slipped backward or forward the proper amount, after reading, without stopping the counter. In this way a reading made from zero could he taken each minute. A similar method has been adopted in the printer. The type-wheel, which is held by friction on the shaft, and is carried with it. is at the beginning of the minute slipped backward or forward until it has the correct relative position, then for the remainder of the minute revolves at the same rate as the large gear-wheel on the same shaft. The method of slipping the type-wheel around is partly shown in Figure '2. The toothed wheel ", driven by the bevel gear, revolves once for every thirty revolutions of the main shaft. The plate or arm h revolves with it when the pawl c is allowed to catch in the teeth of the wheel. As the plate b goes around with the wheel, the pin d overtakes the arm e, which is attached to the type-wheel, and pushes it along slipping the typewheel on the shaft. When h has made one complete revolution, the pin/'on the pawl c comes opposite a notch, not shown in the figure, into which it is pushed by a spring, thus pulling the pawl from the wheel, and also holding h stationary. It takes thirty revolutions of the worm to carry the plate h around, it coming to rest just at the end of the thirtieth revolution. If at this instant the priming hammer were allowed to act, the number 30 would be printed, but if it does not strike till the end of the minute, the type-wheel will move with the large gear-wheel till that time. The striking of the hammer releases the pin f, and the pawl which has a spring of its own is again brought in contact with the teeth and so carries the plate l> around again. If the type-wheel should complete its revolution before the pin J overtakes the arm <-. then the arm e comes against another pin which stops it, until the pin il does overtake it. This is the slipping backward which has been mentioned, if the printer should be driven too fast, so that the type-wheel makes a revolution before the minute is up, or if the hammer is not caused to act for several minutes, then it will, the next time it acts, print a zero, which is the type under the hammer while the wheel is held stationary.

The hammer is released at the end of each minute by means of an electro-magnet operated by a clock, and when the hammer makes its stroke, it opens a switch in the electric circuit while the contact in the clock is still closed, so in this manner the spark on breaking the circuit is made at the switch, and not at the clock contact. The hammer is set again by a pin on the side of a wheel coming in contact with a lever, and this closes the switch again before the next contact of the clock. When the machinery is not running, the switch remains open, and so the battery materials are not consumed in useless work.

The numbers move past a stationary point which, prints its position upon the paper when the hammer makes its blow. If the number printed is exactly opposite the point, then there has been a whole number of revolutions that minute, but, if the number is above or below it, the fraction indicated by its position can easily be determined. Fig. 3, shows how the fraction of a revolution appears on the record. An exact number, 145 is

shown at (a), 145.25 at (b) 145.5 at (c) and 145.75 at (d). Usually enough of either the preceding or following number is printed to easily make it out, and to estimate the fraction, which can be read to one fifth of a revolution. In the original design of this instrument the question of adding a time printing device was considered, for it was thought to be necessary to know just the hour and minute that any part of the record was made. Such an attachment was plauned but was not built, for the reason that in nearly all cases the record made will be referred to within the first few minutes after it is printed. If it is desirable to mark a certain point on the record to identify it, the paper can be displaced, and so leave a blank at that point. The paper is fed through at the rate of fifteen and one-fourth inches per hour, and so when necessary this distance can be taken as a unit of length to measure back, and mark off the hours.

One of the ways in which the record obtained is being used, is to find the ratio of its reading to that of the speed of the dynamo being experimented upon, then from the record calculate the speeds for the different parts of the experiment, but this will only answer when the belt is able to drive the machine without slipping.

There is now being built a spindle which is to be tight geared to the main shaft, and which will run ten times faster, or about 1,400 revolutions per minute. This will be for the purpose of testing a student's ability to obtain a correct reading of speed, for the true speed of the spindle will always be ten times the printed speed for that same minute.

The printer which has been described, has been running for about three months and has not failed in any respect to do the work that was expected of it.

The Secretary :—I have blue prints of a chart prepared by Lient. Parkhurst, of the Watervliet Arsenal, entitled '• Diseases of Dynamos." As there was not time to have it put in type I had blue prints made, which are at the service of members who are interested in the subject.

A recess was then taken until 2 p. M.

Upon reassembling for the afternoon session, Mr. Willyoung presented a paper on "A New Method and Apparatus for Measuring Conductivity."

The following paper by A. H. and C. E. Timmerman on "The Heating of Armatures," was presented by Dr. Nichols.

of the American Institute 0/ Electrical En gineers, Neiv York, May ry, I8qj. President Houston in the Chair.

HEATING OF ARMATURES.

BT A. H. AND C. E. TIMMERMAN.

Although much depends upon the heating of armatures, but little attention appears to have been given to the study of a subject of such importance to the engineer in the design of electrical machinery. The losses in an armature caused by the transformation of electrical energy into heat are by no means negligible quantities, as the total amount of energy that we can obtain from a dynamo depends directly upon the temperature of the armature.

What will be the temperature of an armature when a certain amount of electrical energy is transformed into heat within that armature i How much heat will be liberated per square inch per degree rise in temperature above the temperature of the room \ Will this quantity be different for different temperatures \ Does the field aid in the escape of heat, or does it prevent its radiation ( Also, what effect has the peripheral velocity on the amount of heat radiated ? etc. Such questions as these seem to have passed almost unnoticed. Professor Harris J. Ryan, of Cornell University, has done some work in this direction, and it was through him that the present series of experiments was undertaken. M. itechniewski has also determined by experiments on machines, certain values for the heat radiated. In experimenting with dynamos or motors there is, however, some uncertainty as to the amount of heat generated. The loss in the coils of the armature can, of course be accurately calculated. The lo.sses in the iron core can be calculated approximately, but not accurately; for we do not know that the hysteresis loss due to the rotation of a mass of iron in a magnetic field is the same as that due to a variation in the magnetizing force, nor is the eddy current loss easily estimated. Owing to this uncertainty in the actual amount of heat generated, it was decided to construct a special form of apparatus so as to be able to measure accurately the amount of heat generated. Following is a description of the apparatus designed and constructed and employed in this determination. (See Fig.

The machine consists of a hollow cylinder of brass, mounted on a shaft, so that it ean be revolved between two polepieces, one end of the cylinder being detachable. In the interior of the cylinder is a coil of german silver wire. The coil is built

up in the following manner:—About 1H4 feet of german silver wire, No. 17, A. W. G., was covered its entire length with glass beads; the beads were about f" in diameter and from J" to \" in length. The wfire, thus insulated, was wound in a coil over a hollow core of glass; the completed coil being of" long and about 5" in external diameter. Thisglasscore with the surrounding coil was then slipped over the steel shaft and into place within the cylinder, care being taken to pull one terminal of the coil through a hole in the closed end of the cylinder, this terminal being well insulated from the cylinder. The remaining portion