1. The jurisdiction of this court to review the judgment of the court of appeals is questioned. There is no money in dispute nor anything to which a pecuniary value has been given. Jurisdiction is claimed under the clause of § 8 of the act of February 9, 1893 [27 Stat. at L. 436, chap. 74 U. S. Comp. Stat. 1901, p. 573], which gives an appeal to this court from the final judgment or decree of the court of appeals in cases in which there is drawn in question the validity of 'an authority exercised under the United States.'

By § 483 of the Revised Statutes (U. S. Comp. Stat. 1901, p. 272), the Commissioner of Patents, subject to the approval of the Secretary of the Interior, is empowered to establish from time to time regulations not inconsistent with law, for the conduct of proceedings in the Patent Office. The Commissioner of Patents, exercising the power conferred, established, among other rules of practice, rule 41. It thereby became a rule of procedure, and constituted, in part, the powers of the primary examiner and Commissioner. In other words, it became an authority to those officers, and, necessarily, an authority 'under the United States.' Its validity was and is assailed by the plaintiff in error. We think, therefore, we have jurisdiction, and the motion to dismiss is denied.

2. The issue is well defined between the parties, both as to the right and remedy, in the Patent Office. As to right, petitioner contends that a union by an inventor of process and apparatus claims, which are essentially the same invention, is given by the patent laws, and that rule 41, so far as it takes that right away, is repugnant to those laws, and invalid. As to remedy, that the decision of the primary examiner constituted a final decision upon the case, and petitioner was entitled to an appeal under the patent laws to the board of examiners-in-chief. The latter proposition depends upon the first. Assuming the right in an inventor as expressed in the first proposition, the primary examiner denied the right. True, a distinction can be made between his ruling and one on the merits, if we regard the merits to mean invention, novelty, or the like. But in what situation would an applicant for a patent be? If he yield to the rule he gives up his right of joinder. If he does not yield he will not be heard at all, and may subsequently be regarded as having abandoned his application. Section 4894, Rev. Stat. (U. S. Comp. Stat. 1901, p. 3384). A ruling having such effect must be considered as final and appealable. Whether, however, to the examiners-in-chief or to the Commissioner, and from the latter to the courts, we may postpone answering until we have considered the right of an inventor to join process and apparatus claims in one application.

Section 4886 of the Revised Statutes of the United States (U. S. Comp. Stat. 1901, p. 3382), provides as follows:

'Any person who has invented or discovered any new and useful art, machine, manufacture, or composition of matter, or any new and useful improvements thereof, not known or used by others in this country before his invention or discovery thereof, and not patented or described in any printed publication in this or any foreign country before his invention or discovery thereof, or more than two years prior to his application, and not in public use or on sale in this country for more than two years prior to his application, unless the same is proved to have been abandoned, may, upon payment of the fees required by law, and other due proceeding had, obtain a patent therefor.'

There is nothing in the language of the section which necessarily precludes the joinder of two or more inventions in the same application. But the section does distinguish inventions into arts (processes), machines, manufactures, and compositions of matter, and the earliest construction of the law denied the right of joinder. An exception, however, came to be made in cases of dependent and related inventions.

In Hogg v. Emerson, 6 How. 437, 12 L. ed. 505, it was said:

'The next objection is, that this description in the letters thus considered covers more than one patent and is therefore void.

'There seems to have been no good reason at first unless it be a fiscal one on the part of the government when issuing patents, why more than one in favor of the same inventor should not be embraced in one instrument, like more than one tract of land in one deed or patent for land. Phillips, Patents, 217.

'Each could be set out in separate articles or paragraphs, as different counts for different matters in libels in admiralty or declarations at common law, and the specifications could be made distinct for each, and equally clear.

'But to obtain more revenue, the public officers have generally declined to issue letters for more than one patent described in them. Renouard, Desbrevets D'Invention, 293; Phillips, Patents, 218. The courts have been disposed to acquiesce in the practice as conducive to clearness and certainty. And if letters issue otherwise inadvertently, to hold them, as a general rule, null. But it is a well established exception that patents may be united, if two or more, included in one set of letters, relate to a like subject, or are, in their nature or operation, connected together. Phillips Patents, 218, 219; Barrett v. Hall, 1 Mason, 447, Fed. Cas. No. 1,047; Moody v. Fiske, 2 Mason, 112, Fed. Cas. No. 9,745; Wyeth v. Stone, 1 Story, 273, Fed. Cas. No. 18,107.'

This language would seem to imply that not the statute, but the practice of the Patent Office, required separate applications for inventions, but the cases cited were explicit of the meaning of the statute. Mr. Justice Story, in Wyeth v. Stone, said:

'For, if different inventions might be joined in the same patent for entirely different purposes and objects, the patentee would be at liberty to join as many as he might choose, at his own mere pleasure, in one patent, which seems to be inconsistent with the language of the patent acts, which speak of the thing patented, and not of the things patented, and of a patent for invention, and not of a patent for inventions; and they direct a specific sum to be paid for each patent.'

But he confined the requirement to independent inventions, and his illustrations in dicated that he meant by independent inventions not those which, though distinct, were 'for the same common purpose and auxiliary to the same common end.'

Hogg v. Emerson came to this court again and is reported in 11 How. 587, 13 L. ed. 824. Of one of the objections to the patent the court said:

'It is that the improvement thus described is for more than one invention, and that one set of letters patent for more than one invention is not tolerated by law.

'But grant that such is the result when two or more inventions are entirely separate and independent,—though this is doubtful on principle,—yet it is well settled, in the cases formerly cited, that a patent for more than one invention is not void if they are connected in their design and operation. This last is clearly the case here.'

Many other cases are to the same effect.

Can it be said that a process and an apparatus are inevitably so independent as never to be 'connected in their design and operation?' They may be completely independent. Cochrane v. Deener, 94 U. S. 780, 24 L. ed. 139. But they may be related. They may approach each other so nearly that it will be difficult to distinguish the process from the function of the apparatus. In such case the apparatus would be the dominant thing. But the dominance may be reversed and the process carry an exclusive right, no matter what apparatus may be devised to perform it. There is an illustration in the Telephone Cases, 126 U. S. 1, 31 L. ed. 863, 8 Sup. Ct. Rep. 778. The claim passed upon in those cases was as follows:

'The method of, and apparatus for, transmitting vocal or other sounds telegraphically, as herein described, by causing electrical undulations, similar in form to the vibrations of the air accompanying the said vocal or other sounds, substantially at set forth.'

The claim was held to refer to the art described, and the means of making it useful. The court observed:

'Other inventors may compete with him for the ways of giving effect to the discovery, but the new art he has found will belong to him and those claiming under him during the life of his patent.'

A distinction between the process and the means employed for using it was recognized. It was said:

'The patent for the art does not necessarily involve a patent for the particular means employed for using it. Indeed, the mention of any means, in the specification or descriptive portion of the patent, is only necessary to show that the art can be used; for it is only useful arts—arts which may be used to advantage that can be made the subject of a patent.'

The patent was sustained. It was not attacked because it embraced independent inventions. The fact is not without force. Considering the ability of counsel engaged and the division of the court in opinion, it is a proper inference that no tenable objection to the patent was overlooked.

It is said by Robinson, in his work on Patents, that 'special rules which govern the joinder of arts or processes with each other or with related inventions of a different class, are more stringent in the Patent Office than in the courts.' 2 Robinson, Patents, § 473. And the author deduces the conclusion that under the rules of the Patent Office a process cannot be 'joined with the apparatus that performs it, nor either of these with the product in which they result, unless they are to such an extent inseparable that the existence of some one of them is dependent upon that of the others.' But rule 41 precludes even this.

If there is a divergence of views between the courts and the Patent Office, and the divergence proceeds from a different interpretation of the statute, the views of the courts ought to prevail. If the courts, however, have only recognized and enforced the exercise of a discretion of the Patent Office, the question occurs, What is the extent of such discretion, and can it be expressed and fixed in an inflexible rule such as rule 41? In Bennet v. Fowler, 8 Wall. 445, 19 L. ed. 431, a discretion in the Patent Office was recognized. The question arose upon the validity of two reissued patents for improvements, which 'had been embraced in one, in the original patent.' The court said:

'It may be that if the improvements set forth in both specifications had been incorporated into one patent, the patentee taking care to protect himself as to all his improvements by proper and several claims, it would have been sufficient. It is difficult, perhaps impossible, to lay down any general rule by which to determine when a given invention or improvements shall be embraced in one, two, or more patents. Some discretion must necessarily be left on this subject to the head of the Patent Office. It is often a nice and perplexing question.'

Some discretion is not an unlimited discretion, and if the discretion be not unlimited it is reviewable. In other words, the statute gives the right to join inventions in one application in cases where the inventions are related, and it cannot be denied by a hard and fixed rule which prevents such joinder in all cases. Such a rule is not the exercise of discretion; it is a determination not to hear. No inventor can reach the point of invoking the discretion of the Patent Office. He is notified in advance that he will not be heard, no matter what he might be able to show. His right is denied, therefore; not regulated. Such is the necessary effect of rule 41, as amended.

Without that rule the action of the Patent Office can be accommodated to the character of inventions, and discretion can be exercised, and when exercised, we may say in passing, except in cases of clear abuse, the courts will not review it. But the rule as amended, as we have said, precludes the exercise of any judgment, and compels the separation of claims for a process and claims for its apparatus, however related or connected they may be. And the right denied is substantial. Counsel for petitioner have explained that right by the embarrassments caused by its denial, one of which is that, by disclosing the apparatus in his application for the process, he might lose the right to and a patent for the apparatus; and to sustain that view James v. Campbell, 104 U. S. 356, 26 L. ed. 786, is cited. We are not prepared to admit such consequences nor that James v. Campbell so decides. If the classification of the statute makes a distinction between the different kinds of inventions—between a process and an apparatus—and requires or permits a separate application for each, it would seem to follow irresistibly that an application and patent for one would not preclude an application and patent for the other, and the order of the application could not affect the right which the law confers. James v. Campbell was a case of reissued patent, and by express provision of the statute as to reissued patents no new matter can be introduced in them. In other words, the reissue is to perfect, not to enlarge, the prior patent. Whether the principle of the case applies to related as well as to independent inventions is not clear from its language. The court said:

'Where a new process produces a new substance, the invention of the process is the same as the invention of the substance, and a patent for the one may be reissued so as to include both, as was done in the case of Goodyear's vulcanized rubber patent. But a process, and a machine for applying the process are not necessarily one and the same invention.'

The facts of the case did not call for a more definite ruling. The original patent was for a device for postmarking and canceling postage stamps by a single blow. The reissued patent claimed the act of marking and cancelation, and it was observed by the court:

'The process or act of making a postmark and canceling a postage stamp by a single blow or operation, as a subject of invention, is a totally different thing in the patent law from a stamp constructed for performing that process.'

But without attempting to enlarge the case and extend it to more intimately related inventions, it is enough now to say that there is nothing in the case which decides that if the process had been claimed in an independent application it (the process) would have been adjudged to have been dedicated to the public by the other patent. There is language indicating the contrary. It was said:

'If he [the patentee] was the author of any other invention than that which he specifically describes and claims, though he might have asked to have it patented at the same time and in the same patent, yet if he has not done so, and afterwards desires to secure it, he is bound to make a new and distinct application for that purpose, and make it the subject of a new and different patent.'

The case, however, indicates what embarrassment and peril of rights may be caused by a hard and fixed rule regarding the separation of related inventions. See also Mosler Safe & Lock Co. v. Mosler, 127 U. S. 354, 32 L. ed. 182, 8 Sup. Ct. Rep. 1148, and Miller v. Eagle Mfg. Co. 151 U. S. 186, 38 L. ed. 121, 14 Sup. Ct. Rep. 310.

The Patent Office has not been consistent in its views in regard to the division of inventions. At times convenience of administration has seemed to be of greatest concern; at other times more anxiety has been shown for the rights of inventors. The policy of the office has been denominated that of 'battledore and shuttlecock,' and rule 41 as it now exists was enacted to give simplicity and uniformity to the practice of the office. Its enactment was attempted to be justified by the assumption that the patent laws gave to the office a discretion to permit or deny a joinder of inventions. But, as we have already said, to establish a rule applicable to all cases is not to exercise discretion. Such a rule ignores the differences which invoke discretion, and which can alone justify its exercise, and we are of opinion therefore that rule 41 is an invalid regulation.

3. Having settled the right of appellant, we may now return to the consideration of his remedy. Respondent contends:

'It is fundamental that mandamus will not issue against a public officer, except to compel the performance of some plain, clear, ministerial duty, and will not issue to control his discretion.'

And it is further contended that respondent has acted, and, having acted, cannot be required to refer the case to a lower tribunal in his office. To sustain the contention Holloway v. Whiteley, 4 Wall. 522, 18 L. ed. 335, is cited.

The unity of the inventions claimed by petitioner in the case at bar we may assume. It is not denied by respondent. Petitioner had, therefore, the right to join them in one application. The denial of this right by the primary examiner was a rejection of the application and entitled petitioner to an appeal to the examiners-in-chief, under § 4909 of the Revised Statutes (U. S. Comp. Stat. 1901, p. 3390). That sections provides:

'Every applicant for a patent, . . . any of the claims of which have been twice rejected, . . . may appeal from the decision of the primary examiner to the board of examiners-in-chief; . . .' The § 482 (U. S. Comp. Stat. 1901, p. 272), provides:

'The examiners-in-chief shall be persons of competent legal knowledge and scientific ability, whose duty it shall be, on the written petition of the appellant, to revise and determine upon the validity of the adverse decisions of examiners upon applications for patents, and for reissues of patents, and in interference cases; and, when required by the Commissioner, they shall hear and report upon claims for extensions, and perform such other like duties as he may assign them.'

The procedure on appeal is provided for by the rules of the Patent Office. It is taken by filing a petition praying an appeal with the primary examiner, setting forth the reasons upon which the appeal is based, and it is made the duty of the examiner, five days before the date of hearing, to funish the appellate tribunal and the appellant with a statement of the grounds of his decision. A petition praying an appeal was filed, but the primary examiner refused to answer the appeal, and the defendant in error also refused to direct him to answer it. It is manifest that if an appeal cannot be compelled from the decision of the primary examiner, an applicant is entirely without remedy. And respondent has asserted that extreme. In Ex parte Frasch, 20 App. D. C. 298, the court of appeals of the District of Columbia was persuaded that an appeal was not the proper remedy. In the case at bar it is contended that mandamus is not the proper one. One or the other must be. A suggestion made is that the inventor must await a decision on the merits, meaning by merits 'lack of invention, novelty, or utility,' as expressed in rule 133. But after waiting he would encounter the arbitrary requirement of rule 41. Besides, what would there be to review if the order of the primary examiner were complied with and the claims put into separate applications? There are some observations in Holloway v. Whiteley, which may be quoted. Whitely claimed to be the assignee of a patent, and filed an application for a reissue. The Commissioner declined to entertain it on the ground that Whitely was only assignee of an interest, and not of the entirc patent. He also declined to allow an appeal to be taken from his decision. The supreme court of the District of Columbia awarded a peremptory writ of mandamus commanding the Commissioner to refer the application to the proper examiner, or otherwise examine or cause it to be examined according to law. Error was prosecuted to this court. Under the act of 1836 [5 Stat. at L. p. 117, chap. 357], it was provided that if the Commissioner decided adversely to an applicant for a patent an appeal could be taken to the board of examiners, and by the act of 1837 [5 Stat. at L. 191, chap. 45], that remedy was given to an applicant for a reissue of a patent, and the question in the case was whether that remedy should have been pursued. In other words, whether the remedy was by appeal or mandamus. It was decided that appeal was the remedy. Singularly enough, the Commissioner, in answer to the rule, took the position that the application was not before him because it had not been filed. The court said if that were so 'mandamus would clearly lie to compel the Commissioner to receive it. It was his first duty to receive the application, whatever he might do subsequently. Without this initial step there could be no examination, and, indeed, no rightful knowledge of the subject on his part. Examination and the exercise of judgment, with their proper fruit, were to follow, and they did follow.'

And so the exercise of judgment might follow a hearing of the application under review. It was the duty of the primary examiner to accord a hearing or, refusing to do so, to grant an appeal. It was the duty of the Commissioner to compel the appeal. The Commissioner of Patents is primarily charged with granting and issuing patents. Applications for patents are made to him (§ 4888, Revised Statutes, U. S. Comp. Stat. 1901, p. 3383), and his superintendence should be exercised to secure the rights which the statutes confer on inventors. The first of those rights is a hearing. If that be denied other rights cannot accrue.

The Commissioner justifies his decision by the rules of the Patent Office and a long practice under them. If there is inconsistency between the rules and statute, the latter must prevail. But the primary examiner did not follow the rules. The rules provide that if appeal be regular in form (italics ours) he shall, within five days of the filing thereof, furnish the examiners-in-chief with a written statement of the grounds of his decision on all of the points involved in the appeal, with copies of the rejected claims and with the references applicable thereto. If he decide that the appeal is not regular in form, a petition from such decision may be made directly to the Commissioner. The regularity of the appeal in form is not questioned in the case at bar, and it was the duty of the examiner to answer the appeal by furnishing the examiners-in-chief the statement provided for in rule 135. A petition to the Commissioner was not necessary except to make the examiner to perform his duty.

4. We do not think that petitioner was estopped from insisting upon his application by proceeding with the interference with Duncan after the examiner's letter of December 15, 1899. It would be pressing mere order of procedure and the convenience of the Patent Office too far to give them such result under the circumstances.

The judgment of the Court of Appeals is therefore reversed, with directions to reverse that of the Supreme Court, and direct the Supreme Court to grant the writ of mandamus as prayed for.