(concurring in part and dissenting in part). To the extent the majority opinion generally states the principles of law pertaining to this matter, I have no major disagreement. Nor do I disagree with the court’s view as to the disposition of the “Paul Revere claim” stated in paragraph 5 of the opinion, and hence concur therein. My disagreement with the majority opinion is as to its holding that the misallocation of the burden of persuasion is but harmless error. I would remand this case for reconsideration by the trial judge in light of the majority’s opinion and in light of further considerations set forth briefly here.
This record shows that the plaintiff contested the validity of other fees charged by the law firm as well as the fee of $760,000 discussed in this appeal. Although the latter fee is the only one in issue on appeal, the other (and lesser) fees charged arose out of the same relationships and were contested below as being neither fair nor reasonable. The judge made as to each of such other fees specific findings that they were “fair and reasonable.” In light of these findings contrasted with the finding of the judge that the fee now in issue was not shown to be “unfair, unreasonable or clearly excessive” or that it lies “at the outer limit of the range of reason and fairness” such latter finding cannot be “interpreted” as a finding that the fee was “reasonable,” as the majority suggests. I agree with the majority that the judge made “elaborate findings of fact.” He was also meticulous and precise in the way he approached the matter. I would give him credit also for meaning what he says. Thus, when he further stated that *269he was “not persuaded, and has no firm and definite conviction” that the tax case fee “was unreasonable, unfair or clearly excessive in the circumstances,” I would take that to mean his mind was in relative equipoise on the issue. See Sargent v. Massachusetts Accident Co., 307 Mass. 246, 250 (1940). In such a case the proper allocation of the burden of persuasion on the defendant law firm would appear crucial to the ultimate findings of the judge. Since the judge misallocated the burden of proof on this issue (as the majority and I agree), I cannot agree that this error was “harmless” to the plaintiff’s cause. On this aspect alone, the matter ought to be remanded for clarification or further findings by the judge.
I would go a step further, although that additional step is not crucial to my views. While there was some dispute between the parties as to the nature of the fiduciary relationship between the law firm and the stockholders of Thompson Wire, the judge found that there were involved a number of relationships fiduciary in nature. First, partners of Hale and Dorr were trustees of sixty-one per cent of the 9,000 shares to be sold by Thompson Wire; second, these same partners acted in a fiduciary role as selling agents for the stockholders of Thompson Wire; third, one partner of Hale and Dorr was an officer and director of Thompson Wire for a number of years; fourth, the relationship between the defendant partners of Hale and Dorr and the former stockholders of Thompson Wire as to this fee was a fiduciary relationship; fifth, the entire agreement, escrow agreement and release of funds were managed by Hale and Dorr partners. The record also shows that Hale and Dorr partners managed the entire transaction and retained their own firm to handle the § 531 tax matter. The case is replete with fiduciary relationships and some degree of self-dealing. This is not to say that there was anything unethical involved. But one wonders whether such an arrangement ought to be dealt with, as the judge does and the majority opinion seems to accept, as an arrangement where the fee involved a “voluntary contractual relationship.” While the agreement to retain counsel *270might be viewed as voluntary, the agreement on fees was only that they be “fair and reasonable.” In these circumstances plus the fact (as the judge found) that Thompson Wire was not a major client of Hale and Dorr and the average annual fee billed to Thompson Wire prior to this case was $4,000, a finding of a voluntary arrangement as to a fee of this magnitude might properly be viewed as unwarranted. One need not explore this matter at length here. Suffice it to be said that a healthy degree of skepticism on the claim of “voluntariness” is in order — and that such skepticism is relevant to whether the burden of proof was met by the defendant law firm.
Last, the absence of a contingent fee agreement as required by the rules of this court seems significant on the question of how “voluntary” this fee arrangement was in fact. The findings of the judge create an inference that it was a contingent fee without prior agreement when he states that the $760,000 fee reflects twenty per cent of the saving on the tax in the § 531 case and ten per cent of the saving on interest, or seventeen per cent of the “total saving.” He found further that such fees were often based on the tax saving, and that a fee of fifteen to twenty per cent “undoubtedly would have been fair and reasonable in this case if it had been agreed upon in advance by a knowledgeable taxpayer.” It was not so agreed on.
In light of these factors, the cautious posture taken by the judge, and his marginal findings on the issue, the majority’s view seems open to substantial doubt when it says the judge’s error as to burden of proof was harmless. The more appropriate disposition not only for the benefit of the parties but for the benefit of bench and bar alike would be to remand this matter for further consideration. “It is difficult to conceive of anything more likely to undermine public respect for the administration of justice than a wide spread suspicion that the courts are aligned in aiding the distribution among counsel of excessive proportions of the funds of those who are unfortunate enough to become involved in controversy.” Lewis v. National Shawmut Bank, 303 Mass. 187, 191 (1939).