girder placed upon its side; the horizontal web or diaphragm being only sufficiently thick to ensure lateral stiffness. In this section all the centre lines were axes of symmetry, and consequently intersected each other in its centre of gravity; and the horizontal axes were easily made to intersect the centre of gravity of the web-joints. The chief mass of metal was also placed immediately contiguous to the bars of the web, which transferred the stress to the boom, instead of being at some distance from them, as in the trough-shaped and T-shaped form of boom. The material was likewise disposed in the best possible manner for resisting vibration, while this section gave complete facilities for examination and painting. The ends of each truss rested upon hinged bearings by means of cast iron saddles riveted to the junction of the endmost bars of the truss, rollers being provided at one end.

The means adopted in the designs of these girders to obtain the utmost economy of weight consistent with moderate economy of workmanship were: The closest practicable approximation of the average strength to the minimum strength; the observance throughout of the condition of uniform stress, in order that all the compressed members might be trusted with the least possible weight of stiffening; the preference of riveted web-joints to those formed by single pins, and such an arrangement of the riveting that every bar or plate subject to tension should have its whole width, less the diameter of only one rivet hole, available to resist the tensile force applied to it.

XL P

I

Lastly, the author demonstrated the true value of the condition of uniform stress by an exact comparison of the state of a bar of the top boom of the truss of bridge No. II, when under uniform stress, with that condition of unequally distributed stress that would occur if the boom had a suitable trough-shaped section of equal area, breadth, and depth, and therefore of equal nominal value, the elasticity of the material being assumed as perfect. The first case considered was where the stress was uniform in intensity, and the second in which the stress was unequally distributed. The final result was denoted by the equation p=po ; and applying this formula* to the case of the trough-shaped section of the boom, supposed to be equivalent to the H-shaped section actually used, the following was obtained: The area of section was exactly the same, being 36-17 square inches. The inside depth of the trough, 10 inches, would permit precisely the same disposition of the rivets of the web-joint, so that the centre of pressure was situated at the same perpendicular distance, 5 inches, from the lower *In this formula, p is the intensity of the stress per unit of area at the distance from the neutral axis of the stress which intersects the centre of gravity of the section. Po is the intensity of the stress considered as uniformly distributed over the surface of section; that is, the total pressure P upon the entire surface of section divided by the area of that surface. L is the perpendicular distance of the centre of pressure from the neutral axis, and I is the moment of inertia of the surface with respect to that axis. x is or according as it is measuree on one side or the other of the neutral axis.-C. R.

edges of the trough, as from the edge of the H-shaped section actually used. The centre of gravity was found to be situated at 8.088 inches perpendicular distance from the lower edge of the trough, and 2.537 inches from the top edge. The magnitude of the total stress upon the section was 125 tons. The uniform intensity of the stress was 3.45 tons per square inch, and the moment of inertia with respect to the axis was 336-892. From these data the greatest stress was found to be 12.717 tons per square inch at the extreme edges or corners of the sides, and the least intensity 0.544 ton per square inch along the extreme bottom of the trough. In this result the effect of flexure was purposely omitted.

In summing up the conclusions sought to be established, it was submitted that

First, a comparatively small deviation of the centre of the stress, upon the cross section of any bar of any piece of frame-work from the centre of gravity of that section, produced within the limits of elasticity a very great inequality in the distribution of the stress upon

that section.

Secondly, if it were conceded that the real strength of every structure was inversely proportional to the greatest strain suffered by its weakest member, then the existence of this unequal distribution of the stress must be detrimental to the strength of any structure in which it existed, and which had been designed upon the supposition that the mean intensity of the stress upon any bar was necessarily a correct measure of its strength.

Thirdly, there was no practical or theoretical difficulty in designing a truss or girder in which the stress upon every cross section, of all the important members at all events, should be absolutely uniform.

Fourthly, the condition of uniform stress was perfectly consistent with the utmost economy of material in the structure to which it was applied.

On a new Method of Working Atmospheric Railways.-By FRANCIS CAMPIN, Č.E.

From the London Artizan, Feb., 1865.

The atmospheric system of railway propulson having been, in several localities, tested as to its practicability and found wanting, appears for some years to have been set aside, no attention being paid to it except by those personally interested in its success; and so the locomotive has had a clear field, having no rival except in one or two instances where the gradients are too heavy for the ordinary mode of working, and stationary power is requisite. Recently, however, the subject of atmospheric propulsion has received attention, as it seems suitable, if its peculiar difficulties can be overcome, for the working of underground railways where the presence of the locomotive is frequently inconvenient. The obvious advantages of a perfect system of atmospheric propulsion are briefly, simplicity of machinery, impossibility

of collision, as two trains cannot meet on one line of rails, nor can one overtake that preceding it; and reducing the weight of the train by displacing the locomotive, and the consequent reduction of wear and tear in the permanent way. The difficulties which have hitherto resisted the practical application of the system are, the mode adopted for connecting the propelling piston to the train of carriages; the continuous valve on the air-tube working very unsatisfactorily; the low pressure at which the propelling power is worked, by the reason that only the atmospheric pressure is available, because if compressed air were admitted within the air-tube, the valve would be blown open and the operation of the machinery stopped; and a valve could scarcely be made to open inwards, as, in that case, it would come in the way of the piston within the tube.

By the method which it is the object of the present paper to describe, any reasonable pressure may be employed, and the difficulty of the valve is obviated by doing away with it entirely. It is the invention of Mr. William Lake, C.E., who has devoted considerable attention to this subject, which is now promising fruitful results to those who shall succeed in practically establishing the utility of the atmospheric railway.

I will now proceed to explain the details of the machinery by means of which it is proposed to conduct the traffic according to Mr. Lake's principle, commencing with the tube and gearing on the carriages.

The air-tube laid between the rails, instead of being rigid, as heretofore, is made of some flexible substance, such as india rubber, upon which runs a roller attached to the bottom of the carriage or carriages to be propelled, the roller also being coated with an elastic substance, and fixed at such a level that it will, by pressing on the tube, collapse and close it at the point of contact: Then, if air be forced into the tube, the roller will be forced along, drawing the train with it, the effect upon the roller being the same as if it were placed upon an inclined plane. The annexed sketch may, perhaps, make the action above described more clear. Let A B represent a portion of the elastic air-tube, compressed so as to

be closed by the roller c. If air be forced into the tube at A, then, by reason of the pressure exerted at the point d, or rather along the inclined surface d of the tube, the roller

c is caused to progress along the tube in the direction of the arrow, or from A to B, and vice versa. The impelling force will be proportional to the area of the tube A B, and the pressure of the air forced into it, but care must be taken that the diameter of the roller c is properly proportioned to that of the tube A B, or a point may be arrived at where motion will cease, if the roller be continually reduced in size-the larger it is the easier will it work, as the propelling power will have, as it were, a greater amount of leverage; this leverage being proportional to the distance from the centre of force d, acting on the roller to the bottom point e of the roller:

The first question to be answered as to the practical working of this project on a large scale, is: Will the elastic tubing bear the pressure thrown upon it, and the constant passing of the roller over it? To which it is answered, that not only is it possible to make elastic tubing to fulfil these conditions, but responsible manufacturers are prepared to undertake it, and to guarantee it to a much higher pressure than would ever be required in working, or, in figures, to upwards of 250 lbs. per square inch. Let us now take a case, and determine the size of tube necessary to propel a train at average velocities upon an incline of 1 in 100 rise; the weight of the train will be taken as 50 tons, the speed about 30 miles per hour, and the working effective pressure of the compressed air in the elastic tube 50 lbs. per square inch. In order to be on the safe side, the resistance of the train in rolling friction will be assumed as 20 lbs. per ton.

=

=

The resistance from friction will be 50 tons X 20 lbs. 1000 lbs.; that due to the incline, 50 tons X 2240 lbs. 100 (ratio of gradient) =1120 lbs.; total, 1000 lbs. + 1120 lbs. 2120 lbs. propelling force required; this, divided by the working pressure, gives the area of the working tube. Thus, 2120 lbs. 50 lbs. 42-4 square inches, corresponding to a diameter of 7.375 inches. This size is so moderate that it is evident there is no difficulty about getting the requisite power to propel trains upon any incline-(thus, if the gradient were even 1 in 30, the tube would only need an increase of diameter of from 7.375 inches to 11 inches; and it may here be noted that the roller will be, within certain limits, capable of acting on tubes of varying diameters). Now, let us determine the amount of power requisite to maintain the motion of the train at 30 miles per hour on the gradient of 1 in 100. The propelling force is 2120 lbs., and 30 miles per hour equals 2640 feet per minute, giving for the work per minute 2640 feet 2120 lbs. =5,596,800 foot lbs., requiring for its performance 169 6 horse-power. To run the same train on a level, the work would be 2640 feet X 1000 Ibs. 2,640,000 foot lbs. corresponding to 80 horse power.

We will now pass on to the means to be adopted for supplying the air at the required pressure to the elastic air-tube. Stationary engines placed at certain points on the line would be employed to force air, not at once into the air-tubes, but into an accumulator, which would at the same time act as a governor. This accumulator would consist of a cylinder with a loaded piston, on which the weight would be such as should equal the pressure to be created in the air-tube; then, if the pumping engine worked too fast, the piston in the accumulator in rising would close the throttle valve, and vice versa. In the case of the train leaving one section of tube for another, which would occur at each pumping station, the following course would be adopted to prevent inconvenience arising from the sudden removal of the load from the engine at the previous station :

The train, in passing the end of the section of tube upon which it has been traveling, is caused to act upon a tumbler attached to the permanent way, which tumbler closes the valve admitting air from the accumulator into the elastic air-tube; then the engine pumps air into

the accumulator only, until the plunger therein, by rising, cuts off the steam, and so stops the engine. The same result may be produced in other ways, but the above appears most certain, and, therefore, preferable.

In applying the foregoing method to the propulsion of railway carriages, it is argued that great advantages will be obtained; for the tube is small in size, very durable, and easily replaced; and, further, a uniform speed can be maintained, as the tubes can be increased in size on gradients, the same roller being capable of acting upon tubes of various diameter, as mentioned above, and, by these means, the speed of the train is maintained without raising the pressure in the elastic tube.

Pressure Gauges and other Instruments used in Steam Engineering. By Mr. A. BUDENBERG.

From the London Artizan, February, 1865.

or

Letters patent was granted to Mr. Schaeffer early in 1850, by the Prussian government, for the invention of the "Manometer," Pressure and Vacuum Gauge, which I am about to describe to you, and which has become so great a success, as may be judged by the

fact that, since the 23d of January, 1850, when those letters patent were granted, upwards of 60,000 gauges of that construction have been sold. The well known difficulty to obtain letters patent from the government of Prussia, which is particularly illustrated by the refusal of such for Giffard's Injector, gives some reason to believe