form, so that the the same air-compressor and tanks furnish air for the oil circulation and for cleaning.

DISTRIBUTION SYSTEM

From the station let us go to the cost of the distributing system, which is now becoming an enor

mous item of expense since underground conduits. are required in many places. I give the following curves, showing the cost under Belgian block and materials of

per foot of construction asphalt, and with various of conduit. The cost of underground

construction complete is about four times that of overhead construction

This curve was prepared by the electrical commission of the city of Baltimore, and from our work in Brooklyn these curves have been found to be practically correct. We are using cement-lined pipe only. The most costly construction is the wrought-iron pipe, the upper curve. The lower curves are for Camp & McRoy terra-cotta duct. The upper curves represent three-inch ducts and the two lower ones two-inch ducts.

In the choice of duct and insulation on conductors, again let me emphasize the importance of reliability. With such an enormous investment it is almost out of the question to consider material liable to break down.

In the tests which should be required of the cable two important features should be made prominent. First, the testing apparatus should be at least about fifty kilowatts capacity. The seven and eight-kilowatt testing transformers, which up to within a few years were considered sufficient, were a delusion. When it is stated that a cable must stand a test pressure, it should be required that that test pressure be maintained at least one hour, as the momentary test is not sufficient to develop faults. Make the insulation resistance requirements moderate, but insist on the test pressure from a large transformer being maintained one hour.

In the matter of joints employ only the most skilled labor.

PLANT MANAGEMENT

Cost per unit in what follows refers to 1,000 watts maintained one hour.

The price charged per unit is made up of cost and profit.

The cost may be subdivided as to

I. Generation and distribution; II. Maintenance; and III. Management.

It may again be divided into (a) material and (b) labor; (a) representing (1) fuel, (2) water and (3) supplies; (b) representing (4) wages and (5) super

II. Maintenance is made up of (6) material, (7) wages and (8) superintendence, and

III. Management may be subdivided as to (9) directors, (10) salaries and (11) expenses.

As we compared total cost of plant per kilowatt

and found it on the average to be under $400, calculations as to cost of operation of plant per unit output per year gives that figure to be, on the average, under fifty cents. That is, on the average, it requires fifty cents' worth of plant to generate and distribute one kilowatt-hour per year.

This immediately brings out an important point, namely, load-factor. This feature has been brought

out with advantage by Mr. Hammond from data of the English companies.

Such data are not

obtainable yet in America,* so I quote in full from Mr. Hammond's paper.

"LOAD-FACTOR

"Former writers have not failed to dwell upon the marked effect which the load-factor has upon the costs of production, and in my estimation of the influences which bear upon that problem, I place load-factor directly after output. Without the latter the most favorable load-factor is useless, but combined with a healthy output, a favorable load-factor greatly reduces

costs.

"So convinced is Mr. Arthur Wright of the importance of load-factor, that he offers to supply all his customers at three cents per unit for any consumption beyond an average of one hour per day of their maximum demand throughout the year.

"The advantages arising from a high load-factor are, however, universally acknowledged, and I therefore

Since writing this paper I have learned that Massachusetts has initiated this feature; and I understand that this year data, similar to what follows here for English companies, will be required from every one of the stations in Massachusetts. The data blanks will be distributed by the Gas and Electric Light Commission. The January report for this year is a very interesting report, and, if you have not obtained it, I advise you to write for it. It gives valuable data of all the gas and electric light companies of the

state.

pass on to lay before you the data in connection with it that I have been able to collect.

"These data are not demanded by the Board of Trade, and I am much indebted to those works engineers who have responded to my request for information on this head.

"I use 'load-factor' in the sense of the ratio of the actual units generated to the product of the maximum of the year and the total hours of the year; the 'load-factor,' therefore, of 1896 being obtained by multiplying the maximum load of 1896 by 8,784 (ie., 366 days by twenty-four hours), and of 1897 by 8,760.

Though it appears to be the custom at every works to record the maximum load, the units generated' are in the majority not arrived at by meter, but by calculation, and in some by a process. which one of my correspondents describes 'cooking.'

as

"Under these circumstances I have hesitated about dealing with the data, but, finally, being convinced of the absolute accuracy of a portion, and of the approximate accuracy of the bulk, I have decided to set it out, leaving the members of the Institution to convince themselves in particular cases as to whether the units generated' have been arrived at in an accurate manner."