LAGOMBE, Circuit Judge.
The patent is for a mattress frame, and was issued to complainant November 29, 1898, upon an application filed August 7, 1897. The specification states that the object of the improvement is to construct a simple and inexpensive frame, in which the smallest possible surface will be exposed to the accumulation of dust, dirt, or vermin, and in which the weight put upon the ihattress will be so distributed as to be directed against all parts, and in- which .the greater the weight the more firmly will the frame be held together. The device comprises a corner bracket in which the side rails and end bars are inserted. Of this bracket the specification states:
“The bracket, O, comprises a socket, e, into which the end of the side rail is adapted to snugly fit, while the upper surface of this socket extends backwardly a suitable distance, forming an extension, d, which, when the woven-wire fabric is attached to the end bars, bears firmly down upon the side rails, as hereinbefore mentioned.”
*963The first claim is as follows:
“(1) A socket for a mattress frame comprising a part adapted to engage' tke upper surface of the side rails of the mattress frame, a horizontally extending shoulder formed on one side thereof, a vertically extending bracket at the outer end of said shoulder, a similar bracket at the point whore said socket engages the side rail, and means for screwing the end bar of the mattress to the upper ends of said brackets, substantially as described.”
On November 3, 1896 (upon application filed August 10, 1896), a patent for a design for a corner block for mattress frames was issued to one Taber (design No. 26,245). It is conceded by complainant that the subject-matter of the first claim of the patent in suit is disclosed in this design patent. To avoid the defense which this design patent establishes, it is necessary for complainant to show that his invention antedates the date of his application by several months. The burden of proving this is upon complainant, and we are unable to concur with the judge who heard the cause at circuit that he has borne such burden. Such evidence, especially when it deals with experiments which resulted only in some two or three specimens, which never left the shop, and were seen years before by but a few persons, who, in giving their recollections of dates, are unable to fix such dates by reference to some transaction whose date is susceptible of definite proof, is rarely satisfactory. Before discussing this testimony, it may be noted that it deals with two different groups of devices, an “earlier pattern” and “earlier casting,” and a “later pattern” and “later casting.” The judge at circuit was of the opinion that the earlier device embodied the invention of the first claim. We have reached a different conclusion, not finding in it the “part adapted to engage the upper surface of the side rails,” which the specification refers to as an extension hackwardly a suitable distance, thus securing a hearing surface adapted to give the strength of structure and distribution of weight, which was the object sought for. The crucial fact, therefore, which complainant must establish by persuasive proof, is a date for the “later” pattern or casting prior to November 3, 1896.
Complainant’s application is sworn to July 19, 1897. His concern moved into a new factory on Second avenue in December, 1894. His narrative of what took place between these fixed dates is as follows: Shortly after he got settled in the new quarters (Second avenue), it being a dull season, be made sketches, experimented, and, having finally settled upon a plan, made patterns, which he got completed somewhere about March or April, 1895 (before the busy season, which was usually in April or May). Shortly after the patterns were completed, castings were made of composition. After these were made, he got the tubular side rails ready, and the frame was assembled into a spring (the trade-name for a completed spring mattress), and put together by himself and one of his workmen, Beyers. It then appeared that the tension was not great enough, so the spring was taken apart, and the fabric cut off, in order to put on fabrics with greater tension. By that time they were in the busy season, and thé thing was laid aside, so that they might get out orders, and the “invention lay dormant for some little time.” At that time his idea, as *964he says, was to “make the casting of something stronger than the composition,”-and he finally got them made of malleable iron. There is no suggestion that the model of the castings was then changed;, only the material. This “took some time to get them cast,” and it was late in the fall before they were ready to do much more with it. “When we got time,” says the witness, “after our busy season was over in the fall, we took the matter up again, made a spring that suited us so far as tension was concerned, but not so attractive as we would like.” In this narrative complainant is corroborated by the workman Beyers, who says he believes he saw the pattern in the early part of 1895, and the casting in about May or June; that he very well recollects putting the bed together, which, “to the best of his knowledge, must have been” in the month of June, 1895. He is also corroborated by his son, William J. Lein, who says he saw the pattern before they moved to Second avenue (a fixed date, — December, 1894), and the casting four or five months after moving, when he assisted to construct a metallic bed with the identical casting in it. The evidence abundantly proves the existence of the device embodied in the earlier pattern and casting in the year 1895. As to the device embodied in the later pattern and later casting, however, the evidence is unpersuasive. Complainant’s narrative is that after he had made a bed that suited, using the earlier form of casting in malleable iron, which was late in the fall of 1895, he went to work to make another set of patterns. “These patterns,” he says, “with the sketches, took some time, and, with the business then increasing, it was some time before we got the castings ready to put together. We were, however, so well satisfied with what we had done that I determined to apply for a patent, and did so in July, 1897, to the best of my recollection. The pattern I produce is one that was made at that time.” In this narrative of events it will be observed that the witness gives no date for the production of the later pattern, and his statement seems rather to indicate that such production was nearer in time to the application for the patent than it was to the'completion of the earlier model of spring bed. The man who made the castings is named, but neither himself nor his books produced to fix .the date. At the close of a long direct and cross examination, complainant was asked on redirect the question, ‘When did you receive the later castings?” to which he replied, “The latter part of 1895 or in 1896,” — an answer which witness does not accentuate by reference to any transaction of known dates, and which does not entirely harmonize with his narrative -of events as given on the direct. Complainant’s partner, Irvine, was not called as a witness, but it is stipulated that, if called, he would testify as set forth in a page of the record: . “In the latter part of 1895 Mr. Lein showed me the complainant’s exhibit later pattern, and in the latter part of 1895 or early in 1896 he showed me complainant’s exhibit later casting, and about this time I saw this casting put into a mattress frame which was assembled in the shop by Mr. Lein, and tested by him and some of the workmen.” Here again the date is fixed by the unaided memory without reference to any other occurrence which the witness knows the date of. He does, however, fix it as of the time when the mattress frame was as*965sembled in the shop by complainant and some of the workmen. The. testimony of these two witnesses is not very strong as to date when, these later patterns and castings were produced. It lacks corroboration. Nevertheless, it might be accepted as sufficient to carry the date of invention back prior to November 3, 1896, if it were uncontradicted. Two other witnesses, however, called by the complainant, give testimony which indicates that the date was more recent. Beyers, the workman who has been already quoted as to the assembling of the bed made with the original castings, testified that he put together a mattress with the later castings “right after we moved from Second avenue to Twenty-Third street”; he “thinks it must have been” before they moved; that he first saw the later patterns and castings, but says that it was after removal that they were-assembled; “must have been in July, I think, because I know it was our slack season.” The removal to Twenty-Third street was in 1897. William J. Lein, the foreman of the shop, testified that he saw the later pattern “first just before we moved tó Twenty-Third street, and assisted in constructing one immediately after we moved there.” Asked when he first saw the later casting, he replied, “To the best of my belief, it was about two years ago [he testified September 12, 1899], or when we first moved into Twenty-Third street, where I assisted in constructing a spring of these castings,” and added that they moved into Twenty-Third street in May, 1897.
Upon this state of the testimony, all produced from complainant’s own shop, we are not satisfied that he has shown by a fair preponderance of proof that his invention antedates the issue of the Taber design patent, which concededly discloses the limited improvement sought to be covered by the first claim of the pátent in suit. The decree of the circuit court is reversed, with costs.