Appellant George Moran and two co-defendants were convicted by a jury, after a joint trial, of various drug offenses. Moran was found guilty of conspiring to distribute cocaine and was acquitted on two other counts charging him with specific acts of distribution. All of the defendants have appealed, but the evidence and issues relating to Moran differ from those concerning the other defendants and we decide his case separately. Concluding that the evidence was sufficient to sustain Moran's conviction for conspiracy and finding no other errors, we affirm.

The procedural history can be briefly stated. On August 9, 1990, Moran and a number of others were indicted under 21 U.S.C. § 846 for conspiring to distribute cocaine and, in other counts pertaining to one or more of the defendants, with distribution and related crimes. The co-conspirators charged in the umbrella conspiracy count included Moran, the alleged ringleader Hobart Willis, and others. Before trial, Willis and three others pleaded guilty. Moran and two other defendants were tried in February 1991 and convicted on one or more counts. This appeal followed.

Moran's central argument on appeal is the often made, but rarely successful, claim that the evidence was inadequate to support the verdict against him. In appraising such an argument, we "assess the sufficiency of the evidence as a whole, including all reasonable inferences, in the light most favorable to the verdict...." United States v. Lopez, 944 F.2d 33, 39 (1st Cir.1991). So viewed, we ask "whether a rational trier of fact could have found the defendant guilty beyond a reasonable doubt." Id. In general, issues of credibility are resolved in favor of the verdict. Id. "The evidence ... need not exclude every reasonable hypothesis of innocence; that is, the factfinder may decide among reasonable interpretations of the evidence." Id.

In this case Moran was tried on the charge, among others, that he conspired with Willis and his co-defendants. The "essence" of conspiracy is an agreement to commit a crime, Iannelli v. United States, 420 U.S. 770, 777, 95 S. Ct. 1284, 1289, 43 L. Ed. 2d 616 (1975), here, an agreement between Moran and others to distribute drugs. Such an agreement may, of course, be inferred from other evidence including a course of conduct. United States v. Concemi, 957 F.2d 942, 950 (1st Cir.1992). More than that, while the term "agreement" is customarily used in defining conspiracy and is properly employed in jury instructions, the agreement of the defendant with others may be implicit in a working relationship between the parties that has never been articulated but nevertheless amounts to a joint criminal enterprise.

In this case, taking the evidence most favorably to the government, the jury could have found from direct testimony, telephone recordings and other evidence that Willis was engaged in a drug distribution conspiracy with various persons during 1988. As to Moran, the evidence against him came almost entirely from one Paul Callahan, who cooperated to some extent with the Drug Enforcement Administration. Callahan's trial testimony came freighted with his long criminal record, admissions that he procured false testimony in other proceedings, and his incentive to favor the government in order to secure favorable treatment for himself. Nevertheless, his testimony was not incredible, was corroborated on certain limited points, and was essentially uncontradicted. Thus the jury was entitled to accept some or all of Callahan's testimony.

According to Callahan, he first met Moran in 1981 but had no further contact with him until June or July 1988 when he had a friend give Moran his beeper number. Callahan then met with Moran and sought to purchase cocaine from him in a substantial amount. Moran replied that he would contact the "fat man" (understood by Callahan to be Willis) with whom Moran said he was dealing at the time. At their next meeting, Moran told Callahan that the fat man's prices were too high but that Moran had another source in the North End. Moran also said that he was going to try to get a cheaper price from "Mary," a friend of the fat man later identified by Callahan as a member of Willis' ring. Subsequently, Callahan and Moran met again and Callahan purchased 500 grams of cocaine from Moran, after testing it for purity.

Some weeks later, Callahan again contacted Moran, a further meeting ensued, and Moran told Callahan--in Callahan's words--that he (Moran) was "still looking in to ingratiate with the fat guy." At the next meeting, Moran offered a package of what Callahan took to be cocaine; Moran explained that it came from the fat guy. The contents had a diesel smell and Callahan rejected the package on the ground that his own customers would not accept it. Moran left and then returned several hours later with a kilo of cocaine from an unidentified source. Callahan tested the new package and purchased a half kilo.

The final evidence relating directly to Moran involved two telephone calls between him and Callahan in October 1988. The first call was not tape recorded. According to Callahan, Moran complained during the call that federal agents were scrutinizing him. On cross-examination Callahan indicated that Moran also said during the call, "I saw the Pillsbury Boy a few days ago, but that was just to say hi.... I don't have nothing to do with those guys." Callahan told the jury that the Pillsbury Boy was Willis.

The second conversation occurred a week later, it was tape recorded with DEA assistance, and the recording was offered at trial. In this conversation Moran, referring to his prior questioning by federal agents, said that it had occurred because the agents had seen him with "fatso" two or three times. Callahan said he had heard that the fat guy was being scrutinized by law enforcement agents and Moran replied, "Oh, my God. Unbelievable. I already told him and his first lieutenant, I says, I think somebody made you expendable." At trial Callahan identified the first lieutenant as Mary. Callahan concluded the taped conversation by asking Moran, "Can we do some business?" and Moran essentially agreed (although no evidence of further transactions between them was offered).

This, omitting a few intervening conversations between Moran and Callahan that add nothing pertinent, is the gist of the evidence against Moran. The jury, after hearing this evidence and evidence of Willis' ring, acquitted Moran of the two distribution counts based on the sales to Callahan but convicted him of conspiracy. The reason for the discrepancy is unclear. Possibly the jury hesitated to rely solely upon Callahan to prove the sales, but thought that the tape of the second conversation confirmed Moran's relationship with Willis regarding drug distribution. But the discrepancy does not matter. The question presented now is whether, having heard the evidence, including nuances and intimations that a cold record cannot capture, a rational jury could find beyond a reasonable doubt that Moran was guilty of conspiracy.

No court lightly overturns a jury verdict on the ground that the jury lacked sufficient evidence, for the jury's central role and competence is to weigh the evidence and find the facts. Yet the issue here, or at least the aspect we find troubling, actually poses the "legal" question whether the conduct the jury could reasonably have found to have occurred amounts to a conspiracy under the statute. In our view, the jury here had no rational basis to infer, as it often may in conspiracy cases, that the defendant was effectively an employee or a formal "share partner" in the ring. The most that the jury could find without sheer speculation was that the relationship was what was portrayed on the surface. At this point we are driven back to first principles to determine whether this relationship amounted to a criminal conspiracy.

Our starting point is the legal definition of conspiracy as an agreement by the defendant with another person or persons to commit the crime in question. Iannelli, 420 U.S. at 777, 95 S.Ct. at 1289; United States v. Glenn, 828 F.2d 855, 857 (1st Cir.1987). The evidence in this case, taken most favorably to the government, shows that Willis agreed to supply Moran a package which Moran represented to be cocaine, which Moran tendered to Callahan, and which Callahan then rejected as tainted with a diesel smell. This connection between Moran and Willis is bolstered, or so the jury could have found, by Moran's prior use of Willis as a source of supply, by Moran's unsuccessful initial effort to buy drugs from Willis for Callahan, by Moran's desire to ingratiate himself with Willis, by Moran's encounters with Willis and Mary and by Moran's knowledge that Willis was under federal scrutiny. On appeal, the government argues that the evidence surely demonstrates a conspiracy either as charged (with Willis and others) or, at the very least, between Moran and Willis.

An agreement surely existed between Willis and Moran relating to drugs. But if the evidence showed only an agreement by Willis to sell drugs to Moran, it would not necessarily show them to be co-conspirators in drug distribution. There is substantial law, including cases in this circuit, that a single drug sale does not automatically make buyer and seller co-conspirators. United States v. DeLutis, 722 F.2d 902, 906 (1st Cir.1983) (collecting cases). This "rule" in varying forms prevails or has been intermittently adopted in a number of circuits, including the Second, Fifth, Sixth, Seventh and Eighth. E.g., United States v. Douglas, 818 F.2d 1317, 1321 (7th Cir.1987) ("a mere buyer-seller relationship, without more, is inadequate").

Surprisingly the reason for excluding such buyer-seller cases from the definition of conspiracy is not wholly clear, and some explanation is needed since even an unplanned sale involves an agreement between seller and buyer and the offense of drug distribution (at least by the seller). Some have thought it to follow from the so-called Wharton rule, now much reduced in force by Iannelli v. United States, 420 U.S. 770, 95 S. Ct. 1284, 43 L. Ed. 2d 616 (1975), that a crime legally requiring a plurality of actors (e.g., dueling) should not have a conspiracy charge superimposed upon it. Other courts have felt that a single purchase and sale do not involve the union of two participants in a manner that increases either the likelihood that the individual crime will be committed or that the two will extend their joint endeavor to new crimes. The latter explanation has force in the case of an unplanned spot sale with no agreement beyond that inherent in the sale. It makes less sense where the agreement is to make a sale at a future point, an agreement that does increase the likelihood that the crime will be committed. Yet even in the latter case, the transaction may seem to some to lack the quality of jointness--the hallmark of conspiracy--in the sense that seller and buyer are not part of the same criminal enterprise.

This may seem a fine point but it is one that goes to the root of conspiracy law: conspiracy is treated as a separate crime because of the jointness of the endeavor. A multiplicity of actors united to accomplish the same crime is deemed to present a special set of dangers, either that the criminal end will be achieved, Callanan v. United States, 364 U.S. 587, 593, 81 S. Ct. 321, 325, 5 L. Ed. 2d 312 (1961), or that the conspiracy will carry over to new crimes, United States v. Rabinowich, 238 U.S. 78, 88, 35 S. Ct. 682, 685, 59 L. Ed. 1211 (1915), or both. See 2 W. LaFave & A. Scott, Substantive Criminal Law § 6.4(c) (1986) (summarizing the rationale). It is these dangers stemming from jointness that justify early intervention to stem conspiracies even before they rise to the level of attempts and to impose a separate punishment on the conspirators even if they fail to achieve their ends. This special set of dangers is present if two individuals agree that one of them will sell cocaine and the other will assist; it is arguably not present if one merely sells the same cocaine to another without prearrangement and with no idea of or interest in its intended use. In the latter case, both may be guilty--one of distribution and the other of possession--but without more they are not conspirators. Glenn, 828 F.2d at 858.

At some point the relationships converge. A pattern of sales for resale between the same persons, together with details supplying a context for the relationship, might well support a finding of conspiracy. Id. at 857-58. Even a single sale for resale, embroidered with evidence suggesting a joint undertaking between buyer and seller, could suffice. United States v. Carbone, 798 F.2d 21, 27 (1st Cir.1986). Common knowledge, interdependence, shared purpose and the other ingredients of a conspiracy are matters of degree. Almost everything in such a case depends upon the context and the details. The evaluation of the facts is entrusted largely to the jury.

In this case, taking a practical rather than a formal view of the matter, we believe that the jury was entitled to conclude that the arrangement amounted to a conspiratorial agreement between Willis and Moran for the distribution of drugs. Based on testimony that the jury was entitled to credit, Moran (according to Callahan) admitted that he was dealing with Willis, an admission suggesting that Willis had supplied Moran with drugs in the past. Moran then turned to Willis as his first choice of supplier in seeking to fill Callahan's first order. Although Willis' price was too high for this first transaction, for the second one Moran--after expressing his desire to bolster his relationship ("to ingratiate with the fat guy")--again turned to Willis. This time Moran did acquire from Willis a resale sized quantity, even though Callahan then rejected the shipment. This picture of a continuing sale-for-resale relationship, even if Willis was not the exclusive supplier, was reinforced by Moran's other contacts with Willis and knowledge of his law-enforcement jeopardy. See United States v. Anello, 765 F.2d 253, 261 (1st Cir.), cert. denied, 474 U.S. 996, 106 S. Ct. 411, 88 L. Ed. 2d 361 (1985).

We think that a realistic appraisal of Moran's and Willis' relationship would permit a jury to find that it amounts to an implicit agreement and comprehends the continuing supply by one to the other of drugs for resale to customers. See United States v. Geer, 923 F.2d 892, 895 (1st Cir.1991). Even though Moran was not an employee nor did Willis and Moran formally divide the profits, in this case a jury could conclude that both Willis and Moran had an ongoing stake in the success of Moran's own sales of the drugs Moran acquired from Willis. See Glenn, 828 F.2d at 857-58. From those sales Moran could profit directly and Willis indirectly through the maintenance of the drug distribution channel crucial for a drug network. See generally Direct Sales Co. v. United States, 319 U.S. 703, 717, 63 S. Ct. 1265, 1272, 87 L. Ed. 1674 (1943). Such an arrangement, we think, is not only an agreement within the ordinary conspiracy-law ambit but is one that unites two participants in seeking to accomplish the crime of distribution and involves both of the dangers of conspiracy--increased likelihood of success and extension to other crimes--to which the cases advert. We think that the pragmatic approach of Direct Sales in defining conspiracy foreshadows the result in this case and, given Congress' intent to stamp out drug transactions, it certainly did not mean to narrow the conspiracy concept when it enacted 21 U.S.C. § 846, the statute involved in this case.

We leave for another day the lesser variations on the same theme. Obviously a single sale in resale sized quantities presents one problem and an advance agreement to make a single sale involves another. Where one draws the line is more a matter of discerning congressional policy and intent than an exercise in logic, and the case-by-case approach is for the present the wisest course. As for the classic single sale--for personal use, without prearrangement, and with nothing more--the precedent in this circuit as well as others treats it as not involving a conspiracy. In such cases the jointness element is clearly at a minimum, if it exists at all. Where nothing more is involved, we reaffirm existing authority that such a case is not a conspiracy.

Moran's remaining arguments are less formidable than his attack on the sufficiency of the evidence. Moran first argues that even if the evidence was adequate to prove a conspiracy between Willis and Moran, it was not sufficient to prove Moran to be a member of the larger conspiracy charged in the indictment. This variance, he argues, prejudiced him by associating him with more powerful and extensive evidence against other defendants. We agree that there was probably a variance but find that it was not prejudicial.

The indictment charged a single drug distribution conspiracy, naming as co-conspirators Willis, Moran, six other named defendants, and "other persons both known and unknown to the Grand Jury ..." At trial much of the evidence related to the two other defendants who pleaded not guilty and to Willis and his relations with defendants other than Moran. The references of Mary to one side, nothing linked Moran to any of the ring members other than Willis or any of the transactions charged elsewhere in the indictment other than Moran's own sales to Callahan.

On this record, it is true that the evidence pointing to a Willis-Moran conspiracy is far stronger than evidence of conspiracy between Moran and the Willis ring as a whole. A very serious problem would be presented if the jury had held Moran liable for other substantive crimes committed by the ring. Compare Glenn. The situation is different where the government charges a defendant with a crime (here, conspiracy to distribute) but the facts proven at trial vary somewhat from those charged in the indictment. In that case, it is settled law that a conviction for the crime charged will be affirmed unless the variance as to the facts is shown to have prejudiced the defendant. Berger v. United States, 295 U.S. 78, 55 S. Ct. 629, 79 L. Ed. 1314 (1935); Fed.R.Crim.P. 52(a) (variance not affecting substantial rights may be disregarded). Convictions are often sustained under this principle where the indictment alleges a single conspiracy but multiple conspiracies are actually proved. E.g., United States v. Sutherland, 929 F.2d 765, 772 (1st Cir.), cert. denied, --- U.S. ----, 112 S. Ct. 83, 116 L. Ed. 2d 56 (1991).

In this case, Moran does not and could not claim that the variance deprived him of notice of the charge adequate to prepare a defense. Rather he contends that the disparity in evidence--specifically, the array of witnesses and tape recordings incriminating other defendants--created a "spillover" effect that enhanced the narrower case against him. See Sutherland, 929 F.2d at 772. The enhancement may be assumed; motions for severance are routinely made in conspiracy cases, partly to escape this taint. The question is whether the impact threatened to deprive defendant of a fair trial. We conclude that it did not.

Most of the evidence concerning Moran was distinctly different from the evidence against others. It derived from Callahan's testimony and concerned his transactions with Moran. Similarly, Moran's relationship with Willis was based upon Callahan's description buttressed by Moran's own tape recorded statements. The distinct separation between this evidence and evidence of other Willis-related activities diminished the risk of jury confusion. Indeed, on these facts the risk appears to have been minimal compared to the usual mass conspiracy case. Under these circumstances we do not think the apparent variance even arguably threatened Moran's right to a fair trial.

Finally, Moran argues that error inheres in a supplementary instruction given to the jury during its deliberations. Jury deliberations began on February 14, 1991, and the next day the jury sent in the following written question, as described by the trial judge:

The indictment states, quote, David Elwell, Richard Morretto and George Moran, defendants, combined, conspired and agreed with each other--underlined "with each other"--and with other persons, both known and unknown to the grand jury, close quote. Does the statement mean these three--and circled--people conspired with each other--and "with each other" is again underlined. Your instruction seems to be different from the indictment. Signed by the foreperson.

The judge then re-instructed the jury, reminding them that "first, remember the indictment is only the charge, the accusation. It is not evidence. It is not a statement of the law. On the other hand, my instructions are a statement of the law and are binding on you." The judge then repeated his prior instructions on conspiracy (two or more persons, agreement to commit crime charged, defendant's knowledge of unlawful purpose and knowing joinder). Within the hour, the jury returned its verdict, including the conspiracy conviction of Moran.

On appeal, Moran agrees that "[v]iewed in isolation, the judge's instructions were unobjectionable," and this is clearly so: the response to the jury's question was clear, correct, and precisely answered the question posed. Moran argues, however, that in context the instruction could have led the jury to believe that it could disregard the indictment entirely and convict the defendant of any conspiracy it chose. There is a distinct possibility, says Moran, that the jury convicted him of a conspiracy not charged such as a conspiracy with Callahan or "a conspiracy with Willis, different from that involving Moretto, Polito, and Elwell."

Moran's counsel at trial did not object to the supplementary instruction and any objection is therefore waived absent a showing of serious prejudice. United States v. Maraj, 947 F.2d 520, 525 (1st Cir.1991). No such showing has been made here. Further, we do not think that the instruction invited the jury to disregard the charge in the indictment; indeed, the supplementary instruction reminded the jury that the agreement here charged was "to possess with intent to distribute cocaine." As for the suggestion that the jury convicted Moran for such a conspiracy with Willis, rather than with Willis and others in his ring, this may well be so. But as cases like Sutherland show, such an outcome is not conviction for a "different crime" than that charged but is merely a permissible variance.

Affirmed.