TAFT, Circuit Judge
(after stating the facts as above). The catalogue of the defendant shows that it is offering for sale to the public without restriction the switch and trolley to be used as part of the equipment of an electric street railway. Defendant has not shown, and we infer from the evidence that it cannot be shown, that either the switch or trolley and harp can be used in an electric railway except in the combinations described and claimed in the two patents here in suit. The third claim of patent No. 424,695 is for a combination of an overhead wire for receiving an underneath contact and a switch plate attached to the wire in about the same horizontal plane as the wire. The description of the article sold by the defendant in its catalogue is that of an “under-running adjustable switch,” and it is said to make a perfectly straight under-running approach for the trolley wheel. It is apparent that the switch plate has no practical utility except in such an arrangement of parts as that stated in the third claim. The same thing is true of the fourth and eleventh claims of that patent. The sixth claim of patent No. 495,443 is a combination in an electric railway of (1) a suitable track, (2) a supply conductor suspended above the track, (3) a car provided with (4) a swinging arm carrying (5) a contact device in its outer extremity, and (6) means for imparting upward pressure to the outer portion of the arm and contact, to hold the latter in continuous working relation with the underside of the supply conductor. The evidence sufficiently shows that neither the trolley nor the harp is adapted to be used on electric street railways except in the above combination. Purchasers buy articles for practical use, and would only buy the switch [721]*721and trolley, therefore, for nse in complainant’s patented combinations. One is legally presumed to intend the natural consequences of his act. Hence the defendant, in offering the switch and trolley for sale to the general public, may be reasonably held to intend that they should be used in combinations in an electric railway covered by the claims of complainant’s patents.
It is well settled that where one makes and sells one element of a combination covered by a patent with the intention and for the purpose of bringing about its use in such a combination he is guilty of contributory infringement and is equally liable to the patentee with him who in fact organizes the complete combination. The leading case on the subject is Wallace v. Holmes, 9 Blatchf. 65, 29 Fed. Cas. 79. It was cited with approval in Tie Co. v. Simmons, 106 U. S. 89, 1 Sup. Ct. 52, and the same doctrine was applied and extended by this court in Heaton-Peninsular Button-Fastener Co. v. Eureka Specialty Co., 77 Fed. 288, to a case where the article sold was not even an element of the patented combination, but was an article the use of which in connection with patented combination was a violation of the conditions of a license, and destroyed the protection the license would otherwise have afforded. The cases in the circuit courts where the same general principle has been applied are legion. We cite a few of them: Bowker v. Dows, 3 Fed. Cas. 1070; Richardson v. Hoyes, 20 Fed. Cas. 723; Travers v. Beyer, 26 Fed. 450; Willis v. McCullen, 29 Fed. 641; Alabastine Co. v. Payne, 27 Fed. 559; Celluloid Manuf’g Co. v. American Zylonite Co., 30 Fed. 437. The contention of the counsel for the defendant, if we understand it, is that the effect of the decision of the supreme court in the case of Morgan Envelope Co. v. Albany Perforated Wrapping Paper Co., 152 U. S. 425, 14 Sup. Ct. 627, is to do away altogether with the doctrine of contributory infringement. If this is a proper reading of the judgment of the supreme court in that case it was a somewhat startling departure from previously understood principles in the law of torts. An infringement of a patent is a tort analogous to trespass or trespass on the case. From the earliest times, all who take part in a trespass, either by actual participation therein or by aiding and abetting it, have been held to be jointly and severally liable for the injury inflicted. There must be some concert of action between him who does the injury and him who is charged with aiding and abetting, before the latter can be held liable. When that is present, however, the joint liability of both the principal and the accomplice has been invariably enforced. If this healthful rule is not to apply to trespass upon patent property, then, indeed, the protection which is promised by the constitution and laws of the United States to inventors is a poor sham. Many of the most valuable patents are combinations' of nonpatentable elements, and the only effective mode of preventing; infringement is by suits against those who, by furnú' ig the parts which distinguish the combination, make it possibl ^r others to assemble and use the combination, and who, by aUertisement of the sale of such parts and otherwise, intentionally solicit and promote such invasions of the patentee’s rights. The Morgan Envelope Company Case, so far from departing from the doctrine of contributorv [722]*722infringement, expressly recognizes it and the authorities in which it is announced and enforced. The court held, however, that the doctrine of contributory infringement could not apply in a case in which the claim of the patent embraces a machine and something to be treated thereby which is perishable in its nature, and the alleged infringer furnishes the latter element to be used with the machine and consumed. Whether this holding is really a limitation upon the doctrine of contributory infringement, or is to be regarded only as in effect deciding that such combinations are impossible in a patent, may admit of question.
. In Heaton-Peninsular Button-Fastener Co. v. Eureka Specialty Co., 77 Fed. 288, this court was obliged to consider with much care the Morgan Envelope Case. After quoting at length from Mr. Justice Brown’s opinion, Judge Burton, speaking for this court, said:
“It is true that Mr. Justice Brown, in discussing- the question involved in that ease, assumes that a combination of the machine for delivering the paper with the paper to be delivered was valid. But, before he finishes the argument, he shows that the assumption leads to an absurdity; and the decision, in effect, is that form of argument known as the ‘reductio ad absurdum,’ establishing that his original assumption was not founded in reason. The illustration of the result of such a combination shows that what the court was deciding was that a combination of the machine with an unpatentable paper or material to be operated upon by the machine was an impossibility, and the sale of the machine involved and implied the right of use of the material with which it was to be combined; and this is shown by the case of Wilson v. Simpson, 9 How. 109, which is cited by Mr. Justice Brown as a case sustaining his conclusion. * * * Thus, with respect to the paper holder, the supreme court in effect held that the sale of the paper fastener with the paper in it contained the implication of a right to renew the paper when that paper sold should be exhausted, and did not require the purchase of the paper from the original patentee, the paper itself not being patented.”
We are very clearly of opinion that the Morgan Envelope Company Case does not affect the question of contributory infringement before us.
It is said that no concert of action by defendant with any one for the purpose of accomplishing an infringement of complainant’s patent rights is shown. As already stated, it does appear that defendant is offering for sale articles that can only be used in combinations •covered by complainant’s claims. This is an effort to secure a con•cert of action by which the combinations of complainant’s patents may be assembled. If successful, infringement will follow; hence the preliminary steps which are intentionally taken to bring about the injury may be enjoined.
In considering the same point in Wallace v. Holmes, 29 Fed. Cas. 79, Judge Woodruff said:
“Here the actual concert with the other is a certain inference from the nature of the case, and the distinct efforts of the defendants to bring the burner in question into use, which can only be by adding the chimney. The defendants have not, perhaps, made an actual prearrangement with any particular person to supply the chimney to be added to the burner; but every sale they make is a proposal to the purchaser to do this, and his purchase is a consent with the defendants that he will do it, or cause it to be done.”
How, it is suggested that defendant had the right to sell parts to be used in complainant’s combinations to the licensees of complainant, and to those who, having once bought the articles of the com[723]*723bination from the complainant, it is said, have the implied right to repair and renew parts worn out with use. It being established that defendant is offering for sale articles, intending them to be used in combinations which, if unlicensed by complainant, would be infringements of complainant’s patents, we think that it is the duty of the defendant to see to it that such combinations which it is intentionally inducing and promoting shall be confined to those which may be lawfully organized. We are unable to see why any different rule should be applied in such a case from that applicable to a case in which a defendant makes a patented machine to order. He may make such a machine upon the order of the patentee or a licensee, but not otherwise. Hpon him is the peril of a mistake as to the lawful authority of him who gives the order. So, he may knowingly assist in assembling, repairing, and renewing a patented combination by furnishing some of the needed parts; but, when he does so, he must ascertain, if he would escape liability for infringement, that the one buying and using them for this purpose has a license, express or implied, to do so. What we have said has application only to cases in which it affirmatively appears that the alleged infringer is offering the parts with the purpose that they shall be used in the patented combination. We have found that it does so appear here, and is a matter of certain inference from the circumstance that the parts sold can only be used in the combinations patented. Of course, such an inference could not be drawn had the articles, the sale or offering of which was the subject of complaint, been adapted to other uses than in the patented combination. In the latter case the intention to assist in infringement must be otherwise shown affirmatively, and cannot be inferred from the mere fact that the articles are in fact used in the patented combinations or may be so used. If defendant avers that he is selling to express or implied licensees of the patentee, the injunction should be granted in such form as shall permit the continuance of these lawful sales. In the case at bar the circuit court offered thus to modify the order, but the defendant declined to apply for the modification. It is hardly in a position now to assign for error the failure of the court to insert such a modification.
The conclusion we have reached as to contributory infringement is supported by the decision of the circuit court of appeals of the Second circuit on similar facts in Thomson-Houston Electric Go. v. Kelsey Electric Railway Specialty Co., 22 C. C. A. 1, 75 Fed. 1005, affirming an order of injunction by Judge Townsend in the circuit court. Judge Wallace dissented, but we think the reasoning of the majority more satisfactory. The learned dissenting judge likened the case to one in which the seller of wire or rails should be enjoined from selling them, because the articles might be used in any unlawful combination on an electric railway. Such articles could, of course, be used in so many lawful ways other than in the combinations of the patent that it would be quite unlikely that the sellers could have an intention to promote infringement of particular patents in their sale, and it would be most difficult to prove such intention. But, where the article can only be used in a patented combination, the inference of the intention of the maker and seller [724]*724is certain, and the right of the patentee to injunction ought, we think, to be equally certain.
We come now to the second assignment of error in these cases, namely, that the paténts relied on are void because the inventions covered by them were patented in prior patents to the same patentee. The circumstances make this question different as it arises on the two patents,- and we shall first consider whether patent No. 424.695 is void by reason of the prior issue of patent No. 397,451. We are clearly of opinion that it is not. The application for No. 424.695 was filed more than a year before that for the patent No. 397,451. The inventor expressly states in his specifications in No. 397,451 that his invention is an improvement on the switch and trolley devices and combinations shown in the application which subsequently resulted in patent No. 424,695. An examination of the drawings and specifications leaves no doubt that this statement is true. With respect to the switch or switch plate, in the patent of later application and earlier issue the inventor added to the simple switch box center ribs and side contractions of the extremities of the passageways for the purpose of more certainly directing the trolley wheel in entering and leaving the switch in the way in which it should go. With respect to the trolley arm, the improvement consisted in making it of adjustable length. The claims, all of them,, include and refer to one or the other of these.improvements. Now, it is not material to this discussion whether these improvements are patentable or not. They are expressly claimed as improvements,, and no attempt is made by the patentee to cover anything but them. If inventions at all, then they are separable from the old switch and trolley combinations, and, if they are not inventions, the patents are-void, and cover nothing. Since the case of O’Reilly v. Morse, 15 How. 61, 121, 133, it has been well settled that a patent may issue-for an improvement on an earlier invention either to the original inventor or to a stranger. Of course, no one can use the improvement without right or license to use the fundamental invention; but, on the other hand, the right to use the original invention does not confer the right to use the improvement without license from the tributary inventor. We do not understand this general doctrine to be denied, but it is said that if, by some chance, the application for the fundamental patent is delayed in its course through the patent office until a patent on the avowed improvement has issued, then the-patent on the fundamental invention is void. In cases where the delay in the issuing of the patent for the main invention cannot be-charged to the laches or fraud of the patentee, such a rule would be a hard one; and unless it is required by the express words of the-statute, or by the express holding of the supreme court, we should be inclined, if possible, to avoid declaring it to exist. The contention of counsel for the defendant in this behalf, instead of having the support of the authority of the supreme court, is in the teeth of two decisions of that tribunal.
In Suffolk Co. v. Hayden, 3 Wall. 315, an inventor, applied for a patent on improvements in the interior arrangements of an elongated trunk previously in use for cleaning cotton. While this application [725]*725was pending, lie applied for an improvement on the form of the trunk, i. e. its external form; and in his second application he described the improvement in the interior arrangements of the trunk, without making any claim as to it. A patent issued on the second application before one was issued on the first, and the point at issue before the court was whether the prior patent on the later application describing but not claiming the improvement for which a patent had first been applied for, avoided the later patent on the earlier application for this improvement, and it was held that it did not. In disposing of the contention that the second patent was void, the supreme court said:
“The first point of the plaintiff in error is that the description in the patent of March, of the improvement patented the December following, and on which the present suit is brought, and omission to claim it in such earlier patent, operated as an abandonment or dedication of it to the public, and that for this reason the subsequent patent of 1st December was void. But the answer to this ground of defense is that it appeared that Hayden, the patentee, had pending before the commissioner of patents an application for this same improvement at the time he described it in the specification of the 17th of March, which was doubtless the reason for not claiming it in this patent. The description in no sense affected this application thus pending before the commissioner,' and, while it remained before him, repelled any inference of abandonment or dedication from the omission to again claim it.”
The same question arose again in the Barbed-Wire Patent Case, 143 U. S. 280, 12 Sup. Ct. 443, 450, and is stated and disposed of in Mr. Justice Brown’s opinion in that case as follows:
“The application for the patent in suit was filed October 27, 3873, though the patent was not issued until November 24, 1874. Subsequent to the application for this patent, and on March 14, 1874, G-lidden filed an application for an improvement in wire stretchers for fences, upon which a patent was issued May 12, 1874. It is not perceived how this patent could affect in any way the pending application for the later patent. The patentee abandoned nothing he had claimed before, but sought, as an improvement upon the former, to claim a slotted tube midway between the posts, in which was put a coil spring to spread the wires, and automatically tighten them, and keep them at the proper tension, as against expansion by heat and contraction by cold. If the later application had covered the same invention as the prior application for the November patent, the later patent might have been void, under our ruling in Suffolk Co. v. Hayden, 3 Wall. 315; but this claim was for a combination of wires with the slotted tube, containing a coiled spring, and perched upon a post. In this application he makes no mention whatever of barbs as a feature of his claim, although in describing his invention he mentions the use of two wires, provided at suitable intervals with spurs coiled around them, and which are spread apart between the coils to keep the latter from moving longitudinally upon the wires. But he says of these spurs: T do not claim to have originated the devices known as “spurs” or “prongs” on the wires, they having been used before, but confine myself to the means for holding the spurs at proper intervals on the wires, and to th'e means for attaining a uniform tension of the wires, as claimed.’ This disclaimer, it will be observed, is of spurs or prongs generally (not of the coiled barb either alone or in combination with the twisted wires), and is made with reference to that application only. It is true that this patent was subsequently reissued with a broadly expanded claim for a combination with a fence wire of a barb formed of a short piece of pointed wire, secured in place upon the fence wire, by coiling between its ends, forming two projecting points; but this reissue was held to be unwarranted and void in Manufacturing Co. v. Fuchs, 16 Fed. 661, 667. This attempted reissue, however, did not in any way affect his original application, which stood upon its own merits, and, after being rejected and amended three times, was finally passed with a claim substantially identical with the first claim of the original [726]*726application, and the patent granted. In legal effect, this was a prior patent, since the date of the application, and. not the date of the patent, controls in determining the legal effect to he given to two patents issued, at different dates to the same inventor, and the order in which they axe to he considered. In any event, the reissue in 1876 of one patent would not affect another patent granted in 1874.”
It will be observed that the case at bar is stronger than either of the cases cited, because in the application for the patent on the improvement not only is no claim made for the main invention, but the applicant expressly states that he has an application pending for the main invention necessarily described in describing the improvement, but not claimed, and thus shows beyond peradventure that he has no intention of abandoning or dedicating to the public his main invention. The authority and effect of these two cases counsel for defendant seeks to meet by the claim that the ground upon which he contends that the second patents in the cases before the court must be void was not presented to the court, and was not considered by it. He says that the second patent for the main invention is void, not because it was dedicated to the public in the first patent, but because the effect of the second patent is to extend the monopoly of the first patent beyond the statutory period. The argument runs thus: The monopoly of the first patent, the one for the improvement, can legally be of no longer duration than 17 years. But the improvement cannot be used except as applied to the main invention, and, as the later patent (the one for the main invention) does not expire until some time after the expiration of the patent for the improvement, the monopoly of the improvement patent is in fact extended until the expiration of the later patent (the one for the main invention), and so exceeds the statutory 17 years. It may be conceded that it is doubtful ■whether the case of Suffolk Co. v. Hayden on its facts presented the question thus raised by counsel, because the two patents there under consideration were improvements on different parts of the same machine, and it did not appear that the earlier patent might not have been used without also using the later patent. And, possibly, the same thing is true of the Barbed-Wire Patent Case. However this may be, we cannot yield to the argument based on such a distinction, because we cannot accept its minor premise, to wit, that the later granting of the patent for the main invention extends the monopoly of the earlier improvement patent. The patent for the improvement expires in 17 years. After that any one may use the improvement without infringing the patent issued upon it. If he uses the improvement without a license to use the main invention he is liable for the infringement, not of the patent for the improvement, but of the patent for the main invention; and, in estimating the damages for the same, the value of the main invention, and not that of the improvement, would be the basis for estimating the damages.
It can make no difference in considering this question whether the patent for the improvement issues to the patentee of the main invention or to another. The right of the public to use the improvement when the patent on it expires is exactly the same, whether the patentees of the two inventions are the same or not, because in each case [727]*727the improvement can only be used with the license of the patentee of the main invention. If the patentee of the improvement is a stranger to the main invention, it is manifest that he can derive no benefit from the limitation upon the use of his invention after his patent expires, because of the patent on the main invention. Why, then, does he derive an advantage if he happens to own the main patent? The only advantage conferred by the issue of the patent for the main invention is the legitimate monopoly for the statutory period of that invention, and of no other. Did the personality of the owner of two different patents affect the validity of either, then the anomalous result would follow that the owner of one patent would avoid it by acquiring ownership of another. According to the argument of counsel, the patentees of the earlier improvement patent and of the later main patent being the same person, the main patent is void. Let us suppose that they were different persons, but that, by mesne assignments, the two patents became the property of one person; the effect upon the public use of the improvement patent is exactly the same as if the now owner had been the inventor and patentee of both. Does the unity of title avoid the main patent, which was valid before? It must do so if the argument of counsel for defendant in this behalf is to be followed, for the effect of the unity of title is “to extend the monopoly” of the earlier improvement patent in the sense in which counsel uses that phrase. To our minds, this conclusion is reductio ad absurdum. The fact that a patent for an improvement may expire before the patent for the main invention is the result of several circumstances,—one that a patent may be taken out for an improvement on a patentable invention, another that there is no limitation by statute upon the time within which a patent may issue upon an application after it is filed, provided the applicant is not guilty of violating the two years’ restriction imposed by the statute, and a third that the course of an application for a generic or broad invention may legitimately take longer in its course through the patent office than a comparatively unimportant improvement on that invention.
The case upon which counsel for defendant chiefly relies to support his argument is Miller v. Manufacturing Go., 151 TJ. S. 186, 14 Sup. Ct. 810, where it was held that when two patents issued to the same patentee for the same invention the second patent was void for the reason that the new or later patent would prolong the monopoly, beyond the period allowed by law. As pointed out by Mr. Justice Jackson, this was not a new doctrine and found support in a number of earlier cases cited by him. The patents under consideration in that case were for a peculiar form of spring, which, when connecting the plow beams of a plow with the upright portion of the axle, assisted the operator in lifting the plow beams when above the level, and in depressing them when below it, and the patent first issued covered the spring thus used. The spring thus used had the additional function of increasing its lifting force and action the higher the beams were raised. The second patent was taken out to cover this feature of the spring. In effect, the only difference between the patents was that the earlier patent covered a spring with both a de[728]*728pressing and a lifting function, while the second patent covered exactly the same spring with a lifting function increasing in force as the spring rose above the level. Now, the function of the spring described in the second patent was necessarily present in the spring as it was shown in the earlier patent, for no change in form of the spring was suggested in either patent by which it might serve the purpose and claim of the first patent, and discharge the functions therein described, without also at the same time discharging the function which was supposed to distinguish the second patent and its claim from the first. In other words, the division of the original application into two patents was nothing more than an attempt to patent, as two separate inventions, the same device when discharging different functions. The opinion of Mr. Justice Jackson distinctly recognizes that “where the second patent covers matters described in the prior patent, essentially distinct and separable from the invention covered thereby and claims made thereunder, its validity may be sustained,” and “that an inventor may make a new improvement on his own invention of a patentable character, for which he may obtain a separate patent,” and “that a later patent may be granted where the invention is clearly distinct from and independent of one previously patented.” It is, of course, true that, if an improvement is an invention separable from the generic invention, the latter is an invention distinct and separable from the improvement. Hence it follows from the propositions above quoted from the opinion in the case of Miller v. Manufacturing Co. that a patent for a generic invention is not avoided by the fact that a prior invention has been issued for a distinct improvement on that invention, provided always that the language of the application for the first patent and the circumstances of filing it are not such as to dedicate the generic invention to the public. The case of Miller v. Manufacturing Co., therefore, instead of sustaining the claim made for it, is distinctly an authority to the contrary. We are of opinion that patent No. 424,695 is not rendered void by patent No. 397,451. We are strongly fortified in this conclusion by the most satisfactory opinion of Judge Wallace, speaking for the circuit court of appeals of the Second circuit, upon exactly the same question,— Thomson-Houston Electric Co. v. Elmira & H. Ry. Co., 18 C. C. A. 145, 71 Fed. 396, affirming the decree of Judge 'Coxe in the circuit court in the same case, 69 Fed. 257.
We come now to the question whether patent No. 495,443 is rendered void by the prior issue of patent No. 424,695. This presents much more difficulty than the question just disposed of. In this case the drawings and specifications of the two patents are substantially alike, and show a car, a track, a post on top of the car, a swinging and hinged arm pivoted in the post, with a contact wheel at its outer end. A spring is secured to the lower end of the swinging arm, and to the spring is attached a weight which works in suitable vertical grooves down through the roof to the front platform, within reach of the driver. The spring and weight maintain the contact of the outer end of the swinging arm with the overhead conductor. Switches in the overhead conductor are maintained immediately over the point in the track where track switches occur. The trolley post and [729]*729.arm are in such positions and of such size that the point of contact •of the outer end of the arm and the overhead conductor is back of the front wheels of the car. This is for the purpose of imparting to the trolley wheel, as it enters the switch in the overhead conductor, the ■direction already taken by the front wheels of the car in entering the switches upon the track. The spring and weight working in vertical grooves are intended to keep the trolley arm in the vertical plane of the longitudinal center of the car, and thus to make its contact wheel more certain to follow in the overhead switch the direction of the car as it turns into a track switch. It is shown by the -evidence that the inventor first used in his combination a spring attached to the top of the car to secure contact, and then a spring attached to the foot of the trolley post, and finally the spring and weight arrangement shown in the drawings of the patents. The ■claims of the second patent in question are for the broad claims of a combination in an electric railway of a car, a track, an overhead •conductor, a post and swinging hinged arm on the car, and a tension spring for maintaining contact between the arm and the conductor; and the language of the specifications shows that it was the intention of the inventor to make this cover the generic invention. The claims of the first patent that embrace the whole combination include the weight as part of the means for maintaining upward pressure of the arm against the conductor. There are five claims of the first patent that cover the whole combination, and include a spring •or weight to perform the function of keeping the trolley arm in the center line of the car. Now, this same spring and weight in the •drawings discharge the function of maintaining the upward pressure •of the swinging arm. The contention of the counsel for the complainant is that the first patent was a patent for the special and improved form of the invention, including the spring and weight with their upward pressure and centralizing tendency, and that the second patent, though using the same drawings and specifications, shows by the language of the latter and its claims that it was intended to cover, and did cover, a combination with a spring without a weight in such a position that it need only discharge the function of maintaining the upward pressure of the arm without the centralizing tendency, and that the modification of the drawings and specifications to show such a tension spring is only the work of a skilled mechanic. To the ■objection that the last five claims of the earlier patent are exactly the same as the broad claims of the later patent with the mere statement of a necessary centralizing function of the same spring always present in it, it is answered that the second patent was intended to ■cover springs that had no centralizing tendency, and that the use of the function in describing the spring, therefore, is a limitation of the claim showing it to be a special form of spring. It is argued, therefore, that, as the claims of the first patent do not cover any of the broad claims of the second patent based on a simpler combination of parts than that shown in the drawings, the second patent may be held to be a separate generic invention, while the earlier patent is merely for improved forms of the same invention entitled to a sepa[730]*730rate patent. We think the ease on these two patents much nearer Miller v. Manufacturing Co. than the case on the two patents already considered, but we are unwilling upon an appeal from a preliminary injunction heard upon affidavits, and without a full review of the art, and without a fuller argument and closer consideration of the claims- and specifications, to decide the question mooted. The questions are whether, in determining the separability of the inventions, we may consult evidence dehors the record, and whether, in considering the-gist of the second patent, we may supply such variations in the form of the combination shown in the drawings and specifications suggested by mere mechanical skill as would make it one not covered by the first patent, but a simpler and more generic form, and whether the-claims of the second patent are limited to the devices actually shown by the words “substantially as described.” The main invention is confessedly a meritorious one, and we shall be loath to declare it void, because of a mistake in the patent office proceedings, if we can, by any reasonable construction of the patents, do otherwise. The second patent was sustained as valid after a full hearing on the merits,, before so good and experienced a patent judge as Judge Townsend,, of Connecticut. Thomson-Houston Electric Co. v. Winchester Ave. Ry. Co., 71 Fed. 192. This certainly justified the court below in assuming the validity of the patent on a motion.for preliminary injunction. It is well settled that, on appeals like this, this court will ordinarily look into the case merely to see whether the discretion of the court below in issuing or withholding the order of preliminary injunction has been abused; and that only in exceptional cases, in-which a controlling question of law may be as fully and fairly considered as upon final hearing, and the court has no doubt upon it, will-it finally dispose of the injunction and the case on a hearing like this. Duplex Printing-Press Co. v. Campbell Printing-Press & Manuf'g Co., 16 C. C. A. 220, 69 Fed. 250; Mayor, etc., v. Africa, 23 C. C. A. 252, 77 Fed. 501. Questions on appeals of this character are ordinarily to be treated in this court from the standpoint from which they were viewed by the circuit court, and the decision on the merits by a circuit court of another circuit sustaining the patent is therefore usually of controlling weight here, as it should be in the court below.
In an appeal from a preliminary injunction on this same patent, the circuit court of appeals of the Second circuit affirmed the order without examining or deciding the validity of the patent, justifying its course in this regard by the statement that in another cause, heard upon a voluminous record by Judge Townsend, the patent had been sustained. Thomson-Houston Electric Co. v. Kelsey Electric Railway Specialty Co., 22 C. C. A. 1, 75 Fed. 1005. The orders of preliminary injunctions appealed from in both cases are accordingly affirmed, at the costs of the appellant, the defendant below.