is screwed a piece of copper, H. The pressure of the gas acting on the piston 1 forces the knife into the copper; by mechanical means a similar cut can be produced, and hence the magnitude of the cut gives the measure of the pressure which has produced it. A small cup at c prevents any gas passing the indenting tool.

The great improvements that Major Rodman made in gunpowder are well known. To him we are indebted both for the earliest experiments on the effect of the size of grain on the maximum pressure and for the powder adopted by all nations for large guns, I mean prismatic powder; but it is a question whether he was not in some degree led to these great improvements by an erroneous estimate of the pressures produced, this erroneous estimate being mainly due to the necessity of placing the Rodman gauge at the exterior of the gun; and the effect of this objectionable position would be greatly exaggerated if the powder experimented with were of a brisante nature.

It is curious that so distinguished an artillerist as Major Rodman FIG. 1.-Rodman's Pressure Apparatus.

should never have taken the trouble to calculate what energies the pressures which his instrument gave would have generated in a projectile; had he done so he would have found that many of the results indicated by his instrument were not only improbable but were absolutely impossible.

As an illustration of Major Rodman's method I take an interesting series of experiments made in smooth-bored guns of 7-inch, 9-inch, and 11-inch calibres, and so arranged that in each gun an equal column or weight per square inch of powder was behind an equal column or weight per square inch of projectile. Under these conditions, in each gun, during the passage of the shot along the bore, the gases would be equally expanded, and the energy per unit of column developed at every point in the three guns should be the same, except for slight differences on account of increased temperature and pressure in the larger guns, due to the smaller cooling surface in proportion to the weight of charge.

Major Rodman measured his pressures at the base of the bore and at

every 14 inches along it, and his results are given in the annexed table, which is a most instructive one :

Examining this table, it will be observed, in the first place, that the muzzle velocities of the equal column projectiles are nearly the same; that of the 11-inch gun being, as it should be, somewhat the higher; hence the energies per square inch must be nearly the same, and the mean pressures per square inch, inducing these energies, must likewise be the same.

But, for example, comparing the 7-inch and the 11-inch guns, it will be noted that in the latter gun the pressures are always twice and sometimes more than four times as great as in the 7-inch gun, the mean pressure being nearly three times as great.

The energy should be in the same proportion; hence, if the pressure observations had been correct, the observed velocity should have been 1,570 f.s., instead of 927 f.s.

It will be noted also that the forward pressures not only differ greatly in the several calibres, but, for instance, in the 9-inch gun the pressure at 56 inches from the bottom of the bore is double the indicated pressure measured at 42 inches. Rodman accepts the pressures up to and including 42 inches as correct, but ascribes the irregular pressures in the chase to the vibrations of the metal due to the discharge.

Some experiments made by the earlier Explosive Committee fully explain the cause of the differences between the pressures exhibited by the 7-inch and 11-inch guns.

In the first of the experiments of this Committee, they used simultaneously Rodman's gauge and the chronoscope to which I shall presently advert. In the former case of course the pressure was determined directly. In the latter it was deduced from the motion communicated to the projectile. The results were quite irreconcilable, as a few examples will show.

In an 8-inch gun, with a charge of 32 lb. of Russian prismatic powder and a projectile of 180 lb. weight, fired from a vent a little in advance of the centre of the charge, and called the forward vent, the chronoscope gave a maximum pressure of 204 tons, while the Rodman gauge gave maximum pressures in the powder chamber varying from 267 to 33-7 tons per square inch. In the same gun, under similar conditions, a similar charge of pellet powder gave, with the chronoscope, a maximum pressure of 19.2 tons per square inch, while the chamber pressures given by the Rodman gauge varied from 41.6 tons to 49-2 tons per square inch.

But perhaps more striking discrepancies were exhibited by two series of experiments with R.L.G. of Waltham Abbey make, fired from the same gun, and developing in the projectile approximately the same energies. In the first of these series, with a charge of 20 lb. fired from a forward vent, the maximum chronoscope pressure was 13-3 tons, while 1894.

M M

the Rodman gauge gave pressures varying from 24-6 to 38-9 tons per inch.

square

In the second series, all conditions being the same, except that the charge was fired from the extreme rear, the maximum chronoscope pressure was 143 tons, while the Rodman pressure varied from 31-6 tons per square inch to over 50 tons per square inch, that pressure being the highest which the instrument was capable of registering, every observation in this series with the gauge placed at the seat of the shot being over fifty tons.

Shortly afterwards the Rodman gauges were destroyed, two of them being blown from the gun.

These discrepancies led the Committee to investigate with certain powders the variation in pressure indicated when a gauge was placed at the surface of the bore and at the exterior of the gun as with the Rodman gauge.

For this purpose they used the crusher gauge, which admits of being placed in both positions.

With pebble powder the gauge placed at the interior of the bore gave 14.5 tons; placed under precisely the same conditions at the exterior it gave 27 tons per square inch. With R.L.G. the similar figures were respectively 20 and 57 tons, and with L.G. respectively 19.5 and 45-5 tons per square inch.

The error I have just discussed was due to the position of the gauge ; but Rodman's pressures and the pressures of the Explosive Committee were exaggerated from another cause. It will be readily understood that if a pressure of, say, 20 tons per square inch be suddenly applied to a gauge, and if the resistance to the motion of the knife be initially trifling, a certain amount of energy will be communicated to the piston and knife; and the copper when measured will indicate not only the gaseous pressure, but in addition a pressure corresponding to the energy impressed upon the piston during its motion.

This cause of error can, however, be eliminated by producing beforehand by mechanical means a cut indicating a pressure a little less than that to be expected.

Rodman admits that his chase pressures are erroneous; their exaggeration is no doubt greatly due to the causes I have just pointed out; but in my opinion, based upon long experience, no gauge of this description placed in the chase, where the products of explosion are moving with a very high velocity, can be depended upon to give reliable results.

If we disregard the energy of the moving products and suppose the gauge to be acted on by pure gaseous pressure, with a projectile moving at the rate of 2,500 f.s. (and such velocities are now quite within the range of practical ballistics), the projectile would pass the entrance to the Rodman gauge in something like the 100th part of a second. It is difficult to imagine that the full indentation could be given to the copper in this small fraction of time, and, if it were not so given, the gauge would indicate the pressure at a point considerably in advance of the gauge.

On the other hand, if, as would generally be the case, the products of explosion moving at a high velocity acted on the piston, the energy of these products would be reconverted into pressure, and the gauge would in this case give too high a result.

Major Rodman appears to have considered it impossible that any gauge could rightly indicate a pressure higher than that indicated by another nearer to the seat of the shot. This, however, is not so; nothing is more certain than that, with the powders known as 'Poudres brutales,' and, possibly, in a less degree with all explosives, motion is communicated to the shot by a series of waves or impulses; and it is easy to see that, if the position of a gauge coincided with the 'hollow' of a wave, while that of a more forward gauge coincided with the 'crest,' the latter might easily show the higher pressure. Later on I shall revert to this point.

The crusher gauge is a modification of the Rodman gauge, designed to overcome some of the defects of that instrument, and it is now almost universally used for the direct measurement of pressure: it is shown in the diagram exhibited (fig. 2), and its action is easily understood. The powder gases act upon the base of the piston, compressing the copper cylinder; the amount of crush on the cylinder serves as an index to the maximum tension acting on the piston. It is usual, where possible, to employ in each experiment two or three gauges so as to check the accuracy of the determination. Properly used, very great confidence may be placed in their results; but, as may be gathered from my remarks on the Rodman FIG. 2.-Crusher Gauge.

gauge, this and all similar gauges will cease to give reliable information as to the energy that can be impressed on a projectile, or as to the mean pressure on the surface of the bore, if there be any probability of the products of explosion being projected into them at a high velocity. In such a case the pressure indicated would not be the true gaseous pressure, such as, for instance, would exist were the products of ignition retained in a vessel impervious to heat until the waves of pressure generated by the explosion had subsided. But I defer an examination of the results given by the crusher gauge until I compare these results with those given by the indirect method of deducing the pressure from the motion of the projectile within the bore.

The method I have adopted for this purpose consists in registering the times at which a projectile passes certain fixed points in the bore of a gun. The chronoscope (figs. 3 and 4), which I have designed for this purpose has been so often described that I shall only here briefly allude to it. It consists of a series of thin discs made to rotate at a very high and uniform velocity through a train of geared wheels. The speed with which the circumference of the discs travels is between 1,200 and 1,300 inches per second, and, since by means of a vernier we are able to divide the inch into