The Tubular Post

Pneumatic telegraphs were a Victorian wonder, now largely forgotten

A movable plug in a tube can be driven by the difference in pressure on the two sides of it, while it is guided and protected by the tube. The pressures can be greater or less than atmospheric pressure, to suit requirements. Samuda's atmospheric railway of the 1840's was based on this principle. The sliding piston transmitted the force to the carriages of the train by a member passing through a sealing flap, as it progressed into a vacuum established before it. The vacuum was necessary to hold the flap tightly shut. The flap proved the weak point in the design. There was no success in designing a flap that would have a sufficiently long life. A recent enthusiast on TV said the problem was that rats ate the flap, and that these days a suitable material could be found. No doubt, but the principle would still not be satisfactory. A pipe 15" in diameter could provide only about 1700 lb tractive effort from a vacuum of 20", about most that could be relied upon. This was enough for a few light carriages in 1846, but would be utterly inadequate now. It was, however, a delightful idea.

A different, and more practical, idea was first demonstrated not long afterwards by Ador in the Parc de Monceau in 1852. A carrier was driven along a closed tube by air pressure, carrying a message swiftly from one point to another. Patents were taken out in 1854 for similar systems by Cazalat in France and by Latimer Clark in England. Clark's system was improved by C. F. Varley in 1863, and arranged so that either compressed air or a vacuum could be used. the line was made by soldered lead tubing of 1-1/2" or 2-1/4" internal diameter, protected by being surrounded by sand in buried cast-iron pipes. The soldered connections were carefully made with the aid of a mandrel so that there was no roughness or obstruction in the pipe. The carriers, 5-3/4" long, were made of gutta-percha, the new marvel plastic, covered with felt, and with a piston of felt discs at one end. Messages could be placed within the carriers for transport.

The essential appurtenance was the valve by which a carrier could be sent on its journey, or received. A sending valve was simply a flap that was opened to insert a carrier in a chamber. When the flap was closed, the other end of the chamber opened and the carrier was on its way. A receiving valve kept the pipe closed until a carrier arrived at speed, and then admitted it to a chamber from which it could be retrieved. The construction of the valve depended on whether there was a vacuum or pressure in the pipe at that point. When a carrier was dispatched with pressure, the air had to be admitted behind the carrier. When dispatched by vacuum, the carrier would be sucked into the pipe automatically. The inverse problem had to be solved at the receiving end. Some valves, made by Siemens from 1870, who also entered the field, were even adapted to be used as way stations, on a line passing through. My reference does not say how a carrier was addressed for a particular station, but this does not seem beyond the ingenuity of the Victorian inventor.

One must expect that now and then a carrier would become stuck somewhere along the line. The first attempt to dislodge a carrier was made by sending water into the line instead of air, and this probably did the job in most cases. Water could not be admitted as quickly as air, and so was useless for normal service, but provided a very strong incentive for a balky carrier to move along. When this did not avail, the line had to be dug up and the blockage manually removed. But where was the blockage? An accurate answer to this question would save a lot of money and time. Charles Bontemps came up with an ingenious solution. He arranged a diaphragm at one end of the pipe so that its movements could be recorded by a pen on a rotating drum. Another pen marked seconds, and a third pen was driven by a vibrator that divided the seconds minutely. He then fired a pistol (blank, of course) into the pipe at one end, near the diaphragm, and turned on the recorder. The pressure wave travelled down the pipe to the obstruction, then back to the diaphragm. It then made another trip down the pipe to the obstruction and back. The time interval between the two arrivals then gave the distance to the fault, the pulses travelling at 1074 ft per second. The same method, but now done electronically, is used for finding faults on cables today.

A pneumatic pipe was like a single-line railway. Tyer's block instruments were applied to the control of the traffic, and carriers were signalled back and forth like trains. There was no danger of collison, of course, but the air pressures had to be managed properly.

The supply of compressed air could be obtained from a steam engine and air compressor (or a Westinghouse air brake compressor), and a vacuum from a steam ejector, but this was economical only at a central station with numerous lines where a boiler could be maintained. In many places, compressed air was obtained from the town water supply by allowing the water pressure to compress the air. This was considered too expensive in London, where water rates were high. The problem was completely solved with the arrival of electricity, since a small motor could run a compressor or vacuum pump with very little expense and bother. However, this did not happen until after 1900.

London, Paris and Berlin all had pneumatic telegraphs working in connection with the postal and telegraph services. The Berlin Tubular Post (a popular term, die Röhrenpost or Rohrpost) began service on 1 December 1876. Letters, postcards and telegrams were immediately dispatched to the proper offices, and in some cases could be delivered within an hour of the time of posting. Small packages could be sent to railway stations when there was insufficient time to use a messenger. The contrast with modern postal services is notable.

In New York, the Western Union Telegraph Company put in a pair of 2-1/2" ID brass tubes from their main operating room at Dey and Broadway to the Stock Exchange, and another pair to the Cotton Exchange, in 1876. In 1879, the Times, Tribune, Herald, World, Sun and Staats Zeitung each got a 1-3/4" tube for Associated Press reports, and in 1884 four tubes were sent on different routes, with way stations, to the midtown messenger office at 5th Ave. and 23rd St. The tubes were laid in masonry trenches with manholes at intervals. Pumping engines at the termini supplied pressure and vacuum reservoirs, so the carriers could be moved by vacuum and pressure, or by vacuum only. The 2" diameter carriers were of vulcanized felt, with felt collars on each end, and could hold 100 messages each on the standard tissue forms. A way station was alerted that a carrier was coming by an alarm bell, and a wire screen was rotated into the tube to catch it. A carrier took 2 minutes and 12 seconds to travel a 14,500-foot tube, a speed of 13.5 mph. 10 carriers could be sent per minute.

There were occasional proposals for tube railways driven by air pressure, in which the carriages were inside the pipe. This would eliminate the seal, of course, but it would be difficult to imagine passengers willingly entering the carriages. No such schemes ever saw light. The Atmospheric Railway of Clegg and Samuda was something different, a train propelled by a piston travelling in an evacuated tube, given a full-scale trial in the 1840's between Exeter and Newton Abbot on the South Devon Railway. The longitudinal seal was the fatal problem in this exercise.

The pneumatic telegraph was used widely in offices, shops, factories and mines where it was necessary to send papers or small objects physically from one point to another with more speed than could be provided by a messenger. In some cases, it competed with carriers moving on suspended wires, but was much more flexible than such schemes because it could be routed anywhere, and for considerable distances. An example within living memory is the use of pneumatic tubes in large stores, introduced before the widespread use of cash registers and point-of-sale terminals. When an item was sold, the salesperson would send a sales voucher and the customer's payment to the cashier's office. At this point, the sale was recorded, and the receipt and change were returned in the carrier. In this way, money was handled at one point only, eliminating the chance of defalcation. In a factory, samples could be sent to a central laboratory for analysis, and the results returned for process control.


A.-L. Ternant, trans. R. Routledge, The Telegraph (London: G. Routledge and Sons, Ltd., 1895)

J. D. Reid, The Telegraph in America and Morse Memorial (New York: John Polhemus, 1886), p. 736f

Return to Telegraph Index

Composed by J. B. Calvert
Created 6 May 2000
Last revised 17 April 2001