where the fine line is drawn through the experimental points and the calculated curve is traced by the thick line. P The value for p here obtained was then used to calculate the curve shown in Fig. 2 for a 1,200 watt dynamo of similar type. It will be noticed that a great discrepancy appears in the quality of the iron in the two cases. The only reason which the writer can advance is that the small forgings were supplied at short notice, whilst, with the larger ones, sufficient time was taken to anneal them thoroughly. A similar comparison was finally made for an 18 unit 4-pole machine (see Fig. 3), p of course being redetermined for this type. In addition to the quantities given on the diagrams, the following are necessary in order to verify the results obtained :— It is noteworthy that if we imagine such a thing as complete saturation to be unattainable, then that portion of the magnetic circuit which is directly excited must, at some point, overstep the other portion in its degree of saturation whatsoever may be their relative sections. For, if we suppose the armature indefinitely close to saturation point, no increase in the magnetic induction through it can take place except the pressure at the pole pieces becomes infinite, which means an infinite leakage through p in addition to the armature lines; thus, whatever may be the section of the field cores they must ultimately become more nearly saturated than the armature. The system adopted by Mr. Kapp provides for an infinite resistance in the armature core at saturation point, consequently by his formula the fields will always be found to saturate first. The determination of p therefore presents no difficulty other than a correct assumption for the quality of the iron, for the horizontal tendency of the experimental curve may be very accurately judged and a maximum value for E, obtained at which the field cores will be saturated. and We have then Z1, p1 and Z, as known quantities, whence 2 Ρι = Pi P It may also be remarked that the fall of potential in the armature and series winding in a compound machine on open external circuit being practically negligible, the proper exciting power due to the shunt winding may be taken directly from the calculated curve; thus, if makers can but obtain iron of reliable quality, the dynamo can be calculated in its entirety with all practical accuracy by this very valuable method. 9, MONTSERRAT ROAD, PUTNEY, December 14th, 1886. GUY C. FRICKER. ELIOT, MAINE, U.S.A., Sept. 29th, 1886. MY DEAR PROFESSOR HUGHES, I have taken great interest in your researches into the phenomena of self-induction as manifested in different metals, and in the practical application of these discoveries; and I trust that you will pardon me for bringing to your notice one or two curious facts that have come under my own observation. When I had built the Fire Alarm Telegraph in Boston, some time in the summer of 1852, I think it was, one of the signal-circuits was struck, and one or more of the electro-magnets in the signalboxes had the wire with which they were wound (No. 23 B.W.G.) burned off. After testing the circuit carefully, and having mended all the breaks, we were still unable to charge the signal-magnets for some long time, perhaps three or more hours, the magnets seeming to have been paralysed so as to be unable to receive a charge of magnetism, but after a time, perhaps before the next day, they recovered their tone, so to speak, and worked well afterwards. I am not quite sure whether this occurrence was repeated or not; if it was, I could possibly find account in my notes at the time. Another interesting occurrence was in the City of New York, in the year 1870, if I mistake not. The city was then having many miles of compound steel and copper telegraph wire strung from one signal-box to another; three of these wires ran over the Deaf Mute Asylum at Carmansville, but had not yet been connected to the signal-boxes for which they were designed. The lightning struck the building, cut off a cast-iron water pipe that was some three inches in diameter, and let all the water out of the tank that was in the attic of the four-story building. It also burned off one or more of these three wires, and from one, at least, it stripped all the copper for a distance of more than one hundred feet, leaving most of the copper in small bits, more or less than half an inch in length, lying on the ground, but leaving the steel wire unharmed, and as bright as a new tin pan. The compound wire had been made by drawing the steel wire through a bath of melted tin, then drawing a ribbon or skelp of copper over the tinned steel. Lastly, the copper was sweated to the steel by drawing the compound wire through melted tin. This stripped steel wire, with its coating of tin, remained bright for more than a day, but after that time it began to rust, and was soon replaced with new wire. I remember that somewhere about that time an iron wire in the eastern part of Massachusetts was struck by lightning, and for many rods it was melted into short pieces, which dropped upon the ground beneath the wire, and were found lying there by the line repair-man. I remember that in my early experiments in electric lighting, about the year 1859, I was trying a wire of aluminum for incandescent purposes. I was using a piece of wire that was perhaps five or six inches in length, and possibly threehundredths of an inch in diameter, or a little more. The wire was suspended in form of a loop, and by means of a rheostat the current was increased until the light was very bright. I noticed that it was quite flexible, or rather flexile, for it had no elasticity, behaving in this respect in an entirely different manner from a wire of platinum. I heated and cooled it two or more times. I then noticed that it seemed shrivelled, somewhat like in appearance a shed snake-skin. I punctured the wire with a fine steel needle, and withdrawing the needle there followed a stream of melted aluminum to a length of rather more than the quarter of an inch; this metal cooled immediately, and remained sticking out from the wire after the current had been withdrawn. I suppose that the outer surface of the heated wire became oxidised, and perhaps hydrated, and in this condition had sufficient tenacity to hold together with its interior in a state of fusion, while its exterior was solid. While this loop was in this condition, solid without but fused within, I presented a small steel magnet towards it, and the loop was immediately attracted in virtue of the magnetic effect of the current; the presented pole caused the loop to twist so that its sides came together, short-circuiting itself, and thus burning off at the junction. I tried other wires, some of platinum, some of copper, and some of iron, but could not succeed in fusing them in the interior and have the surface remain solid. I did not try magnesium, not at that time having any on hand. I have often thought it would be interesting to try it. I found the following empirical formula for the maximum current which a platinum wire would withstand was extremely useful to me in my early, and also later experiments, viz. :— |