About the same time Mr. J. H. Greathead completed the City and South London Railway in London. It consists of two tunnels, made of cast-iron segments, 10 ft. 6 in. in diameter each, and 3 miles long. Compressed air was only necessary for short distances at three points, as it was principally driven through London clay.

Previous to the construction of the St. Clair Tunnel compressed air had been used in the Hudson Tunnel, which still remains unfinished owing to financial difficulties. The greater portion of its length is constructed of cast-iron rings of 19 ft. 8 in. diameter.

In 1891 Messrs. Pearson and Son's tender for the construction of the Blackwall Tunnel, amounting to 871,000l., was accepted by the Council, and the work was commenced in 1892. Mr. D. Hay and the author were appointed as resident engineers under Mr. A. R. Binnie, and Mr. E. W. Moir took charge of the works for the contractors.

The Blackwall Tunnel is much larger than any tunnel yet constructed by the methods adopted. The outside diameter of the St. Clair Tunnel, which is the largest one at present, is 21 ft., while that at Blackwall is 27 ft. in external diameter.

The tunnel is level under the river, and the gradient on the north side is 1 in 34, and on the south side 1 in 36. There are four vertical shafts, two on each side of the river, and varying in depth from 75 ft. to 100 ft. below ground level. Each shaft is a wrought-iron caisson of 58 ft. external diameter at the bottom, and 48 ft. internal diameter throughout, and lined with brickwork. Each caisson consists of two wrought-iron skins, 5 ft. apart, braced together, and terminating in a cuttingedge. Two circular holes, which are temporarily plugged while sinking, are left in each caisson to give way for the tunnel through the shaft, and provision is made for an airtight floor above the level of the tunnel when necessary. The space between the two skins is filled with concrete. Two caissons have been sunk, and the two others are in course of being sunk.

The tunnel is constructed of cast-iron rings 2 ft. 6 in. long, and each ring consists of fourteen segments and a key-piece. The thickness of metal is 2 in. and each segment has flanges 12 in. deep, and both longitudinal and circumferential joints are planed.

The shield used for the construction of the tunnel is 19 ft. 6 in. long, and is 27 ft. 8 in. in external diameter. The outer shell consists of four -in. steel plates. The shield is divided into a front and back portion by two vertical diaphragms at right angles to its axis. It is thus possible, when necessary, to have a higher air pressure in the working face of the shield than in the completed portion of the tunnel. The space between these two diaphragms forms an air-lock, both diaphragms, of course, being provided with doors, by which access to the working face is obtained. At the back of this air-lock the shield consists only of the outer shell, which always laps over and outside at least one completed ring of the tunnel, and inside of which all the rings are built. The space of 4 in. left outside the rings when the shield is shoved forward is filled with grout, forced in by air pressure through screwed holes made in each segment for the purpose. Everything is, therefore, quite solid at the back of the cast-iron lining. At the air-lock and in front of it there is an inner shell, connected stiffly to the outer shell by circular girders and in other ways, and both joining together at the cutting-edge. The working face is divided into four horizontal floors and twelve working chambers by vertical and horizontal diaphragms in the line of the axis of 3 c

1894.

the shield. A hanging iron screen in each compartment about 6 ft. back from the cutting-edge forms a safety chamber at its back, where men could stand with their heads above water in case of a rush of water in the face due to air blowing out suddenly, or from other causes. Provision is made for using iron poling boards at the face, shoved forward by jacks, when in ballast, if necessary. The shield, which weighs about 250 tons, is shoved forward by twenty-eight hydraulic jacks fixed at the back and butting against the cast-iron lining, and able to exert a total pressure of over 3,000 tons.

At the present time the portion of the tunnel between shafts No. 4 and No. 3 has been constructed. The shield was started from No. 4 shaft in June 1893, the first permanent ring being erected on June 9. When first starting the face was nearly all clay, there being a little ballast at top and a little fine sand at the bottom. It was therefore decided to go on without compressed air for some time, and in three months seventy-seven rings, being equivalent to a length of 192 lineal feet, were erected. This length was done with practically no timbering, except in the top ballast for a short length, in which a good deal of water was met. A little previous to this time the cutting-edge at the bottom of the shield got bent by shoving the shield against a layer of rock which was met in the clay, and it was found now to have got much worse. It was therefore considered better to drive a timbered bottom heading in front of the shield, and put in a concrete foundation on which the shield could be shoved forward to the next shaft. This was done for the bottom 6 ft. of the shield, all the excavation above this level being done without timbering. In this way by the middle of December, six months after starting, 191 rings, equivalent to 477 lineal feet, had been erected.

As the tunnel was going down a gradient of 1 in 36, and the top of the sand remained nearly level, the latter was now up to the middle of the shield, and was carrying a great deal of water. On December 16 there was a rush of water and sand, which filled the heading with sand, and the water rose 15 ft. in the tunnel at the shield.

As the shield was now close to No. 3 shaft, which was not yet sunk to its full depth, it was considered safer not to continue the tunnel until the shaft was sunk, and in the meantime to make preparations for using compressed air when work was again started. A concrete bulkhead was therefore built across the tunnel, with air-locks in it, a little back from the shield, and when the shaft had been sunk to its full depth, work was recommenced in the tunnel on March 23, under air pressure of about 15 lb. to the square inch.

No further trouble from water was then experienced, and by going on in the same way as above described with a bottom heading, the shield was got into No. 3 shaft in May. It is now being strengthened and repaired there, preparatory to starting under the river from No. 3 to No. 2 shaft.

In the deepest portion of the river, where the top of the tunnel comes within 7 ft. of the river bed, which is all gravel at this point, it is proposed to lay a bed of clay 10 ft. thick in the bottom of the river, to prevent the compressed air blowing out; and with this precaution it is hoped that no difficulty will be met with in this portion of the work which cannot be overcome by the experience already gained in the portion between shafts No. 3 and No. 4.

1. On Methods that have been adopted for measuring Pressures in the Bores of Guns. By Sir A. NOBLE, K.C.B., F.R.S.-See Reports, p. 523.

2. On the most Economical Temperature for Steam-engine Cylinders; or, Hot v. Cold Walls. By BRYAN DONKIN, M.Inst.C.E.

The author calls attention to the important question of the most suitable temperature of cylinder walls to obtain the maximum economy by reducing condensation in steam cylinders to a minimum. To diminish condensation his experiments prove that it is essential to reduce the difference of temperature between the incoming steam and the cylinder walls. In most cases the cylinder walls are much colder than the steam, and often one-half the weight of steam is condensed during admission. Many experiments show that the nearer the temperature of the cylinder wall is to that of the entering steam the greater is the economy. With walls 40° to 60° Fahr. colder than the steam, as is often the case, the consumption is greatly increased, owing to the large amount of condensation. On the other hand, the cylinder metal can be made too hot and heat wasted at exhaust. This has also been experimentally proved by the author. The best results in steam economy have been obtained when the temperature of the internal surfaces is a little higher than that of the entering steam.

Cylinder walls can be heated in various ways. The most usual is to raise their temperature and that of the covers, by means of boiler steam introduced into the jacket spaces. Another method, and one used a good deal on the Continent, is to work with superheated steam. This steam has been employed in some cases with advantage in the jackets as well as in the cylinders. Smoke jackets have generally proved a failure. When the jacket is connected to a condenser having a good vacuum, economy of steam is also obtained, but not so much as with steam in the jackets. To secure the maximum economy also, it is important not only to diminish the volume of the clearance spaces, but especially to reduce as much as possible the area of the clearance boundary surfaces. In this way the weight of iron heated and cooled per stroke so many degrees can be materially diminished. The character of the internal surfaces, whether rough from the sand or turned, and their position, horizontal or vertical, have also some influence on the transmission of heat and condensation of steam, as verified by recent trials.

Experiments have also shown that the cylinder wall in any working steam engine is divided thermally into two parts; the outer portion remains at a constant temperature, and the inner or periodic portion is heated with each steam, and cooled with each exhaust stroke. The relative proportion of these two parts, or the depth to which the heat penetrates into the metal, depends largely upon the speed of the engine, and on the temperature of the cylinder relatively to that of the entering steam. In a non-jacketed engine, with one-inch cylinder walls, working at thirty-five revolutions per minute, the depth to which the heat penetrates and fluctuates per stroke is at least about eight to nine millimetres from the internal surface. The depth of heat-penetration for the same speed is much less with hot than with cold walls; a less weight of metal is heated per stroke; and condensation is found to be much reduced.

The author gives some of the results of eighty experiments on a small vertical engine, with cylinder 6 in. diameter and 8 in. stroke, made expressly for experimental research. The details are to be published in Proceedings of the Institution of Mechanical Engineers.' The engine was worked with very different temperatures of walls, and many tests were made, condensing and non-condensing, jacketed and non-jacketed, single and double acting, at different expansions, and with both saturated and superheated steam in the cylinder and jackets. Care was taken in these experiments only to vary one set of conditions at a time. Results of two experiments are added, one with hot and the other with cold walls, or one with steam and the other with air in the jackets. Both tests were made condensing, double-acting, with 50 lb. steam pressure, cut-off, and at a speed of 220 revolutions per minute. The walls were some 30° Fahr. hotter with than without steam in the jackets; the steam consumption per I.H.P. per hour was reduced from 414 lb. to 284 lb., or about 31 per cent.; the thermal efficiency was increased from 5.7 to 8.1 per cent., or 40 per cent., and the rate of initial condensation was reduced from 460 lb. to 217 lb per square foot per hour, or by more than

one-half. The percentage of steam present in the mixture during expansion was also increased by about 50 per cent.

Throughout these experiments an increase of economy with the hotter walls was always verified; the thermal efficiency was higher, the initial condensation less, and the percentage of steam present during expansion always increased.

The tests on this engine point to the practical conclusion that the range of steam temperature in the cylinder per stroke has much less effect on the steam consumption than the temperature of the walls.

The paper may be shortly summarised thus:-The most uneconomical results were always obtained with the cylinder walls colder than the entering steam. Under these conditions considerable initial condensation was produced, drops of all sizes up to 3 mm. diameter being formed and running down the cold surfaces. The heat also penetrated into the colder walls to a considerable depth, a certain quantity being given up by the steam at every stroke, to raise the temperature of the internal surfaces after exposure to the condenser temperature. Both at cut-off and release there was a great deal of water in the cylinder compared with the weight of steam present. On the other hand, the most economical results were always obtained when the cylinder walls were at about the same temperature as that of the entering steam. Under these conditions the rate of initial condensation was very much lower, and the drops of water formed much smaller in size. The heat penetration into the walls was also much less, a smaller amount of steam sufficing to heat the internal surfaces after being cooled by the condenser. The percentage of steam present in the mixture, at cut-off and release, was also very much increased.

If engineers and others using steam engines wish to work economically and with smaller boilers, they must arrange to keep their cylinders and covers as hot as the steam entering the cylinder; otherwise the cylinder becomes unintentionally an efficient condenser, with a large area of cooling surface.

Properly applied steam jackets are economical, because they raise the temperature of the walls touched by the steam. Those who cannot steam-jacket the whole cylinder should at least jacket the two covers, which are the most important surfaces.

Well-arranged jackets with proper-sized pipes for entering steam and exit water, without places for air to collect, are an excellent investment, and pay a good interest on the small additional cost.

MONDAY, AUGUST 13.

The following Papers were read :-

1. On Signalling through Space. By W. H. PREECE, C.B., F.R.S.

2. On Some Advantages of Alternate Currents.

By Professor SILVANUS P. THOMPSON, F.R.S.

3. Continuous-current Distribution of Electricity at High Voltage at Oxford. By THOMAS PARKER, F.R.S.E., M.Inst.C.E., M. Inst.M.E., M.Inst.E.E.

The Central Station at Oxford was started in the middle of the year 1892, and is equipped with high-tension continuous-current dynamos driven by means of belts from triple-expansion vertical engines; it is placed 1,500 yards away from the area of lighting. The current is distributed by means of a network, which is fed by motor generators transforming from 1,000 to 105 volts. These motor generators are started, stopped, and regulated from a central switch station placed in the area.

The main feature of the system is the complete control of the motor

generators from the switch station; and the number connected on to the network being varied to suit the load, it is possible to always work the transformers at a high efficiency. A small battery situated at the switch station is used to supply the small day and night loads, thus enabling the Central Station to be entirely shut down for a great portion of the twenty-four hours.

The figures for 1893 show that the total efficiency of the system was 61.62 per cent., and the efficiency of the motor generators was 74.44 per cent., including losses in mains and resistances. This is not so high as it will be in the future, when the lamps are more evenly distributed over the area. The battery efficiency was 50-64 per cent. The actual coal used throughout the year works out to 6.83 lb., or 718 penny per unit sold, which is a very good result, as only slightly over 100,000 Board of Trade units were metered. The oil, waste, water, and general engine-room stores work out to 0657 penny per unit metered. The total number of lamps installed at the beginning of the year was 4,041, which increased to 7,012 by the end of the year. As the great proportion of supply is taken up by colleges, the term time is the only part of the year when anything like a load can be obtained, and the load factor is only 6:31 per cent.

The revenue during the year under notice was 10s. 11d. per 35-watt lamp installed.

4. On a Special Chronograph.

By HENRY LEA, M.Inst.C.E., and ROBERT Bragge.

In his capacity as electric inspector for the city of Birmingham the duty of one of the writers is to test the accuracy of the electric meters used in that city. The test involves an accurate measurement on the one hand of the current of electricity passing through the meter, and on the other hand of the period of time during which the current of electricity effects a certain number of revolutions of the meter armature.

For the electrical measurements he selected a Crompton potentiometer, but as his chronograph was not accurate enough he drew up a specification upon which the English Watch Company built the chronograph exhibited to the meeting. The ideal chronograph is one with a perfectly divided dial, an absolute prompt start, a definite stop at zero, and a continuous or running motion of the hand.

The test-room chronograph under notice is not intended for use as a watch; it is purely and simply a micrometer applied to time, and therefore it has only two pointers-one from the centre, giving seconds and their fractions, the other placed halfway between the centre and circumference, showing minutes up to sixty. The dial of aluminium, being engine divided, is mechanically perfect; and being of large size, 24 inches of sight, the divisions of ten to the second are clearly defined.

The prompt, absolute start is obtained by dispensing with all intermediate gearing and relying on the main train of wheels impelled by the ordinary mainspring. The starting lever not only releases the balance-wheel, but at the same time by the momentary contact of a light spring blade impels the wheel in the required direction.

The absolute stop at the end of an observation is obtained by the frictional contact of the aforesaid spring blade with the edge of the balance-wheel.

The absolute stop at zero is obtained by fixing upon the axis of the seconds wheel a collet carrying a blade. Upon the top of the movement there is a lever, which, during the period of an observation, is held back; but when the observation is concluded it is set free by a push of the button, and falls to a point where it crosses the path of the blade already referred to, and stopping the wheels brings about the absolute stop we require.

The last thing requisite is for the seconds hand to have a continuous run. Everything tried up to the present has had a tendency to accelerate its rate, and is therefore inadmissible. The lever escapement has been adhered to, and by using hard stone jewellings, and by observing great accuracy in the angles of the pallets, the makers are enabled to double the ordinary number of beats-36,000