thinks the information contained in the appended series of tests (whilst not altogether within the scope of this treatise), taken to arrive at the working cost of two or three private installations, by Mr. Thomas Hanning, A.M.I.C.E. (who has made a special study of this subject, and who has given the author the results of his valuable experience), may be of considerable use.

In towns supplied by electric energy from central stations the question to be considered is whether it would be more economical for the consumer to lay down a private generating plant or take his current from the town supply. The prices charged by electric lighting companies vary from 41d. to 8d. per Board of Trade unit, or 1,000 Watts hours; whereas it will be found, from the annexed tables, that current can be privately generated at something under 1d. per unit-that is, for fuel consumption alone. It should, however, be mentioned that all these cases may be termed favourable for private generating, since they represent long working hours-a most important point in comparing public lighting with private installations.

In the first test, for instance, the plant works all the hours of darkness in the year, with the exception of Sundays, and practically has as large a load as possible under ordinary conditions. Even 1,000 hours' lighting a year is equivalent to an average of about four hours per day for the nine months when artificial illumination may be necessary. In engineering shops the annual period during which artificial light is used is not more than 300 hours unless overtime is worked.

The fuel consumption is greatly increased with varying loads, because of (1) the waste in getting up steam and keeping boilers banked up; (2) the decrease in mechanical and thermal efficiency of the steam engine with light loads. The steam consumption per horse-power increases as the average load factor decreases.

It may also be seen that with gas engines, although the cost per unit generated remains fairly constant, the lighter load factor greatly increases the total average cost per unit, the amount set aside for interest on first cost and depreciation being a fixed charge.

The following are the results and conclusions arrived at by Mr. Hanning in his report :—

It will be interesting to consider the facts arising out of these tests in rotation as they appear in the tabulated list.

Consumption of Gas.-In relation to the lamps burning, the consumption is lower in No. 3 test than in No. 2, due probably to the fact that the engine is more nearly proportioned to the work it has to do.

Cubic feet of gas consumed per I.H.P.—A remarkable and noticeable feature in the perusal of these figures is that in the whole of the tests the gas consumption-in relation to the I.H.P.-is lower at a certain light load than at full load. This must be in some way due to the regulation of the gas valve, which at these ́loads was not full open, but shut off to a certain

extent.

In No. 4 test (Olympia) the engine was fired electrically; therefore there was no gas consumed for firing, as in the other

cases.

Indicated Horse-power.-There is a striking similarity in the figures of power required to do the full load in steam engines and gas engine plants Nos. 1 and 2 tests, which will no doubt surprise many who contend that the steam engine is more efficient than the gas engine. A comparison of the power required to generate energy for a certain number of lamps by a large engine and small is shown on tests Nos. 2 and 3, or in No. 2. 40 I.H.P. is required for 200 16 c.P. lamps, and 48 for 300; whereas in No. 3 30 I.H.P. suffices for 220 lamps, and 39 I.H.P. for 300.

Number of lamps burning per I.H.P.-There is a great regularity about these figures which should almost make them standards for the purpose of calculation.

Consumption of gas in cubic feet per lamp hour.-The same may be said of this.

Reading on switch-board in volts. It would appear as though the machines in Nos. 1 and 4 tests were over-compounded.

Reading in ampères.—In No. 1 test 112 ampères is required

I.

Steam Engine Plant (exclusive of Boiler).--Installed in 1891

III. Tangye' Gas Engine Plant.-Cylinder, 13 inches diameter; stroke of engine, 20; revs. per minute, 160. (Installed in 1893.)

ENGINE

DYNAMO

Watts per I.H.P. full load = 514-84. Watts per B.H.P. full load = 638.4.

Gas at. 1s. 8d. per 1,000 cubic feet.

for 200 lamps, and 125 in No. 2. This is no doubt due to a difference in the resistance of the lamps and wiring.

Watts consumed per lamp.-The fluctuation in these figures is difficult to account for, and point to probable unintentional switching on and off of the lamps by the workpeople.

Output of Dynamo in kilowatts.-The higher output in Nos. 2 and 3 tests is due to the higher voltage to which the machines are run to overcome loss over large area of circuits.

Electrical output in H.P.-In proportion to the number of lamps burning this is fairly consistent throughout.

Combined efficiency.-These figures are important to all concerned in electrical engineering, as they give results from actual every-day practice, and are not derived from specially prepared plant, which is the case in a great many of the tests which have up to the present been published.

The steam engine plant is slightly more efficient than gas engine plants Nos. 2 and 3, but No. 4 stands highest—due in a great measure to the higher efficiency of this dynamo.

Cost of Fuel, &c.—From the results it would appear that with a steam driver plant the cost of water and coal is about half the cost of gas (at 1s. 7·8d. per 1,000 cubic feet) consumed by a gas engine, and that the total cost of generating a unit of electricity is about 20 per cent. less in favour of the former (see notes as to rates of interest, &c.). But against this is to be put the increase in rate which would be chargeable to the steam plant due to interest and depreciation on boiler, repairs to same, &c., cost of stoking and carting away clinkers, ashes, &c., which has not been estimated for. The cost of gas is noticeably constant in these tests.

Number of lamp hours.-This is one of the most important factors to be considered in comparing the cost of lighting as between obtaining energy from supply company and consumers generating their own. In No. 1 test the mill worked from Monday to Saturday, night and day, and light was required all the hours of darkness in week days throughout the year. Consequently the number of hours during which the lamps burned are heavy--namely, 3,500. The working hours of Nos. 2 and 3 plants are 1,000 and 700 per annum respectively. Reference