits and behaviour of gunpowder and gun-cotton were investigated, by determining the velocities imparted to bullets weighing 444 grains (28-8 grammes), and having a diameter of inch (17 millimètres), fired with charges of 123 grains (8 grammes) of powder, or 44 grains (2·86 grammes) of gun-cotton, in gun-barrels of ten different lengths, varying from 4 calibres to 64 calibres. The results of these experiments may be represented by curves, the abscissæ being the spaces traversed by the projectile, drawn full size, and the ordinates, half its 'vires vivæ,' drawn to a suitable scale on section paper. The inclination of tangents to the curve thus described, where the W W VdV WdV ordinate is

V2 and the abscissa s, is

g

ds

=

gdt

which gives the motive force imparted to the projectile; or the resultant of the expansive force, the resistance to alteration of shape, friction, and displacement of air. By marking off abscissæ at definite intervals, drawing ordinates to meet the curve, and

assuming the portion of the curve intercepted between each ordinate to be a straight line, the curve becomes a polygon, the sides of which correspond to the inclinations of the required tangents. These are easily calculated, and the values of the forces imparted to the projectile at any part of its passage through the barrel are in this manner readily found. The results thus arrived at approach much nearer to the truth than do those obtained by M. Pothier's graphic method of drawing normals, applied by him in working out the results of certain experiments conducted with cannon in 1869. The conversion of the curve into a polygon implies that the force exerted on the projectile may, without great error, be considered constant in the space between each ordinate, and that the motion of the projectile may be taken to vary uniformly between these points. On this assumption the following relations are obtained between s, the space traversed; t, the time of transit; V, the velocity acquired; and F, the force :

1

239

W

These formulæ show the total time occupied by the standard bullet in traversing a barrel 31 feet (1 mètre) long to be, with the standard charge of gunpowder, 35 second, and, with gun-cotton, 3 second. They also show that the velocity acquired at the muzzle should be 1,247 feet (380 mètres) with gunpowder, and 1,260 feet (384 mètres) with gun-cotton, the actual velocity being 1,234 feet (376 mètres) for both. The maximum motive force, exerted by the explosive, is to the mean motive force capable of producing the same velocity, as 2:39 to 1 with gunpowder, and 6.02 to 1 with gun-cotton. These investigations also demonstrate that the velocity imparted by gun-cotton to the projectile at the commencement of its course is greater than that derived from gunpowder, but that the difference in time in passing over equal spaces becomes less as the projectile advances, and just before leaving the barrel the times are the same for each. The velocity imparted to a projectile by gunpowder increases continuously, though more slowly, as it advances; but with gun-cotton the velocity actually decreases before it leaves the barrel, owing to the condensation of the water, which is one of the products of the combustion. L. V. H.

Experimental Researches on Explosive Substances.
By MM. Roux and SARRAU.

(Comptes-rendus de l'Académie des Sciences, Oct. 5, 1874, pp. 757-760.)

It had been shown in a former communication that dynamite might be exploded by two methods. Simple explosion is caused by

the ordinary ignition of the substance; detonation-by the percussion of a strong priming of fulminate of mercury. By these two kinds of explosion very different pressures are produced, and the Authors have endeavoured to measure the relative intensities of these pressures, by the quantities of each explosive substance respectively required to rupture bomb-shells identical in form and dimension. They have shown, further, from recent experiments, that this property of double explosiveness belongs' to the greater number of other explosives besides dynamite. The charge of gunpowder, necessary to produce rupture, was 200 62 grains (13 grammes)-by simple explosion. The ratio of 13 grammes to the rupturing charge of another substance is a measure of the force of the substance, the force of gunpowder by simple explosion being taken as 1. The subjoined Table contains the explosive force, thus experimentally obtained, of various substances, together with the proportion of permanent gases produced by simple explosion, in percentages of the weights of the substances, and the quantity of heat disengaged by 1 kilogramme and 1 lb. of the substance, in French and English units respectively. It is shown that the simple explosive force of gunpowder is more than quadrupled by detonation; that the simple explosive force of a substance is proportional to the product of the weight of gases disengaged by the heat; and that the detonating forces, for six of the substances, are nearly proportional to the heat disengaged.