the readers of "Nature," through the columns of that periodical, whether any experiments had been made with a telephone in circuit with a selenium galvanic element arranged as in Sabine's selenium battery; 20 and suggested that it was not unlikely that sounds would be produced in a telephone by the action of light of variable intensity upon a selenium element in circuit with it.

In September, or October, 1878, Mr. A. C. Brown, of London, submitted to me, confidentially, the details of a most ingenious invention of his, of which we may yet hear more. This invention, although entirely different from my own, involved the use of selenium in circuit with a battery and telephone, and the production of articulate speech by the action of a variable light. I am also aware that Mr. W. D. Sargent, of Philadelphia, has had some ideas of a similar nature, the details of which I do not know. I understood from Mr. Sargent that he proposed submitting selenium to the influence of an oscillating beam of light, which should be sent on and off the selenium by the action of the voice. If this is so, the effect produced would be only of an intermittent character, and a musical tone, not speech, would be heard from the telephone in circuit with the selenium.

Although the idea of producing and reproducing sound by the action of light, as described above, was an entirely original and independent conception of my own, I recognize the fact that the knowledge necessary for its conception has been disseminated throughout the civilized world, and that the idea may therefore have occurred, independently, to many other minds.

I have stated above the few facts bearing upon the subject that have come under my observation.

The fundamental idea on which rests the possibility of producing speech, by the action of light, is the conception of what may be termed an undulatory beam of light in contradistinction to a merely intermittent one.

By an undulatory beam of light I mean a beam that shines continuously upon the receiver, but the intensity of which upon that receiver is subject to rapid changes corresponding to the changes in the vibratory movement of a particle of air during the transmission of a sound of definite quality through the atmosphere. The curve that would graphically represent these changes of light

20 Nature, Vol. XVII, p. 512, Apr. 25, 1878.

would be similar in shape to that representing the movements of the air. I do not know whether this conception had been clearly realized by J. F. W., of Kew, or by Mr. Sargent, of Philadelphia; but to Mr. A. C. Brown, of London, is undoubtedly due the honor of having distinctly and independently formulated the conception and of having devised apparatus, though of a crude nature, for carrying it into execution.

It is greatly due to the genius and perseverance of my friend, Mr. Sumner Tainter, of Watertown, Mass., that the problem of producing and reproducing sound by the agency of light has at last been successfully solved. For many months past we have been devoting ourselves to the solution of this problem and I have great pleasure in presenting to you to-night the results of our labors.

Researches of Sumner Tainter and Alexander Graham Bell.

The first point to which we devoted our attention was the reduction of the resistance of crystalline selenium within manageable limits. The resistance of selenium cells, employed by former experimenters, was measured in millions of ohms, and we do not know of any record of a selenium cell measuring less than 250,000 ohms in the dark.

We have succeeded in producing sensitive selenium cells measuring only 300 ohms in the dark and 150 ohms in the light. All former experimenters seemed to have used platinum for the conducting part of their selenium cells, excepting Werner Siemens, who found that iron and copper might be employed. We have discovered that brass, although chemically acted upon by selenium, forms an excellent and convenient material; indeed, we are inclined to believe that the chemical action between the brass and selenium has contributed to the low resistance of our cells by forming an intimate bond of union between the selenium and brass.

We have observed that melted selenium behaves to other substances as water to a greasy surface, and we are inclined to think that when selenium is used in connection with metals not chemically acted upon by it, the points of contact between the selenium and the metal offer a considerable amount of resistance to the passage of a galvanic current, and thus serve to increase the apparent resistance of the selenium.

By using brass we have been enabled to construct a large num

This cell consists of two brass plates insulated from one another

by a sheet of mica. The upper plate has numerous perforations, and brass pins attached to the lower plate pass through these orifices, so that their ends without touching the upper plate are flush with its surface.

The annular spaces between the pins and the plate are filled with selenium. The whole arrangement forms part of a galvanic circuit, and it will be observed that the current can only pass from the plate to the pins through the selenium rings.

It will also be seen that ow

ing to the conical shape of the perforations, the points of closest

approximation between the pins and the plate are on the upper surface. As the effect produced by light upon selenium is chiefly a surface action, this arrangement is found to be of great advantage.

The second typical cell is cylindrical in form for the purpose of being used with a concave reflector instead of with a lens (see fig. 3).

This cell is composed of a large number of metallic disks separated by disks of mica slightly smaller in diameter. The spaces between the brass disks over the mica are filled with selenium, and the alternate brass disks are metallically connected. The arrangement practically consists of a large number of annular selenium cells united in multiple arc.

The mode of applying the selenium is as follows: The cell is heated, and when hot enough a stick of selenium is rubbed over the surface.

In order to acquire conductivity and sensitiveness the selenium must next undergo a process of annealing.

The method we first adopted was the following:

FIG. 4.

The selenium cell was placed with a thermometer in the interior of the cylindrical annealing chamber shown in fig. 4.

This was inserted in a pot of linseed oil, and the latter stood upon glass supports within another similar pot containing linseed oil. The whole arrangement was then placed over a gas stove and heated to a temperature of about 214° C., which was found to be the temperature of maximum conductivity for the selenium used.

This temperature was retained for about twenty-four hours, and the pots, with their contents, were then packed in a box so arranged as to retard radiation of heat.

The selenium took from forty to sixty hours to cool down to the temperature of the air.

A powerful battery current was passed through the selenium during the whole process of heating and cooling, in accordance with our theory that the current exerted a powerful influence in causing a set of the selenium molecules, and in retaining them in position until fixed by crystallization.

A shunted galvanometer was introduced into the circuit for the purpose of observing the changes of conductivity. We subsequently found this tedious process to be unnecessary, as during the course of our experiments we discovered a method of preparing sensitive selenium in a very few minutes.

We.now simply heat the selenium over a gas stove and observe its appearance. When the selenium attains a certain temperature, the beautiful reflecting surface becomes dimmed. A cloudiness extends over it, somewhat like the film of moisture produced by breathing upon a mirror.

This appearance gradually increases and the whole surface is soon seen to be in the metallic, granular, or crystalline condition. The cell may then be taken off the stove and cooled in any suitable way. When the heating process is carried too far, the crystalline selenium is seen to melt.

Our best results have been obtained by heating the selenium until it crystallizes as stated above, and by continuing the heating until signs of melting appear, when the gas is immediately put

out.

The portions that had melted instantly re-crystallize, and the selenium is found upon cooling to be a conductor, and to be sensitive to light. The whole operation occupies only a few minutes. This method has not only the advantage of being expeditious, but it proves that many of the accepted theories on this subject are fallacious.

Early experimenters considered that the selenium must be "cooled from a fused condition with extreme slowness." Later authors agree in believing that the retention of a high temperature -short of the fusing point-and slow cooling are essential, and the belief is also prevalent that crystallization takes place only during the cooling process.

Our new method shows that fusion is unnecessary, that conductivity and sensitiveness can be produced without long heating and slow cooling; and that crystallization takes place during the heating process. We have found that on removing the source of