The opinion
FERNANDEZ, Circuit Judge,
concurring and dissenting:
I agree with Parts I and with much of Parts III and IV of the opinion. I do not, however, agree with Part II nor, of course, with the conclusions in Parts III and IV which are influenced by the majority’s conclusion in Part II.
My difficulty with Part II is with its construction and rejection of Rosette v. Rainbo Record Mfg. Corp., 354 F.Supp. 1183 (S.D.N.Y.1973), affirmed, 546 F.2d 461 (2d Cir.1976), and with its concomitant explication of copyright law.
While Rosette may present some difficulty, it does grapple with the conundrum posed by White-Smith. See White-Smith Music Pub. Co. v. Apollo Co., 209 U.S. 1, 17-18 , 28 S.Ct. 319, 323 , 52 L.Ed. 655 (1908). If a record (or piano roll) is not a copy of the underlying *817 work, how can it result in publication of that work? See Corcoran v. Montgomery Ward & Co., Inc. 121 F.2d 572, 573-74 (9th Cir.) (records treated like piano rolls), cert. denied, 314 U.S. 687 , 62 S.Ct. 300 , 86 L.Ed. 550 (1941). Rosette held that because a record is not a copy, it is not a publication. 354 F.Supp. at 1191-92 . 1
In that, I think Rosette was quite logical, indeed correct. 2 I think it takes some very fancy footwork to duck White-Smith and, thus, convert a recorded performance into a publication of the underlying work. Like Rosette , I do not believe it can be done. If that is true, I, like Rosette , must confront the argument that it might appear that a common law copyright holder has greater rights than a person who has actually registered his copyright under the 1909 Act. Rosette avoided that difficulty by deciding that the actual protective part of copyright — the right to recover for another’s use of recordings of the musical work — would have to await the registration of the work, as contemplated by 17 U.S.C. § 1 (e) (1909). Thus, the author of the copyrighted work really could not obtain any greater protection than that available to the author of a registered work.
Of course, Rosette did not directly speak to the length of protection under the 1909 Act, once an author actually did register. That is a problem which confronts us and which causes the majority to eschew the Rosette analysis. But it seems to me that the reasoning process which drove Rosette supplies the answer to the length issue as well. That is, the author cannot recover the § 1(e) royalty beyond the time that an author could have recovered it had the author registered the work when he first released a recorded performance of it. Thus, the author who does not register in a timely fashion cannot artfully extend the time during which he can exploit his work. No doubt, had the author actually published the musical work — as opposed to releasing a recording of its performance — the issue would be a different one, but we need not deal with that issue in this case.
I do not think that either Rosette or my gloss upon it makes a tidy package. I do, however, think that it is the proper way to make White-Smith and the 1909 Act compatible. No other view seems better to me. If a record is not a copy, then placing a copyright notice upon the record itself would do no good at all because the notice is to be affixed to “each copy.” 17 U.S.C. § 9 (1909) (In 1947 this became § 10.) Thus, under the majority’s view the result is that a record is a publication for the purpose of divesting the author’s copyright protection in the underlying musical work, but it would not be a publication for the purpose of investing that musical work with protection. The fact is that Congress devised a scheme to protect authors by allowing them to collect royalties from others who did not copy the author’s musical work when they made records. That should not be taken as an intent by Congress to divest authors of protection for their own musical works when they issue records, which, perforce, are not copies of the musical work either. At least the untidy package I suggest protects authors, and I see no reason to accept the untidy package wrapped by the majority, which protects those who would poach on the author’s creativity.
For the reasons outlined by the majority, the difference between my approach and that of the majority makes no difference in the result regarding the recorded 1948 and 1950 versions of Boogie Chillun. Moreover, there will be no difference as to the 1970 version if a copyright under the 1909 Act was otherwise secured when that version was released. *818 However, our difference could affect the result if no steps were taken to obtain a copyright under the 1909 Act at or before the time that the recorded 1970 version of Boogie Chillun was issued. 3
Thus, I dissent in part but concur in the remainder of the opinion.
. As Rosette indicated, both the fact that the United States became a party to the Universal Copyright Convention, which adopted the view that a recording is not a publication, and the fact that Congress could have, hut did not, abolish White-Smitk when it created the 1909 Act buttressed the no-publication conclusion. Id. at 1189, 1191-92 .
. It seems to me that Rosette's reasoning was preferable to the ipse dixits of a number of other courts that opined, without explaining the reason, that the publication of the record (a non-copy of the underlying work) still managed to publish the underlying work itself. See, e.g., McIntyre v. Double-A Music Corp., 166 F.Supp. 681, 682-83 (S.D.Cal.1958); Mills Music, Inc. v. Cromwell Music, Inc., 126 F.Supp. 54, 69-70 (S.D.N.Y.1954); Shapiro, Bernstein & Co., Inc. v. Miracle Record Co., Inc., 91 F.Supp. 473, 474 (N.D.Ill.1950).
. I express no opinion on other issues or defenses that might affect the decision in this case.