1. The Hungarian publisher, the proprietor at the time of the copyright registration on September 14, 1936, was a citizen of a foreign country with which the United States has a treaty extending copyright protection to Hungarian citizens in accord with § 8(b).1 As publication in Hungary occurred on November 11, 1935, the registration followed publication, and therefore § 9, not § 11, applied. As on the date of publication the author was a citizen of Hungary, and the song had then been published solely in a foreign state, there was compliance with § 12, as amended in 1914, by the deposit of one complete copy.2 The trial judge correctly found that “no printed copies * * * were ever distributed, offered for sale, sold or disposed of in the United States.” The letter of November 4, 1940, from Cummins to Pasternak, enclosing a copy of the song, was not a publication or offering for sale in the United States.3 Nor were the playings of the song here,4 nor was the filing of the copy in the copyright office.5 The sales of imported copies in this country were not shown to have been authorized by the then proprietor. It follows that the mistake of date in the notice of copyright was not, on any theory, a violation of §§ 9 and 18; for § 9 merely requires that the notice be affixed to each copy “published or offered for sale in the United States by authority of the copyright proprietor.” We construe the statute, as to a publication in a foreign country by a foreign author (i.e., as to a publication described in the 1914 amendment), not to require, as a condition of obtaining or maintaining a valid American copyright, that any notice be affixed to any copies whatever published in such foreign country, regardless of whether publication first occurred in that country or here, or whether it occurred before or after registration here.6

It seems to be suggested by some *487text-writers6a that, under the 1914 amendment, where publication abroad precedes publication here, the first copy published abroad must have affixed to it the notice described in § 18. Such a requirement would achieve no practical purpose, for a notice given by a single copy would obviously give notice to virtually no one. There is no doubt textual difficulty in reconciling all the sections, as has been often observed; the most practicable and, as we think, the correct interpretation, is that publication abroad will be in all cases enough, provided that, under the laws of the country where it takes place, it does not result in putting the work into the public domain. Assuming, arguendo, that plaintiffs publication in Hungary did not do so, it could not affect the American copyright that copies of his song were at any time sold there without any notice of the kind required by our statute, and it would therefore be of no significance, in its effect on the American copyright, if copies sold in Hungary bore a notice containing the wrong publication date. On that assumption, there would be no need to consider whether, had the notice with the mistaken date been affixed to copies published or offered for sate in the United States by authority of the proprietor, that mistake would have invalidated the copyright,7 especially in the light of § 20. We do not know whether the publication in Hungary was such as to amount to dedication in that country, but, as we are affirming the dismissal of the complaint for other reasons, it is not necessary to decide that question.

2. In a suit like this, plaintiff, to make out his case, must establish two separate facts: (a) that the alleged infringer copied from plaintiff’s work, and (b) that, if copying is proved, it was so “material” or “substantial” as to constitute unlawful appropriation.8 Plaintiff here must lose for failure to establish the first of these facts.

The evidence by no means compels the conclusion that there was access; on the other hand, it does not compel the conclusion that there was not. Consequently, copying might still be proved by showing striking similarity. Here similarity exists; indeed, a passage in Franchelti’s “verse” is identical with one in plaintiff’s “chorus.” Mere similarity is not enough; but here one finds more; both to the eye and ear, the identity is unmistakable, as defendants virtually concede. But defendants explain this fact by saying that, quite independently, both composers utilized a common source — either Dvorak’s composition or the older commonplace theme which Dvorak had adopted and adapted.

*488As, however, both optically9 and aurally, plaintiff’s treatment is distinguishable from Dvorak’s and also from the older commonplace theme, that explanation would not wash, were plaintiff’s -contribution highly original.10 In an appropriate case, copying might be demonstrated, with no proof or weak proof of access, by showing that a single brief phrase, contained in both pieces, was so idiosyncratic- in its treatment as to preclude coincidence. In such circumstances, stimulation by the same stimulus would not serve as a defense: Buchanan tells us that Kekulé’s “idea of the carbon-ring came out of the lurid imagery of a morning after a party”; 11 many a chemist had had a like experience without such a fruitful result. Hamilton reported of his great mathematical discovery that “the Quaternions started into life, or. light, full grown, on the 16th day of October, as I was walking with Lady Hamilton to Dublin, and came up to Brougham Bridge”; no other mathematician who had observed a bridge when strolling with his wife in mid-October had made the same discovery.12 Nor would it be alone enough that the passage in question is brief13 or that the identical matter in plaintiff’s song is found in the “chorus,” and, in Franchetti’s, in the “verse.” Nor would Franchetti’s musical reputation and achievements answer,14 for Handel ruthlessly plagiarized;15 we do not accept the aphorism, “When a great composer steals, he is ‘influenced’; when an unknown steals, he is ‘infringing.’ ”16

On the issue of copying, it was proper for the trial judge to avail himself of (although not to be bound by) expert testimony. He heard the experts of both sides. In effect, he found that plaintiff’s method of dealing with the common trite note sequence did not possess enough originality, raising it above the level of the banal,17 to preclude coincidence as an adequate explanation of the identity. We cannot say that the judge erred.18 Whether, had he reached a contrary conclusion, we would have affirmed, we do not consider. Affirmed.

(concurring in the result).

1. The opinion holds that American copyright is secured by publication abroad without the notice of copyright admittedly required for publication here. This novel conclusion, here suggested for the first time, seems to me impossible in the face of the statutory language that the person *489thereto entitled “may secure copyright for his work by publication thereof with the notice of copyright required by this title,” § 9 of the Copyright Act, 17 U.S.C.A. § 9, and § 18, defining the “notice of copyright required by section 9 of this title,” with the provision that as to a work of the character here involved “the notice shall include also the year in which the copyright was secured by publication.” It is against the view of such expert copyright judges as Hough, J., in Italian Book Co. v. Cardilli, D.C.S.D.N.Y., 273 F. 619, and Universal Film Mfg. Co. v. Copperman, D.C.S.D.N.Y., 212 F. 301, affirmed 2 Cir., 218 F. 577, certiorari denied 235 U.S. 704, 35 S.Ct. 209, 59 L.Ed. 433, and Woolsey, J., in Basevi v. Edward O’Toole Co., D.C.S.D.N.Y., 26 F.Supp. 41,1 and apparently the universal assumption of text writers. See Howell, The Copyright Law, 1942, 73; Ladas, The International Protection of Literary and Artistic Property, 1938, 698; Ball, The Law of Copyright and Literary Property, 1944, 217; Copyright Protection in the Americas (Law & Treaty Series No. 16) 66; 18 C.J.S. Copyright and Literary Property, § 66, p. 190.

While the ground of the decision is not made clear, apparently it is based upon the second part of § 9, reading as follows: “and such notice shall be affixed to each copy thereof published or offered for sale in the United States by authority of the copyright proprietor, except in the case of books seeking ad interim protection under section 21 of this title.” But this deals with the preserving of the copyright after the original publication lias secured it, Sieff v. Continental Auto Supply, Inc., D. C.Minn., 39 F.Supp. 683; Fleischer Studios v. Ralph A. Freundlich, Inc., 2 Cir., 73 F.2d 276, certiorari denied Ralph A. Freundlich, Inc. v. Fleischer Studios, 294 U.S. 717, 55 S.Ct. 516, 79 L.Ed. 1250; Basevi v. Edward O’Toole Co., supra; Record & Guide Co. v. Bromley, C.C.E.D.Pa., 175 F. 156; 18 C.J.S., Copyright and Litefary Property, § 71, p. 193, and is indeed the only direct requirement for notice of the already acquired copyright. Other sections rest upon such a requirement, e.g., § 18 as to the form of notice, § 19 as to its location on the publication, and § 20 dealing with the effect of accidental omission of notice from a copy or copies. The second part of § 9, therefore, does not destroy the effect of what is said in the first part of the same section.

There is nothing in § 12 to support the stated thesis. That section requires deposit of copies before an action of infringement is brought, but explicitly applies only “after copyright has been secured by publication of the work with the notice of copyright as provided in section 9 of this title.” Hence its amendment in 1914, to require only one copy (instead of two) of a work by a citizen of a foreign state published abroad, while perhaps affording some additional evidence that § 9 was intended to include publication abroad, 18 C.J.S., Copyright and Literary Property, § 66, p. 191, contains nothing to suggest the exception here read into § 9.2 Moreover, the reference in § 9 to hooks seeking ad interim protection under § 21 is significant; the latter section affords protection to a limited class of publications — books first published abroad in the English language — under a special procedure; all others must follow the general procedure and preserve their copyright in America by affixing the required notice to copies published or offered for sale. The provision does include, at least by implication, the rule settled by United Dictionary Co. v. G. & C. Merriam Co., 208 U.S. 260, 28 S.Ct. 290, 52 L.Ed. 478, that notice of copyright must be carried only on copies published or offered for sale here; but it does not suggest an exception, operating against American authors, in the process of originally securing the copyright by publication.3

*490The opinion seeks further support because the requirement would achieve “no practical purpose.” There perhaps may be some doubt as to the utility of any notice; it is said not to be required in “most foreign countries.” Howell, 73. But if Congress thought it a necessary requirement for the literary monopoly it granted, common fairness would seem to suggest that it apply also to publication abroad, or at least that foreign publication be not made notoriously easier and more profitable than domestic publication. And the required notice does furnish a certain amount of information and warning to competitors and possible infringers, perhaps enough-to warn them away from infringement in many an obvious case. That more drastic requirements might have accomplished more does not justify elimination of those which were specifically retained.

Since, therefore, the record shows incorrect dating of the Hungarian publication, we must face the decision in Baker v. Taylor, C.C.S.D.N.Y., Fed.Cas.No. 782, that statement of a later than the actual date invalidates a copyright, and the general view that it still represents the law. See, e.g., American Code Co. v. Bensinger, 2 Cir., 282 F. 829, 836; Howell, 66; Ladas, 746; Shatter, 98. True, in Baker v. Taylor,. supra, there was more than a mistake, for the error was persisted in after it was discovered. But it is difficult to get away from the rationale suggested in the cases that stating a later date is a more serious mistake than stating an earlier date, for the former may be a wrong to the public as extending the term, while the latter at most only penalizes the copyright owner. Callaghan v. Myers, 128 U.S. 617, 654, 657, 9 S.Ct. 177, 32 L.Ed. 547. Sec. 20 of the Act does not help, for this is not the case of omission of the notice by accident or mistake “from a particular copy or copies,” but a mistake in the publication by which the original copyright was secured. The liberalizing trend shown by the Act of 1909 does not appear to have extended to this requirement, and I fear that on the authorities we must hold the copyright invalid. On that ground alone, and with some hesitation, I would support the judgment below.

2. Except for the question as to the copyright I would think the plaintiff quite entitled to judgment, since, notwithstanding the apparently trifling value of the verse, as distinguished from the refrain, of popular music, the evidence is strong as to copying and access and defendants’ answering evidence is significantly weak. Franchetti’s verse was a “Chinese copy” of plaintiff’s song in its significant parts, and so similar in the others that, as the trial judge said, substantial identity does not appear to be seriously controverted. As to the reference to Dvorak’s “Humoreske,” since plaintiff’s song is quite different in rhythmic values, setting, and arrangement from the latter, the similar musical progression found in the midst of certain measures of the latter (in itself simple enough before touched by Dvorak’s genius) does not prevent copyright by plaintiff; and the only use of the material is in the attempt — clearly unsuccessful — to make Franchetti’s disclaimer of infringement seem “plausible.” Sheldon v. Metro-Goldwyn Pictures Corp., 2 Cir., 81 F.2d 49, 54, certiorari denied Metro-Goldwyn Pictures Corp. v. Sheldon, 298 U.S. 669, 56 S.Ct. 835, 80 L.Ed. 1392; Fred Fisher, Inc. v. Dillingham, D.C.S.D.N.Y., 298 F. 145.

Proof of access on the part of Pasternak was direct and immediate while both were in Hungary; and again there was direct evidence of submission of the song to him at the very time use of Franchetti’s song was becoming important in November, 1940. True, Franch'etti claims a somewhat indefinite date of composition which would be earlier, and shows a transfer of “Dancing Doll” to his lyricist in April, 1940. But defendants’ testimony is convincing on the point noted above, that the refrain is practically everything, the verse being essentially only filler for the sheet music trade. And Franchetti in his April, 1940/ transfer warranted in explicit detail that his song was original, which can be true' only on the assumption, supported by the surrounding circumstances, that his verse was not written until the need for it arose. In the light of this, the weak general denials of his interrogatories and the strange absence of supporting testimony from Pasternak and his lyricist seem to me pretty *491conclusive. Even though the judge did not see or hear the important witnesses, we cannot disregard his findings for that reason, Advisory Committee’s Note to F.R. 52(a) ; 3 Moore’s Federal Practice, 1945 Cum.Supp. 116; but I think we are entitled to reject his conclusion because he gave no weight to defendants’ failure, with the witnesses obviously available to them, to rebut plaintiff’s persuasive case.

I may add that if in Arnstein v. Porter, 2 Cir., 154 F.2d 464, copying could be suggested by mere sound in the case of dissimilar compositions, this exact reproduction, together with evidence of access, would seem to me to show copying a fortiori. This the opinion at first seems to accept, in its various steps refuting the grounds taken below — particularly the claimed defeat of copyright because of similarity to the “Humoreske” — and it supports this result with some intriguing historical allusions, which, however, have relevancy only in support of a reversal. But ultimately it comes to rest upon a point directly opposed to that taken below, the banality of the music, though the District Court had said of the melody shared between these composers, “There is no lack of ingenuity here.” 51 F.Supp. 233. Surely, if the Arnstein case teaches us anything, it must be that banality is no bar to a claim for plagiarism. That results at once so divergent and so musically astonishing as the decisions in these two cases can occur simultaneously I can attribute only to the novel conceptions of legal plagiarism first announced in the Arnstein case and now repeated here. By these the issue is no longer one of musical similarity or identity to justify the conclusion of copying — an issue to be decided with all the intelligence, musical as well as legal, we can bring to bear upon it — but is one, first, of copying, to be decided more or less intelligently, and, second, of illicit copying, to be decided blindly on a mere cacophony of sounds. Just at which stage decision here has occurred, I am not sure.