ALDRICH, Chief Judge.

This is a diversity action brought by plaintiff-appellant Burns in the district of Massachusetts against Massachusetts Institute of Technology and a number of its employees, hereinafter collectively MIT, and Research Corporation. It is conceded that Massachusetts law governs.

Plaintiff alleges that in 1943 he was negotiating with an agency of the United States known as NDRC with respect to certain secret ideas of his; that in 1944 he was informed by MIT that it had been requested to assess his ideas and furnish NDRC with an opinion; that MIT stated that "plaintiff's disclosures would be held in confidence and that the plaintiff's interests would be protected in the hands of the defendant MIT"; that relying thereon plaintiff communicated his ideas to MIT, and that "MIT acted in a fiduciary relationship toward the plaintiff, and received the information and ideas so given to [it] by the plaintiff in utmost confidence, and under a duty not to disclose the same, nor to use the same to the advantage of * * * MIT * * * and to act in good faith toward the plaintiff in all respects, particularly in his quest for governmental sponsorship of the information and ideas. * * *" The complaint further alleges that MIT gave NDRC an unfavorable report on the plaintiff's ideas but proceeded, without his knowledge, to develop them itself, and made improper disclosure thereof; that as a result it obtained a certain patent known as Jarosh, No. 2,752,791, issued July 3, 1956, and assigned it to Research; that in furtherance of its misuse MIT maliciously drove plaintiff out of the engineering profession; that the defendants concealed what they had done, but that in 1958 plaintiff learned of the Jarosh patent; that "upon examination of said patent, some further investigation, and reflection, the plaintiff became aware of the several breaches of trust herein complained of," and promptly made demands upon MIT, which a high officer thereof promised to investigate, and that, following interim communications, he was informed in 1961 that MIT would do nothing for him. In his prayers for relief plaintiff prays that the court "will impress a trust on the Jarosh patent and any other inventions or devices, patented or not, resulting from the unlawful use * * * of * * * ideas disclosed in confidence * * *" and for damages.

The suit was brought in 1965. Defendants moved for dismissal or summary judgment, relying, inter alia, on the statute of limitations. Affidavits and cross-affidavits were filed. The court denied the motions, rendering a memorandum opinion. Thereafter the defendants filed additional affidavits and renewed their motion. Either feeling that the further affidavits met the doubts previously expressed, or, as plaintiff contends, changing its mind, the court ordered summary judgment for all defendants and dismissed the complaint. Plaintiff appeals.

Much of plaintiff's brief is devoted to the contention that the additional affidavits made no material change, and that since they did not, the granting of summary judgment was erroneous because the first ruling constituted the law of the case. This is singular logic. We agree with the plaintiff that the defendants would have had no ground of complaint if the court had refused to consider a second motion for judgment. However, the court was not bound to perpetuate error if it later believed it had committed such. See Dictograph Products Co. v. Sonotone Corp., 2 Cir., 1956, 230 F.2d 131, 231 F.2d 867, petition for cert. dismissed 352 U.S. 883, 77 S.Ct. 104, 1 L.Ed.2d 82. Moreover, regardless of what action the court took, it could not be that the first ruling became the law of the case so as to be impervious to review on appeal. Seemingly plaintiff would have it that we first determine that the second order was wrong because it departed from the first, and then remand for trial even if it be apparent that any judgment for the plaintiff would have to be vacated. We accept no such necessity. "[N]o assumed 'law of the case' can require us to reverse a just and ultimately necessary judgment, so that the District Court may arrive at it in a more roundabout way." Frost v. Bankers Commercial Corp., 2 Cir., 1952, 194 F.2d 505, 508; cf. Vandenbark v. Owens-Illinois Glass Co., 1941, 311 U.S. 538, 61 S.Ct. 347, 85 L.Ed. 327; White v. Higgins, 1 Cir., 1940, 116 F.2d 312, 317.

Coming to the merits, the sole question is whether the statute of limitations had run. For breach of contract the Massachusetts statute is six years. Mass. G.L. c. 260 § 2. At least at first blush the statute would appear to have started to run, at the latest, in 1958 when the plaintiff discovered that MIT had misappropriated his ideas in breach of its agreement, and not upon the final communication received in 1961. In the ordinary case a plaintiff's discussions with a defendant do not postpone the accrual of the cause of action until he is satisfied that the defendant is going to do nothing for him. Aetna Casualty Surety Co. v. Bell, 1 Cir., 1968, 390 F.2d 612. Plaintiff asserts that this is not the ordinary case. Rather, he says, MIT was a trustee, and with respect to breaches of trust the statute of limitations does not begin to run until there has been repudiation of the trust to the knowledge of the plaintiff, which knowledge he did not acquire until 1961. We will consider the several parts of plaintiff's contention separately.

In the first place, while it may be questioned whether the complaint alleges a true trust, as distinguished from a fiduciary relationship, in his brief plaintiff asserts that "the case at bar is either one of express trust or one of resulting trust," constantly refers to MIT as a trustee, and cites many authorities involving trusts. We think it clear, however, that there was no such trust. Indeed, on plaintiff's basic position there could not be. Plaintiff did not approach MIT to entrust it with his ideas; MIT approached plaintiff, informing him that it was acting for NDRC. Whatever duties MIT owed, it to NDRC. It could not become a trustee for plaintiff with respect to the same property; such an undertaking would have been a clear conflict of duties. In fact, MIT did not in any fashion receive "title" to plaintiff's ideas, see fn. 1, supra, the basic requirement for a trust. The mere assertion that plaintiff reposed "trust and confidence" in MIT did not create the legal relationship of trustee and cestui. Plaintiff's cases to the effect that the statute of limitations does not run against a trustee until there has been "repudiation" of a trust, e.g., Hanrihan v. Hanrihan, 1961, 342 Mass. 559, 174 N.E.2d 449, and Akin v. Warner, 1945, 318 Mass. 669, 63 N.E.2d 556, are not in point. The rule does not apply to a trust imposed as a remedy by the court. Currier v. Studley, 1893, 159 Mass. 17, 33 N.E. 709; Scott, Trusts §§ 219.1-2, 409, 481.1 (2d ed. 1956).

Plaintiff is on possibly firmer ground when he points out that the word "trust" is used in another sense, and that there may be a looser, but nonetheless recognized, fiduciary relationship which will have an effect upon the statute of limitations. Such a relationship occurs when one party reposes, to the other's knowledge, trust and confidence under circumstances in which the other's failure to make disclosure would be inequitable. E.g., Reed v. A.E. Little Co., 1926, 256 Mass. 442, 152 N.E. 918; Broomfield v. Kosow, 1965, 349 Mass. 749, 212 N.E.2d 556. Breach of the resulting duty of disclosure is a substitute for the active fraud normally required to constitute fraudulent concealment within the Massachusetts tolling statute, Mass. G.L. c. 260, § 12. Stetson v. French, 1947, 321 Mass. 195, 72 N.E.2d 410, 173 A.L.R. 569, and cases cited. A duty of this sort must sometime come to an end. Cf. Naukeag Inn, Inc. v. Rideout, 1966, 351 Mass. 353, 220 N.E.2d 916. We would think that where MIT's sole undertaking was to assist plaintiff in his dealing with NDRC and, so far as appears, he never heard from it again after this ended in 1944, plaintiff could not reasonably have considered MIT to have further obligation to him (except the negative one of nondisclosure of his secret ideas) long before 1958. However, for the purposes of this opinion, we will assume in plaintiff's favor both that MIT in fact had assumed a fiduciary relationship towards him, and that it continued until 1958. Even on that assumption plaintiff cannot succeed.

Plaintiff asserted in his complaint, as previously quoted, that in 1958 he learned "of the several breaches of trust herein complained of." The affidavits and counter-affidavits, construed most favorably in plaintiff's favor, do not detract from, but amplify this admission. Any relationship was then ended. Furthermore, there was no longer any concealment, fraudulent or otherwise. See Brackett v. Perry, 1909, 201 Mass. 502, 87 N.E. 903; cf. Tracerlab, Inc. v. Industrial Nucleonics Corp., 1 Cir., 1963, 313 F.2d 97. The fact that, as plaintiff now says, he may not have known the full extent of the use MIT may have made of his ideas, cannot mean that the cause of action was not apparent. Although thereafter plaintiff may have explored, and MIT considered, the possibility that MIT would review the matter, such discussions subsequent to a known breach do not toll the statute. Aetna Casualty Surety Co. v. Bell, supra.

As a matter of interest we consider briefly special questions with regard to Research. This defendant did not, at least expressly, enter into a contractual relationship with the plaintiff, or undertake a duty towards him. Having no duty, it is difficult on the facts alleged to see the basis for charging fraudulent concealment by this defendant, or for tolling the statute even if MIT was subject to that disability. This would seem so even if the patent had been held by MIT, as Research's assignor, in trust. Merriam v. Hassam, 1867, 96 Mass. (14 Allen) 516, 522 (assignee had record notice of trust; beneficiary has no right to await "repudiation" by assignee); cf. Harlow v. Dehon, 1872, 111 Mass. 195, 199 (constructive trustee liable only "in simple contract"). A fortiori it would seem so if MIT was not a trustee, but had a mere fiduciary duty. Bates v. Preble, 1894, 151 U.S. 149, 14 S.Ct. 277, 38 L.Ed. 106. In that case the court, having before it a broker sued in tort for conversion after having sold shares on the orders of a thieving fiduciary with knowledge of the theft, held that, unlike the fiduciary, the broker had no duty to disclose under the Massachusetts fradulent concealment statute. We add, without need of further consideration, that if Research's alleged liability as assignee lies strictly in tort, and if Research's wrongdoing occurred, as appears to be the case, only in 1956, the Massachusetts statute was not six, but two years. Mass. G.L. c. 260, § 2A.

The sum of all this is that under any view of the facts suit was brought too late. The judgment of the district court is affirmed.