affined wood afforded absolute insulation or not, since it insulated sufficiently not to short-circuit the secondary.

If I had intended to measure the electric resistance of the dielectric, it would indeed have been necessary to have perfect insulation; for break-down tests, however, it does not matter whether some micro-Hinperes leak through here and there.

I also beg to disagree with Dr. Williams's statement, that an insulation is worthless if it lets an electrostatic charge escape, because work is done upon the dielectric, and it, therefore, must break down ultimately.

The electrostatic charge represents such an infinitely small amount of energy, that it may leak through the dielectric in practically no time, and still do no harm, for lack of energy. Dr. Williams leaves out of consideration here, that besides the work done upon the dielectric by the leakage current, at least under alternating stress, a very much greater amount of work is done upon the dielectric, outside of all leakage, by what I call dielectric hysteresis, a kind of molecular friction in the dielectric under changing electrostatic stress, anaiogons to magnetic hysteresis.

The effect of dielectric hysteresis upon the insulation no electrostatic tests show, and therefore electrostatic tests can not give us results reliable for alternate current practice.

Take for instance a long concentric cable, conveying alternating currents of say 10,000 volts. Perhaps 1,000, or even 2,000 watts are continuously exerted upon the insulation by dielectric hysteresis. This cable will certainly not break down sooner, if besides the 2,000 watts lost by dielectric hysteresis 10, or 20 watts leak through the insulation. Now suppose you replace the insulating material of this cable by a dielectric of much inferior insulating resistance, but of very small dielectric hysteresis, as oil. An electrostatic charge will disappear almost instantly, and perhaps some hundred watts leakage current act destructively upon the dielectric. But if the dielectric hysteresis of this cable is very low, the total amount of energy exerted upon the dielectric will be much less than in the first cable, and while electrostatic tests show in the first cable an almost absolute insu'ation, and show the second cable as " worthless," still in practice the first cable will break down very soon, the second stand for years and years.

Hence :—Leakage through the insulation is not the only work done upon the dielectric, and in an inferior insulating material not even the greatest, but more work may be expended by dielectric hysteresis, and though air is much superior to oil in insulating resistance, it is very much inferior to oil in dielectric hysteresis, and in disruptive strength; and, therefore, air cannot compete with oil as high potential insulation.

Of all this, electrostatic tests show nothing, but give only one quality, and not the most important one either.

The insulation of my transformer was very fair, the wood boiled in paraffin, until all the moisture was expelled, and cooled in the paraffin until it began to solidify, so that the pores were completely filled, and the wood at the outside covered with a layer of paraffin. In consequence thereof, nothing was noticeable when touching one terminal. If, however, the connecting wires between transformer and electrodes were laid on the table, from either electrode small sparks could be drawn by the finger. But •till the results were exactly the same, as was to be expected.

Now with regard to the discrepancies noted with air in Table XV, I entirely agree with Prof. A. Macfarlane, that the shape of the electrodes has a considerable influence, especially in a certain range, and have therefore excluded all the tests with pointed electrodes, and brought only some tests with balls as electrodes, without laying great stress upon them. I cannot, however, accept the influence of the shape of electrodes as explanation of the discrepancies, because in this case the differences should be greater for greater sparking distances, and the different curves converge toward each other at small sparking distances, while rather the opposite is the case.

Calculating the curves of electrostatic force between the elec trades shows, that even if the distances of the plates is equal to their radius, the distortion of the field is not yet so great, and in agreement herewith I found that even at a sparking distance of 2.2 cm. the disruptive discharge took place about as often near the centre of the plates as at their edges, so that this can not account for the discrepancies, which amount to over 100 per cent, between the different observers. Besides, if we cancel the tests made at greater sparking distances, the tests up to V— 10,000 xolts alone, already give the parabolic curve.

The agreement found with liquid dielectrics—which are known to have a very small dielectric hysteresis,—and the disagreement of the tests made with air—which has a very large dielectric hysteresis—rather point to an influence of dielectric hysteresis upon the sparking distance, and once more make tne application of tests made under conditions where dielectric hystersis is absent, upon the alternating practice of suspicious value.

The values calculated from Prof. Macfarlane's formula agree with the parabolic formula also,

S — 19.1 F+ 1.13 F2

with anaveagc difference of .011 cm., or about 4 per cent.

With regard to the extrapolations from my empirical formula, they certainly claim no reliability whatever, an empirical formula being reliable only within the range covered by the tests. I have given the potential calculated for a lightning stroke merely as a curiosity, of interest only in so far as it gives a very much lower value than former extrapolations of this kind, and therefore is perhaps somewhat nearer the truth than the former values, of 5,000,000,000 volts, etc. It is of interest, however, to note, that in the E. Thomson high-frequency transformer a sparking distance of about 2 metres has been reached, which according to my formula would correspond to about 400,000 volts, while from the

ratio of transformation the E. M. F. has been estimated to about

half a million volts.

Furthermore, comparative tests of the sparking distance in

air and in oil (A. A. Campbell, London Electrician) gave:
2 inch oil equivalent 7.875 inch air.
1$ "" " 5.125 ""

1 "" " 2.75 ""

extrapolating for air by means of the parabolic formula, gives for

oil:

that is, the sparking distance in oil is nearly proportional to the potential, a result which is in agreement with tests made at lower potential difference, and thereby confirms the parabolic airformula to a certain extent.

[remarks By Dr. J. B.williams, At Meeting Of March 2i, 1893.]

Dr. Williams :—Mr. President: Before the meeting closes I would like to say a word or two; for it is evident that I have been misunderstood—probably from not having been sufficiently explicit—both to my remarks on electrostatic testing, at the last meeting, and to the statements made in my paper on "Oil vs. Air as an Insulating Medium," which was read at the general meeting June 8th, 1892.

At the commencement of my paper, (p. 601 of the Transactions) I say, " The relative powers of the oils and air to resist disruptive discharges, and also to insulate alternating currents, will not be considered at this time, as the writer proposes to reserve those subject for a future paper."

Thus it will be seen that my statements were made with reference to direct currents only.

During the discussion at the February meeting, as I began by referring to direct currents, I supposed that it would be understood that my remarks would be confined to the consideration of electrostatic tests, with respect to direct currents. I should not like to be placed on record as one who does not know the difference between the values of the oils and the air as insulators for alternating currents; nor as one who claims that the electrostatic tests heretofore exhibited or described are suitable for insulated conductors, intended to convey alternating currents.

AMERICAN INSTITUTE OF ELECTRICAL
ENGINEERS.

New York, March 21st, 1893.

The seventy-fifth meeting of the Institute was held this date at 12 West Thirty-first Street, and was called to order at 8.30 P.M., by President Frank J. Sprague.

The Secretary—At the dinner of the Institute, which took place last May, on the occasion of the Annual Meeting, it was voted that an annual dinner be made a regular feature of the Institute, and in accordance with that voted the Council has today appointed Messrs. Phelps, Hamilton and Pickernell as a committee on the dinner to take place on the occasion of the Annual Meeting May 10th.

At the Council Meeting this afternoon the following Associate Members were elected:

Mcci.i'f.r, C. E. Superintendent, First District, So. G. A. Hamilton.

Bell Telephone and Telegraph Co., F. A. Pickernell.

Richmond, Va. Thos. D. Lockwood.

Mottram, William T. M. Electrical Engineer, New Or- Wm. J. Hammer.

leans Traction Co., 102 Canal St., Francis R. Upton.

New Orleans, La. Oscar T. Crosby.

President, The A. C. Robertson M. J. Wightman.

Co., Electrical Engineers and Con- H. Bergholtz.

tractors, Wilkesbarre, Pa. Ralph W. Pope.

Assistant in Mechanical Department, John C. Lee.

American Bell Telephone Co., 2 H. V. Hayes.

Craigie St., Cambridge, Mass. Thos D. Lockwood. Professor of Electrical Engineering, Geo. A. Hamilton.

University of Vermont, Burling- Geo. M. Phelps.

ton, Vt. F. A. Pickernell.

Robertson, A. C.

Spencer, Theodore

Storrs, H. A.

REPORT OF MEETING OF BOARD OF EXAMINERS.

March 15TH, 1893.

Present—Messrs. W. B. Vansize, Chairman, E. T. Birdsall, G. A. Hamilton

C. O. Mailloux and E. P. Thompson.

R. W. Pope, Secretary, present ex officio.

The President :—I take pleasure in announcing that the paper of the evening is by Dr. Charles E. Emery on " The Cost of Steam Power produced with Engines of different types under Practical Conditions; with Supplement relating to Water Power." Dr. Emery needs no introduction to his associates; he has been for a long time one of the leading members of the American Society of Mechanical Engineers, and a Vice-President,