(after stating the facts). The production of an article which became known as “featherbone,” to be used in lieu of whalebone, for “corset stiffeners” and analogous purposes, originated with Edward K. Warren, one of the patentees, in the year 1883. .It was then made of quills, or quill splints stripped of the feathers, and bound together forming a rib or stiffener for various articles of dress, as described in letters -patent No. 286,749, issued to the inventor, October 16, 1883. Several subsequent patents were obtained for various alleged improvements in methods and product, which are mentioned in reference to this origin of the name “featherbone,” in the opinion of Judge Jenkins, speaking for this court, at the present term, in the case of Warren Featherbone Company v. American Featherbone Company, 141 Fed. 513.

The contention in defense of the present suit, however, that the product of this patent (No. 559,827) is identical with that described in the original patent (No. 286,749), or in one or the other of the last-mentioned patents (now expired) does not impress us as tenable. While it is true that quill splints are alike the basic element, thus justifying retention of the name featherbone, the introduction of sizing, reheating, and pressure was a new utilization of that element, and resulted in a new homogeneous material which has attained commercial success. The prior discovery that quills could be used as such element in a substitute for whalebone does not deprive this improvement of recognition as another useful discovery. Nor are we impressed with force in the contentions, either of anticipation of the process claims {1 and 2) in the prior patents introduced, or of noninfringement of such claims in the light of the prior art.

Thus viewing the invention of the patentees as meritorious, we are brought regretfully to the consideration of the evidence of its public use prior to the application for patent. The statute (section 4886, Rev. St. [3 U, S. Comp. St. 1901, p. 3382]) conditions the grant of a patent for an invention that it shall not have been “in public use or on sale in this country for more than two years prior to” the application; and the policy of this provision and its strict construction against the patent are well settled. Egbert v. Lippmann, 104 U. S. 333, 336, 26 L. Ed. 755; 10 Notes U. S. Rep. 176. Nevertheless, the evidence of such public use to defeat the patent must be clear and convincing; and if the rule stated in Morgan v. Daniels, 153 U. S. 120, 123, 14 Sup. Ct. 772, 38 L. Ed. 657, is applicable as well to this issue, every reasonable doubt should be resolved against the defense. While any well-defined case of public use more than two years before the application is filed bars the patent, use “if made in good faith, solely to test the qualities of the invention, and for the purpose of experiment” only, is not within the inhibition. Egbert v. Lippmann, supra. The evidence upon the issue in the present case is voluminous, and the direct testimony is conflicting in reference to the time when practical operations commenced in the use of the patent process. Such use for a year and several months prior to the application is substantially conceded, but the witness Charles K. Warren testifies, in effect, that the process was not perfected until the “summer or fall of 1893”—the application being filed March *6595, 1895—and was not used commercially before that time; that all prior use was experimental only, in perfecting the process. Another witness corroborates this view. If this testimony may be credited— not considering the circumstantial evidence and the conduct and interest of the witness—strengthened by the presumptions in fayor of patent-ability, it may be doubted whether the testimony of the 20 or more opposing witnesses, tending to show public use long prior to March, 1893, would be deemed sufficient to overcome its force. We are constrained, however, to the opinion that the evidence, direct and circumstantial, is decisive that the process and product of the patent in suit were in public (commercial) use in the factory more than two years before a patent was applied for, and as early as the year 1887 or 1888; and that the version given by the witness Warren is not only inconsistent with the circumstances and exhibits in evidence, but is further discredited by the stress of circumstances under which it arose, by the indefiniteness of the story, and by the apparent want of candor on the part of the witness.

Review of the evidence in point is unnecessary, as the admission which was ultimately brought out, of public use during and after the summer of 1893, narrows the controversy; and a few of the indisputable facts are sufficient to establish such use beyond the two-year limit. No explanation is offered for this conceded long delay in applying for the patent, and it is surely not unreasonable to infer that it was due to a wish to prolong the period of monopoly, rather than to want of familiarity with such applications. The monopoly under the original feather-bone patent (No. 286,749) for the product “formed of quills or quill splints” was then unexpired, and its use by the Warrens was both large and exclusive, while the number of other patents theretofore taken out by these patentees indicated an intimate acquaintance with patent applications. The introduction of glue and the methods of utilizing it to make a new product of quill splints constitute the invention in question. Glue or sizing did not enter into the product of the earlier patents, but the manufacture of featherbone under those patents was continuous and extensive in the factory of the complainant, up to the adoption of the present invention.

Upon these premises a well authenticated series of featherbone exhibits in evidence establishes commercial use of the final invention long prior to the time thus conceded, and beyond possible escape from the limitation. If confirmation of these product exhibits were needful, it is supplied by the cumulative testimony of the witnesses before mentioned in reference to the use of glue in the process of manufacture; and the earlier correspondence between the complainant and certain of its customers, introduced in evidence, is additional confirmation of such use.

In the course of taking testimony for the defense, numerous exhibits of old dress waists or bodices were introduced, containing feather-bone used for stiffening, and witnesses were tendered to prove dates and circumstances of making. These exhibits were examined by counsel for the complainant and by the patentee, E. K. Warren, according to a statement appearing in the record, and thereupon this admission was entered of record: “That each and all of the bodice exhibits above *660mentioned contains stiffening strips of featherbone, and that the several bodices referred to were all made and used more than two years prior to the filing of the application for the patent here in suit, and that the strips of featherbone that now appear in them were in them at the time when they were so first made and used.” None of the witnesses, therefore, were called for authentication of the exhibits. Subsequently the witness C. K. Warren (not the patentee) testified—over vigorous objections—in substance, that the featherbone in all these exhibit dress waists, except two called the “Newberry Exhibits,” was made under the old process, and not under the patent, and that the featherbone in the Newberry exhibits was made not earlier than 1896, for reasons stated. Comment on this and other courses in the order and character of the testimony is needless for the present consideration, as the-Newberry exhibits referred to were unmistakably identified by several credible witnesses as made in 1887 and 1889 respectively (for occasions which fix the dates beyond doubt) together with a third exhibit made in 1892. It was not only undisputed, but obvious on inspection, that the featherbone in each of the first mentioned exhibits was the product described in the present patent, and entered into the original make of the garments as the witness testified. Without reference to other exhibits and circumstances in the case of like effect, or to the failure of either of the patentees to testify upon this issue when no reasonable excuse appears for their absence, these instances of use are sufficient to defeat the patent.

The defense which was thus made out encountered unfair difficulties in the methods of introducing the various phases of the testimony, in misstatements of fact, direct and indirect, and other devious ways, involving much research and expense in overcoming them, and these means cannot be passed over without censure. Particular mention of one bold instance will suffice. The testimony of the witness Warren in reference to the time when the new process entered into use requires no further comment; but in support of his version, by way of showing the marked departure in the increase of business under the new process at the date so fixed—indicative as well of its advance over the earlier invention—tables were introduced to show the comparative sales of featherbone from the year 1888 to 1902, with a great increase after 1893. The testimony of Edward K. Warren in a prior suit was then discovered and introduced by the defense, and plainly refutes the story of these tables. So that it is now conceded 'that the tables omitted a very large proportion of the sales of the earlier years. In other words, they were worthless for any legitimate object, and were calculated to deceive; and but for the vigilance of the defense would have deceived the court. None of these censurable means referred to are attributed in any sense to either of the counsel in the case, but we are satisfied that they were devised by Charles K. Warren, the secretary and chief witness for the complainant, and he deserves the censure of the court accordingly. ‘

The decree of the Circuit Court is reversed, with direction to dismiss the bill for want of equity.