FEDERAL DEPOSIT INSURANCE CORPORATION, as Successor in Interest to New England Merchants Leasing Corporation, etc., Plaintiff, Appellee, v. OGDEN CORPORATION, et al., Defendants, Appellants, 202 F.3d 454

SELYA, Circuit Judge. The district court ordered a law firm, Dickstein, Shapiro, Morin & Oshinsky, LLP (“Dickstein”), not itself a party to the underlying action, to produce documents that the appellants, Ogden Corporation and Ogden Martin Systems of Haverhill, Inc. (collectively, “Ogden”), claim are within the attorney-client privilege. The ap-pellee, Federal Deposit Insurance Corporation (“FDIC”), asserts that the so-called “joint client exception” trumps the privilege and, thus, legitimates the order.

After providing necessary context, surmounting a jurisdictional obstacle, and charting the parameters of both the privilege and the exception, we affirm the turnover order. I. BACKGROUND In 1978, Citicorp North America, Inc. (“Citicorp”) and New England Merchants Leasing Corp. (“NEMLC”) formed a general partnership (“SBR Associates”) to develop a refuse-to-energy facility in Haver-hill, Massachusetts. Each nominated a wholly-owned subsidiary to serve as a general partner: CIC Omega Lease, Inc., for Citicorp, and NEMLC Alpha, Inc., for NEMLC.

In the early 1980s, the partners (hereinafter, with their parents and successors, sometimes collectively called “the banks”) designed and built the facility and leased it to an independent operator, Refuse Fuels, Inc. (“RFI”), on condition that RFI purchase insurance policies (“the efficacy insurance”) to protect against operational glitches leading to shortfalls in revenue. The hedge proved prudent; from the moment that the facility went on line, it was plagued with problems. In an effort to protect their investment, the banks terminated the arrangement with RFI and brought Ogden into the venture.

The details of the transaction are unimportant at this juncture, save to say that by virtue of a series of complicated agreements, Ogden acquired the banks’ interests in SBR Associates and assumed sole control of the business on December 23,1986. The operational difficulties that the facility encountered had given rise to claims under the efficacy insurance, and the parties sought to tie up this loose end. They entered into a specific agreement (“the restated assignment agreement”) with regard to those claims.

Under that agreement, Ogden was to direct the recovery effort against the efficacy insurers and pay portions of the realized proceeds (net of fees and expenses) to the banks. Betimes, Ogden would keep the banks apprised of progress. Finally, the agreement contained a mechanism whereby the banks could redeem Ogden’s interest and take direct control of the recovery effort should Ogden wish to consummate a settlement with the efficacy insurers that the banks deemed unacceptable. Ogden retained Dickstein to handle the claims against the efficacy insurers.

Meanwhile, it continued to operate the Haverhill facility, incurring additional losses (which furnished a basis for further insurance claims). The facility shut down sometime in 1990. On January 6, 1991, NEMLC’s parent company, Bank of New England, N.A., was adjudged insolvent, and the FDIC was appointed as receiver (thus becoming, in effect, successor in interest to NEMLC and NEMLC Alpha). By mid-1996, Dickstein had recovered $18,700,000 from the efficacy insurers. On August 2, 1996, a Dickstein partner, Leslie Cohen, wrote to the banks, notifying them of their allocable shares of the funds collected.

Both Citicorp and the FDIC protested the proposed, allocation, arguing that they were being shortchanged and that the terms of the restated assignment agreement were not being followed. Each demanded substantially more money. Ogden balked. Dickstein continued to prosecute the underlying litigation — at the time the parties submitted their appellate briefs, the total amounts recovered on the insurance claims exceeded $60,000,006— but it refused to become entangled in the internecine squabble over the allocation of the proceeds.

Citicorp and the FDIC sued Ogden in the district court for breach of contract and unfair business practices. In due course, the FDIC served Dickstein with a subpoena duces tecum that, inter alia, commanded production of communications between it and Ogden. Dickstein objected, citing the attorney-client privilege. The FDIC moved to compel, contending that no privilege attached because Dickstein had represented Ogden and the banks jointly in connection with the litigation against the efficacy insurers. The district court granted this motion by endorsement.

Ogden appealed the order,, and the district court stayed production pending resolution of the appeal. II. APPELLATE JURISDICTION There is a threshold issue here. Ogden premises appellate jurisdiction on 28 U.S.C. § 1291, which provides for jurisdiction over appeals taken from “final” decisions and orders of the district courts. Since the order from which Ogden purports to appeal does not conclude the litigation on the merits, it is not final in the stereotypical sense. See United States v. Metropolitan Dist. Comm’n, 847 F.2d 12, 14 (1st Cir.1988).

Nevertheless, some orders that do not themselves end litigation are deemed final (and thus immediately appealable) under the collateral order doctrine. See Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 546-47, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949). To qualify for this sanctuary, an order must conclusively resolve an important question distinct from the merits and yet be unreviewable, as a practical matter, in a conventional end-of-case appeal. See Cunningham v. Hamilton County, 527 U.S. 198, 119 S.Ct. 1915, 1920, 144 L.Ed.2d 184 (1999); Swint v.

Chambers County Comm’n, 514 U.S. 35, 42, 115 S.Ct. 1203, 131 L.Ed.2d 60 (1995). The compass of this exception is “narrow,” Quackenbush v. Allstate Ins. Co., 517 U.S. 706, 712, 116 S.Ct. 1712, 135 L.Ed.2d 1 (1996), and discovery orders generally are not thought to come within it. See Insurers Syndicate for the Joint Underwriting of Medico- Hosp. Prof'l Liab. Ins. v. Garcia, 864 F.2d 208, 210 (1st Cir.1988).

One reason that most discovery orders do not fall within the collateral order exception is because they do not meet the “otherwise effectively unreviewable” requirement; the party resisting the discovery order “can gain the right of appeal ... by defying it, being held in contempt, and then appealing from the contempt order, which would be a final judgment as to [him].” Corporacion Insular de Seguros v. Garcia, 876 F.2d 254, 257 (1st Cir.1989). This praxis—insisting upon disobedience followed by contempt as a condition to reviewability—is commonly called the Cobbledick rule. See Cobbledick v.

United States, 309 U.S. 323, 328, 60 S.Ct. 540, 84 L.Ed. 783 (1940). The rule serves efficiency interests because it encourages reflection both by the party seeking discovery and by the party resisting it. See 15B Charles Alan Wright, Arthur R. Miller & Edward H. Cooper, Federal Practice and Procedure § 3914.23, at 154 (2d ed.1992). The rationale underlying the Cob-bledick rule is distorted, however, when a discovery order runs to someone other than an adverse party (a phenomenon that occurs when, say, a court enforces a subpoena duces tecum served upon a non-party).

Since a third person “presumably lacks a sufficient stake in the proceeding to risk contempt by refusing compliance,” Church of Scientology v. United States, 506 U.S. 9, 18 n. 11, 113 S.Ct. 447, 121 L.Ed.2d 313 (1992), this circumstance justifies a different approach. Under what has been termed the Perlman rule, a discovery order addressed to a non-party sometimes may be treated as an immediately appealable final order vis-á-vis a party who claims to hold an applicable privilege. See id.; see also Perlman v. United States, 247 U.S. 7, 12-15, 38 S.Ct. 417, 62 L.Ed. 950 (1918).

Courts frequently have invoked Perlman when a client (who is herself a party or a grand jury target) seeks to appeal an order compelling her attorney (who is neither a party nor a target) to produce allegedly privileged materials. See, e.g., In re Sealed Case, 146 F.3d 881, 883 (D.C.Cir.1998); In re Grand Jury Subpoenas, 123 F.3d 695, 699 (1st Cir.1997); Conkling v. Turner, 883 F.2d 431, 433-34 (5th Cir.1989); In re Grand Jury Subpoena Duces Tecum, 731 F.2d 1032, 1036 n. 3 (2d Cir.1984).

On its face, this appeal appears to fit the classic Perlman mold: the FDIC directed a subpoena duces tecum at the law firm (a non-party); the district court enforced the subpoena and ordered the firm to produce the disputed documents; compliance with that order will let the cat out of the bag, thus rendering an end-of-case appeal nugatory; and the client, although a party to the case, has no way of testing the order by allowing itself to be held in contempt. ' In this sense, the client (Ogden) is at the mercy of its quondam counsel, and an immediate appeal offers the only vehicle by which it can gain effective review of the privilege issue.

Despite this apparent match, we proceed with'circumspection. Some tension exists in our precedents as to whether the availability of immediate review in cases such as this should be gauged by the Perlman rule or by the more encompassing Cohen collateral order doctrine. In the past few years, two different panels of this court have furnished divergent answers to this question. Compare United States v. Billmyer, 57 F.3d 31, 34 & n. 1. (1st Cir.1995), with In re Grand Jury Subpoenas, 123 F.3d at 696-99.

Although these decisions do not fit tongue in groove, the distinction between them does not affect reviewability in this case. Under either approach, a substantial privilege claim that cannot effectively be tested by the privilege-holder through a contemptuous refusal ordinarily will qualify for immediate review if the claim otherwise would be lost. .This is such a case. Moreover (as we shortly shall explain), the scope of review here essentially involves what are more nearly categorized as clear-cut questions of law, reviewable no matter whether the appeal is viewed through the lens of Cohen or Perlman.

Thus, Bill-myer, fairly read and applied, does not remove the order sub judice from the class of orders that are immediately appealable under In re Grand Jury Subpoenas and the Perlman rubric. Consequently, we have jurisdiction to hear and determine Ogden’s appeal. III. STANDARD OF REVIEW Trial judges enjoy broad discretion in the handling of interstitial matters, such as the management of pretrial discovery.

As a result, an appellate court will intervene in such matters “only upon a clear showing of manifest injustice, that is, where the lower court’s discovery order was plainly wrong and resulted in substantial prejudice to the aggrieved party.” Mack v. Great Atlantic & Pacific Tea Co., 871 F.2d 179, 186 (1st Cir.1989). Ogden invites us to abandon this abuse-of-discretion standard in favor of plenary review because the district court granted the motion to compel by endorsement, without elaborating upon its thinking. We decline the invitation.

Although a lower court’s elucidation of its reasoning invariably eases the appellate task, motions often are decided summarily. We have thus far refused to insist upon a rigid rule to the contrary. See, e.g., Camilo-Robles v. Hoyos, 151 F.3d 1, 8 (1st Cir.1998), cert. denied, -— U.S. -, 119 S.Ct. 872, 142 L.Ed.2d 773 (1999); Domegan v. Fair, 859 F.2d 1059, 1065-66 (1st Cir.1988). While it is sometimes necessary to remand for specific findings when confronted with an opaque ruling, see, e.g., Francis v. Goodman, 81 F.3d 5, 8 (1st Cir.1996); Pearson v.

Fair, 808 F.2d 163, 165-67 (1st Cir.1986) (per curiam), we are aware of no authority that would allow us automatically to vary the standard of review depending on whether a district court has taken the time to explain its rationale. In all events, the question is academic here, as the record before us permits a clear understanding of why the district court ruled as it did. IV. THE MERITS In a discovery dispute, the burden to establish an applicable privilege rests with the party resisting discovery. See United States v. Construction Prods. Research, Inc., 73 F.3d 464, 473 (2d Cir.1996).

If the privilege is established and the question becomes whether an exception to it obtains, the devoir of persuasion shifts to the proponent of the exception. See McMorgan & Co. v. First Cal. Mortgage Co., 931 F.Supp. 699, 701 (N.D.Cal.1996). We look to Massachusetts law to determine the scope of both the asserted privilege and the exception in this case. See Fed.R.Evid. 501; 6 James Wm. Moore et al., Moore’s Federal Practice § 26.47[4] (3d ed.1999).

As to matters about which the Supreme Judicial Court of Massachusetts has not spoken, we take a predictive approach and seek guidance from other persuasive case law, learned treatises, and pertinent public policy considerations. Cf. Blinzler v. Marriott Int’l, Inc., 81 F.3d 1148, 1151 (1st Cir.1996) (endorsing such an approach for use in diversity cases). The privilege at issue here— the attorney-client privilege — serves important ends.

Its root purpose is “to encourage full and frank communication between attorneys and their clients and thereby promote broader public interests in the observance of law and administration of justice.” Upjohn Co. v. United States, 449 U.S. 383, 389, 101 S.Ct. 677, 66 L.Ed.2d 584 (1981). The privilege springs from the attorney-client relationship.

In Massachusetts, such a relationship comes into being “when (1) a person seeks advice or assistance from an attorney, (2) the advice or assistance sought pertains to matters within the attorney’s professional competence, and (3) the attorney expressly or impliedly agrees to give or actually gives the desired advice or assistance.” DeVaux v. American Home Assur. Co., 387 Mass. 814, 444 N.E.2d 355, 357 (1983); accord Sheinkopf v. Stone, 927 F.2d 1259, 1264 (1st Cir.1991) (applying Massachusetts law); Bays v. Theran, 418 Mass. 685, 639 N.E.2d 720, 723 (1994).

The privilege that attends the attorney-client relationship “extends to all communications made to an attorney or counsellor, duly qualified and authorized as such, and applied to by the party in that capacity, with a view to obtain his advice and opinion in matters of law, in relation to his legal rights, duties and obligations.” Hatton v. Robinson, 31 Mass. 416, 14 Pick. 416, 421 (1833). Despite its venerable provenance, the attorney-client privilege is not absolute. One recognized exception renders the privilege inapplicable to disputes between joint clients. See Beacon Oil Co. v.

Perdis, 263 Mass. 288, 160 N.E. 892, 894 (1928). Thus, when a lawyer represents multiple clients having a common interest, communications between the lawyer and any one (or more) of the clients are privileged as to outsiders but not inter sese. See Eureka Inv. Corp. v. Chicago Title Ins. Co., 743 F.2d 932, 936-38 (D.C.Cir.1984); 8 John Henry Wigmore, Wigmore on Evidence § 2312 at 603-09 (McNaughton rev. ed.1961).

As one leading treatise explains, When two or more persons, each having an interest in some problem, or situation, jointly consult an attorney, their confidential communications with the attorney, though known to each other, will of course be privileged in a controversy of either or both of the clients with the outside world, that is, with parties claiming adversely to both or either of those within the original charmed circle.

But it will often happen that the two original clients will fall out between themselves and become engaged in a controversy in which the communications at their joint consultation with the lawyer may be vitally material. In such a controversy it is clear that the privilege is inapplicable. 1 Kenneth S. Broun et al., McCormick on Evidence § 91 at 335-36 (4th ed.1992).

In determining whether parties are “joint clients,” courts may consider multiple factors, including but not limited to matters such as payment arrangements, allocation of decisionmaking roles, requests for advice, attendance at meetings, frequency and content of correspondence, and the like. See McMorgan, 931 F.Supp. at 702; In re Colocotronis Tanker Secs. Litig., 449 F.Supp. 828, 830-32 (S.D.N.Y.1978); Connelly v. Dun & Bradstreet, Inc., 96 F.R.D. 339, 342 (D.Mass.1982).

In addition, the joint client exception presupposes that communications have been “made in the course of the attorney’s joint representation of a ‘common interest’ of the two parties.” Eureka, 743 F.2d at 937. The term “common interest” typically entails an identical (or nearly identical) legal interest as opposed to a merely similar interest. See, e.g., McMorgan, 931 F.Supp. at 701; NL Indus. v. Commercial Union Ins. Co., 144 F.R.D. 225, 230-31 (D.N.J.1992). Thus, the proponent of the exception must establish cooperation in fact toward the achievement of a common objective. See Shamis v.

Ambassador Factors Corp., 34 F.Supp.2d 879, 893 (S.D.N.Y.1999). In this case, it cannot be gainsaid that Dickstein and Ogden enjoyed an attorney-client relationship in respect to the war being waged against the efficacy insurers. The record attests that the banks also were Dickstein’s clients in that struggle. They, like Ogden, had a pecuniary interest in the avails of the insurance— interests that were identical in character (albeit different in amount). They, like Ogden, desired to press for those proceeds, through litigation if necessary.

The banks pooled their interests with Ogden’s and authorized Ogden to secure the services of a law firm to mount a unified offensive to prosecute their joint claims against the insurers. The assistance that Ogden sought on behalf of itself and the banks fell well within the purview of Dick-stein’s professional competence, and Dick-stein actually rendered the desired services. In undertaking the enterprise, Dickstein unequivocally committed itself to joint representation.

When it brought suit against the efficacy insurers it entered an appearance not only for Ogden, but also for the wholly-owned subsidiary of NEMLC (and it still represents both in that litigation). These entries of appearance themselves constitute persuasive evidence of a joint client relationship. Dickstein’s subsequent actions confirm that impression. Early on (in February 1987), it wrote to both NEMLC and Citi-corp, describing an executive

summary of the strategy that it proposed to pursue on behalf of Ogden and the banks in the unified litigation. The words “ATTORNEY CLIENT PRIVILEGED COMMUNICATION” were emblazoned at the head of the first page of each letter. The text of the letter sent to NEMLC (which was materially identical to the one sent to Citi-corp) stated in relevant part: This firm has been engaged to represent the beneficiaries of a proposed claim against the efficacy insurers covering the referenced Project. We have prepared a substantive Memorandum and Executive

Summary detailing our review of the matter and the potential for recovery under the efficacy policies. Before we provide this material to you, however, in order to protect and maintain its privileged and confidential nature, we need to confirm, and obtain your acknowledgment, that with regard to the proposed efficacy claim, there exists, between New England Merchants Leasing Corporation (“NEMLC”) and NEMLC Alpha Inc. on the one hand and this firm on the other, an attorney-client relationship.

As you are aware, under the terms of the [restated assignment agreement], NEMLC and NEMLC Alpha Inc. stand to benefit from any recovery from the efficacy insurers. As such, we have been retained to represent the interests of NEMLC and NEMLC Alpha Inc. as well as all other beneficiaries under the referenced agreements.

The letter went on to request that the recipient “acknowledge the applicability of the attorney-client privilege to our relationship, and provide assurances that the contents of the referenced documents, and any others prepared by this firm, will be treated as confidential and not disclosed to anyone who is not a ‘client’ of this firm in this matter.” Finally, the letter provided a means for signifying the recipient’s agreement to the formation of the attorney-client relationship.

Both NEMLC and Citicorp executed and returned copies of the engagement letters to Dickstein, which then furnished the strategic assessment to them. If more were needed — and we doubt that it is — a surfeit of other evidence indicates the existence of a joint attorney-client relationship. We offer three examples.

The record contains (1) copious notes taken by a Citicorp representative during a telephone conference with Attorney Leveridge (a Dickstein partner) on March 30, 1988, in regard to litigation strategy; (2) copies of correspondence from Leveridge to the banks asking for assistance in responding to discovery requests and soliciting suggestions for dealing with the efficacy insurers; and (3) a copy of a letter written by a Dickstein lawyer, Paul Taskier, on September 16, 1991, responding to an inquiry from Citi-corp and furnishing a detailed

interpretation of the allocation provision contained in the restated assignment agreement (an

interpretation starkly at odds with the

interpretation that Dickstein urged upon the banks in August of 1996). This evidence points clearly and convincingly to a joint client relationship. See Bays, 639 N.E.2d at 723; DeVaux, 444 N.E.2d at 357. The entries of appearance and the engagement letters alone constitute powerful proof, and the correspondence evinces a coordinated legal strategy sufficient to lead a reasonable person standing in NEMLC’s shoes to infer that Dickstein had become its attorney. See Sheinkopf 927 F.2d at 1265; Shamis, 34 F.Supp.2d at 893.

Courts customarily determine the existence vel non of an attorney-client relationship by evaluating whether the putative client’s belief that such a relationship existed was objectively reasonable under all the circumstances. See Sheinkopf, 927 F.2d at 1265; Sky Valley L.P. v. ATX Sky Valley, Ltd., 150 F.R.D. 648, 651-52 (N.D.Cal.1993). Here, the reasonableness of the banks’ professed belief that Dickstein had become their attorney in regard to the efficacy insurance litigation is manifest. Ogden’s attempt to persuade us to a • different view lacks force.

It claims that no joint client relationship could have been forged because of the uncertainty about the eventual allocation of the litigation proceeds. This “intrinsic” adversity, Ogden says, destroyed the requisite identity of interests. See McMorgan, 931 F.Supp. at 701 (explaining that mere similarity in interests is not enough to establish a joint client relationship). This argument is spun from whole cloth.

At the crucial time — the time when Dickstein and the banks tied the attorney-client knot — no one had expressed the view that the allocation provision in the restated assignment agreement was inscrutable, and there was no reason to believe that the beneficiaries of the insurance would

part company when it came time to divide the proceeds. For aught that appeared, Ogden’s interest was entirely congruent with NEMLC’s and Citicorp’s. No more was exigible: the mere possibility of a future dispute did not prevent the formation of a valid joint client relationship. See Sky Valley, 150 F.R.D. at 662. In a related vein, Ogden asseverates that even if a joint client relationship initially existed, the relationship was dissolved once it (and, presumably, Dickstein) realized that its interests had become adverse to the banks. This asseveration rests on a false premise.

A joint attorney-client relationship remains intact until it is expressly terminated or until circumstances arise that readily imply to all the joint clients that the relationship is over. See Flynt v. Brownfield, Bowen & Bally, 882 F.2d 1048, 1051-52 (6th Cir.1989). Here, the record reveals no hint of any circumstances existing prior to Dickstein’s August 2, 1996 letter (proposing a particular allocation of proceeds) from which the banks reasonably could have inferred that their interests had become inimical to Ogden’s.

In sum, the FDIC has adduced substantial evidence to support its assertion that Ogden and the banks sought and received legal advice from Dickstein with regard to a common interest — the efficacy insurance litigation — thereby establishing that all three had forged a joint attorney-client relationship with Dickstein. See Sheinkopf, 927 F.2d at 1264; Bays, 689 N.E.2d at 723. Since that relationship remained whole until the banks received Dickstein’s letter of August 2, 1996, the joint client exception to the attorney-client privilege applies. See Beacon Oil, 160 N.E. at 894.

It follows inexorably that the claim of attorney-client privilege is impuissant with respect to documents generated on or before August 2, 1996. Affirmed. . The crux of the dispute appears to be whether, under the restated assignment agreement, the parties are to share only the proceeds of claims accrued against the efficacy insurance as of the closing date of their transaction, or also the proceeds of claims that arose thereafter. . At an earlier stage of this litigation, the FDIC filed a request for production of documents in Ogden's possession relating to the efficacy insurance litigation. See Fed.R.Civ.P. 34.

Ogden objected to furnishing communications between it and Dickstein, asserting attorney-client privilege. The FDIC moved to compel production, and the district court granted the motion. We dismissed Ogden's attempted appeal without prejudice for want of appellate jurisdiction (though the disputed documents have yet to be produced). This is a perfect example of a discovery order that is not immediately appealable: Ogden, after all, can refuse to comply with the order and thus invite a finding of contempt (or some equivalent sanction). .

The principal difference lies in the fact the Perlman rule arguably contains no limitation on the scope of review, while review under the collateral order doctrine arguably is limited to “clear-cut legal error” as opposed to challenges that seek to test either factual determinations or the application of a settled legal rule to the particular facts. Billmyer, 57 F.3d at 35. But, as the panel’s actions in Billmyer evince, see id. at 35-37 (addressing waiver issue on the merits), the scope-of-review limitation is flexible. .

With regard to the tension between Billmyer and In re Grand Jury Subpoenas, we note that two recent developments favor the latter. Billmyer relied in part on In re Oberkoetter, 612 F.2d 15 (1st Cir.1980), a decision that this court subsequently overruled in In re Grand Jury Subpoenas, 123 F.3d at 697 n. 2, 699. Billmyer likewise relied to some extent on the decision in In re Grand July Proceedings, 43 F.3d 966, 969-70 (5th Cir.1994), a decision that the Fifth Circuit since has qualified or abandoned. See In re Grand Jury Subpoena, 190 F.3d 375, 384 n. 11 (5th Cir.1999), petition for cert. filed, - U.S.L.W. (U.S.

Dec. 20, 1999) (No. 99-1046). . Ogden argues that Dickstein merely intended these engagement letters to protect certain documents from discovery by the efficacy insurers — but it has cited no respectable authority for this kind of “limited” or "nominal” attorney-client privilege, and we dismiss the notion out of hand. A law firm that says one thing and induces confidences as a result cannot later be heard to profess that it meant another. . Contrary to Ogden’s importuning, Eureka does not stand for a different rule.

There, the communications at issue were made after the interests of the joint clients diverged and their attorney, aware of the divergence, undertook separate representation of one client, distinct from the joint representation. See 743 F.2d at 937. The court’s holding addresses whether the content of these communications pertained to the common interest that formed the basis of the joint representation. See id. We decline to read Eureka as holding that a joint client relationship evaporates whenever one client unilaterally determines that its interests have diverged from those of its co-clients. .

In its present posture, this appeal requires us only to pass upon Ogden’s global claim of attorney-client privilege. To the extent (if at all) that other grounds for resisting production attach to particular documents (say, that a given item is wholly unrelated to Dickstein’s representation of the joint clients’ common interest or originated after August 2, 1996), the district court, in its discretion, may consider those objections and may, if necessary, review specific documents in camera.