The operation and advantages of the system are thus stated: The dynamo gives its output at various train speeds; the current passes through a circuit in which are the lamps, and also a storage battery to keep the lamps supplied when the dynamo is not running. In the same circuit there is a switch and also the coils of a solenoid. A train does not always run at tire same speed, therefore there must be some method of regulation when the speed increases above normal. The coils of the field magnet of the dynamo are shunt wound; within the circuit formed by the leads from the field magnet there is a rheostat of the type which is known as a carbon-pile resistance. The bottom plate of this pile resistance is fixed; the upper *991plate is movable, so as to regulate the strength of the resistance. An insulated arm projecting from the upper plate is connected, mechanically, not electrically, with one arm of a pivotally mounted lever. That lever, properly equipped with a retractile spring (and preferably with a dashpot), has its other arm mechanically connected with the plunger of the solenoid. As increased speed of the car increases the current in the main circuit, the plunger of the solenoid is drawn down, and, in consequence, the insulated arm and upper plate are drawn up, thus diminishing the pressure on the pile of carbon disks and increasing the resistance of the field of the dynamo. The specification says:

“The variations of resistance of the pile of carbon disks to slight variations in pressure are so great that a very slight movement of the solenoid will be sufficient to react strongly upon the dynamo and to compensate for the increase in speed. At low train speeds the carbons will be pressed together by tlio action of the retractile spring (owing to the rise of the plunger and the consequent depression of'the arm and top plate), thus enabling the dynamo to build up rapidly.

“It will be apparent that this system provides means whereby a source of energy adapted to supply a current for any desired use may be caused to react upon itself in order to regulate tile amount of energy supplied thereby tending always to bring the output of such energy to normal. The advantages of such self-regulatable system, both general and specific, will be clear without elaboration. Certain advantages to be noted, however, reside in the fact that with the proposed system I am enabled to regulate satisfactorily the output of a shunt wound dynamo, with a field produced by a single circuit only, thereby doing away with the complicated wiring of other systems, wherein the field is a compound one embodying differential series coils and tlie like. This facility of regulation is due, among other things, to the i>eculiar advantages of the variable resistance device which I use, and to the fact that the shunt circuit in which said resistance is introduced is taken off from the main circuit, as shown, at points nearer to the dynamo than the point at which tho solenoid is introduced in the main circuit, so that the field circuit may be said to be independent of the controlling device in the main circuit.”

The claims are:

“1. In combination, a shunt wound generator having serially included in its field circuit a plurality of contacting electrodes and positively acting electromagnetic means controlled by the current output of the generator for varying tho. i>ressure with which said electrodes contact.

“2. In a car axle lighting system, in combination, a dynamo adapted to be driven from a car axle, said dynamo having a single field circuit shunt wound from the main circuit, a variable resistance device included in said field circuit, said device embodying a material wiiose resistance varies with pressure, a controlling device, in series with the main circuit of the dynamo, said controlling device adapted to be operated by variations of current in said main circuit, and connections'from said controlling device for varying the pressure on said resistance device.

3. In a car axle lighting system, in combination, a dynamo adapted to be driven from a car axle, said dynamo having a field circuit shunt wound from the main circuit, a series of carbon disks interposed in said field circuit, a controlling device adapted to be actuated by variations in current in the main circuit of the dynamo and in series therewith, a connection between said controlling device and said disks such that pressure on said disks is decreased as the current in the main circuit increases, and a spring acting upon said disks in opposition to said controlling device to increase the pressure thereon.

“4. In a car axle lighting system, in combination, a dynamo adapted to be driven from a car axle, said dynamo having a field circuit shunt wound from the main circuit, a series of carbon disks included in said field circuit, a solenoid in series with the main circuit of tho dynamo and independent of the *992field circuit, a connection between the plunger of said solenoid and said disks, such that the pressure upon said disks is decreased as the current in the main circuit increases, and a spring adapted to act upon said disks to increase the pressure thereon in opposition to said solenoid.”

Claim 1, like the others, must be limited to a dynamo in a lighting system driven from a car axle. It is quite evident from the patent and also from the record of the prior art that the device we have here is a specific one. Its elements are these: (a) An axle-driven generator, (b) having a shunt field .winding; (c) a rheostat of the kind known as carbon-pile resistance; (d) such resistance being in series with the shunt field winding; (e) control of such resistance automatically by a solenoid; (f) the plunger of which is connected positively and directly with the movable part of the resistance; (g) the solenoid being in series with the current to be regulated. The testimony was talcen by deposition; it is voluminous, and accompanied with many documents from the prior art. The District Judge has discussed the testimony and the prior art at considerable length. Inasmuch as we are generally in accord with him, it will be unnecessary to go over the ground which he has covered.

Defendant opens his brief with the assertion that what is known as- the Moskowitz commercial system of car lighting “completely anticipates” the patent in suit and that the “claims in suit read upon” this Moskowitz structure. Examination of the record, however, shows that,these words, “completely anticipates,” are.used in the sense frequently attributed to them by experts, who when they find a specific combination of enumerated elements in the claim of a patent, and also find the same combination with one change of element in the earlier art, insist that the one combination anticipates the other, on the theory that any one skilled in the art with the earlier combination before him would naturally malee the single change required to produce the later combination. This is a confusion of two defenses, which are quite distinct — anticipation and lack of patentable novelty. This very Moskowitz device differs from that of the patent in element (d) above' recited. In the Moskowitz device the carbon pile resistance is in shunt with the shunt field winding, while in the patented device it is in series with the shunt field winding„ In like manner another device from the prior art has the combination of specific devices claimed by Thomson, (a) to (g) supra, with the single exception of (c); its rheostat being of the type known as “wire” or “step by step” resistance. Now it may be that changes of this sort will sometimes not involve patentable invention; it may be held that there was no invention required in changing one well-known type for another, or*in locating a part differently relatively an electric circuit; but in the face of such changes it is incorrect to 'say that the one combination “completely anticipates” the other.

Judge Hazel evidently reached the same conclusion, for he did not defeat the patent because the prior art showed an anticipation, but because he was satisfied that invention could not be predicated in placing carbon-pile resistance in series with the shunt field winding instead of in shunt therewith, as in the prior Moskowitz structure. He says:

*993“My conclusion is that at the date of filing of the application in suit it did not involve any invention to place the carbon pile in series with the field winding, which was an essential feature of complainant’s car-lighting system.”

That is the fundamental question in the case, and, as so frequently happens with electrical patents, it is a difficult question for a court to decide. On that question the experts are quite positive in support of their respective sides, and there is not much to choose between their arguments. Results accomplished by a change of some element, seemingly slight and unimportant, will sometimes throw much light on a situation otherwise obscure. The District Judge expresses this in his opinion:

“If the proofs substantiate plaintiff’s claim (that the patentee remedied defects and inefficiencies and did) solve the problem of overcoming difficulties in tiie way of efficient car lighting, then beyond doubt his improvement is of great merit and value and the protection of the patent laws should not be withheld from him. * * If it were proven that Thomson succeeded where Moskowitz failed and made railroad car lighting systems more efficient by removing objectionable features, such as fluctuation of the current and constant flickering of the lights, then his patent was not devoid of merit, but was of superior value and entitled to corresponding praise and the protection from appropriations by others notwithstanding Ferrand’s suggestions.”

Instances are not infrequent — there have been several before this court — where what is seemingly a slight change of structure runs quickly into great demand, while alongside of it older systems, at one time selling readily, gradually sink into insignificance. The plaintiff contends that we have here a conspicuous instance of such commercial success; he has set forth the evidence quite fully, and represented its results by a chart. Considering merely the figures of sales and machines in use, the showing is a strong one. The commercial date of the Thomson device may be taken as 1905. In 1900 a device known as the Consolidated, employing a wire rheostat, was on the market. From 1900 to 1903 it sold 400 equipments; from 1903 to 1905, 800 equipments. From that date on their sale dwindled gradually, so that from 1909 to 1913 only 300 equipments were sold, 50 of them only in 1911, 1912, and 1913. In 1903 the Adams & Westlake device, also having a wire rheostat, was put on the market. It seems to have been well received lor automobile cars; 4,000 equipments therefor havi ing been sold. For railroad cars its sale has always been small, of late years gradually diminishing substantially to zero — manufactured only when specially ordered. The Moskowitz commercial type, in use prior to Thomson, began sales about 1903. It was the type which differed from Thomson by having the resistance in shunt with the shunt field instead of in series therewith. Sales began about 1903, they increased slowly to not more than 100 in one year (1908) and then gradually fell off to an insignificant amount, although, of course, some of those already sold were still in use. Contrasted with these are two other types. One is the Safety Company’s device, marketed under the patent to Thomson, of which it is the assignee; the other is the modified Moskowitz, in which the location of carbon rheostat has been changed from shunt to series. The sales of the Safety type, beginning in 1905, rose steadily and rapidly year by year until in 1913 (when the testimony closed) 500 were sold and there were *9942,300 in use. The sales of the modified Moskowitz type began in 1910 and increased with greater rapidity; in 1913 the sales were 1,500 and the total number in use was 3,000. '

It may be noted, also, that sales of equipment of this sort are not made to ordinary individual consumers, where business enterprise and ingenious advertising are sometimes the inducing cause of sales and a purchaser buys something about which he knows little, because he is confronted by an alluring description of its merits every time he takes a seat in a trolley car. The purchasers are mainly railroad corporations, which have well-organized equipment and supply departments, with skillful and experienced engineers to study carefully the merits of every new device they are invited to purchase. The showing is a strong one; the dominant success of the two types, “Safety” and “improved Moskowitz,” has reduced the sales of competing systems to insignificance. Of course there are still thousands of the other types in existence; they last some years and with the aid of repair parts may be kept going, without the expense of scrapping them for something better. But the record does seem to indicate that for new equipment the two types, “Safety” and “improved Moskowitz,” dominate the field. If this were all, one might feel convinced that Thomson’s new combination solved existing and baffling problems and was promptly recognized by those skilled in the art as supplying a long-felt want.

But there are two weak points in the argument. The sales of the type known as “improved Moskowitz” greatly exceeds those of plaintiff’s “Safety” type. That the “improved Moskowitz” type infringes the Thomson patent we have no doubt; it has all the specific elements, (a) to' (g), supra, of the combination, but it has other things besides, independent of and additional to the combination, some of which other devices the Thomson patent does not disclose. This is well shown in a simplified drawing of defendant’s equipment (Exhibit Kennelly’s Simplified Drawing), in which the parts making up the combination of Thomson’s patent are printed in black and the parts not included in such combination are printed in red. The red parts are substantially as numerous as the black. How much of these rapidly increasing sales from 1910 to 1913 are due to the presence of the black parts alone, and how much to the presence of the red parts, we have no means of knowing. ■ •

Moreover, the “Safety” type, the sales of which have also increased since 1905, is not now exclusively the combination of the patent. Added elements were introduced in 1910 and 1912, independent of and additional to the elements, (a) to (g), of the specific claims of the patent. It must be assumed that these added elements in some way improved the equipment or they would not have been introduced. To what extent the sales of “Safety” type equipment are due to the presence of these added elements we have no means of knowing.

To sum up, then, the mere change from a resistance in shunt with the shunt field to a resistance in series therewith was a slight one; so also the mere substitution of the well-known carbon-pile resistance for the equally well-known step by step resistance was a slight one. *995To one informed of the electrical art merely as a court is, such a change seems one within the ordinary skill of the calling; it so seemed to Judge Hazel and so seems to' us. The evidence of the experts as to presence or advance of invention is conflicting and does not clearly establish patentable invention. The test, which not infrequently turns the scale when a court is considering these devices of an obscure art, viz., a measure of success which indicates the solution of a different problem, is not found persuasive. We are not inclined, therefore, to dissent from Judge Hazel’s exhaustive and careful discussion of the questions presented.

The decree is affirmed, with costs.