The conclusions to be drawn from the diagrams and tables are that if an amount of exhaust steam can be constantly used up to abont 80 to 85 per cent. of the whole amount exhausted from a high pressure engine, the most economical plant to put in would be a special form of compound engine; but if more than 80 to 85% of the exhaust could be used for heating purposes, then the proper type would be the high-pressure non-condensing. The condensing engine, running with a portion high-pressure, comes between the compound and non-condensing in running expense below 75 per cent. of the amount of exhaust steam used, and above 75 per cent. used it becomes a regular non-condensing engine.

If the amount of exhaust steam used were a variable, but averaged more than would allow for equal cylinder on a compound engine, the proper type of engine to use would still be the noncondensing. If the average of the variable amount fell below that amount which would allow for equal cylinders on a compound, then the proper type to use would be the compound engine. (For these amounts see later on.)

There is one advantage of the compound over the non-condensing with variable amounts of exhaust steam used, viz.: The low pressure cylinder, being arranged for a variable cut-off, can control the variation, thus making use of all the steam, decreasing the amount used in the high pressure cylinder and preventing any wasteful and unpleasant blowing-off of exhaust steam.

The practical limit of the average proportion of exhaust steam which can be used and still employ the compound system, when the quantity required is variable, is when that proportion requires equal cylinders on the compound engine, and this limit is established by the ability to control the steam exhausted from the high pressure cylinder.

I would like, however, to reproduce right here two indicator cards (Fig. 13), which were taken from the compound engine at the Lower Pacific Mills, with the 44 inch cylinder running high pressure and the 32 cylinder low pressure. 17 per cent. of the steam exhausted from the 44 inch cylinder being taken into the 32 inch, and 83 per cent. was taken from the receiver for various heating purposes. Similar cards to these have been taken day after day from the engine. These cards show an extreme case of this method of running. It is not intended to be the regular way, but under the conditions then existing, it was the most economical way to run.

II. RELATIVE AREAS OF CYLINDERS.

When steam is taken from the receiver for other purposes than power, those purposes to which it is put will determine the average pressure which should be maintained in the receiver, and the average quantity of steam which shall pass into the low pressure cylinder.

It will thus be seen that the most economical pressure in the

Fig. 18.

Crank End- 44 inch Cylinder.

Crank End-32 inch Cylinder.

receiver cannot be considered to any extent nor the equalization of power in the two cylinders, and that the best form mechanically for an engine for this kind of work is a pair of tandem engines, although there is more care and trouble running the tandem than with the cross-compound.

The size of the low pressure cylinder as compared with the high pressure will depend then upon the average portion of steam exhausted from the high pressure cylinder which is to go into the low pressure cylinder, the pressure at which it enters the high pressure cylinder, and upon the pressure at which it is to go into

the low pressure cylinder. All of these conditions are variable within certain limits, and an average must be determined as nearly as possible for each.

To arrive at the proper proportions of the cylinders, we must first consider an engine where no steam is used for other purposes than power, and determine the proper ratio of areas of cylinders for different receiver pressures.

A very large number of the high duty pumping engines of the compound type have relative areas of cylinders of about 1 to 4 for about 100 pounds boiler pressure, and this proportion has been adopted by some of the builders as proper when steam is used for power only. The double compound marine engines which gave very economical results, when not compared with the triple compounds, have a ratio of areas of cylinders of about 1 to 4 for 90 pounds boiler pressure, but these engines usually carry a higher receiver pressure than would be required when steam is taken from the receiver for other purposes than power.

Some of these engines whose ratios are 1 to 4 have given remarkable economic results, and so also have those of smaller ratios, and there is probably not very much difference in the economy to be obtained between engines whose ratios of areas of cylinders are anywhere between 1 to 3 and 1 to 4 with ordinary receiver and boiler pressures. There is a difference, however, in the cost of the engines, those with the smaller ratios being less expensive than those of higher ratios, and thus the charge for interest on plant and depreciation of same is less for smaller ratios.

Those engines in the above table which have been tested have shown excellent results.

The test of the engine at the Wetamoe Mills, Fall River, by Mr. Barrus, made on an engine with unjacketed cylinders, gave 16.28 lbs. of water per hour for the running time, and this based on the stipulated evaporation of 10 lbs. of water per lb. of coal would be 1.63 lbs. of coal per I. H. P. per hour. The boiler pressure was about 94.5 lbs., and the receiver pressure about 6.7 lbs. in this test.

The test on the Nourse Mill engine by Mr. Henthorn, showed the remarkable result of 1.63 lbs coal per I. H. P. per hour, including all coal or wood used for starting and banking fires for a week's

run.

The test lately made at the Atlantic Delaine Mills showed also

* For details of last four engines, see "London Engineering," July 20, 1888, Mechanics, August number.