DENISON, Circuit Judge
(after stating the facts as above). Infringement of both claims is alleged; and, if they receive the construction which we think they must have, infringement is conceded.
[1] Several earlier uses are set up in order to anticipate or otherwise invalidate the second and broader claim. We agree with the District Judge that the claim is invalid; but we think this conclusion is most safely based upon the earlier so-called segmental eutter. Curtis made his constructive reduction to practice by filing his application in October, 1900. He claims to have conceived the invention in June of that year; but the evidence of any definite early conception of the invention embodied in these claims is vague, and there was no dis*94closure until in the course of preparing the application. The defendant’s predecessor experimented in the summer of 1899 with a machine of this general class, and in December of that year let a contract for the building of a commercial machine. This commercial machine was completed in the spring of 1900 and operated by hand sufficiently to demonstrate its effectiveness. This is proved with all the certainty necessary to establish a claim of anticipation. The rolls were not oper-' ated by power at that time, because they were intended to be only a part of the complete machine which should cut and expand, and the expanding part was not ready; but there was no necessary connection between the two parts, and we consider what was done to be a sufficient reduction to practice of the idea involved. It was the same as that later developed by Curtis, with two exceptions: (a) The notches were placed upon the corresponding (say left-hand) edges of each cutting roll, upper and lower, instead of upon both edges of the upper rolls only, as in Curtis’ preferred form; (b) instead of cutting notches in the edges of the body of the roll, defendant’s predecessor made the roll of less diameter than the adjacent spacing disc, and then built it out with a series of so-called segments until it was of the larger diameter required for the cutting roll. Each segment contained a portion of the cutting edge having one notch (two half notches) therein, and was bolted fast to the body of the roll. When they were all in place, they constitute a continuous roll-cutting edge interrupted by notches. This awkward arrangement was adopted because it was thought the cutting edge must be of harder steel than was advisable for the body of the roll, and because it was thought that separable parts would facilitate sharpening and repairs. It was anticipated that notches would break down and wear out, and in this way one part could be replaced without affecting the rest of the roll. When the entire machine was later completed, it was employed for some time to produce the commercial result. It was finally discarded and replaced by the Curtis construction, because the making in this segmental way and the necessary replacements and repairs proved to be too expensive.
[2, 3] This segmental cutter, although it had never been patented or described in any printed publication and was not known to Curtis, yet was a part of the prior art, with reference to which the character of his variatibn therefrom must be judged. If some one else had completely invented and used this segmental cutter, and if what Curtis did involved no invention over what had thus been done, certainly Curtis was not entitled to a patent as the first inventor of the thing which he produced. One cannot be an inventor, if there is no invention. See Package Co. v. Johnson Co. (C. C. A. 6) 246 Fed. 598, 601, 159 C. C. A. 568; Lemley v. Dobson (C. C. A. 6) 243 Fed. 391, 395, 156 C. C. A. 171; and see Buser v. Novelty Co. (C. C. A. 6) 151 Fed. 478, 492, et seq., 81 C. C. A. 16. This is not a case where a prior use, exact or suggestive, had been abandoned, and so perhaps should go into oblivion instead of into the prior art; the segmental cutter was in complete and active existence, when Curtis made his conception, and in commercial use when he filed his application. If the segmental cutter is not generally different from Curtis’ form, then it follows that *95the substance of claim 2 — a square-edged cutting roll notched on one edge only — has been in continuous use by defendant and its predecessor ever since a date earlier than Curtis’ invention, for change in form within the limits of the claim would not interrupt the continuity of the use.
[4] We find thus presented the controlling question whéther Curtis’ integral square-edged cutting roll with notches on the edge involved invention as compared with defendant’s sectional square-edged roll with notches on the edge. The precise question may be clearer by transposing to the more familiar circular saw. L,et it be supposed that saws composed of an integral disc with teeth cut into the edge were common; that defendant had devised a peculiar kind of teeth, and had attached these by removable means to the edge of the disc, thereby producing a saw with these peculiar teeth, each tooth separately removable; and that Curtis then put these same peculiar teeth into the edge of an integral saw by cutting them therein. We are confident that this would not, and that what Curtis did with these cutter rolls did not, involve invention. We have a clear instance of that merely making integrally what had before been made in two parts and then put together, rather than an instance of that uniting into one part which dispenses with a function or gets a new result. See Gould v. Cincinnati Co. (C. C. A. 6) 194 Fed. 680, 685, 115 C. C. A. 74. Curtis’ cutter roll was stronger and simpler than defendant’s, but only as an integral structure is always stronger and simpler than a sectional one. The defendant had passed through and beyond the Curtis form. Curtis simply dropped off what he thought a useless and troublesome refinement, and adopted the remainder, the substance of defendant’s structure. This is especially illustrated by the fact that a sectional view of this segmental cutter and adjacent spacing disc, taken on any selected diameter, cannot be distinguished from a sectional view of the Curtis roll and spacing disc taken on the same diameter, save by the presence of the attaching bolt and the seam between the permanent body and the removable edge.
So far as concerns claim 2, the decree of the court below dismissing the bill must be affirmed.
[5] Claim. 1 presents a different problem. We think it must be interpreted as of specific intent, directed to and limited to Curtis’ preferred form, whereby he put all his interrupting- notches upon the upper rolls, leaving the lower rolls plain, and arranged these notches on the opposite edges of the same roll in staggered relation to each other. Defendant has adopted and used this precise form. It seeks to escape liability by alleging that, if given this specific construction, the claim involves no patentable distinction from claim 2, and is invalid for the same reason, but that it ought not to be so limited, but should be constructed so broadly that it is equivalent to, and is defeated by the same facts which defeat, claim 2.
Putting these notches upon both edges of the same disc, and not opposite each other but staggered, was entirely new and had obvious advantages over the plan of notching one edge of both upper and lower rolls. In the latter case, the ultimate alternate relation of slit *96and bond must be secured by a rather nice relative rotary adjustment of upper and lower rolls on their respective shafts and also by regulating the driving means for each shaft so that the notches will leave the bonds at the right spots. Either wear in connecting gears or inaccuracy in replacing rolls after they have been removed tor grinding would produce bad results. With the Curtis preferred form, the proper relation of slit and bond in the two adjacent rows was necessarily and automatically preserved. When the rolls were placed on the shaft, originally or after removal, it was immaterial what circumferential position the notches happened to have, so long as the notches in the series of rolls on that one shaft were aligned with each other. It is true these results and these advantages are not testified to by any expert; but they are obvious, and there is no testimony tending to dispute their apparent value. We cannot say that this specific improvement involved no invention.
{6, 7] Since a device which does not bring the bonds and the slits into alternate position will not make a useful product, it is said that the thought expressed by “the notches being relatively staggered” must be read into the second claim as well, and that the notches are “relatively staggered” even in the form of machine which notches only one side of the roll but uáes notched rolls on both upper and lower shafts, because these are so adjusted that the notches strike the metal in a staggered relation to each other. Viewed in this way, the claims are equivalent and of a breadth which makes them both invalid. We cannot accept this point of view, for the sufficient reason that a limitation found in one claim and by which alone it substantially differs from another cannot be read into that.other, but the expressed distinction must be preserved. There is the further reason that, although by resort to all the possible meanings of the phrase in question, we can conceive that in the upper and lower notched roll construction, the notches are “relatively staggered,” yet this was plainly not what Curtis meant when he selected the words; and a patentee is at liberty to supply his own dictionary; the terms of a claim should be taken in the sense given in the lexicon of the specification. Baker, C. J., in Kennicott Co. v. Holt Co. (C. C. A. 7) 230 Fed. 157, 160, 144 C. C. A. 455. He shows, in Fig. 5, one form having the notches alternately arranged on the opposite edges of the same roll, and, in Fig. 6, a form having the notches upon the same edge, and upon one edge only, of upper and lower rolls. He says:
“I prefer to form these notches upon both sides of all the rings as shown in Fig. 5, arranging the notches in staggered fashion. * * * X do not notch the (lower) rings 30 as that is unnecessary if the (upper) rings 20 are notched upon both sides as shown at Fig. 5; hut, instead of that construction, the upper rings may be notched upon one side only and the lower rings should, in that case, also be notched upon one side. * * * Either construction works well.”
The word “stagger” is not found in any form in the specification, except in the sentence just quoted; and this reference to the authoritative lexicon makes it entirely clear that the first claim was intended to cover this preferred form, and the second claim was intended to be broad enough to cover either form. Defendant’s argument upon this *97subject, based upon the contents of the file wrapper, is not convincing. As the application was filed, the second claim was in the same form as when issued; but the first claim was still broader and reached generally this arrangement of cutting rings, provided only that they had “notches on their side faces.” This would cover notches extending transversely entirely across the periphery and was old, as appeared by the reference cited by the Patent Office. Curtis then amended by substituting claim 1 as issued, and pointed out that there was invention in making notches in the edge only instead of all the way across, because by this change he succeeded in producing bonds which were necessarily opposite each other. The patent then issued.
[8] It results from these considerations that there must be the usual interlocutory decree based on claim 1; but for accounting only — the patent has expired. We appreciate the difficulties attending an accounting, but they are not sufficient to justify the refusal of that remedy, particularly in view of the present rule that damages may be measured by reasonable royalty, if there is no better means of establishing profits or damages.
The judgment is reversed, with costs of this court, and the case remanded for further proceedings.