L. HAND, Circuit Judge.

Judge Byers has stated the facts in this cause with a completeness which makes unnecessary any restatement by us; his opinion is reported, and we shall discuss the points raised upon the appeal, assuming an acquaintance with its contents. Moreover, we shall confine our discussion to whether the claims in suit are too broad, and whether the patent was forfeited by prior use. The plaintiff says that the claims are invalid because they are "functional," and so Judge Byers held. In form that is not true, for they speak in definite terms, referable to the specifications. However, when we look to the specifications, we find all the terms defined in whole or in part only by what they will do. The first of the five is a "resinaceous material," which is to be a resin or wax, or a mixture of both, that shall have nine so-called "characteristics." The first "characteristic" is that it shall "repel" water; the second, that it shall "resist" acid; the fourth, that it shall not oxidize "readily"; the fifth, that it shall be so friable that, when mixed with the "filler," a "very fine powder" will result; the seventh, that it shall "soften" within given limits; the eighth, that it shall liquefy when heated only a little above those limits; the ninth, "that it may be relied upon to supply the foregoing characteristics in a uniform manner." We have omitted the third and sixth "characteristics" because it may be doubtful whether these should be called "functional." Thus, the specifications leave it to the art to select those resins or waxes, or their mixtures, which in practice will perform all these seven duties; and the record is silent as to what "resinaceous materials" will do so, and what will not.

The "filler" is the next ingredient, and its first "characteristic" is that, when mixed in the prescribed proportion with the "resinaceous material" "the layer of powder retains substantially its original form and position without spreading or flowing over the surface of the plate," when the resin is heated to its softening point. In addition, it must itself repel water and resist acid so that the etching solutions shall not penetrate it, for it makes up more than half the body of the patented powder. Furthermore, the specifications go on to say that "these characteristics should be found when the filler material is pulverized to such an extent that it all passes through a one hundred mesh screen and the greater proportion passes through a two hundred mesh screen" (page 3, col. 1, lines 9-13). This language we understand to mean that a suitable "filler" will result, if "material" is ground to that degree of fineness; but we do not understand it to mean that all "fillers" must be as finely ground; and, if so, there may be suitable "fillers" of coarser grain. Thus the definition is in terms of "function."

There are three "secondary ingredients," all of which are necessary to all the claims in suit except "Claim Nine," and some one of the three is necessary to that claim. Moreover, since "Claim Nine" does not say which of these three is necessary, it must cover a combination of the "primary ingredients" with any one of the three "secondary ones, and it must be valid for all three, if it is to be valid at all. The first of the three is the "adhesion powder," which, although it need not be "gummy," should adhere to the sides of the "channels" as a powder will adhere because of an "electrostatic charge." Accordingly, "adhesion powders" are described only as "light, fluffy powders which adhere readily to dry surfaces against which they are brushed," like those which stick electrostatically as contrasted with those "which adhere by reason of a gummy or tacky texture." This is even more "functional" than the definition of the "filler."

The specified "lubricant powder" is talc, but it need not be that. It is added to the "acid resist powder" because, although the powder must coat the sides of the channels, their bottoms must be freed from it, else subsequent etchings will not deepen them. The "lubricant" must therefore contribute to the whole powder the capacity of being brushed away from the bottoms of the channels. This the specifications describe as follows: "By the addition thereto of a lubricant powder of the desired properties, this tendency is satisfactorily counteracted, so that when the powder is brushed on * * * it adheres firmly to the vertical sides or shoulders, * * * and * * does not adhere to the flat or horizontal areas * * * thus leaving such flat areas exposed to the free action of the etching solution." Disregarding the fifth ingredient — the "coloring material" — it therefore appears that the first ingredient is "functional" in greater part, and that the last three are wholly so.

A vast deal has been written about "functional" claims and it must be owned that much of it cannot be reconciled with the rest; yet, in spite of the fact that the latest decisions of the Supreme Court have declared with great strictness against them, we do not think that form is inevitably determinative of validity. The question always is whether such claims extend the monopoly beyond the "invention," and that is not to be determined so simply. An applicant for a patent must make "a written description" of "his invention or discovery" "in such full, clear, concise, and exact terms as to enable any person skilled in the art * * * to make, construct, compound, and use the same;" and he must "particularly point out and distinctly claim the part, improvement, or combination which he claims as his invention or discovery." If the claims were limited to the "concise and exact terms" in which the specifications ordinarily describe a single example of the invention, few, if any, patents, would have value, for there are generally many variants well-known to the art, which will at once suggest themselves as practicable substitutes for the specific details of the machine or process so disclosed. It is the office of the claims to cover these, and it is usually exceedingly difficult, and sometimes impossible, to do so except in language that is to some degree "functional"; for obviously it is impossible to enumerate all possible variants. Indeed, some degree of permissible latitude would seem to follow from the doctrine of equivalents, which was devised to eke out verbal insufficiencies of claims. Since by virtue of that doctrine a claim will cover whatever will accomplish substantially the same result by substantially the same means, it cannot be that a claim becomes invalid when it states expressly what the courts would in any event imply.

"Functional" claims certainly fulfill one of the offices of claims in general, which is to advise the art of the scope of the monopoly; at least they do so unless, as in United Carbon Company v. Binney Smith Company they are too vague to be understood at all. Their vice is not that, but that they extend the monopoly beyond the proper limits of the "invention"; and to ascertain what those limits are, we always have to look to the contribution of the disclosure to the art. Almost all inventions are combinations of old elements, whose selection as a new unit gives them their only importance. Their combination is the end or purpose of the "invention": its "nature and design" which the applicant must state. The elements of the combination are the means by which that "nature and design" is realized; and nobody invades the patent who does not appropriate both end and means. To the extent to which variants, which will be serviceable as substitute means, are known to the art, and at once suggest themselves without need of further substantial experimentation, they are equivalents, and to extend the monopoly to them is not only justifiable but necessary to the protection of the inventor. However, although for these reasons it is possible for claims, "functional" in form, to be valid, verbally they do leave at large the means by which the "invention" is to be practised; they do not "distinctly claim the part, improvement, or combination which" the in ventor "claims as his invention or discovery." On that account the patentee must be prepared to prove, if he would support them, that all practicable means, comprehended within the general language he has chosen, were in fact known and accessible to the art, and that their substitution for the specific details of the specifications, would at once suggest itself without further substantial experimentation. That is, as we understand it, the doctrine of Holland Furniture Company v. Perkins Glue Company, supra, 277 U.S. at page 257, 48 S.Ct. 474, 72 L.Ed. 868.

The defendant made no such proof. There are at least one hundred resins — perhaps more — and, although one witness for the defendant — Grosvenor — ventured the belief that of these "you probably could pick between half a dozen and a dozen fairly readily that it would be worthwhile starting on," in doing so he excluded, not only "thermosetting" resins, but "thermoplastic resins which do not fuse to a clear liquid." The specifications do not require a resin which will fuse to a clear liquid; the nearest approach is that the specific example of the resin is said to be "perfectly clear and amber in color" (page two, col. 2, lines 28, 29); but the eighth "characteristic" has no such requirement. How many of the one hundred are not "thermosetting," the record does not disclose; and we should have no warrant for saying (unless we were in addition to lay aside all which do not fuse to a clear liquid) that they number only a dozen. Whatever their number, the defendant has not shown that it is possible without further substantial experiment to pick out those which have the nine "characteristics." Besides, Grosvenor did not even say that all the dozen or half dozen would serve; all he did say was that they would be "worthwhile starting on" — a very different thing. The same defects apply even more plainly to the "filler," the "adhesive powder" and the "lubricant." It may be that the patentee did discover a patentable invention, and it is quite possible it was broader than his examples; but the defendant has not met the difficulty, and we agree with Judge Byers and with his reasons, that the claims in suit are invalid.

Moreover, they are invalid for another reason. The invention was in public use for more than two years before the application was filed on May 9, 1938, and the right to a patent had been forfeited, unless this delay is excused. The excuse offered is that the application was a "continuation" of an earlier application which the examiner had rejected because its disclosure was inadequate; and the Court of Customs and Patent Appeals held in Whittier v. Borchardt that a second application is not a "continuation" of an earlier one if the earlier one was inadequate to support any claims whatever. That decision was an extension of an earlier doctrine, which the same court laid down in Lavin v. Pierotti, that, if an application is rejected because the disclosure is inoperative, a later amended application with an operative disclosure is not a "continuation" of the first. So far as we can find, the notion of a second application as a "continuation" of an earlier one was a contrivance of Commissioner Newman in 1920, who discussed it in an opinion of which we quote the important part in the margin. Assuming that his language is to be taken as authoritative, we must own that it is not quite clear to us why a second application, which corrects an inoperative disclosure, or which supplies what is lacking in an earlier one, should not be regarded as disclosing "a substantial part of the subject-matter of the prior application"; and we are by no means sure that as a new matter we should have so held. Be that as it may, the question is only as to the proper procedure in the Patent Office: i.e., as to what kind of application shall be regarded as being identical with an earlier one. The Court of Customs and Patent Appeals is the court of last resort — except for certiorari — upon such questions, and it seems to us proper, save in a case of very clear difference of opinion, to follow its rulings. We therefore consider Whittier v. Borchardt, supra, as controlling, and so did Judge Byers.

On the other hand, he thought that the second application had been filed under circumstances which made that decision inapplicable. The trouble with the first application was that the only specific disclosure of the resin was by reference to the name of a proprietary article. The examiner held this inadequate, although the application recited the same nine "characteristics" which are contained in the second application. The applicant tried to correct the defect by describing the proprietary preparation by its contents, but the examiner was not satisfied, and once more rejected the application, that time "finally," adding that "under these circumstances it appears that the applicant's only recourse if he knows the composition of some suitable resin, is to file a continuation in part which includes such a composition." The applicant was still not satisfied and tried once again to amend the first application. Again the examiner rejected it, though he considered the amendment on the merits, and he once more said: "it is believed that the applicant's only remedy is a continuation in part" etc. The applicant this time accepted the suggestion instead of appealing from the decision. So far as appears, there was nothing to forbid the amendment of the first application, nor do we understand that the examiner denied the applicant that privilege; he merely held inadequate the amendment actually made. However that may be, the new application was not a "continuation," as we have said, because the original disclosure as it stood had not been enough to support any claims whatever. The whole coil of procedure does not impress as having been necessary, and, as we have said, we do not quite see why the second application did not disclose "a substantial part of the subject-matter" of the first; on the other hand, whatever injustice the applicant may have suffered, we cannot see why the examiner's remark should help out the patent. In the first place his ruling, if he did make any ruling that a second application would be a "continuation," was not authoritative, for he had no authority in the premises. But even if he had had, we cannot see that he meant to exercise it. To tell the applicant that his "only recourse" was "to file a continuation" was not to assure him that the new application would be valid as such. All that the applicant could fairly infer was that that was the only possible escape for him; and that he might try it if he chose. The applicant knew, or at least he was charged with knowledge, that the examiner's ruling was subject to review, and he chose to abandon his right to review it. If the result is harsh and over strict, it arises from the law as to "continuations."

Judgment affirmed.