L. HAND, Circuit Judge.

This is the usual suit in equity for. the infringement of a patent for a method of making a dielectric condenser and for the condenser so made. The art had made such condensers in the past by winding alternate sheets of tin-foil and paper into a cartridge or cylinder, and attaching one terminal to one sheet of foil, and the other, to the second, the paper being the dielectric. The difficulties with this were that though the cartridge was heavily compressed, there would remain air spaces which caused “breakdowns.” To correct this the whole cartridge while in the press was put into a vacuum pump, to withdraw the air so far as possible, and was “impregnated” with paraffin or other similar waxes heated so as to flow into the interstices. This worked very well, but it was found that, as the wax cooled, it contracted and still left spaces, though not so great as when the condenser was not impregnated at all. The patent stood upon the art at this stage, and added its own contribution. It was to dip the condenser when so “impregnated” and while the wax was still hot into a bath of oil “having high dielectric strength, nonhygroscopie and free from moisture” (page 2, lines 38-40). Several of such oils were mentioned. The patent proved successful, and we may assume, though we indicate no opinion as to it, that the invention was patentable. The only question which we need decide is whether it had been in publie use for more than two years before the original application was filed, February 28,1929.

Scheeker, the inventor, in 1926 and 1927, was an employee of the Dubilier Condenser Company. He was the supervisor in charge of winding such condensers, and appears to have discovered the added step by a happy chance. It was certainly practised by the Dubilier Company in the autumn of 1926, so that the only relevant issue is whether the use was publie in the sense of the statute, which here means whether it was secret or experimental. On that issue the patentee has the burden, once the use is proved, and he must establish it by stronger proof than in ordinary civil suits. We should have supposed this settled, were it not for language to the contrary in Austin Machinery Co. v. Buckeye Traction Ditcher Co., 13 F.(2d) 697,

700 (C. C. A. 6). The decision was by a divided court, but the reputation of the judge who wrote the opinion, particularly in patent law, is so high that no one can venture to differ from him without doubts, even when as there the question was not a turning point in the result. The rule which seems to us established has its rise in the language of Matthews, J., in Smith & Griggs Mfg. Co. v. Sprague, 123 U. S. 249, 269, 8 S. Ct. 122, 31 L. Ed. 141, and can hardly be thought to have been unnecessary to the result. The court apparently relied upon it in Root v. Third Avenue Railroad Co., 146 U. S. 210, 226, 13 S. Ct. 100, 36 L. Ed. 946, though only by quotation; but so far as we can find the point has not come up again. However, several Circuit Courts of Appeal have treated the language as authoritative. We have ourselves done so three times. Thomson-Houston Electric Co. v. Lorain Steel Co., 117 F. 249, 262; Eastman v. New York, 134 F. 844, 857; Schrader’s Sons v. Wein Sales Corp., 9 F.(2d) 306, 308. So has the First Circuit. Swain v. Holyoke Machine Co., 111 F. 408; Westinghouse, etc., Co. v. Stanley Instrument Co., 133 F. 167,174 (semble). And so also has the Third, Wendell v. American Laundry Mach. Co., 248 F. 698, 700; and the Seventh, American Ballast Co. v. Davy Burnt Clay Ballast Co., 220 F. 887, 889, 890. There is indeed something to be said for requiring an infringer to prove all the elements of the defence, not only the actual use of the invention, but its public nature; the statute does not distinguish between the two elements. On the other hand certainly when the putative use is by the patentee, it is fairer to put the burden on him, since it is he who has access to the evidence. For example, in the ease at bar, Scheeker was privy to the use, for he was employed by the Dubilier Company during all the period in question. But whatever the reasons a priori, there has now grown up such a consensus of opinion that until the Supreme Court decides otherwise, we accept this part of the opinion in Smith & Griggs Mfg. Co. v. Sprague, supra, 123 U. S. 249, 8 S. Ct. 122, 31 L. Ed. 141, as authoritative.

In the autumn of 1926, the Dubilier Company filled turn orders for condensers, one for the Victor Talking Machine Company, the other for the Willard Company. The number made for the first doés not appear, though there is1 no doubt that they were oil-cooled. For the Willard Company the total number was about one hundred thousand “sections,” that is, separate cartridges, seven being used to a condenser. Nearly all of these were oil- *862 cooled, and the only question is whether they wore sold experimentally, or secretly. A small earlier order to the Western Electric Company shipped in the spring of 1926 was unquestionably only experimental, and still earlier experimental work had been done in 1925. The Dubilier Company had tided various ways to supplement a single paraffin “impregnation” to avoid the defects which we have already mentioned. Schecker’s was one of these; Weiss’s was another. Weiss was employed by the company as a “researcher”, and his experiments went along with Scheeker’s. He used petrolatum, always mixed with some oil, and when heated, having a resemblance to Seheeker’s oil. Seheeker condensers showed some tendency to leak oil, and Weiss thought that by using a heavier unguent this might be avoided. But the only testimony we can find that the sales were in any sense experimental, is that of an employee, named Menut, that the Willard shipments were made as “a sort of test” and were “worked out to find its feasibility as a manufacturing proposition.” Dubilier referred, we think, to the earlier shipment to the Western Electric Company or to the confessed experiments earlier than that; so did the other witnesses, or at least it is not clear that they did not. While it is true that Weiss did not originally have any faith in Seheeker’s method, and that after the Willard and Victor shipments, petrolatum for a season took up the company’s attention, it is not true, at least it is not undisputed, that Seheeker’s method was abandoned, to be resumed only after February, 1927. Apparently it continued to be used to some extent, although hopes were held that petrolatum might turn out to be better. This trial of a substitute was eertainly no evidence that the sales were ex- • perimental when made. Nor was that substitution an indication that Sehecker’s process had been abandoned. It stood at hand in ease Weiss failed; indeed there is some evidence that it became factory practice before March, 1927, though we do not so hold.

The true situation appears to have been that the invention was not in the experimental stage, if by that is meant that the inventor was testing it with an eye to perfecting it. His work was done; indeed he had no technical knowledge with which to proceed, for he merely chanced on the idea. At most all that remained was to see whether it would work well under service conditions. There are occasional expressions in the books which read as though the period of experiment is only that during which the inventor is trying to reduce the invention to a stable form, to adapt it completely to its purposes; but we should hesitate to hold that his privilege is so limited. For example, in Elizabeth v. Pavement Co., 97 U. S. 126, 24 L. Ed. 1009, it did not appear that Nicholson, the inventor, delayed for any other reason than to learn how well his pavement would wear; apparently it was already as good as he hoped to make it. At any rate we shall assume that an inventor may wait longer, may wait until he learns whether his invention is of enough value to justify an application for a patent. On this view he may test it, not only to put it into definitive form, but to see whether his ideas are worth exploiting.

But this added privilege has its limit. If in so doing he does in fact exploit the completed invention commercially he takes a chance that he may lose his patent. It is not enough that he intends incidentally to test its value; he may not so reserve his rights; his primary purpose must be the test, though incidentally he may get what profit he can. This is the doctrine of Smith & Griggs Mfg. Co. v. Sprague, supra, 123 U. S. 249, 256, 8 S. Ct. 122, 31 L. Ed. 141, and it has never been questioned. No doubt it is hard to apply, but so axe many legal doctrines. Some such limitation is clearly necessary if the privilege of experimenting is not to bo abused; the policy of the statute is to prevent an inventor from reserving his monopoly while the art is becoming dependent upon the invention. It gives him two years after he has begun to exploit it, and that exploitation is to be judged by his major purpose; its effect cannot be cancelled because at the back of his mind he hopes to gain more knowledge of the extent of his success.

Tested by this doctrine, the plaintiff has failed to carry its burden. The very substantial shipments in 1926 were in ordinary course of business; they were commercial in that they were for profit; the customer was not advised that they were experimental; the new process was used indiscriminately with the old. Though it may have been true that the manufacturer hoped in this way to get some light upon how well they worked, the main object was apparently not that. At least the primacy of that purpose is doubtful; the plaintiff has certainly not shown by “full, unequivocal and convincing” proof that the sales were mainly to test out the invention. It has failed to meet the undoubted prior use.

That the use was not secret is clearer. There was no effort to keep the process from the employees of the Dubilier Company, a1 *863 large manufacturer with a numerous corps of workmen. Even the roping off of those parts of the rooms where it was practiced when certain visitors come, was not for secrecy, but because the floors were slimy and slippery. Moreover it is at least open to question whether the process could not be detected by inspection of the condensers themselves after they came on the market. That was eoneededly possible later, when the art had become familiar with them, and it was a sufficient disclosure. Hall v. Macneale, 107 U. S. 90, 97, 2 S. Ct. 73, 27 L. Ed. 367; Grasselli Chemical Co. v. National Aniline & Chemical Co., 26 F.(2d) 305, 306 (C. C. A. 2). The plaintiff argues that this was not true in 1926, but the record does not affirmatively bear this out; once more it has failed to carry the burden of proof.

Decree affirmed.