CHAPTER III.

THE SYSTEM AND APPARATUS OF THE BATCHELLER PNEUMATIC TUBE COMPANY.

=General Arrangement and Adaptability of the System.=—The experience gained in the construction and operation of the Philadelphia post-office tubes has naturally suggested improvements that can be made in future construction, and, furthermore, it has taught us what the requirements will be of an extensive system of tubes laid in the streets of our cities, both for the transmission of mail and for a general commercial business. Since the Philadelphia post-office tube was completed, we have been busily engaged in working out all the details of a system of many stations so connected together that carriers can be despatched in the most direct manner possible from any station to any other. It is the purpose of the present chapter to describe this system.

While the Pneumatic Transit Company has ample field in the State of Pennsylvania to carry out the work which it has mapped out, a field broad enough to yield a good profit for the capital invested, there is no reason why the system should be limited to one State. So, in order to obtain a broader charter, covering all places where pneumatic service may be needed, a new corporation was formed and styled the BATCHELLER PNEUMATIC TUBE COMPANY.

It is impossible to lay down a rigid system equally well adapted to all places and purposes. We must accommodate ourselves somewhat to circumstances. For example, the post-office department may require one size of tube, arranged to operate in a particular way, while the requirements of a parcel delivery business would be utterly different. The geographical location of the stations will have much to do with the general arrangement; also the condition of the streets. Some of the streets of our large cities are so filled with water- and gas-pipes, electrical conduits, sewers, steam-pipes, etc., etc., that it is almost impossible to find space for pneumatic tubes, especially of large diameter. Railway or water facilities have much to do with the location of a central pumping station, on account of the coal supply. All of these and many other things have to be taken into consideration in planning a system for any locality.

We have an example of a peculiar location and conditions in a proposed line of tubes over the New York and Brooklyn bridge connecting the main post-offices of those cities. This would be in many respects a unique plant. Two air-compressors would be used, one at each office.

In order to give a general idea how a large number of stations can be connected into one system, the diagram Fig. 21 has been drawn.

We have already referred to the attempts of Clay and Lieb to devise means whereby several stations could be located along a main line and carriers be sent from any station to any other through the main line. Their method was to use branch tubes leading off from the main line with switches at the junctions. They deflected the air-current into the branch by placing an automatic closed valve in the main line just beyond the junction, returning the air from the branch to the main line just beyond the valve. The carriers were to open and close this valve automatically as they passed.

The branch and switch system has many attractions for the inventor, and upon first thought it would seem the most feasible solution of the problem. It has been the dream of more than one inventor, as the records of the patent-office show, but no one has succeeded in working it out. The current of air cannot be divided; carriers passing from the branch into the main line must not collide with other carriers running in the main line; a certain minimum distance must always be maintained between the carriers in the same tube; when a carrier is despatched it must go directly to the station for which it is intended without further attention from the sender and it must not interfere with other carriers; expense of manufacture prohibits the use of any but round smooth tubes up to eight inches in diameter, hence projections cannot be placed upon the carrier to give it an individuality and cause it to operate a switch at any particular point along the line; the carrier is free to rotate in a round tube about its longitudinal axis, therefore, its individuality must be indicated by some symmetrical marking about this axis, if it is to be automatic in its operation; the speed of the carrier is so high that electrical contacts placed in distinctive positions on the carriers cannot be used while it is in motion, for mechanism having inertia could not be moved during the short time that the electric circuit would be closed; only the simplest attachments can be made to the carrier, for constructional reasons and because of the rough usage that they receive. These and numerous other reasons make the problem most difficult. We have not attempted to solve it by the use of branch tubes and electrically operated switches, but have adopted the simpler and equally effective method of carrying the main line through each of the stations that it unites. In our system each carrier has an individuality determining the station at which it will be discharged from the tube. By a simple attachment to, the front end of the carrier, consisting of a circular metal disk, the sender so marks the carrier that it will pass all stations until it arrives at the station for which it was destined and will there pass out of the tube. In addition to this a method has been devised whereby carriers can be inserted into the tube without the possibility of collision with carriers already running in the tube.

Referring now to the diagram, Fig. 21, we have here an imaginary system which we will suppose to be located in some large city. The two large squares I and II indicate central pumping stations, and the small squares A, B, C, D, etc., indicate receiving and sending stations. Some of the stations, such as A, B, C, D, E, F, and Y, which do a large amount of business and may be supposed to be large retail stores, are connected directly with the central station by double tubes, one for sending and the other for receiving carriers. Two smaller stores, such as G and H, may be located on the same line. At I, J, K, and L we have four stations, all connected by the same double line of tubes. These stations we will imagine to be located in the residence section of the city. Carriers containing parcels of merchandise or other matter destined for private residences would be sent from the stores A, B, C, etc., to the central station I, where they would be transferred to the line 2 and be adjusted to stop at the station nearest the residence to which the parcels were addressed. From this station the parcels would be delivered by messengers to the residences. If a carrier is to be sent from the central station I to station K, it will be so adjusted before it is put into the tube that it will pass stations I and J, but be discharged automatically from the tube when it arrives at station K. In a similar manner carriers can be despatched from station L to station I or from station J to station L. In passing through the central station the carriers are manually transferred from one line to another.

In another part of the city we may have another central pumping station, II; and the two central stations may be connected by a double trunk line, 3. Again, we have lines radiating from this central station, as shown by station Y. There will be some localities where it will be an advantage to arrange the stations upon a loop, as shown in circuit 4, where stations S, T, U, V, W, and X are connected together in this way. Or we can combine the two arrangements of loop and direct line, as shown in circuit 5. Stations O and R are on the double line, but from O a loop is formed including stations N, M, P, and Q. Here it is supposed that the stations O and R do a much larger business with the central station II than the stations N, M, P, and Q, this being the principal reason for placing them on the double line. All carriers must be returned to the station from which they were sent, or others to replace them, otherwise there will be an accumulation of carriers at some of the stations. It is like a railway: there must be as many trains despatched in one direction as the other, each day. Station O can receive a carrier from the central station and return it directly, but when station N receives one it must be returned via M, P, Q, O, and R, a much longer route than that by which it was received. This disadvantage is compensated, when stations N, M, P, and Q do only a small amount of business, by the less cost of laying a single line. If a carrier is to be sent from M to N, it must go via P, Q, and O, being manually transferred at O from the “down” to the “up” line. P can send directly to Q, but Q must send to P via O, N, and M. R can send directly to O and O to R. Similarly in circuit 4 the carriers must all travel around the loop in the same direction, shown by the arrows. Station S can receive carriers directly from the central station, but they must return via U, W, X, V, and T.

Again, we may have a double-loop line, as indicated in the diagram by circuit 6. Here five stations, _a_, _b_, _c_, _d_, and _e_, are connected by a loop consisting of two lines of tube, in which the air circulates in one direction in one line and in the opposite direction in the other. Here _b_ can send directly to _c_, _c_ directly to _b_, and _e_ to _b_ via _d_ and _c_, or via central and _a_. This is an arrangement that would be used where there is a large amount of business between the stations on the loop. As stated before, the best arrangement for any particular locality depends entirely upon circumstances.

=Size of Tubes.=—The pneumatic-tube system that we are describing is not limited to any particular size of tube. The size is usually determined by the number and size of packages to be transported. A small tube, two or three inches in diameter, is best suited for telegrams and messages; mail, parcels, etc., require a six- or eight-inch tube, while mail pouches and bulky material, a thirty-six inch or possibly larger tube. We divide tubes into three classes, according to their size, naming them small, large, and very large tubes. By small tubes we mean those not larger than three or possibly four inches in diameter. Large tubes are those having a diameter more than four inches and not more than eight inches. Very large tubes include all that are more than eight inches. This classification is for convenience, but it has a deeper significance. For example, in the transportation of mail, it must either be handled in bulk, that is, in pouches, or in broken-bulk, that is, loose or tied up in small packages. There are many advantages in transporting it in broken-bulk, in fact, there are very few places where it could be handled in any other way. For this service six- or eight-inch tubes—not larger—are best suited. The carriers are light enough to be easily handled; they are not so large in capacity as to make it necessary to wait for an accumulation of mail to fill them; they can be delivered from the tube on to tables at any point in the building where the mail is wanted, for cancelling, distribution, or pouching, thus rendering a very rapid service; the mail is kept moving in an almost constant stream, keeping the postal employees more uniformly employed; special carriers can be despatched with “special delivery” letters. In other words, the most rapid service can be rendered by this size of tube.

If a larger than eight-inch tube is to be used for mail service, it should be not less than thirty-six inches. Carriers larger than eight inches cannot be handled: they are too heavy. They are also too heavy to slide through the tube, hence, must be mounted upon wheels. It is not practical to make a carrier on wheels less than eighteen or twenty-four inches, and the carrier must be at least twenty-four inches to contain a large mail-pouch. Now, if we are going to despatch mail-pouches through a pneumatic tube we must send more than one in a carrier, otherwise the service will be too slow. Such large carriers could not be despatched oftener than once or at most twice in a minute. Suppose we were to transport the mail from a railway station to a main post-office. A train arrives with, say, sixty pouches. If only one pouch could be put into a carrier and the carriers could be despatched at half-minute intervals, it would take thirty minutes to despatch all the pouches. Now, suppose we make the tube thirty-six inches. The carriers will be eight feet long and will contain from twelve to fifteen pouches. Five carriers would contain the entire train-load of mail, and they could be despatched in four or five minutes.

=System of Very Large Tubes.=—The cross-section of a thirty-six-inch tube is shown in Fig. 22. It is built flat on the bottom and sides, with an arched top. The floor is of concrete containing creosoted ties; the side walls and top are of brick, plastered with cement upon the interior. The two tubes may be built one above the other or side by side, depending upon the condition of the streets, but one common separating wall will serve for both. The carriers, one of which is shown in Fig. 23, run on two rails laid close to the sides of the tube. At curves a guard-rail is placed upon the side wall, making it impossible for a carrier to leave the track. The carriers are made of hard wood with an iron frame, and are as light as consistent with the service required of them. They are open on top. Their outside dimensions are thirty-four inches by thirty-four inches by eight feet. The sending and receiving apparatus for these very large tubes have to be specially designed for each particular station, so no attempt will be made here to describe them. The air-pressure required depends upon the length of the line. If it were not more than six or eight ounces a fan would be used to maintain the air-current, but for pressures above this, up to a pound or two per square inch, some form of positive blower would be used.

At the stations considerable floor space or “yard room” would be required for side tracks, switches, etc. Usually the basement of a building would have to be utilized for the termination of such a tube. There are but few places in our large cities where the streets are so free from pipes, sewers, conduits, etc., that it would be practicable to build a thirty-six-inch pneumatic tube. When the service can be rendered by an eight-inch tube, the cost of installation favors its adoption. Steep grades cannot be ascended by these very large tubes, while the eight-inch tubes can be placed vertically. We do not say that there is no use for eighteen- and twenty-four-inch tubes, but the demand for them would be in special cases and we will not discuss them here. For ordinary mail and parcel service we recommend the use of six- and eight-inch tubes. An eight-inch carrier is twenty-four inches long, about seven inches inside diameter, and will contain five hundred ordinary letters. It weighs about thirteen pounds empty, and one can be despatched every six to ten seconds. We estimate that eighty per cent. of all the parcels delivered from a large retail department store could be wrapped up to go into these carriers. The minimum radius of curvature of an eight-inch tube is eight feet.

=General Arrangement of Apparatus in the Stations. Two-Station, Two-Compressor Line.=—We will now proceed to a description of our system in detail. Figs. 24, 25, and 26 are diagrams showing how the tubes, air-compressor, tanks, sending and receiving apparatus are connected together at the stations. These diagrams are drawn to represent an eight-inch tube, but essentially the same arrangement would be used for smaller tubes.

Fig. 24 represents a line of two stations with an air-compressor at each station. Such an arrangement is proposed for the line of postal tubes over the New York and Brooklyn bridge, or for any two stations located a very long distance apart, say six or eight miles.

Referring to the diagram, we have at station A an air-compressor, _c_, which draws its supply of air from the tank _e_, and delivers it, compressed to the necessary pressure, into the tank _d_. From the tank _d_ the air flows to the sending apparatus, _a_, and thence through the tube _f_ to the station B. Upon arrival at B it flows through the receiving apparatus _m_, and then by the pipe _l_ to the tank _j_. A second air-compressor, _o_, is located at station B, and it draws its supply of air from the tank _j_. The tank _j_ has an opening to the atmosphere, _i_, through which air can enter when the air-compressor draws more than is supplied from the pipe _l_. The opening _i_ in the tank _j_ serves as an escape for air when the air-compressor at station A is started before that at station B. Stations A and B are similar in their arrangements. At B the air-compressor _o_ delivers its compressed air to the tank _k_, from which it flows to the sending apparatus _n_, and thence through the tube _g_ back to station A. Upon its arrival at A it passes through the receiving apparatus and enters the tank _e_, which is open to the atmosphere at _h_. The tanks _d_ and _k_ serve as separators to remove from the air any dirt and oil coming from the compressors, and they form a cushion, deadening, to some extent, the pulsations of the compressors and making the current of air in the tubes more steady and uniform. The tanks _e_ and _j_ form traps to catch any moisture, oil, or dirt coming out of the tubes.

Carriers are placed in the tubes and despatched by means of the sending apparatus _a_ and _n_. They are received from the tubes and delivered on to tables by means of the receiving apparatus _b_ and _m_. It will be seen that the arrangement is such that the air flows through one tube and returns through the other, the same air being used over and over again. Any air that escapes at the sending and receiving apparatus is replaced by an equal amount entering the tanks _e_ and _j_ from the atmosphere. By thus keeping the same air circulating in the tubes we prevent an accumulation of moisture in the tubes.

The air is at its maximum pressure in the tanks _d_ and _k_. The pressure falls gradually as it flows along the tubes and is down to atmospheric when it enters the tanks _e_ and _j_. The pressure at the receivers, _b_ and _m_, is just sufficient to push the carriers out on to the tables. The construction of the sending and receiving apparatus will be described in another place.

=Two-Station, One-Compressor Line.=—Fig. 25 is a diagram showing two stations, A and B, connected by a double line of tubes, both operated by one air-compressor located at station A. This is the arrangement used in the Philadelphia post-office line, and is the arrangement that will ordinarily be used for all two-station lines except where unusual conditions require something different. Station A is arranged precisely like station A in Fig. 24, so it need not be described again. The air flows from the sending apparatus _a_ through the tube _f_ to the receiving apparatus _p_ at station B. From the receiver _p_ it flows through the pipe _l_ to the sending apparatus _n_ and thence through the tube _g_ back to station A. The receiver _p_ at station B is what we will call a closed receiver,—_i.e._, it delivers the carrier from the tube on to the table without opening the tube to the atmosphere. The use of this form of receiver is made necessary by the fact that the air-pressure in the tube at this station is considerably above atmospheric. The air-pressure is at a maximum in the tank _d_. It falls gradually along the tube _f_, and when the air arrives at the receiver _p_, at station B, the pressure has fallen nearly to one-half its maximum amount in the tank _d_. On its return journey through the tube _g_ the pressure continues falling until it reaches the atmospheric pressure when the air enters the tank _e_ at station A.

The entire line of tube, going and returning, is operated by air at a pressure above the atmospheric. There is no exhausting in the return tube. It is distinctly a _pressure_ system.

=Three- to Eight-Station Line.=—Thus far we have described only two-station lines. In Fig. 26 we have a diagram of three stations connected together by a double line of tubes, and the arrangement would be similar if it were extended to four, five, six, seven, or eight stations. The stations are called A, B, and H. Station A is arranged exactly the same as stations A in Figs. 24 and 25, therefore, needs no description. Station B, being an intermediate station, is quite differently arranged from any of the preceding. From station A the air flows through the tube _f_ to station B, where it enters the automatic receiving and transferring apparatus, _s_. From this it flows through the tube _f_{´}_ to the sending apparatus _r_, and thence through the tube _f_{´´}_ to the next station, which may be another intermediate station, C, or the terminal station H. Station H is arranged like station B, Fig. 25. The air from the tube _f_{´´}_ enters the receiver _p_, and is then returned, through the pipe _l_, to the sending apparatus _n_. From the sending apparatus _n_ it continues on its return journey through the tube _g_ to the intermediate station B, where it enters the receiver and transfer apparatus _t_, then passes to the sending apparatus _q_, and through the tube _g_{´´}_ back to the receiver _b_ at station A. Thus we have followed the air-current out through one tube and back through the other. The current is kept circulating by the compressor located at station A. The pressure is at a maximum in the tank _d_, and falls gradually as the air flows along the tube until it returns to the tank _e_, when the pressure has fallen to atmospheric. A carrier is despatched from station A, and after passing through the tube _f_ arrives at station B, where it stops momentarily in the automatic receiver and transfer apparatus _s_. If the carrier is intended for station B, and was properly adjusted when it was despatched at A, it will be discharged from the apparatus _s_ on to the table _u_. But if it were intended for some other station and were so adjusted, after the delay of two or three seconds in the apparatus _s_, it will be automatically transferred to the tube _f_{´},_ pass through the sending apparatus _r_, and go on its journey through tube _f_{´´}_ to the next station. If it is not discharged from the tube at any of the intermediate stations, it will finally arrive at the terminal station H and there stop. Just how the carriers are adjusted and the details of the receiving and transfer apparatus will be described hereafter. Carriers arrive at station B from H, or other stations on the line, through tube _g_, in the apparatus _t_, which either discharges them on to the table _u_ or sends them on through the tube _g_{´}_ and _g_{´´}_ to station A. Carriers are despatched from station B to station A by means of the sending apparatus _q_, and from station B to other stations along the line, C, D, E, F, G, and H, by means of the sending apparatus _r_. Thus, from B carriers can be sent and received in either direction. In order to prevent the possibility of a collision of carriers by attempting to despatch one at station B at the instant another is passing through the sending apparatus, an automatic lock is attached to each sending apparatus. Just outside the station B, say three hundred feet on each side, are located manholes, and in these manholes boxes are attached to the tube containing an electric circuit-closing apparatus, so arranged that when a carrier passes it will close an electric circuit leading to the sending apparatus in the station. These manholes and circuit-closers are shown and located on the diagram at _v_ and _w_. Wires _x_ and _y_ lead from them to the sending apparatus _r_ and _q_. When a carrier from station A passes the box _v_, it closes the electric circuit _x_, which sets a time-lock on the sending apparatus _r_, holding this apparatus locked, so that it is impossible to despatch a carrier for, say, twelve seconds, a sufficient time for the carrier coming from the station A to pass station B and get three hundred feet beyond it. After the twelve seconds have elapsed the sending apparatus is unlocked and a carrier can then be despatched. In a similar manner a carrier coming from station H, in passing the box _w_, closes the electric circuit _y_ and locks the sending apparatus _q_ for a sufficient length of time to let the carrier pass the station. This resembles, in some respects, the “block system” as used on railroads. A “block” about six hundred feet in length, depending upon the speed of the carriers, is made at each intermediate station with the station in the centre of the block. Whenever a carrier enters this “block” the sending apparatus at the station is locked, and a carrier cannot be inserted into the tube to collide with the one which is passing. It will be noted that a carrier in passing out of the “block” does not unlock the sending apparatus; this is done automatically at a definite time after the carrier entered the block. The unlocking is entirely independent of the carrier after it has entered the block, and the reason it is so arranged is this: suppose that a second carrier enters the “block” before the first one leaves it; if the first carrier unlocked the apparatus when it left the “block,” then it would be unlocked with the second carrier in the block and a collision might occur, but by arranging it as we have done, if a second carrier enters the “block” before the first has passed out, the sending apparatus remains locked for a period of time beginning with the arrival of the first carrier in the “block” and ending, say, twelve seconds after the arrival of the last carrier, which is sufficient time for the last carrier to pass out of the block. Of course, if a carrier becomes wedged in the tube a collision may occur, but this very seldom if ever happens. The details of the locking apparatus will be described in another place.

If stations A, B, ... and H were arranged on a loop, as shown in circuit 6, Fig. 21, then station H, Fig. 26, would be at the central, or station A. If it were a single loop, like circuit 4, Fig. 21, then there would be only one sending apparatus and one receiving and transferring apparatus at the intermediate stations.

A telephone circuit will include all stations, in order to give orders to the station attendants and to signal to the central station in case of an accident, when it might be necessary to stop the air-compressor. The telephone wires, in the form of a lead-covered cable, are laid in the same trench with the tubes and fastened to them.

=The Sending Apparatus.=—We have, in the preceding pages, frequently spoken of the sending apparatus, and have described it as mechanism by which carriers are inserted into the tube. In the Philadelphia postal line this apparatus consisted of a large valve, operated by hand. For an eight-inch tube such a valve would be too large and heavy to be manually operated. Furthermore, that type of apparatus is not suited to an intermediate station, where carriers have to pass through it. To meet all of these requirements we have designed an apparatus, of which Fig. 27 is a side elevation, Fig. 28 a longitudinal section, and Fig. 29 a cross-section. Referring to the longitudinal section, Fig. 28, the sending apparatus is shown inserted into the line of a pneumatic tube, A, A. We have a movable section of tube, B, that can be swung about the large bolt, G, at the top, into and out of line with the main tube, A, A. When the section of tube B is being swung to one side, the air-current has a by-pass through the slots E and F and the U-shaped pipe D. The joints at the ends of the movable section B are packed with specially-formed leathers. Referring to the cross-section, Fig. 29, when the movable section of tube B is swung out of line with the main tube, another and similar tube, C, takes its place. The two movable tubes, B and C, are made in one piece, so that they must always move together. They are connected together at each end by plates, M, that serve not only as connecting-plates, but covers for the ends of the main-line tube while the tubes B and C are being moved. The tubes B and C swing between four plates or wings, L, that extend out on each side of the apparatus. They serve as guards, and, at certain positions of the swinging tubes, prevent the air from escaping.

We will, for convenience, call the system of swinging tubes B and C, with their supports, etc., the swing-frame or simply the frame. This frame is moved or swung from one position to the other by means of a cylinder and piston, H, placed in an inclined position under it. A lug, N, is cast on the tube B, to which the connecting-rod, O, is attached. The cross-head, P, slides upon an inclined guide, Q. On top of the cylinder is placed a controlling valve, made in the form of a piston slide-valve. The piston in the cylinder H is moved by the pressure of the air taken from the main tube through the pipe I. The apparatus is operated by a hand-lever, K. When this lever is pulled, it moves the sliding-head R, and this, through the spring S, moves the controlling valve, if the valve is not locked. If it is locked, pulling the lever simply compresses the spring S. When the controlling valve is moved to the right the air in the cylinder H escapes through the passage V and the port J to the atmosphere, and compressed air from the main tube flows through the pipe I, the passages T and U, to the cylinder H, under the piston, causing the piston to move up the inclined cylinder and swing the frame until the tube C is in line with the main tube. Carriers are despatched by placing them in the tube C, then pulling the lever K, and swinging the frame until the tube C is in line with the main tube. The carrier is then taken up and carried along by the current of air in the main tube.

Replacing the hand-lever K in its original position returns the frame to its normal position.

=Sending Time-Lock.=—In any system of large pneumatic tubes a short time should elapse between the despatching of carriers, in order that they may not collide in the tube, and to give the receiving apparatus at the stations time to act. To insure the impossibility of having carriers despatched too rapidly, we place on the sending apparatus a time-lock that will automatically lock it for a determined length of time after each carrier is despatched, the time-lock being adjustable for any desired time. The time-lock, W, is shown attached to the sending apparatus in Fig. 27. When the swing-frame is swung to despatch a carrier, it pulls up the rod X by means of a link and bell-crank, Y, thereby locking the controlling valve of the cylinder H and starting the time-lock W, which will unlock the controlling valve after the required time has elapsed. The details of this time-lock are shown in Fig. 30. It consists of a long vertical cylinder, A, containing a piston, B, and a spiral spring, C, that tends to force the piston to the bottom of the cylinder. The cylinder is filled with oil, and holes, D, in the piston allow the oil to pass freely through it when it is moved upward in the cylinder. When the piston moves downward an annular collar, E, forming a valve, closes the holes in the piston and prevents the oil from passing through. Extending from one side of the piston to the other is a by-pass, F, in the wall of the cylinder. When the piston moves downward the displaced oil is forced to flow through this by-pass. A small cock, G, is arranged in the by-pass to throttle the stream of oil flowing through it. The opening in this cock, or the amount of throttling, is indicated on the outside by an index and dial, Z (see Fig. 27). When the piston B is raised and allowed to descend by the force of the spring C, it forces the oil through the by-pass F and the cock G. If the latter is wide open the piston will descend quickly, but if it is nearly closed the piston will descend very slowly. In other words, the time of descent can be regulated by opening and closing the cock G. The reading on the dial Z can be made seconds of time that elapse while the piston is descending.

Above the cylinder is a cross-head, H, that moves up and down between vertical guides. This cross-head is moved by the rod X, also shown in Fig. 27, that receives its motion from the swinging frame of the sending apparatus. A piston rod, I, attached to the piston in the cylinder, extends up through the travelling cross-head but is not attached to it. On the piston-rod are two enlargements, J and K, one made a solid part of it, the other formed by two nuts. The travelling cross-head H carries a pawl, L, that engages under the shoulder formed by the nuts K. This pawl is kept against the piston-rod by the spring M. The enlargement, J, on the piston-rod forms a shoulder that bears against the bell-crank, N, that connects with the bolt, O, which locks the controlling valve. In the present down position of the piston and piston-rod, the enlargement J, by pressing against the bell-crank N, holds the bolt O in an unlocked position. When a carrier is despatched the cross-head H is lifted by the rod X, and carries with it the piston and piston-rod, compressing the spring C. This upward movement of the piston-rod allows the bolt O to be thrown by a spring, not shown in the figure, and so lock the controlling valve of the sending apparatus. As the cross-head continues its upward movement, the pawl L comes in contact with the end of the screw P and disengages the piston-rod. This allows the piston to descend as rapidly as the oil can pass through the by-pass and cock G. When the piston has reached nearly to the bottom of the cylinder, the shoulder J, on the piston-rod, engages the bell-crank N and withdraws the bolt O, thereby unlocking the controlling valve. The time that the sending apparatus is locked depends upon the time required for the piston to descend. While the sending apparatus is locked against the sending of another carrier, it is not so locked that the swing-frame cannot be returned to its normal position and another carrier inserted ready to be sent as soon as the necessary time expires. This time is usually not more than ten seconds. Not only may the second carrier be placed in the tube C, Fig. 29, ready to be sent, but the handle K may be pulled and fastened in the notch _a_, thereby compressing the spring S, which, as soon as the controlling valve is unlocked, will move the valve and automatically despatch the carrier. The controlling valve is locked by the passage of a bolt through the hole _b_, in a block carried on the end of the valve stem, when it returns to the normal position shown in the figure. Usually little or no time will be lost in thus locking the sending apparatus, for the small amount of time that the apparatus is locked will be needed in handling the carriers.

=Intermediate Station Time-Lock.=—We have another time-lock attached to the sending apparatus that has been already referred to in describing the “block system” used at intermediate stations; a time-lock to prevent carriers being inserted into the tube at intermediate stations while another carrier is passing that station. This time-lock is shown in Fig. 27 at W´, and is shown in detail by a sectional drawing, Fig. 31.

When a carrier closes an electric circuit in passing one of the boxes located in a manhole about three hundred feet from an intermediate station, it indicates its approach to the station by exciting the electro-magnet A, Fig. 31. This magnet pulls down its armature B and raises the small piston valve C, which admits compressed air to a small chamber, D. The air is supplied to this chamber from the main tube through the pipe E. In one end of this chamber is fitted a piston, F, held to one end of its stroke by a spring, G. When compressed air is admitted to the chamber D, this piston is moved to the left, and by such movement throws the controlling valve of the sending apparatus into its normal position (shown in Fig. 29) and holds it there. This forms a positive lock, and, no matter in what position the sending apparatus may be, it puts the tube B, Fig. 29, into line with the main tube so that the approaching carrier can pass through the apparatus. The piston-rod H, Fig. 31, is connected to the finger _d_, Fig. 29, and by rocking this finger moves the controlling valve, or prevents it being moved by the handle K.

Returning now to Fig. 31, we have on the top of the chamber D, in addition to the electro-magnet A and its armature B, a differential cylinder and piston, K, L, whose function is to close the valve C when the chamber D is filled with air. The piston K is smaller than the piston L, and sustains a constant air-pressure, supplied through the small pipe M M, from the pipe E, which leads to the main tube. When the chamber D becomes filled from the pipe E through the valve C, the pressure in the chamber moves the piston L upward against the pressure on the piston K, because of the greater area of the piston L. This movement of the differential piston raises the lever I, which passes through a slot in the stem of the differential piston, and thus closes the valve C. The air in the chamber now gradually escapes to the atmosphere through a small orifice Q; in fact it has been escaping here all the time while the chamber was being filled, but the opening through the valve C is so many times larger than the orifice Q that the escape of air was not sufficient to prevent the chamber from filling. Now, however, that all supply to the chamber is shut off, the air in the chamber is gradually being discharged through the orifice. When nearly all the air has escaped, the piston F will return to its normal position, shown in the figure, and unlock the controlling valve. The time required for the air to escape from the chamber, D, is the time that the sending apparatus will be locked, and this time can be regulated by varying the size of the orifice Q. The opening of the orifice, or the time that the sending apparatus is locked, is indicated by an index and dial, P.

This locking mechanism is secured to a bracket on the side of the large cylinder H, Fig. 27, in a position where it can be easily inspected. The moving parts of the electro-magnetic valve—for such is the valve C, with the magnet A, Fig. 31—are made very light, in order that they may respond easily and quickly to the closing of the electric circuit.

It is a disadvantage to have stations too numerous upon the same line, especially if they do a large amount of business, for each station will delay the sending of carriers from the others more or less, and the interference will be greatest during the busiest hours of the day. This condition is inherent in any system of large tubes where carriers have to be run a certain minimum distance apart, and cannot be overcome by any mechanism. But the disadvantage is greatly overshadowed by the advantage of being able to connect several stations by one line, instead of having to run independent lines from each station to the central, especially when the business of the individual stations is not sufficient to occupy a separate tube all the time. It makes it possible to have stations where otherwise the business would not warrant the cost of installation and expense of operation. We recommend the establishing of not more than eight stations on a line, and usually a smaller number than this, depending, of course, upon the amount of business to be done at each station.

=The Electro-Pneumatic Circuit-Closer.=—There is one piece of mechanism used in connection with the sending apparatus that we have yet to describe, and that is the circuit-closing device located in the manholes in the street. Since the carriers travel at a high rate of speed, they should not be made to operate any mechanism by impact with fingers or levers protruding into the tube when it can be avoided, even though the work to be done is so slight as the closing of an electric circuit, for the repeated impacts cannot fail to work injury to the carriers and the mechanism to be operated, no matter how carefully they are designed. To avoid such impacts, we have designed the electro-pneumatic circuit-closer, shown by the drawing in Fig. 32. It is operated by a passing carrier, but pneumatically rather than mechanically. In the figure we have a pneumatic tube, A, A, in which a carrier, B, is moving in the direction indicated by the arrow. At two points, about twenty or thirty feet apart, two small holes are tapped into the tube and pipes, C and D, are screwed in. These pipes lead to two chambers in a cast-iron box, F, separated by a diaphragm, E. This diaphragm is insulated electrically from the box supporting it, and is connected with the wire G. Just out of contact with the diaphragm is an insulated screw, H, connected with the wire I. These wires lead to the time-lock, already described, on the sending apparatus at the station. When no carrier is passing, the air-pressure is the same on both sides of the diaphragm, but when a carrier enters that part of the tube between the two points where the pipes C and D are connected, the equality of pressure on opposite sides of the diaphragm is destroyed. There is always a slightly greater pressure in rear of the carrier than in front of it, equal to the frictional resistance of the carrier in the tube. It is this difference of pressure in front and in rear of the carrier that moves it through the tube. When the carrier is in the position shown in the figure, the same difference of pressure will exist on opposite sides of the diaphragm, and it will be deflected into contact with the screw H, thereby closing the electric circuit. When the carrier has passed, equality of pressure on opposite sides of the diaphragm is established and the diaphragm takes its normal position, out of contact with the screw H. This apparatus is easily attached to the tube, and it contains no mechanism to get out of order.

=The Open Receiver.=—Wherever the pressure in the tube is down nearly to atmospheric, we can use an open receiver to discharge the carriers from the tube. This is a receiver that opens the tube to the atmosphere and allows the carrier to come out. Such a receiver is used at the main post-office in the Philadelphia postal-line, and was described in the last chapter. The present receiver is similar in operation, but contains some improvements in details. Fig. 33 is a side elevation of the apparatus, Fig. 34 is a longitudinal section, and Fig. 35 is a cross-section through the cylinder and valve, showing the sluice-gate.

Referring to the longitudinal section, the apparatus is attached to the end of a pneumatic tube, A. The current of air from the tube A flows through the slots B into a pipe, C, that conducts it to a tank near the air-compressor. About the centre of the apparatus is a sluice-gate, E, that is raised and lowered by a piston in a vertical cylinder, F, located just above the sluice-gate. This piston is moved by air-pressure taken from some part of the system. When a carrier arrives from the tube A, it passes over the slots B and runs into the air-cushion D, where it comes gradually to rest. Checking the momentum of the carrier compresses the air in front of it considerably, and this excess of pressure is utilized to move a small slide-valve that controls the movement of the piston in the cylinder F, so that as soon as the carrier has come to rest the sluice-gate rises and allows the carrier to be pushed out with a low velocity on to a table. The small pipe G conducts a small portion of the air compressed in front of the retarded carrier to the controlling valve, H, seen in Figs. 33 and 35. Referring now to the section of the valve and cylinder, Fig. 35, the pipe G enters the top of a small valve-cylinder containing a hemispherical piston, I, that is held up by a spiral spring, J. This spring has just sufficient tension to hold the piston I up against the normal pressure of air in the tube. When a carrier arrives and compresses the air in the air-cushion, the excess of pressure forces the piston I down against the spring J, and moves the piston slide-valve K. This change of position of the slide-valve allows the air in the cylinder F to escape to the atmosphere through the passage L, passage P, and pipe M, while compressed air from some part of the main tube enters through the port N and passage O to the under side of the piston in the cylinder F. This moves the piston up, carrying with it the sluice-gate E.

There is just sufficient pressure in the tube in rear of the carrier to push the carrier past the gate and on to the table. As the carrier moves out it raises a finger, Q, Fig. 34, that projects into its path. Raising this finger extends the spring R, Fig. 33, and rotates the lever S, bringing the pawl T under the end of the controlling valve-stem. When the carrier has passed out and the finger Q is free to descend, the spring R rotates the lever S back to its original position, and thereby raises the controlling slide-valve, which causes the sluice-gate to close. By having the upward motion of the finger Q simply extend the spring R, and the downward motion, by the force of the spring, move the valve, we are enabled to have several carriers pass out of the tube together without having the sluice-gate close until the last carrier has passed out. If raising the finger Q moved the valve, then when the first carrier passed out, the gate would close down upon the second. Attached to the receiving apparatus and extending beyond it is a tube, U, cut away upon one side so that the carriers can roll out of it on to a table, and having in the end a buffer to stop the carriers if by any accident they come out of the tube with too much speed. This buffer consists of a piston covered with several layers of leather and having a stiff spring behind it. The whole apparatus is supported from the floor upon suitable standards, and, for an eight-inch tube, occupies a floor-space twelve feet long by two feet wide, not including the table.

This is the simplest form of receiving apparatus. Owing to conditions of pressure already explained, its use is confined principally to the pumping stations. The only care that it requires is an occasional cleaning and oiling.

=The Closed Receiver.=—Next we will turn our attention to the closed receiving apparatus used at all terminal stations where the pressure in the tube is considerably above the pressure of the atmosphere, so much so that the tube cannot be opened to allow the carrier to pass out without an annoying blast of air and a high velocity of the carrier. This apparatus is similar to the receiver used in the sub-post-office of the Philadelphia postal line, but contains several modifications and improvements tending towards simplification. Fig. 36 shows it in elevation, and Fig. 37 in longitudinal section. As in the open receiver just described, the air from the tube A is deflected through slots B into a branch pipe, C, that conducts it from the receiving apparatus to the sending apparatus and return tube. The carriers arrive from the tube A, pass over the slots B, where the air makes its exit, and run into an air-cushion, D. This air-cushion is a tube about twice the length of the carrier, closed at one end, and supported upon trunnions. When the carrier has been brought to rest, this closed section of tube is tilted by the movement of a piston in a cylinder to an angle that allows the carrier to slide out; the tube then returns to its original position. If the end of the air-cushion was closed perfectly tight the carrier, after coming to rest, would rebound and might be caught in the joint between the stationary and movable parts of the apparatus, when the air-cushion tube tilted. To prevent the rebounding of the carrier a relief-valve, E, has been placed in the head of the air-cushion tube. It is held closed against the normal pressure in the tube by a spiral spring, but the excessive pressure created by checking the momentum of the carrier opens the valve and allows a little air to escape through the passage F and pipe G, down the pedestal H, to the atmosphere. When the air-cushion or receiving tube D is tilted to discharge a carrier, the circular plate I covers the end of the main tube. In order to prevent carriers sticking in the receiving tube when it is tilted, and to insure their prompt discharge, the pipe J is provided. In the tilted position of the receiving tube, the end of this pipe coincides with the end of the main tube, from which it receives air to hasten the discharge of the carrier. A check-valve, K, prevents the air from flowing backward in this pipe when a carrier is being received in the air-cushion chamber. The opening of this check-valve can be adjusted by a screw, thereby regulating the speed of ejection of the carrier.

The carrier is discharged down a chute, L, which has a buffer at the bottom, and from the chute it rolls off on to a table. The buffer is made similar to the buffer in the open receiver already described. The cylinder and piston M, that operate to tilt the receiving tube D, are supported upon the base of the apparatus under the closed end of the receiving tube. The cross-head of the piston- and connecting-rods travels between guides that are made a part of the upper cylinder head. The movement of the piston in the cylinder M is controlled by a piston slide-valve exactly similar to the one shown in Fig. 35. The slide-valve is moved, in the same manner, by the air compressed ahead of the carrier when it is brought to rest in the air-cushion D. The air is conducted from the air-cushion to the controlling slide-valve through a small pipe, N, Fig. 36. This pipe leads to one of the trunnions, where it has a joint to allow for the tilting of the receiving tube. When the carrier is discharged from the receiving tube, it raises a finger, O, Fig. 37, located just outside the tube. Raising this finger pulls the rod P, Fig. 36, extends the spring Q, turns the lever R, and catches the pawl S, under the end of the controlling valve stem. When the carrier has passed down the chute and allowed the finger O to drop down, the spring Q turns the lever R back to its original position and moves the controlling valve. This causes the receiving tube to return to a horizontal position, where it is ready to receive the next carrier.

At first this apparatus may seem a little cumbersome, but nothing could work better. It is certain in its action and almost noiseless. Carriers are received, discharged, and the receiving tube returned to its normal position in four seconds, and it can be done in less time if necessary.

=The Intermediate Station Receiving and Transfer Apparatus.=—One other form of receiving apparatus remains to be described, and this is the apparatus used at intermediate stations to intercept all carriers intended for that station and to send the others on through the tube to the next station. A side elevation of the apparatus is shown in Fig. 38 and a vertical section in Fig. 39. The tubes are led into an intermediate station, carried upward, and then, with a bend of one hundred and eighty degrees, are connected to the top of the receiving and transfer apparatus, as shown in the diagram, Fig. 26. The object of this arrangement will be seen as we describe the apparatus. Referring to the sectional drawing, Fig. 39, the connection of the tube A is seen at the top. As in the other receivers, the current of air arriving from the tube A is deflected through slots, B, into a passage, C, made in the frame of the apparatus. From this passage it enters the tube D through the slots E. The tube D leads to the sending apparatus and on to the next station, as seen in Fig. 26. The carriers are received in a closed section of tube F, which forms an air-cushion, similar to the closed receiver last described. This receiving tube F is made a part of what we might term a wheel. This wheel fits accurately into a circular casing and is supported by two trunnions or axles, upon which it revolves. The wheel has a broad flat rim, G, that covers the end of the tube at H when the wheel is revolved, and, in the normal position in which it is shown in the figure, covers the interior openings I, J, K, and L, in the casing. Leather packing is provided around each of the openings to prevent the escape of air between the face of the wheel and the interior face of the casing. From the bottom of the receiving tube F a passage, M, leads past a check-valve, N, to the tube D. When a carrier arrives from the tube A, it descends into the receiving tube F, compressing the air in front of it. This compressed air begins to escape through the passage M, but the high velocity of it closes the check-valve N as much as possible. A stop on the stem of the valve prevents it being closed entirely. The small opening past the valve allows some of the air to pass, thereby preventing the carrier from rebounding on the air-cushion. As soon as the carrier has come to rest, the check-valve N, by its own weight, opens wide, and the carrier, by its weight, settles gradually down to the bottom of the receiving tube. The wheel containing the receiving tube and the carrier will then be revolved by the cylinder and piston O, which is operated by compressed air taken from the tube through the pipe P. If the carrier is for this station, the wheel will rotate through an angle of forty-five degrees and discharge the carrier through the opening J, down the chute Q, from which it will roll on to a table arranged to receive it. If, however, the carrier is intended for some other station, the wheel will rotate through an angle of ninety degrees and discharge the carrier through the opening K into the tube D, and it will go on its way to the next station. This selection of carriers is brought about in a comparatively simple manner. At the bottom of the receiving tube F there are two vertical needles, R and S, shown upon a larger scale in Fig. 40. The needles R and S are contained in tubes having an insulating lining which keeps them out of electrical contact with the frame of the apparatus. Wires _a_ and _b_ make connection with the needles through metal plugs that form a guide for the needles, and through the springs U and V. Directly below the needle R is an insulated spring clip, W, held by two bolts and connected to the wire _e_.

The end of a carrier is represented at T. As the carrier settles down to the bottom of the receiving tube, it comes in contact with the ends of the needles and presses them down, they being supported by two springs U and V. As the needle R is moved down, it makes contact with the spring clip W, located just below it, and closes an electric circuit that includes the electro-magnet X, Figs. 38 and 39, on the valve of the rotating cylinder O. When this electro-magnet is excited it attracts its armature and moves the piston slide-valve Y, that admits air to the top of the piston in the cylinder O, and allows the air under the piston to escape to the atmosphere. The piston moves downward and revolves the wheel by means of a connecting rod.

Upon the end of the carrier T is placed a thin circular metal disk, _f_, which may be copper, brass, tin-plate or any metal that is not easily oxidized. The diameter of this disk of metal determines the station at which the carrier will be discharged from the tube. Disks of various diameters, that may be attached to the carrier, are represented by dashed lines, _g_, in Fig. 40. When the carrier comes in contact with the two needles R and S, if the circular metal disk on the front end of the carrier has a diameter sufficient to span the space between the two needles, in the position in which it is held, then an electric circuit, made by the wires _a_ and _b_, will be closed through the needles and the metal disk on the carrier. The metal disk makes a short-circuit from one needle to the other. If the metal disk is not large enough to span the distance between the two needles, then the electric circuit remains broken.

Returning again to Fig. 39, we have the opening J, where the carriers are discharged, closed by a sluice-gate. This gate is opened and closed by a piston moving in a cylinder, _h_, shown in Fig. 38. A piston slide-valve, _i_, similar in all respects to the valve on the cylinder O, controls the movement of the piston in this cylinder and the sluice-gate to which it is attached. The slide-valve is moved in one direction, that opens the sluice-gate, by an electro-magnet in the circuit of the wires _a_ and _b_, Fig. 40.

When the electric circuit made by these wires is closed by a disk on the front end of a carrier, short-circuiting the two needles, the valve is moved by the electro-magnet in the circuit, and the sluice-gate is opened. As the wheel, including the receiving tube and carrier, revolves, a lug, _j_, Fig. 38, on the outside of the wheel comes in contact with the open sluice-gate and the wheel can rotate no farther. A blast of air through the valve L, Fig. 39, assisted by gravity, pushes the carrier out of the receiving tube, through the opening J and down the chute Q, on to the receiving table.

Had the disk on the front end of the carrier been too small to span the distance between the two needles, the circuit would not have been closed, the sluice-gate would not have been opened, no obstruction would have been placed in the path of the lug _j_, on the wheel, and the wheel would have continued its rotation through ninety degrees until the receiving tube F came in line with the tube D. During the latter part of the rotation, a pin on the wheel engages a lever, _k_, Fig. 38, and turns a valve, _l_, Fig. 39, stopping the flow of air through the passage C, compelling it to take another route through the passage _m_, and the receiving tube F, taking with it the carrier into the tube D. When the carrier leaves the receiving tube and passes through either of the openings J or K, it engages one of the fingers, _n_ or _o_, that lie in its path. These fingers are connected by rods and levers to the valves on the rotating and sluice-gate cylinders. The ejected carrier pushes these fingers to one side, and after it has passed the fingers return, by the force of a spring, to their former position and move the valves, causing the sluice-gate to close and the wheel to rotate backward into its normal position ready to receive the next carrier. The connection between the fingers and the valves is similar to the mechanism on the open and closed receivers, so need not be described in detail here.

The speed with which the carriers are ejected from the receiving tube through the opening J and down the chute Q is regulated by the valve L, which can be opened or closed by a hand-wheel, _p_. Before the wheel and receiving tube can be rotated, the needles must be withdrawn from the receiving tube, and this is accomplished by a small cylinder and piston, _q_, shown in Fig. 40. The needles and their encasement are attached to a cross-head, _r_, on the end of a hollow piston-rod, _s_. When air is admitted to the top of the piston in the rotating cylinder O, Fig. 39, it is also admitted through the pipe _t_, Fig. 38, to the cylinder and upper side of the piston _q_, Fig. 40. This moves the piston _q_ down against the force of a spring, _u_, and withdraws the needles from the receiving tube. This takes place after the needles have served their purpose and before the wheel is rotated. The piston _q_ has much less inertia than the wheel, therefore it moves much quicker. When the wheel begins to rotate it closes a valve, _v_, in the pipe _t_, Fig. 38, confining the air in the cylinder _q_, and preventing the needles from being raised by the spring _u_ before the wheel returns to its normal position. If by any accident the needles should be raised, no serious harm would result, for their ends would simply bear against the face of the wheel. If this took place constantly, grooves might be worn in the face of the wheel; for this reason the valve _v_ is provided.

In order to facilitate the inspection of the needles and electric contact springs W, they are contained in a cylindrical brass case, _w_, that is held in place beneath the receiving tube by two bolts. By removing the nuts from these bolts the entire mechanism can be removed, examined, and cleaned. It also gives easy access to the receiving tube. The receiving tube is long enough to receive two carriers, if it should ever happen that two arrive at the same time.

To show how the apparatus at the various stations is arranged to correspond with the disks of various sizes attached to the front of the carriers, a diagram, Fig. 41, has been made, in which the needles at the bottom of the receiving tubes of the apparatus at six intermediate stations are represented at A, B, C, D, E, and F. Six disks of different sizes are represented at _a_, _b_, _c_, _d_, _e_, and _f_. The needles are placed farthest apart at station A and nearer together at each succeeding station until we arrive at station F, where they are nearest together. If we wish to send a carrier to station A from the central, we place the largest disk, _a_, upon the front end of it. When it arrives at station A, it closes the electric circuit between the needles and is discharged from the tube. Should we wish to send a carrier to station D, then we place the disk _d_ upon the front end of it. When the carrier arrives at the station A, the disk is not large enough to span the needles; therefore the sluice-gate is not opened and the carrier is sent on in the tube. When it arrives at stations B and C, the same thing occurs again, but when it reaches station D, the needles are sufficiently close together so that the disk makes an electric circuit between them, and the carrier is discharged from the tube, as was intended when despatched. Since the carriers always travel in the same direction in a tube, the first station at which they arrive where the needles are near enough together to have both touch the disk, will be the station at which the carrier was intended to stop. Carriers can be despatched from any station, but if we wish to send from say D to A, they must either travel around a loop or be sent through a return tube in which the needles are arranged in the reverse order. If no disk is placed on the carrier, it will go to the last station on the line.

There are other attachments that might be made to the front end of the carriers in order to have them stop at any desired station along a line. We have worked out two other systems which are entirely mechanical in their operation, not using electric circuits and electro-magnets to move the valves. While such a mechanical system has some advantages over the present combined mechanical and electrical system, yet there is one great advantage in the latter, and that is the simplicity of the attachment made to the carrier. A round flat disk of tin-plate is attached to the front end; it is something that is not in the way; it does not prevent standing the carriers on end in racks to fill them; it is not easily injured, and only those who have had experience can realize the rough usage that the carriers receive; it is quickly and easily attached to the carrier, and it is so cheap that when bent it can be thrown away.

=Carriers.=—The carriers are similar in all respects to those used in the Philadelphia postal-line, that have been described in the preceding chapter and illustrated in Figs. 18, 19, and 20. When there are intermediate stations upon the lines, means are provided for attaching disks to the front end of the carriers. The disks have a central stem that secures them to the bolt in the centre of the head, and are so arranged that they can be quickly attached or removed.

Many experiments have been made to find the best material for bearing-rings, but thus far nothing better than a specially-prepared woven fabric has been found. These rings will run about a thousand miles, when they become so reduced in diameter that they have to be replaced by new ones.

The most essential elements of a carrier are strength, lightness, and security of the contents. Aluminum has frequently been proposed as a suitable material for the bodies of carriers, but for the same weight steel is much stronger, especially in thin rolled sheets, and for this reason it has been used.

One of the most perplexing problems that presented itself in working out the details of the system was to design a secure and reliable lock for the lids of the carriers. We believe that the one which has been adopted fulfils all requirements in a satisfactory manner.

Some experiments have been made with carriers that open on the side, but structurally they are weak and unsuited to stand the blows that carriers frequently receive. They are not so easily and quickly filled and emptied as those that open on the end. These remarks apply to carriers for large tubes. In small tubes for the transportation of cash in retail stores, carriers with side openings are found convenient.

When United States mail is sent through tubes not used exclusively for postal service, carriers with special locks can be used, so that they can be opened only by post-office employees.

=Air Supply.=—This completes the description of the special apparatus used in this system, but we have yet to say something regarding the machines that supply the air. In Paris the water from the city mains has been used to compress or exhaust the air used in small tubes, but to operate large tubes in most of our cities steam is the only available power. Except in isolated cases, an independent steam plant will be erected to supply the air for a system of tubes. This plant should be designed with a view to obtaining the maximum economy in coal consumption, labor, water, cartage, and incidental expenses. We might say that the same general rules of economy which govern the design and construction of electric-lighting plants should be applied to the plans and construction of air-compressing plants.

Three types of blowing machines are used,—viz., centrifugal fans, positive blowers, and air-compressors.

=Fans.=—Very large tubes of moderate length can be operated by ordinary centrifugal fans. These fans are capable of supplying air under a pressure not exceeding ten or twelve ounces per square inch with very good efficiency. They are the simplest and most inexpensive of all blowing-machines.

=Blowers.=—When tubes have a length and diameter that require a pressure from one to four pounds per square inch, some form of positive blower of the Root type can be used with economy. Their construction is familiar to nearly every one at all interested in machinery, so we need give no space to their description here.

=Air-Compressors.=—By far the greater number of our tubes require an air-pressure of more than five pounds per square inch. For such air supply we recommend some form of air-compressor, and usually this is driven by a steam-engine, which forms a part of the compressor. In making our selection we should bear in mind the conditions under which the compressor will run. Usually it must be kept in constant operation at least ten hours per day, and frequently for a much longer period. This makes it important that the compressor be substantially built and supported upon a solid and firm foundation. The bearings should be broad, of good wearing material that has a low coefficient of friction, and provided at all times with ample lubrication. If poppet valves are used in the air-cylinders, and they are most common, the speed in revolutions per minute should not be high. Duplex are better than single cylinder compressors, because they deliver the air in a more steady stream,—the pulsations are less. For constant running, economy of steam is an important item; therefore some good type of cut-off valve should be provided. The air-cylinders should not be water-jacketed unless the pressure is above twenty-five pounds per square inch. It is better to use the air as warm as possible, for it will soon be cooled after entering the tube. A speed-governor should be provided with compressors which are to run at constant speed, but usually they will be run to maintain a constant pressure in the tank, and to this end a good and reliable form of pressure-governor should be provided, together with some reliable safety device to stop the engine when the speed exceeds a safe limit. But most important of all is to have the valves of the air-cylinders large in area; otherwise the efficiency of the machine will be very low. With machines working under eighty pounds pressure, a difference in pressure of one pound on opposite sides of the valves has but little effect, but when the machine is only compressing to five or ten pounds, one pound is a very large proportion of the total pressure and reduces the efficiency. Besides these few suggestions, only the requirements of good engineering need be demanded. In Figs. 42, 43, 44, 45, 46, and 47 we show a fan, two blowers, and an air-compressor suited to the requirements of pneumatic-tube service that can be found in the market, and that are built by responsible concerns. We believe they are all good of their kind, but do not recommend any particular make.

=The Tube, Line Construction, etc.=—Up to the present time we have found no material better suited for the straight parts of pneumatic tubes than cast iron, machined upon the interior. It gives a smooth and accurate tube. It can be made in most convenient lengths. It is strong and not easily deformed. The bell-joint, calked with lead and oakum, having the tubes fitted together male and female at the bottom of the bell, is the best joint yet devised for pneumatic tubes. It is slightly yielding, accommodating itself to slight changes of length of tube due to changes of temperature, and it allows slight bends to be made at each joint. The joints are very accurate, presenting no shoulders to obstruct the passage of carriers. The joints can be made by men accustomed to laying water- and gas-pipe. The cast iron is so stiff that it is not distorted in calking, as may be done with wrought-iron tube. The principal objections to its use are the expense of boring and the readiness with which it corrodes upon the interior.

We are always hoping that wrought-iron or steel tubes will be so much improved in uniformity of dimensions and smoothness of interior that we can use them, but our experiments thus far have been discouraging. It may be that some of the new processes of making tubes will give us what we want, but we have not yet found it.

Small tubes and the short bends of large tubes are made of brass, it being the most suitable material. It would be very difficult to bend iron tubes without involving great expense. The thickness of the bent portion of an eight-inch tube is usually three-sixteenths of an inch and never less than one-eighth of an inch.

Where the ground is firm, no other support is needed for the tubes than to tamp the earth solidly about them. In order to economize space in the streets, it is customary to lay the tubes one above the other; and it is very convenient, although not necessary, to separate them by cast-iron saddle brackets. Such an arrangement has to be frequently departed from in order to overcome obstructions in the streets and to get through narrow passages. At all low points in a tube line, traps are provided to catch any moisture that may accumulate. These traps are made accessible for frequent inspection by means of man-holes or otherwise. The tube is usually laid about three feet below the pavement. This distance has frequently to be varied, but it never becomes so small as to render the tubes liable to injury from heavy trucks passing over the pavement.