"The Excelsior Electric Company seems to have come to the conclusion that the incandescent light has its sphere of usefulness as well as the arc light for it has lately given considerable attention to the subject of incandescent lighting. Besides making shunt-wound dynamos specially adapted to this purpose, the company proposes to manufacture its own incandescent lamp and assessories. This lamp (Fig. 18) differs from the others in the form given to the carbon loop, which, in this case, assumes the semblance of a letter T. This form of loop is preferable to any other form we have yet seen, for it makes the light-producing focus seem more spherical. The carbon loop is formed of vegetable fibre of a peculiar kind, which is found to last remarkably long. The weak spot in incandescent lamps is generally in the imperfect connection between the wires passing through the glass and the carbon filament. It is claimed that in these lamps the connections are made in an improved manner which leaves nothing to be desired.

"Fig. 19 shows a plain form of holder used with these lamps. The three bent wire springs embrace the lower portion or neck of the lamp, and the small platinum wires protruding through the glass are pinched between the split extremities of the circuit, connecting by means of screws, one on each side. Fig. 20 shows one of these holders mounted on a hand-stand and supporting an Excelsior incandescent lamp. In Fig. 21 we have a more ornamental form of holder, in which the bent wire springs are replaced by sheet brass springs, forming part of a casing.

"The device represented in Fig. 18 is called an 'automatic substituter.' The circuit is divided at first into two branches, one of which includes the electro-magnet shown in the base, together with one of the lamps. The other includes the other lamp, but remains closed only as long as the armature is retracted. When the current is put on, therefore, as soon as the magnet is excited, the second circuit is opened and only the first lamp remains in circuit. If this lamp breaks, however, the armature instantly falls back and closes the circuit through the second lamp. There are many situations where it would be very desirable to have the incandescent light in connection with the arc light. For instance, there are often places where a few incandescent lights would be more convenient and useful than one arc light. To meet this want attempts have been made to run incandescent lights in the same circuit with arc lights. For instance, it seems quite plausible that if the current in circuit is ten amperes and each incandescent lamp requires a current of one ampere, by dividing the circuit into ten branches, each of which includes a light, the lamps ought to burn successfully. But the trouble is that if one lamp breaks the nine others will receive the ten amperes among them, or each will receive one-tenth of an ampere more than is 'healthy' for it. This may lead to the rupture of another, and this again causes an increase of one-tenth of an ampere. In short, the breakage of one lamp endangers the rest, and, according to our experience, it generally causes their total destruction, thus breaking the circuit. This is evidently unpractical.

"Mr. Hochhausen, however, has attacked the problem anew, and he has indeed produced a very ingenious solution of it, the principle of which is simple enough. When a lamp breaks or is cut out, he proposes to shunt the surplus current through an equivalent resistance. If two lamps break he diminishes this resistance. Finally, if there is danger that the lamps may burn up altogether, they are all cut out of the circuit. These results are all accomplished by means of an automatic regulator (Fig. 22).

"The switch shown on the right-hand side is normally pulled toward the cover by a spring not shown in the figure. In this condition, the incandescent lamps are cut out of the arc light circuit. The contact piece nearer to the cover of the case is connected to the armature of an electro-magnet within. This electro-magnet is included in the main circuit, and when the current is passing it attracts its armature, and causes the contact piece outside the case to project out from the surface of the case. If, now, the switch is moved to the right, it is prevented from moving back by the insulated end of this protruding contact. In this condition, the incandescent lamps are included in the circuit—that is to say, the circuit divides at one point into several branches, in which are included the lamps—one in each branch. But there is one of these branches which, instead of a lamp, includes a relay magnet of rather high resistance. Now if a lamp breaks the current increases in all the remaining branches stated above. The relay feels this increase, and its armature comes against a contact which closes the circuit through another branch, including an equivalent resistance placed in the case, and consisting of wire wound on wooden reels, every turn of the wire being accessible to the air so as to admit of ready cooling. If now another lamp breaks, the armature of the relay is attracted further, and closes a circuit through a lamp placed on top the regulator case to warn the attendant, who immediately proceeds to replace the two broken lamps. But meanwhile another lamp still may break, and as the relay cannot make further compensation there is imminent danger that the other lamps will break.

This is prevented by a kind of automatic cut-out. The relay is provided with a supplementary armature, which is attracted when the current increases beyond the limit compensated for by the relay itself. The motion of this armature short-circuits the electro-magnet in the main circuit, whose function is to make the contact piece protrude on the outside of the case. Instantly this contact piece moves back close to the surface of the case again, and the switch is quickly pulled back by its spring. As soon as it touches the metallic surface of this contact once again the incandescent lamps are short-circuited and the arc lights are left alone in the circuit. It is evident that the principle of this regulator may be extended not only to one but to several multiple series, and the number of lamps compensated for may be regulated quite easily. The only objection that can be made against this mode of regulation is that there is no economy of energy when only a few of the lights are burning, because the compensating branches absorb the energy previously expended in the lamp in producing light. However, as the number of lights accidently put out would usually represent only a trifling proportion of the total in use, this objection is not serious. Besides, in cases where this procedure is resorted to, expediency or convenience is generally a more important consideration than economy. It may interest some of our readers to know that the incandescent lamps we used in some of our tests and experiments had an average resistance of 80 ohms cold and approximately 40 ohms when hot. In conclusion we wish to express our thanks to Mr. Hochhausen, electrician of the company, Mr. Oswald, his assistant, and to Mr. Geo. D. Allen, general manager, for courtesies received and facilities placed at our disposal in the preparation of these articles."

Acknowledgment
The writer is grateful to Charles A. Crider for providing a photocopy of the advertisement of the Excelsior Electric Company, which appeared in The Electrical World, on January 5, 1884, page viii.

References
1) C. O. Mailloux, "The Electric Lighting System of the Excelsior Electric Company," Part 1, The Electrical World, Vol 3, No 1, Jan 5, 1884, pp 1-3.
2) C. O. Mailloux, "The Electric Lighting System of the Excelsior Electric Company," Part 2, The Electrical World, Vol 3, No 2, Jan 12, 1884, pp 9-10.
3) C. O. Mailloux,"The Electric Lighting System of the Excelsior Electric Company," Part 3, The Electrical World, Vol 3, No 3, Jan 19, 1884, pp 17-19.