PHILLIPS, Circuit Judge:

Giuliano Giunta appeals his convictions of conspiracy to import into and to distribute within the United States more than one kilogram of heroin in violation of 21 U.S.C. §§ 963, 846, and 841(b)(1)(A). We conclude that the evidence was insufficient to convict him of conspiracy on either count and reverse his convictions.

I

Sometime in early 1987, Giunta became a target of an elaborate government sting operation designed to infiltrate organizations importing Sicilian white heroin into the United States and bring members of the organizations to justice. Effective undercover work by government agents and their informers succeeded in identifying and convicting some of the operation's targets for substantive trafficking offenses. The efforts specifically directed at Giunta, however, failed to produce any substantive trafficking offense, leaving only the possibility of prosecuting him for conspiracy to commit those offenses.

The portion of the sting operation, titled Crimson Sky, in which Giunta became involved was centered on Greensboro, North Carolina. Its principal operators were Gregory Calles, an undercover agent, and a local businessman, one Henry ("Duffy") Mazzeo, who recently had been hired by the FBI to work with Calles, apparently because of his already existing social connections with a circle of suspects who had business interests in Greensboro and were either residents or occasional visitors there. The sting operation's cover story was that Calles was an Hispanic who represented a Miami cocaine cartel that was interested in getting into the heroin trade, and Mazzeo was his middleman, seeking out heroin suppliers. Mazzeo's principal role was to find suppliers and introduce them to Calles who would then conduct negotiations.

At the times in issue, Giunta was an Italian citizen in his mid-forties who was employed by a bank in Pozzallo, Sicily. During the early 1980's he started a restaurant business, the Cafe Royal, in Greensboro, in partnership with one Giovanni Carandola. In connection with this venture, he made occasional trips to Greensboro, where he moved in a circle that included several eventual targets of the sting operation. He apparently first met Mazzeo in 1985 at the Cafe Royal on a purely social occasion during which drugs were not discussed and at a time before Mazzeo had been employed by the FBI. His next encounter with Mazzeo occurred by chance in June of 1986 in a cafe lounge in his hometown of Pozzallo, Sicily, where Mazzeo, who by then had been hired by the FBI, was visiting "on a holiday" with several targets of the investigation including Carandola and one Salvatore D'Angelo. On this occasion, Mazzeo reintroduced himself to Giunta and they chatted socially, again without any mention of drugs.

The first time drug trafficking was mentioned between Giunta and either of the Crimson Sky investigators was on a trip to New York in January of 1987. Giunta flew there from Greensboro en route to Rome with two friends. On the same flight were Mazzeo and D'Angelo. Unbeknownst to Giunta, Mazzeo and D'Angelo were going to New York to complete a sale of heroin to agent Calles. When the five became stranded in New York by bad weather, they wound up sharing a hotel suite overnight. During the forced layover, the drug transaction between Mazzeo and Calles and D'Angelo was completed out of Giunta's presence, but at some point during the layover Mazzeo told Giunta that he, Mazzeo, was in the drug business. According to Calles, Giunta later approached him, indicated that he knew that Calles and Mazzeo were in the drug business and that he might be able to help Calles in some way.

Soon after this episode in New York, a situation developed that enabled Mazzeo to deepen his relationship with Giunta by doing him a considerable favor. It grew out of Giunta's falling out with Carandola, his restaurant partner, and Carandola's resulting agreement to buy him out for $100,000. Unable to collect from Carandola, Giunta complicated his situation by involving another member of the Greensboro connection, one Umberto Pitino, in his financial difficulties. When Pitino directed Giunta to wire transfer funds from Pitino's account in Giunta's bank in Sicily to Greensboro for Pitino's use in buying another restaurant, Giunta took $15,000 of the requested amount for his own use, intending to repay Pitino in Greensboro out of the money owed him by Carandola. In the showdown, when Carandola did not pay Giunta, so that he could not pay Pitino, Pitino reacted by taking Giunta's passport to prevent his return to Sicily, and demanded immediate payment.

In this plight, Giunta appealed to his friend Mazzeo for help at a meeting in the Greensboro airport sometime in the early spring of 1987. Mazzeo told his FBI employers of the request and it was decided that he should lend Giunta the $15,000 in order to enhance Mazzeo's credibility with Carandola and Pitino, who were principal targets of the investigation. Mazzeo made the loan in May of 1987, taking a 3-month note for its repayment. This enabled Giunta to relieve his immediate problem by paying off Pitino, though it did not resolve his situation with Carandola.

The loan was not conditioned on Giunta's helping Mazzeo find a heroin source. In fact, it was made principally to aid the investigation of Pitino, who was then a major target. But from this point on until Crimson Sky was terminated in March of 1988, Mazzeo relentlessly pressed Giunta to find a heroin source and consummate a sale in this country to Calles. Their contacts during this period were mainly by telephone calls between Mazzeo in this country and Giunta in Italy, though the two also met occasionally in both countries. More than 40 telephone calls, most initiated by Mazzeo, were recorded by the government.

The most critical contacts and events during this period can be summarized as follows. After several inconclusive conversations between Mazzeo and Giunta during the late spring of 1987, Giunta reported in June 1987 that he had made contact with a person who might have a connection with a source. Though he did not identify him at the time, it turned out to be one Pietro Fedino. Giunta told Mazzeo that he could meet this possible connection in Italy toward the end of June or sometime later in the United States.

Following up on this Mazzeo and Calles went to Europe in late June 1987, but did not make any connection. In mid-August 1987, Mazzeo and Calles went together to Italy to make contact with Giunta's claimed connection. There they met Giunta without difficulty, but though they were there for eight days, Giunta never produced any one. At this time, Giunta did identify Fedino as the man he had hoped they would meet, and he repaid the $15,000 loan from Mazzeo. Otherwise, the agents went away empty-handed, after expressing their annoyance to Giunta.

Shortly after the agents' return to the United States, Giunta called Mazzeo and introduced him over the telephone to Fedino. Fedino discussed with Mazzeo the possibility of meeting Mazzeo in New York, and tentative plans were made. On a follow-up call by Mazzeo to Giunta, Giunta gave Mazzeo information on how to meet Fedino in New York.

Mazzeo and Calles then met Fedino in New York on September 4, 1987, according to plan. Giunta was not present. In a confusing and inconclusive discussion of some forty minutes, the participants came to no agreement. Fedino represented himself not as a supplier but as one who had connections with certain "people" who could supply. He talked vaguely about a plan under which Calles would pay money up front to be held by either Fedino or Giunta until Calles was satisfied with the quality and quantity of heroin supplied, at which time the money would be paid to the suppliers. Among other problems on which the discussions foundered was that of where delivery would be made, whether in Italy or the United States. Calles wanted delivery in the United States (was actually not permitted to bring heroin into the United States), Fedino wanted delivery in Italy. The meeting broke up with Fedino saying that he would stay in touch through Giunta who would contact Mazzeo.

Mazzeo then returned to Greensboro, from where he continued to call Giunta in Italy from time to time. One such call was on October 6, 1987, during which Giunta indicated he knew of the problem concerning the place of any delivery from Fedino's people. Mazzeo asked Giunta to find out what quantity of heroin would have to be purchased in order to get delivery in the United States. Giunta promised to try to find out and told Mazzeo how he could be contacted in the future.

Not having heard anything from Giunta, Mazzeo called him again on October 15, 1987, to inquire what he had found out. This time Giunta promised to call back with the information in three days, insisting that Mazzeo then tell him "yes or not." This conversation too was confusing and inconclusive as to any specific deal, with Giunta indicating at the outset that "for now it's no way because . . . it's too difficult." Indeed, there was confusion as to who Giunta was now talking to as possible suppliers. When asked specifically about Fedino, he pointed out that he had "other connection," and later on remonstrated that "I am no connection with these people . . . I just know somebody" and was acting "cause you're my friend." This conversation ended with Giunta saying that he might go to Milano and that if he did he might call Mazzeo from there.

The next contact was on November 17, 1988, when Giunta called Mazzeo to tell him that Fedino was with him, had "maybe 80%" arranged everything, and was coming through New York in a few days on his way to South America. After indicating that "a way" might be found to resolve the place of delivery problem, Giunta noted that it would require the payment of a substantial sum of money up front, and that it would be better for Mazzeo to discuss the plan directly with Fedino since it would require a "special arrangement" between the two of them. Giunta then turned the telephone over to Fedino and took no further part in the ensuing conversation. Fedino opened by stating that "I can make the arrangement to . . . have the stuff in the states . . . but . . . [would] have to pay in advance because it is a risk of transportation." Mazzeo indicated that he understood the risk and the two then concluded the conversation by arranging a meeting in New York.

As events developed, the New York meeting did not occur at the time arranged, but about a week later, on November 25, 1987, after Fedino had completed his South America trip. This meeting involved only Mazzeo and Fedino; neither Giunta nor Calles was present.

Again the conversation between the two was confusing and in the end inconclusive; no deal was struck. Fedino first proposed an arrangement in which Mazzeo would advance both transportation costs to cover the risks of delivery in New York, and fifty percent of the cost of the heroin to be delivered. The amount of transportation costs to be advanced varied from time to time in Fedino's estimate from ten to twenty thousand dollars; the cost of the heroin itself bogged down in confused discussions of the lira-dollar exchange rate and was never firmed up beyond an estimate that Mazzeo's fifty percent advance payment would come to around $108,000 on current exchange rates. Later, Fedino seemed to revise the original proposal to one in which Mazzeo would only advance transportation costs, and Fedino would make the fifty percent down payment on the purchase price.

Fedino described his role in the arrangement as that of a middleman for the sellers, not as the seller. Again in this conversation he indicated that either he or Giunta would protect Mazzeo from loss by holding the advance payment until satisfactory delivery was made. At one point Fedino digressed to suggest that he might be able to arrange a purchase in South America, though he noted the special dangers of dealing with those "people." Again, he digressed to press Mazzeo at length on how he could transfer money out of this country without running afoul of Customs. When discussion returned near the end to the proposed deal, Fedino told Mazzeo that he, Mazzeo, must pay "something" to Giunta; that Fedino would realize his profit from the "man" in Italy: "The man pay me and you pay Guiliano [Giunta]." The conversation concluded with Fedino telling Mazzeo that he would call him to arrange a meeting with the suppliers' courier in New York once that was arranged. He admonished Mazzeo not to give any money to the courier, but only to him or to Giunta as Fedino might then direct.

Nothing having come of this conversation with Fedino, Mazzeo called Giunta in Italy on January 12, 1988. This conversation was a particularly confusing one in light of what had gone before. When asked whether "details" had been "worked out," Giunta answered that they had, but he then outlined a proposal that obviously was not the one suggested by Fedino in New York, for it involved a delivery in Italy at a price of $215,000. When Mazzeo raised his ongoing objection to delivery outside the United States at such a high price, Giunta remained firm on this aspect of his proposal, though he cryptically remarked that this price was "too much." And when pressed by Mazzeo as to whether the suppliers would be "taking care" of him, Giunta responded that he didn't know; that he hadn't asked for anything and simply told them he was acting for a friend for whom he was trying to find the best product. This conversation ended inconclusively with Giunta agreeing to check to see if his people would make delivery at the indicated price in the United States. Giunta then remarked in concluding that he was coming over to the United States the next week and that they could "meet Peter [Fedino] one way and . . . meet the other people for the other way . . . and take the best."

Giunta did not come to the United States the next week, but called Mazzeo from Italy on January 22, 1988. He reported that he had met his "people," that they were "good," and that they said they had the best. He indicated that he was coming to the United States during the current week, but was fearful of bringing a sample. When Mazzeo asked if his people could deliver in the United States, Giunta responded that he thought he might arrange it, would try, but "don't promise." Neither Fedino nor Fedino's New York proposal was mentioned in this conversation.

The next encounter (so far as the record reveals) between Mazzeo and Giunta did not occur until March 4, 1988, in Greensboro and Winston-Salem, North Carolina. Giunta had just arrived from Italy when Mazzeo enlisted him to act as translator in negotiation by Mazzeo and Calles with old friend Umberto Pitino concerning a drug swap. Giunta performed the service, the deal was struck, and at its conclusion Giunta suggested that he should receive some payment for his services. Fedino was not mentioned as being involved in any way in this deal between the agents and Pitino. Giunta agreed that he would continue to help in the Pitino transaction by calling Calles once Pitino had obtained the heroin for the agreed swap.

On March 6, 1988, Calles and Mazzeo met Fedino at an airport in New York to pursue their solicitation efforts with him. Giunta had nothing to do with arranging this meeting and was not present. This meeting lasted about 45 minutes and wound up as inconclusively as had the earlier ones. Fedino told Calles that he was still working on a contact, but that his people were hesitant to deliver in the United States and wanted to deliver elsewhere. Upon inquiry by the agents about Giunta, Fedino stated he was like a brother to him, that they were partners, and that Giunta must be paid for his part "in this." The meeting concluded with Fedino saying that if he was able to work out a deal Giunta would contact Mazzeo.

At the time of this last meeting with Fedino, it had been decided, as the agents knew, to close down the Crimson Sky operation and all related heroin investigations on March 31, 1988. In the short time remaining the agents therefore made two last-ditch efforts to close deals with Giunta and Fedino.

On March 17, in the course of a short telephone conversation Mazzeo told Giunta that he and Calles "want the deal" and asked if they were now talking about Fedino's or Pitino's deal. Giunta responded that he was talking about Pitino's. This was the agents' last contact with Giunta.

Sometime before March 31, Calles called Fedino in Venezuela in a last attempt to get him to bring some heroin into the United States before the operation's close-down. This was the agents' last contact with Fedino.

Nothing came of either of these final contacts, and no heroin was ever brought into the United States by Giunta or Fedino or by others through their efforts. As planned, Operation Crimson Sky was closed down on March 31, and several arrests of other targets of the overall operation than Giunta and Fedino were made on that and the following day.

Giunta and Fedino were jointly charged in two counts of a four-count indictment as co-conspirators to import and distribute more than one kilogram of heroin, in violation of 21 U.S.C. § 963 and §§ 846 and 841(b)(1)(A), respectively, and Giunta was separately charged in two counts with telephone facilitation violations under 21 U.S.C. § 843(B). Giunta was arrested in and extradited from Italy to face charges on the two conspiracy counts. Fedino was not apprehended.

As indicated, the jury convicted Giunta on both conspiracy counts. This appeal followed.

II

Giunta's principal challenge, and the one we find dispositive, is to the sufficiency of the evidence to convict him on either of the conspiracy counts, an issue raised by his assigning as error the district court's denial of his motion for acquittal made at the conclusion of all the evidence.

We review the sufficiency of the evidence under the familiar standard of Jackson v. Virginia, 443 U.S. 307, 319, 99 S.Ct. 2781, 2789, 61 L.Ed.2d 560 (1979), which inquires whether "any rational trier of fact could have found the essential elements of the crime [charged] beyond a reasonable doubt," and requires us in applying the standard to construe the evidence in the light most favorable to the government, assuming its credibility, drawing all favorable inferences from it, and taking into account all the evidence, however adduced. Id.

The elements of the crime of conspiracy in general are simple: (1) an agreement between two or more persons (not including government agents), (2) to commit in concert an unlawful act. Morrison v. California, 291 U.S. 82, 92, 54 S.Ct. 281, 285-86, 78 L.Ed. 664 (1934). The essential elements of the particular conspiracies charged to Giunta here were (1) an agreement by Giunta with one other identified person, Pietro Fedino, and "others unknown," (2) to import and distribute a kilogram or more of heroin in the United States.

The critical question is that specifically raised by Giunta: whether the evidence sufficed under the Jackson v. Virginia standard to prove such an agreement between him and Fedino. In assessing the sufficiency of the evidence in that particular respect, we are guided by certain other rules especially pertinent to proof of conspiracies in general and those charged here in particular.

Proof of a conspiracy may of course be by circumstantial evidence; it need not and normally will not be by direct evidence. United States v. Brown, 856 F.2d 710, 711-12 (4th Cir. 1988). But circumstantial evidence that proves nothing more than association between two persons, even if one has a fixed intent known to the other to commit an unlawful act, is not sufficient to permit the inference of the requisite agreement between the two to act in concert to commit the act. See United States v. Tyler, 505 F.2d 1329, 1332 (5th Cir. 1975); Miller v. United States, 382 F.2d 583, 587 (9th Cir. 1967). Because conspiracy requires agreement between at least two persons to take concerted action, if the purposes of alleged co-conspirators are different, though both are aimed at unlawful goals, there is no conspiracy. See United States v. Mancillas, 580 F.2d 1301, 1307 (7th Cir. 1978). Therefore, when one of two alleged co-conspirators merely pretends to agree, the agreement requisite to conspiracy cannot exist. United States v. DeBright, 742 F.2d 1196, 1199 (9th Cir. 1984).

III

The evidentiary record we review here is a shadowy one. The government's evidence concerning the intentions and purposes of Giunta and his alleged co-conspirator, Fedino, consists primarily of the testimony of its investigative agents, Mazzeo and Calles, respecting their intermittent contacts over a year-plus period with the two targets during the course of a widespread investigation involving a number of more important targets. This oral testimony is supplemented with tape transcriptions of some of the telephone and face-to-face conversations between different ones of the four principals. In the setting in which these conversations took place, with both sides either deliberately falsifying their true identities and purposes (the government agents) or failing to be fully forthcoming (the targets), the significance of what was being said at any particular time is inevitably a murky proposition. And when to this contextual difficulty the obvious language barrier that the transcribed recordings reveals is added, what results is a basic problem for anyone attempting to assess from these circumstances the actual motives, intentions and purposes of these undoubtedly venal targets. As indicated, these conversations, which led to nothing in the end, are essentially the government's case. There is little if any evidence about conduct independent of the conversations to give any color to what was said in the ultimately inconclusive sparring between agents and targets. There is no smoking gun in these conversations or in any results from them that points to the requisite criminal agreement.

In this circumstance, the danger of guilt being found on the basis of speculation from mere association between criminally disposed people, related criminal behavior, and like considerations is acute. The attempt to assess the evidence on such a record brings home the continued validity of some cautionary observations from times past about conspiracy prosecutions in general and drug conspiracy prosecutions in particular. In one of the more famous judicial ruminations on the subject, that of Justice Jackson concurring some forty years ago in Krulewitch v. United States, 336 U.S. 440, 445, 69 S.Ct. 716, 719, 93 L.Ed. 790 (1949), he noted what to him was already in those days a "drift" of the "elastic, sprawling, and pervasive offense" of conspiracy so advanced as to demonstrate the "tendency of a principle to expand itself to the limit of its logic." This led him to the rueful comment that

The unavailing protest of courts against the growing habit to indict for conspiracy in lieu of prosecuting for the substantive offense itself, or in addition thereto, suggests that loose practice as to this offense constitutes a serious threat to fairness in our administration of justice.

Id.

That sober assessment, made long before the drug scourge began significantly to afflict the justice system along with society in general, was picked up some thirty years later by Judge Goldberg in United States v. Caro, 569 F.2d 411, 418 (5th Cir. 1978), a drug conspiracy case. Writing with the national "war on drugs" now heavily underway in the federal courts, Judge Goldberg, in the course of reversing the conviction for insufficiency of the evidence, found guidance for the task of assessing the evidence in Justice Jackson's foreboding about the special threat to fairness posed by conspiracy prosecutions in general. As Judge Goldberg noted, those forebodings about the pervasive use of conspiracy counts had been borne out in the ensuing thirty years during which conspiracy had remained a "potent and oft-used weapon in the prosecutorial arsenal." Id. For him, the special threat long since recognized by Justice Jackson and borne out in intervening experience required a special way of viewing the "slippery facts and the speculations necessary to uphold [the conspiracy] conviction" there at issue. Id.

In the decade since Judge Goldberg wrote in Caro, conspiracy has of course continued a "potent and oft-used weapon in the prosecutorial arsenal," particularly in connection with the drug trafficking prosecutions that increasingly dominate federal criminal dockets. And Justice Jackson's concerns about the special risks of unfairness in conspiracy prosecutions in general have simply been magnified in drug conspiracy prosecutions as understandable outrage and frustration have charged the "war on drugs" presently underway as avowed national policy.

By these introductory observations, we do not of course suggest that the judiciary may under any circumstances properly view conspiracy prosecutions with general disfavor, nor skew the judicial function at the trial or appellate level to reflect that disfavor. Conspiracy is a crime no less than are "substantive" offenses; the Jackson v. Virginia standard applies equally to assessing proof of both. Our assessment of the sufficiency of the evidence to convict of conspiracy here must therefore of course proceed as it would in respect of any substantive offense. There is no special, more stringent standard in conspiracy. What is appropriate, however, is the recognition, carrying forward Justice Jackson's in Krulewitch, that the very "elasticity" of the conspiracy concept may and frequently does tend to produce more "slippery" facts in proof than do prosecutions for less "elastic" substantive offenses, with a corresponding greater danger that speculation rather than reason may be required to make the necessary leaps of inferences to find guilt. Heightened vigilance to guard against the increased risks of speculation, though not a heightened standard, is warranted in conspiracy prosecutions.

Our assessment of the evidence here reveals exactly the degree of "slipperiness" of factual proof that Justice Jackson feared as a general phenomenon and that Judge Goldberg found present in Caro. Our conclusion — a raw judgment call at odds with that of the district court — is that because of its cumulative weakness in various respects, the evidence here could only lead to a finding of guilt by an unacceptable process of raw speculation rather than by a reasoned process of inferring guilt beyond a reasonable doubt.

Because our ultimate assessment is one of cumulative weakness of the evidence in various respects, we analyze it in terms of those specific weaknesses. We precede the analysis proper with the observation that our earlier account of the facts has been set out in considerable detail and in light of the appropriate evidence assessment standard, precisely to serve as a framework for analysis. That is to say, the facts as recited are essentially those developed by the evidence considered in the light most favorable to the government.

The first step in the analysis proper is to identify the exact conspiracy charged to Giunta: to participate in a conspiracy with Pietro Fedino (and "others unknown") to import, then distribute, a kilogram or more of heroin in the United States. As indicated, proof of each of these conspiracies requires proof of an agreement between these two to commit in concert the specific unlawful acts charged — here importation and distribution.

Given the facts of this case and the government's apparent theory of prosecution, it is important to note that proof that an alleged conspirator (here Giunta) has assisted a willing (ostensible) buyer (here Mazzeo or Calles) in finding an ostensibly willing seller of drugs (here Fedino), would not, standing alone, establish any agreement between the facilitator and the seller to do anything in concert with the drugs, including their importation and distribution. See United States v. Tyler, 758 F.2d 66, 69 (2d Cir. 1985) (such evidence held insufficient to "establish the existence of an agreement between the facilitator and the seller"). Closely related, an agreement between two alleged co-conspirators (here Giunta and Fedino) to assist a willing (ostensible) domestic buyer of drugs (here Mazzeo or Calles) to find a willing foreign seller of drugs (here the various unknown "people" referred to from time to time by Giunta and Fedino) would not, standing alone, establish an agreement between the two facilitators or brokers themselves to act in concert to import or distribute the drugs. See United States v. Melchor-Lopez, 627 F.2d 886, 892 (9th Cir. 1980) ("The law requires more than a conspiracy to attempt to arrange a purchase. . . .").

Here, there was extensive evidence that Giunta was a reprehensible person, prepared to assist Mazzeo and Calles locate a foreign seller of heroin which the agents might buy for importation and distribution in the United States. There was evidence that Giunta had actually located such a person in Fedino, who may have been prepared himself to sell heroin to the agents for importation and distribution, or to arrange and facilitate a purchase by the agents from other "people" than himself if satisfactory terms could be arranged. There was some possible evidence that Fedino and Giunta had agreed to share in any commissions or fees that might be realized from either the ostensible buyers or the sellers with whom they were dealing as middlemen or facilitators. But as to whether Giunta and Fedino ever agreed that they would themselves, in concert, actually import and distribute any heroin, whether directly or by selling it on their own account for delivery to the agents in the United States, the evidence was either nonexistent or so tenuous as to permit nothing but raw speculation.

Our assessment of the insufficiency of the evidence to support a finding that beyond a reasonable doubt Giunta and Fedino had entered into the specific agreement charged rests essentially upon a parsing of the evidence earlier summarized. But its insufficiency on the critical point is made most apparent by a passage of Mazzeo's testimony as the principal government witness. It came, with obvious unexpectedness, during Mazzeo's direct examination by government counsel concerning the November 25, 1987, meeting in New York between Mazzeo and Fedino, and related specifically to Fedino's proposal for an upfront fifty percent payment for heroin to be supplied by his purported source in Italy.

Q: About how long had you been working with the FBI in your undercover capacity at this time.

A: A couple of years.

Q: Is this proposal of Mr. Fedino's the type of proposal that you had become accustomed to dealing with?

A: No, sir; it was not.

Q: What was different about it?

A: My opinion — Mr. Fedino was just going to rip us off for the money.

J.A. 187-88.

This assessment by Mazzeo of Fedino's intentions was later repeated as government counsel pursued the obviously troubling earlier response. Asked whether after having a chance to evaluate his conversation with Fedino, Mazzeo was "optimistic about continuing working with him," Mazzeo responded: "No; like I told you — I think he was going to rip us off." J.A. 188.

And when, predictably, defense counsel picked up on this damaging assessment on cross-examination, Mazzeo stood firm.

Q: Is it fair to say that these are all the reasons why you thought [Fedino] was going to rip you all off, as you put it?

A: That's what I think.

J.A. 250. This is of course just one passage of testimony in a voluminous record. The government attempts to downplay its significance on that basis, but on this confused record, though it is short in content it is unavoidably critical in significance. On any fair assessment of this record, the person best able contemporaneously to assess what Fedino and Giunta were up to — individually or jointly — was Mazzeo. He was the prime mover in establishing the relationship. By virtue of his contacts with the two, his was the prime opportunity to fathom the reality behind the confused, ambiguous, inconclusive series of contacts and conversations between these four people — all of whom, for different reasons, were in varying degrees falsifying or concealing their actual plans and motives from each other in all that they said. The government's obvious effort, in the direct examination just quoted, to qualify Mazzeo as a credible witness on the significance of Fedino's proposals best illustrates the point. Mazzeo's response, of course, completely undercuts the existence of the agreement specifically charged to Giunta and Fedino. If this was Fedino's real purpose, whether shared with Giunta or not, there could have been no agreement between them of the kind charged.

We may of course consider this consistently maintained, undisputed testimony from the government's witness, though unfavorable to the government's case, for purposes of assessing the sufficiency of the evidence to convict. See Melchor-Lopez, 627 F.2d at 888 n. 4, 891 (testimony of government informant on cross-examination that no agreement reached between co-conspirators demonstrated insufficiency of evidence to convict); see generally 2 Wright Miller, Federal Practice Procedure: Criminal § 467 (1982). When we do so in conjunction with the tenuousness of any positive evidence supporting a finding of the requisite agreement, the overall insufficiency of the evidence on this critical element is confirmed.

Accordingly, we must reverse the convictions of conspiracy on both counts.

REVERSED.