L. HAND, Circuit Judge.

This case arises on an appeal by the defendant from a decree upon the usual bill in equity for the infringement of claims 1, 2, and 5 of patent No. 1,389,067, and of claims 1 and 4 of reissue patent No. 17,903. In Winget Kickernick Company v. Kenilworth Mfg. Co., 11 F.2d 1, we held the same claims of the first patent valid and infringed, and for the purposes of this appeal we shall again assume that they are valid. The invention is for improvements in women's drawers, and claim one — which will serve for all three — is as follows: "A woman's undergarment comprising a back member having the central portion of the bottom edge concavely curved throughout its length, a front member having its centre cut along a straight line a predetermined distance, and an insert sewed along one edge to the edges of said central cut, and along another edge to the curved edge of the back member." It is impossible to describe such things verbally so that they shall be understood, and for this reason we annex two exhibits, Nos. 80 and 49, which graphically picture the disclosure and the alleged infringement respectively:

Exhibit 80 shows the three pieces which make up the drawers: "front piece," "insert," and "back piece." The "insert" is a long "rectangular" piece, folded in the middle along its shorter axis, one long edge being sewed to the two edges of the slit in the front, as shown in the middle figure of the upper row. The lower edge of the back piece is not straight, though at either end, from its juncture with the vertical edge, it is straight for a distance and at a slightly obtuse angle to that edge. These two straight portions are connected by the sector of a circle, which is sewn to the other edge of the "insert" as shown in the figure marked, "Relationship of Front to Back Pieces." To complete the garment one or both of the outer edges of front and back are sewn together, as indicated by the unmarked central figure in the lower row. The defendant's garment can be readily understood by following Exhibit 49. It has no "Insert," and its front and back are each made in two halves, shown in the top row as still unjoined. At the lower inner edge of each half of both front and back appear two members, separated by dotted lines from the main fabric. These might well — better, indeed — have been incorporated into the halves themselves; and would have been, except that the defendant uses bolts of cloth from which it is more economical to cut four pieces separately, and to sew them to the halves along the dotted lines. There is not the least reason to suspect this explanation, and the result is quite the same as though the four pieces were never separate at any stage of manufacture. The defendant then sews together the halves of both front and back along their inner edges above these pieces, and sews the pieces themselves together at their top, as appears in the left figure of the lower row. This makes the garment ready for final assemblage by sewing together one or both of the outer edges of the front to a corresponding edge of the back, as indicated by the dotted lines, CC', FF', GG', and HH' of the figure just mentioned; and by sewing together the open edges of the four pieces, as indicated by the lines, BB', DD', and EE'.

We will assume arguendo that, if we had nothing but the disclosure, the claims would cover the defendant's garment; for although, if the halves of the back were sewn together before the pieces were added, the bottom edge of the resulting piece would have no curve like that of the patented back, this is the same difference which did not avoid infringement in Winget Kickernick Co. v. Kenilworth Mfg. Co., supra, 11 F.2d 1, and that case would rule. The record there contained, however, no references of any significance to limit the scope of the claims; nothing closer than Morris, No. 1,098,960, and Johnson, No. 850,013; and particularly it did not contain the prior publications of one Rosenfeld, whom the defendant called as a witness in the case at bar, and two editions of whose "Practical Designer," one in 1911 and the other in 1918, now put the whole case in an entirely different light. These show that, so far as the defendant can be said to have borrowed from the patent at all, it was in features that had existed earlier. Once more, the garment can be best understood from a diagram, and we have annexed a page from the 1918 edition of the "Practical Designer," showing a pair of women's bloomers, and Exhibit 63, which makes this page a little plainer (in our discussion we shall use the publication):

The figure at the right is the left half of the front; that at the left, the left half of the back, the front being superimposed. The edge, XJNI, of the left front, is sewed to the similar edge of the right front, not shown; similarly the edge, KJIM, of the left back, is sewed to the similar edge of the right back, also not shown. There then results a single front and a single back; between the legs of each of which there is a flap which points inward. (See the left-hand figure of the lower row of Exhibit 63). Next the edge, IT, of each half of the front is sewn to the corresponding edge, M 4, of the back, and the outer edge, QUC to the other edge, PU 2, on both, or on only one, side, as may be desired. The result is precisely the same as the defendant's garment, except that it is made differently, the crotch pieces not being first cut separately and then sewn to the halves of the front and of the back.

Rosenfeld's "Practical Designer" need not indeed be construed as an anticipation of the patent in suit, and, as we have said, we shall not so construe it; to do so we should have to say that the four areas of cloth, never separated from the fabric of the front and back, were the equivalent of the single, separate, "rectangular insert" of the disclosure; also that the two-piece front and back were an equivalent of the disclosed front and back. We may assume arguendo that they are not; but the same reasoning by which the patent thus escapes anticipation narrows the claims so as to exclude the defendant's garment; because, as we have seen, its front and back are in two pieces and its "insert" is in four; and it is thus much nearer to Rosenfeld's bloomers than to the patent. For instance, if we suppose that from each half of Rosenfeld's front a piece were cut by continuing the line, XJH, to the bottom, and that the same were done to each half of the back by so continuing the line, RJ, four pieces would emerge, which for practical purposes would be the same as the four pieces which the defendant sews together to make its supposed "insert." (This process is indicated by the dotted lines on Exhibit 63.) It is true that these four pieces, if sewn together, would not make a strictly "rectangular insert"; even less so than if the defendant's pieces were so sewn. The edge, JNIT, of Rosenfeld's front half, is longer than the continuation of the edge, XJ, from J to the edge, TVC; similarly the edge, JIM 4, is longer than the continuation of the line, RJ, from J to the edge, 1, 2, 3, 4. If all four pieces were sewn together, the resulting "insert" would have two peaks, as it were, rising from its plane — one at the front, and the other at the back. But that would also be true of the defendant's "insert," if the four pieces were sewn together; in each case the fold of the "insert" along its shorter axis is not straight, but is a "V," though it is more marked in Rosenfeld's case. This appears by a comparison of Exhibit 49 with Exhibit 63. In a patent like this, such a difference of degree cannot support invention; if all that the defendant did was to make an "insert," which when folded had less of a "V" than Rosenfeld's, the claims will not reach its garment. On the other hand, if the plaintiff seeks to distinguish Rosenfeld's bloomers because they have no "insert" at all, that is, because the four pieces which might make one, always remain parts of the front and back, then infringement must rest upon the fact that the defendant first cuts out four pieces separately, and then sews them to the halves of the front and of the back before sewing together the halves. Obviously there can be no invention in details of manufacture like that; especially when the pieces are cut separate only for economy, and when the claims are not for a process. Thus, the patent must depend upon an "insert," cut in a single piece, as opposed to one built up of several pieces and never existing as a single, separate member at all. Perhaps, if in the earlier record we had had before us what we now have, we might have declared the patent invalid then; but, however that may be, it will not avail the plaintiff in the case at bar.

The Reissue Patent.

The reissue patent is of an earlier date than patent No. 1,389,067. The original application was filed on July 22, 1920, but was allowed to lapse; it was renewed on December 5, 1927, and the original, No. 1,680,267, issued on August 7, 1928. The application for reissue was filed on July 24, 1930. Claims 1 and 4 in suit are for a process of making women's drawers, and can be best illustrated by Exhibit 81 annexed hereto:

The front piece is the same as in patent No. 1,389,067; but there is no insert, and the back — here shown in two pieces — is longer than the front, to gain the necessary fulness. The straight bottom edge of the back is sewn directly along the edges of the slit in the front, as shown in the figure marked, "Relationship of Front and Back Pieces." (Obviously this edge must be exactly twice the length of the slit.) In order to make an parallelism possible between this disclosure and the defendant's garment, so that the claims will even read verbally upon it, we must assume that the two pieces which are in fact sewed to the two halves of the defendant's front before the halves are sewed together, are never sewn to them, but instead, are sewn to the two other pieces which have already been sewn to the halves of the back. The first two may then be regarded as making the bottom edge of the back, which can be gathered up and sewn to the edges of the slit in front. The claims being for a process — a sequence of steps — the defendant must follow that sequence, if it is to infringe. There can be no pretence that it does follow the disclosed steps literally; and, if resort be had to equivalents, what we have already said disposes of this patent even more certainly than it does of patent No. 1,389,067. We have no disposition to question the validity of the claims, but they cannot be stretched to cover the defendant's garment.

Decree reversed; bill dismissed.