This criminal litigation involves laboratory studies, using rats as subjects, undertaken to determine the safety of certain products for human use. Alleged falsification of data and irregular procedures created doubts about the validity of the reported study conclusions. After a trial that lasted six months and produced more than 17,000 pages of transcript, several defendants associated with the laboratory were convicted of various counts of mail fraud, wire fraud and making false statements, although there were also acquittals on some of the counts charged. Over-length briefs totaling more than 500 pages were then added to this already voluminous mass of record material.1 The defendants appeal, claiming insufficiency of the evidence, and a multitude of trial errors. We affirm.
Industrial Bio-Test Laboratories, Inc. (“IBT”) is a contract research laboratory which conducts animal toxicity studies of various drugs, pesticides and chemicals for manufacturers in order to determine their safety and effectiveness. Defendant More*684no Keplinger was IBT’s Manager of Toxicology. Paul Wright was Section Head of Rat and Dog Toxicology at IBT until October 1972, at which time he left IBT to become Manager of Toxicology for Monsanto Corporation and served as liaison to IBT. James Plank served as IBT’s Group Leader of the Rat and Dog Toxicology Department and was later Assistant to Keplinger.
The defendants here were convicted of various counts involving fraud and false statements in connection with two studies run by IBT.2 The first animal toxicity study was performed for Monsanto Corporation (“Monsanto”) and concerned Trichlorocarbanilide (“TCC”), an ingredient in deodorant soaps. Defendants participated in preparing three allegedly fraudulent written documents, which came out of this research: (1) a final study report submitted by IBT to Monsanto and then by Monsanto to the FDA; (2) a written explanation to the FDA from IBT answering questions about the TCC final report; (3) and an early version (the first draft) of the final report, which was submitted to Monsanto with an explanation of the TCC study. The alleged fraud involved suppression of relevant information or falsification of data in the following categories: a) underreported mortality of research animals and commingling of original and later-started research animals; b) omission of histopathological data on research animals; c) omission of data on post-mortem examination of animals; d) failure to report the original opinion of a Dr. Gordon that TCC had a significant effect on research animals at the lowest dose level; e) failure to disclose a pathology report by a Dr. Ribelin with respect to the effect of TCC on the epididymis of research animals.
The second study involving allegations of mail fraud was performed by IBT on Naprosyn, an anti-inflammatory drug used in the treatment of arthritis. The alleged fraud here involved a report sent to Syntex Corporation (“Syntex”) including blood and urine data and an Appendix II relating to gross pathology. The blood and urine data was allegedly wholly fabricated and Appendix II was allegedly falsified. A written explanation of the Naprosyn report submitted to the FDA also allegedly contained false statements.
I. Sufficiency of the Evidence
In reviewing a conviction, “the factfinder’s role as weigher of the evidence is preserved through a legal conclusion that upon judicial review all of the evidence is to be considered in the light most favorable to the prosecution.” Jackson v. Virginia, 443 U.S. 307, 319, 99 S.Ct. 2781, 2789, 61 L.Ed.2d 560 (1979) (footnote omitted) (emphasis in original). Therefore, the decision facing this court is “whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Id. (emphasis in original). Only when the record contains no evidence, regardless of how it is weighed, from which a jury could find guilt beyond a reasonable doubt may an appellate court overturn a verdict. E.g., United States v. Brown, 739 F.2d 1136, 1149 (7th Cir.), cert. denied, - U.S. -, 105 S.Ct. 331, 83 L.Ed.2d 268 (1984). In applying this standard, the court must leave to the jury the weighing of the evidence, the drawing of reasonable inferences and determinations of credibility.
A. Trichlorocarbanilide (TCC)
We first examine the issue of underreporting of rat mortality in the TCC study. Accurate mortality data is important because it provides evidence of a substance’s impact on test animals and also because it can provide evidence of environmental problems which might compromise the validity of a study. Apparently, the TCC *685study was plagued from the outset by a high death rate, which, the government contended, defendants attempted to conceal.
The government’s proof that the mortality rate was falsified included evidence that the first draft of the TCC report and the version eventually forwarded to the FDA reported different mortality data for the same period of time. Another IBT document, an interim status report, contained still another set of mortality data for the same time period. The mortality shown in all of the documents made the subject of the charges in the indictment is lower than that shown in internal IBT documents from which mortality can be calculated. Further, laboratory technician Smith testified that during the first six months of the TCC study, he specifically observed animal deaths which were not accurately recorded. Laboratory technician Penner also testified that he observed deaths which were not reported during the same time period. Particularly when buttressed with contradictions in other data, these facts provide sufficient basis for the jury to find that the mortality data was false.
The defendants argue that these facts do not provide adequate proof of fraud and that the government must prove the true rate of animal mortality in order to establish that the mortality figures in the TCC final report are false. Defendants argue that there was no testimony which established that the data “substantially underreported” mortality. They contend that the testimony of technicians Smith and Penner did not specifically tie the mortality rate in the problematic animal watering room (the “AWR”) to the TCC study and that the witnesses’ recollections of deaths were speculative. Defendants attempt to discount the contradictions in the data reported in the first draft, an interim status report and the version supplied to the FDA by focusing on the mortality rate over the entire course of the study, suggesting that there was no substantial intentional underreporting.
The government contends that it does not have to prove the true rate of mortality, but rather only that the reported mortality rate was too low and the defendants’ reports must have been false. Holdings in similar cases support the government’s position that there is no need to show the “true” figures. See, e.g., Coil v. United States, 348 F.2d 573, 576 (8th Cir.), cert. denied, 382 U.S. 821, 86 S.Ct. 48, 15 L.Ed.2d 67 (1965). And defendants cite no authority for their proposition that the government must prove the real rate of mortality. The very nature of the allegations suggests the inherent impossibility of showing the “true” rate of mortality, since the accurate underlying data was never recorded. Smith and Penner, who worked on the TCC study, testified that they observed more deaths than were listed in the reports. In challenging the testimony given by Smith and Penner by characterizing it as recollection, impression and speculation, defendants attempt to reweigh credibility determinations that were properly left to the jury. We think that these technicians’ testimony supports the inference that the mortality data supplied by the defendants was falsified, even if these witnesses cannot establish the true mortality rate. If totals for some periods were false (6 and 12 month status reports showing no deaths), then the tables including this data are necessarily false, at least in part. At the conclusion of the TCC study, the report included mortality data that was different from the 6 and 12 month status reports. The final report, submitted to the FDA, had still another set of mortality data. All three of these sets of data substantially underreported mortality. This evidence adequately supports a jury determination that the mortality data supplied by the defendants were false.
The high mortality of rats in the TCC study presented concerns going beyond the alleged underreporting of mortality data, however. If an insufficient number of rats survived to the end of the study, IBT would be unable to make any valid evaluation of the safety of the tested substances. Given the unusually high mor*686tality rate, the government charged, defendants therefore chose to commingle with existing research animals new animals which had not been fed the test material for as long a period as the original animals. This obviously is not an acceptable procedure, for it compromises the validity of a study.
In the instant case, adding together the number of animals that were sacrificed during the study and the number of animals that died while the study was in progress produced a total of more animals than there were at the start of the study, thereby implying that commingling occurred. Defendants’ subsequent knowledge of that commingling was established by both direct and circumstantial evidence, and the commingling was concealed from the FDA. Intent to defraud may be inferred from the totality of the circumstances surrounding the actions of the defendants. E.g., United, States v. Brown, 739 F.2d at 1149.
The government proved four incidents of commingling in two ways: 1) through the accumulated testimony about the acts of individual lab technicians, and 2) by a documentary review of the number of animals used in the study. The first incident of commingling was established by technician Garrett who testified that early in the study he filled up empty cages with fresh rats. Defendants assert that this early instance of commingling of “control” rats was not significant since the indictment focuses on the alleged commingling of late-started research animals. But this argument overlooks the fact that the control commingling alone establishes the falsity of assertions in the final report about the length of time animals were on the study. Additionally, this first incident, specifically ordered by defendant Plank, establishes his knowledge that at least some commingling was occurring and evidences a scheme to respond to high mortality by commingling new animals. Further, this early commingling rebuts defendants’ claims that technicians later in the study could have distinguished still newer animals from purportedly older ones already in the cages.
A second incident of commingling took place about March 1972 when technician Penner added new animals to the gang cages. Subsequently, other technicians under the direction of technician Thompson placed some of those animals into the individual cages to fill vacancies caused by deaths. No starting study dates were entered on the relevant cage cards, a fact which partially accounts for the fact that the technicians seemed to have been unaware of the commingling. In August 1972 Thompson observed the addition of more animals to the study. The starting dates were in this case marked on the cage cards of this group of research rats. Records of animals on the study about March 1972 also established that Garrett commingled the animals and other technicians commingled animals added by Penner.
A third instance of commingling involved the necropsy log and blood data, which were used by pathologists and toxicologists to draw conclusions about the effect of TCC. Among other things, after August 1972 Thompson sent dead animals to the laboratory for analysis without distinguishing between animals that started in August and those that had started earlier.
The fourth instance of commingling involved the final sacrifice when all the animals were killed. For the study to have validity, a required number of male and female rats in each study group had to be available for examination at the end of the study. As it turned out, precisely the right number of male rats were available in each group. Either this was a remarkable coincidence or late-started animals were added to the group cages when necessary to provide the requisite number of males. The third and fourth cited instances of commingling suggest that the data from animals started in August 1972 found its way into the final version of the TCC report submitted in 1976 by IBT to Monsanto. The commingling evidence is sufficient to support the jury’s determination that the TCC final report was false.
*687Defendants challenge the validity of the commingling evidence by noting that each of the technicians denied intentional commingling. They dismiss Garrett’s testimony about the first instance of commingling as irrelevant to the alleged commingling of late-started animals. Additionally, by attempting to characterize Garrett’s testimony as uncorroborated, defendants essentially argue that the jury’s credibility determinations may be reweighed. Of course, this is incorrect. See United States v. Grabiec, 563 F.2d 313, 316 (7th Cir.1977). The mere fact that technician Thompson, for example, did not know that “control” animals were being commingled does not preclude his having played an unwitting role in the process.
Defendants also argue that the technicians would have easily identified late-started animals because those animals would have been younger and smaller. But, as noted, this is an inaccurate assumption since Plank was attempting to obtain older and more mature rats and Garrett had introduced new animals into the study before Penner. That the technicians’ commingling was inadvertent appears from their own testimony and from the omission of starting dates on cage cards.
Dr. Lijinsky (an animal testing expert) analyzed documents from the study and concluded that commingling had occurred, because the total number of animals sacrificed during the study and those that died while the study was in progress exceeded the number that originally started. This evaluation was based on an interpretation of notations in the animal weight books (for example, interpreting dashes as meaning deaths) which was supported by the technicians’ testimony that when an animal died a dash appeared where its weight would ordinarily be entered. The defendants attack Dr. Lijinsky’s findings on various bases, including the fact that he placed the great bulk of the deaths and replacements in the first eight months of the study where their significance was diminished. Defendants also claim that there were errors in weighing and substitution procedures, and that Dr. Lijinsky misinterpreted the significance of various notations. But the evidence suggests that, if anything, Dr. Lijinsky underestimated deaths. For example, in at least one treatment group, the number of animals used exceeded the number of animals that started the study by 30%. We think the evidence of commingling is more than sufficient.
Further, the evidence showed that, when commingling was discovered, it was brought to the defendants’ attention before they made their reports and submissions. IBT employees Smith, Garrett and Reyna, who worked on the TCC study, knew about the excessive mortality rate and discussed the bad AWR conditions causing that problem with Plank and Wright, which at least implied that these defendants knew they had a potentially serious problem threatening the validity of the study. Further, there is testimony that Plank explicitly directed Garrett to commingle late-started animals. In addition, Keplinger and Wright made the decision to begin a late-started group of animals. All three defendants attended a February 1975 meeting that revealed, if they did know it before, that commingling had occurred. At another meeting in December 1976, commingling was again confirmed through inconsistencies shown to Keplinger and Plank in the numbering systems that Phil Smith had used to record animal data. Defendants also knew that the study data included histopathology from commingled animals. The data presented in the 1976 version of the final report was asserted to have come from animals that had been on the study for the full 24 months. Keplinger was well aware of the falsity of this statement. Knowledge of this falsity alone is sufficient. See United States v. Outpost Development Company, 552 F.2d 868, 869-70 (9th Cir.), cert. denied, 434 U.S. 965, 98 S.Ct. 503, 54 L.Ed.2d 450 (1977).
The next issue of concern is the misstatement of Dr. Gordon’s conclusions. Dr. Gordon was responsible for microscopic examination of tissues from the study animals to determine whether ingestion of *688TCC had any toxic effects at certain dosage levels including the lowest (“T-l”).3 This examination initially focused on the organ that toxicologists believed would be affected first, in this case the testis.
Gordon conducted histopathological examinations (by microscope) of tissues taken from animals sacrificed at the end of the study and of those from treatment group animals, as well as control animals, that died earlier. Based on these examinations, Gordon concluded that there were treatment-related changes (which he called degeneration) at all three levels of dosage (including T-l) while there was little testicular degeneration in the control animals. According to Dr. Gordon, TCC damaged the testis and T-l was an “effect level.” The defendants for obvious reasons resisted these conclusions. They directed Gordon to examine more slides; but in order to do so, he had to look at tissues from late-started animals because there were no more rats available from the original study. Examination of these late-started animals supported his original conclusion that T-l was an effect level.
Still dissatisfied with these results, Wright sent selected slides, omitting postmortem and late-started animals, to Dr. Ribelin, an independent consultant. After reviewing the slides, Ribelin also found that a no-effect level had not been reached and that there were treatment-related effects from TCC. Ribelin, however, discovered degeneration in the epididymis, the excretory duct of the testes, and not in the testes themselves. Defendants presented a great deal of evidence to show that the testes and the epididymis are different and that since the testes were the target organs of the study, a finding of an effect level in the epididymis did not support a conclusion that T-l was an effect level for the testes. But the jury was entitled to accept the government’s alternative view that if treatment-related effects were found in any part of the rat, a safe level of dosage had not been established, and so Ribelin’s conclusion was not consistent with the position that T-l was a safe level for TCC. The jury was entitled to find that the defendants’ suppression of this report denied the FDA important information about the safety of TCC.4
Despite the evidence that both Gordon and Ribelin believed T-l to be an effect level (although they perceived different types of effects), the defendants argue that these researchers agreed that T-l was a no-effect level. The government contends that Gordon was pressured into approving a change in his opinion, exonerating TCC, but the defendants argue this change was voluntary. The jury decided this question, and we will not reconsider it.
The government also contends there was fraud in the revisions made in the final report to the FDA. That report stated that post-mortem autolysis (deterioration of tissue after death) precluded meaningful evaluation of the testes of animals that died during the study. Yet defendants do not deny that Gordon was in fact able to. examine these animals, did so, and made histopathologic findings regarding these animals. Those findings were omitted from the report, implying that the animals were not examined at all. There was testimony that Keplinger told Wright that there were animals available for post-mortem evaluation but Wright edited the final report to indicate that decay made such animals unavailable. Although defendants offered testimony that there was a valid reason for omitting the data, the jury could reasonably have concluded instead that Gordon’s findings were omitted because they were adverse to the conclusions defendants wanted to draw. Additionally, the govern*689ment presented evidence to show that data collected from gross pathology examinations (opening the animals and examining tissue without a microscope) was changed so that the T-l level would appear to be a no-effect level. Earlier versions of the report, however, indicated that there were treatment-related effects at all three dosage levels. Whether the omitted histopathology and post-mortem data could have affected the FDA’s decision about the safety of TCC was a judgment properly left to the jury.
(7] The government showed that the final 1976 report was backdated by two years to 1974. The defendants argue that this was done to avoid confusion and consolidate all the results, but the government contends that the purpose was to conceal all the conflicts in the interim reports, to mask the falsifications in the data and to cover up the important omissions. Once again defendants’ arguments indicate only that they presented evidence disputing the government’s interpretation of various events; they do not persuade us that the jury could not reasonably accept that interpretation.
There are, of course, numerous other aspects of the TCC evidence which might be discussed if time or space permitted, but we believe we have pointed to sufficient evidence from which a jury could find a scheme to defraud. Defendants’ arguments essentially seek a reweighing of the evidence, which the jury found, and which we conclude was sufficient, to establish materially false statements and omissions in the study reports beyond a reasonable doubt.
B. Naprosyn
The second study examined the long-term toxicity effects of Naprosyn, a drug used for the treatment of arthritis. That study was conducted by IBT for Syntex in 1971 and 1972. The government identified two problem areas in the reports from that study: 1) alleged fabrication of blood and urine data, and 2) alleged fabrication of much of the gross pathology data appearing in Appendix II of the report.
In testing the evidence to determine its sufficiency, we focus first on the blood and urine data reported, which the government charged was never actually collected. The government presented evidence that the study was extended beyond 18 months, and as a result of the final sacrifice was postponed and the blood and urine tests scheduled to occur just before the sacrifice were canceled and never thereafter rescheduled. There was testimony that if blood and urine tests were canceled, they would not be performed later unless specifically rescheduled. At the 22-month point, when the final sacrifice did take place, the defendants for various reasons were under pressure to end the study quickly. The blood and urine tests were not rescheduled.
The government contends that these tests were never conducted, which explains why the “raw data” for them could not be located. Technician Smith testified that he could not find raw blood and urine data to include in his draft of the Naprosyn report and that he brought this to Wright’s attention along with his surmise that the work had not been performed. Wright did not take issue with this conclusion.
Technician Garrett testified that his responsibilities included setting up the rats for the blood and urine tests but that he had not done so, implying that the tests had not been performed. Yet ultimately, the final report contained blood and urine data for animals long dead at the time the blood and urine analysis would have been made.
Defendants argue that the government’s proof rests solely on the absence of the raw data. In particular they contend that the inability to locate data is circumstantial evidence insufficient to establish that the data did not exist and was fabricated. Besides attacking the credibility of technician Garrett, the defendants maintain that any apparent errors in the reported blood and urine data indicate mere negligent recording of the actual data, not falsification of it. Proof of absence of records that would ordinarily exist if a *690particular event had occurred is properly admitted to show that the event did not occur. See, e.g., Federal Rule of Evidence 803(7); United States v. Zeidman, 540 F.2d 314, 319 (7th Cir.1976). Despite the defendants’ arguments, the jury had ample evidence to support a finding of falsification.
Appendix II was a report which contained allegedly fabricated gross pathology data for animals examined post-mortem. The government presented evidence to show that the relevant log completely contradicted or left unsubstantiated findings authored by Plank in Appendix II. Technicians Smith, Garrett and Penner, among others, identified their records for specific animals and reported those animals that died during the study and were discarded without findings of any kind because they were too badly decomposed (“TBD”). Yet Plank included various gross pathology findings about these animals in Appendix II. The necropsy log frequently conflicted with results as reported in Appendix II in other respects as well. Presumably, if Plank’s conclusions had been based on actual gross observations, they would have been confirmed by the necropsy log.
In order to justify Plank’s Appendix II, defendants attempt to persuade us that Garrett performed necropsies that yielded the data reported by Plank. Yet Garrett and other technicians denied performing necropsies. Defendants also argue that several technicians never worked on the Naprosyn project, and therefore could not say whether or not necropsies were performed on TBD animals or whether or not results of gross examinations of these animals were recorded on the backs of cage cards. Dr. Gordon testified that cage cards were not considered raw data and were not kept after a project was completed. Therefore, defendants assert it was unreasonable for the prosecution to contend, or for the jury to conclude, that, five years after the Naprosyn study was finished, IBT’s failure to produce cage cards supports the inference that the cards contained no gross pathology data. But these matters, in light of all the other evidence, presented jury questions and there was sufficient evidence to support the jury’s conclusions.
The government argues that Keplinger had knowledge of the falsifications of the blood and urine data and Appendix II because he never looked into the matter of missing data when given reason to do so. The defendants contend that there is insufficient evidence that Keplinger failed to do so, and, alternatively, even if he did not look into missing data, it is unreasonable to infer that he had knowledge of the falsification. Although failure to investigate the question of the missing data, considered in isolation, may not be sufficient to demonstrate Keplinger’s participation in the scheme to defraud, in combination with other evidence, it entitled the jury to infer that Keplinger knew of the falsifications. We conclude the evidence on the blood and urine data and on Appendix II was sufficient.
C. Mailings
We next turn to the issue whether the government’s evidence was sufficient to support the jury’s findings that the Naprosyn and TCC reports were mailed, and that defendants caused those mailings. The use of the mails may be proved by direct or circumstantial evidence. E. g., United States v. Brooks, 748 F.2d 1199, 1202 (7th Cir.1984). The government points out that evidence of office custom and practice here may establish the fact of mailing. See e.g., United States v. Joyce, 499 F.2d 9, 15 (7th Cir.), cert. denied, 419 U.S. 1031, 95 S.Ct. 512, 42 L.Ed.2d 306 (1974); United States v. Flaxman, 495 F. 2d 344, 349 (7th Cir.), cert. denied, 419 U.S. 1031, 95 S.Ct. 512, 42 L.Ed.2d 306 (1974). Circumstantial proof often consists of evidence of office practice, and such testimony as to office practice is sufficient proof of mailing. E.g., United States v. Ledesma, 632 F.2d 670, 675 (7th Cir.), cert. denied, 449 U.S. 998, 101 S.Ct. 539, 66 L.Ed.2d 296 (1980).
*691Defendants, citing United States v. Wolfson, 322 F.Supp. 798 (D.Del.1971), aff'd, 454 F.2d 60 (3d Cir.), cert. denied, 406 U.S. 924, 92 S.Ct. 1792, 32 L.Ed.2d 124 (1972), argue that the prosecution must show that the office procedure is “invariable.” However, Wolfson is not binding on this court, and we have in fact subsequently held that proof of regular or usual business practice is sufficient. See, e.g., United States v. Joyce, 499 F.2d at 15. Since the government is under no duty to negate all possible innocent inferences from a set of circumstantial facts, see United States v. Graves, 736 F.2d 850, 854 (2d Cir.1984), it should not be required to present proof that the custom is “invariable.” Instead, it is sufficient to prove that mailing is the sender’s regular business practice. The inference that the sender acted in accord with its ordinary practice is reasonable, and the absence of a recollection of departure from that practice strengthens the inference that the practice was followed.
There was evidence that IBT’s Naprosyn submission to the FDA, dated November 11, 1976, was sent to lawyers for Syntex by Attorney Merrill Thompson with a cover letter, dated November 15, 1976. Thompson testified that his normal and customary business practice was to mail such letters to their addresses on or about the date they bore, together with any attachments. Defendants contend that Thompson's additional statement that such documents “would normally be sent” is too equivocal and leaves room for the inference that non-mail methods might have been used. Defendants cite United States v. Hart, 693 F.2d 286, 289 (3d Cir.1982) for the proposition that “send” is too “generic” and may, but does not necessarily, mean “mail”; reports can be sent (without being mailed) by private courier service. Therefore, defendants assert that Thompson’s statement that the report was “sent” is not sufficient to establish that an alleged mailing in fact occurred and that, at most, only a probability of mailing is indicated. Notwithstanding this argument, proof of a company’s office practice and its “usual course of business,” while circumstantial, has been found sufficient to carry the question of mailing to the jury. E.g., United States v. Scott, 730 F.2d 143, 146-47 (4th Cir.), cert. denied, - U.S. -, 105 S.Ct. 572, 83 L.Ed.2d 512 (1984). Here the government offered evidence of ordinary business practice and testimony by Thompson denying knowledge that the letter was sent by a method other than mail.5
With respect to the mailing of the TCC final report, technician Smith testified that the business practice was to mail a copy of the final report to the sponsor of the study. Defendants contend that Smith had no personal knowledge of the IBT’s business practices and that his testimony was too “off-hand” to establish that the report was mailed. We believe that the testimony was sufficient to carry this question of mailing to the jury. Defendants presented other arguments on the mailing question which included evidence of a transmission by air courier service on May 8th about the time the TCC report was allegedly mailed, discrepancies in dates on the cover letter and final report and a Xerox mailing instruction sheet containing handwriting (which was not usually found on an IBT final report). Since the FDA’s copy of the report did not contain any handwriting, defendants argue that the government’s proof of mailing is inadequate.
The government showed that the May 8 air courier delivery was sent to Dr. Wright at Monsanto. Therefore, this delivery did not involve the mailing of the TCC final report charged in the indictment. That mailing was sent to a different Monsanto employee, Dan Roman, on or about May 12. As the government notes, the fact that IBT’s mailing instruction sheet for the final report is dated May 12, a day after the date that a cover letter indicates Monsanto sent *692the report to the FDA, is explainable in several ways. The most likely inference from the evidence presented is that Monsanto had obtained a copy of the report before the copy sent to it on May 12th (perhaps through the May 8 courier delivery that defendants so painstakingly proved). The government also rebuts the contention that the handwriting on the IBT report was unusual by pointing to circumstantial evidence that the writing was added to the file copy after the report was actually mailed. In sum, the government offered evidence to show both that it was IBT’s regular practice to mail copies of final reports to sponsors, and that IBT mailed the TCC report to Monsanto as charged in the indictment. The mere existence of conflicting evidence (which the government explains in a rational manner) does not establish an insufficiency of proof. See, e.g., United States v. Goss, 650 F.2d 1336, 1343 (5th Cir.1981). We are of the opinion that the government presented enough evidence to take to the jury the issue whether the Naprosyn and the TCC final reports were mailed.
II.
Defendants next complain that a multitude of exhibits were erroneously admitted pursuant to the business records exception to the hearsay rule, Fed.R.Evid. 803(6). The government intended to seek admission of about 194 documents which it characterized as IBT business records; these documents were listed on Government Exhibit WT-1. Generally, the IBT records consisted of interdepartmental memos, reports, data compilations and letters between IBT and its sponsors. In order to avoid the delay and confusion that would no doubt have arisen if the admissibility of each document were determined separately at the time it became relevant, the trial court approved a procedure whereby a government witness, Warren Thompson, provided foundation testimony at one time for all of the documents listed on WT-1. The court did not immediately admit all of the documents into evidence, but over defense objections allowed government witnesses to testify about some of them, based on the government’s representation that it would provide further foundation for their admission. Another document witness, Philip Smith, was later called to testify about the IBT records. After entertaining the parties’ arguments as to admissibility and considering each document individually, the district court eventually overruled defendants’ objections that the government had not provided sufficient foundation for admitting the IBT documents as business records. Defendants now contend that this “wholesale” approach to admission of the various exhibits was improper, and that about 70 exhibits did not meet every requirement and so were improperly admitted as evidence.6
Initially, we note that there is nothing inherently objectionable about the “wholesale” approach adopted by the trial court. Under that procedure the government witnesses provided foundation testimony about all of the documents at the same time and testified about large groups of documents by type (for example, memoranda and correspondence with sponsors). In a case of this magnitude, we see no need to require the government to elicit essentially identical foundation testimony separately for each exhibit offered. To the extent that defendants found it necessary to explore information relevant only to a particular document, they were able to do so. The district court considered each document separately. The court was quite understandably concerned that an approach other than eliciting “wholesale” foundation testimony would impede a sensible presentation of the evidence, and its decision to proceed in the manner it did was well within its discretion. See United States v. Evans, 572 F.2d 455, 490-91 (5th Cir.), cert. *693 denied, 439 U.S. 870, 99 S.Ct. 200, 58 L.Ed.2d 182 (1978).
The business records exception to the hearsay rule, Fed.R.Evid. 803(6), provides that certain records of regularly conducted activity are not excluded by the hearsay rule. In light of the recognition that ordinary business records tend to be unusually reliable, the exception is designed to avoid the burdensome process of “producing as witnesses, or accounting for the nonproduetion of, all participants in the process of gathering and transmitting, and recording information____” See Notes of Advisory Committee on Proposed Rule 803. In order to be admissible, a record must be:
[a] memorandum, report, record, or data compilation, in any form, of acts, events, conditions, opinions, or diagnoses, made at or near the time by, or from information transmitted by, a person with knowledge, if kept in the course of a regularly conducted business activity, and if it was the regular practice of that business activity to make the memorandum, report, record, or data compilation, all as shown by the testimony of the custodian or other qualified witness, unless the source of the information of the method or circumstances of preparation indicate lack of trustworthiness____
The determinations whether a proper foundation has been laid for application of the business records exception to a particular document and whether the circumstances of the document’s preparation indicate untrustworthiness are within the discretion of the trial court. E.g., United States v. Cincotta, 689 F.2d 238, 243 (1st Cir.1982), cert. denied, 459 U.S. 991, 103 S.Ct. 347, 74 L.Ed.2d 387 (1983).
Defendants contend that the government failed to satisfy at least one of the primary foundation requirements for each document admitted pursuant to Rule 803(6) (or rather for each group of documents, since defendants also treat the admissibility question by category of documents, rather than individually). First, defendants argue that neither of the two government document witnesses should properly have been considered a “custodian or other qualified witness,” since neither witness had sufficient knowledge of IBT’s record-making procedures. We disagree. Warren Thompson had worked at IBT in various capacities for eleven years, and testified that through his recent position as an auditor of studies, he was familiar with the types of documents created and maintained by IBT and offered as exhibits by the government. Philip Smith worked at IBT between 1971 and 1977 as an assistant toxicologist, responsible for scheduling rat studies, preparing study reports and auditing studies. He had seen about 50 to 60% of the documents in issue while working at IBT, and was generally familiar with the creation and maintenance of IBT records. Smith testified that each of the types of documents that the government sought to admit as business records (memoranda, correspondence with sponsors, and raw data) was created and maintained as part of the regularly conducted business activity at IBT; and was prepared by or from information transmitted by people with personal knowledge of the facts described in them, and was written about the time of the events described.
Although restated in various ways, defendants’ arguments really boil down to the proposition that because Thompson and Smith did not personally participate in the creation of some of the documents, they were not qualified to say whether the preparers actually had personal knowledge of the facts described, when the documents were prepared in relation to the time of the events and like questions. But the business records exception contains no requirement that a “qualified witness” must have personally participated in the creation or maintenance of a document, see, e.g., Itel Capital Corporation v. Cups Coal Company, Inc., 707 F.2d 1253, 1259-60 and n. 13 (11th Cir.1983); Rosenberg v. Collins, 624 F.2d 659, 665 (5th Cir.1980); United States v. Rose, 562 F.2d 409, 410 (7th Cir.1977), nor even know who actually recorded the information, see United States v. Verlin, 466 F.Supp. 155, 158 (N.D.Tex. *6941979). Obviously, such a requirement would eviscerate the business records exception, since no document could be admitted unless the preparer (and possibly others involved in the information-gathering process) personally testified as to its creation. Rather, the phrase “other qualified witness” in Rule 803(6) is to be given “the broadest possible interpretation” — the witness need only be someone who “understands the system.” 4 J. Weinstein & M. Berger, Weinstein’s Evidence H 803(6)[02] at 803-178. See also United States v. Wables, 731 F.2d 440, 449 (7th Cir.1984) (witness must only have knowledge of procedures under which records created); NLRB v. First Termite Control Company, Inc., 646 F.2d 424, 427 n. 5 (9th Cir.1981) (court quotes Notes of the Committee on the Judiciary Senate Report No. 93-1277, U.S.Code Cong. & Admin.News 1974, p. 7051, which indicate that party seeking to introduce evidence must show it was “regular practice” of the activity to base documents upon information provided by a person with knowledge). In sum, a “qualified witness” need only be someone with knowledge of the procedures governing the creation and maintenance of the type of records sought to be admitted; and this requirement was met here.
Defendants’ other attacks on the government’s foundation testimony fare no better and do not merit extended discussion. Although interoffice memoranda are sometimes excluded where they are not created in the course of regularly conducted business activity, see, e.g., Niederkrome v. Commissioner of Internal Revenue, 266 F.2d 238, 242 (9th Cir.1958), cert. denied, 359 U.S. 945, 79 S.Ct. 725, 3 L.Ed.2d 678 (1959), they are, not surprisingly, admissible where they are created in the regular course of business, see, e.g., United States v. McGrath, 613 F.2d 361, 367-68 (2d Cir.1979), cert. denied, 446 U.S. 967, 100 S.Ct. 2946, 64 L.Ed.2d 827 (1980), and the testimony here indicated the latter. Similarly, the government witnesses testified that the documents were maintained in the course of a regularly conducted activity; the fact that some documents were not kept in IBT’s central files does not refute the government’s evidence that the documents were regularly maintained. See United States v. Hedman, 630 F.2d 1184, 1197-98 (7th Cir.1980), cert. denied, 450 U.S. 965, 101 S.Ct. 1481, 67 L.Ed.2d 614 (1981). The evidence that the pathology sheets and necropsy logs were prepared by persons with personal knowledge was also sufficient. A witness testified that the pathology sheets, which were the sources of data for the necropsy logs, were prepared by animal laboratory technicians. Thompson was an animal lab technician, and Smith wrote or audited reports that used the pathology sheet data; Smith testified that raw data was prepared by persons with personal knowledge. In the absence of any indication to the contrary, this testimony and the very nature of the data are sufficient to support the conclusion that the data were generated by someone with personal knowledge. See United States v. McGrath, 613 F.2d at 368 (evidence itself may afford the basis for inferring that it rests on firsthand knowledge). Letters created by another business but regularly received, maintained and relied upon by IBT, as was demonstrated here, may also constitute admissible business records of IBT. See United States v. Flom, 558 F.2d 1179, 1182-83 (5th Cir.1977); United States v. Pfeiffer, 539 F.2d 668, 670-71 (8th Cir.1976).
Finally, various inaccuracies in the body weight books, cited by defendants as evidence of their lack of trustworthiness, are insufficient to demonstrate that the trial court abused its discretion in admitting them. Generally, objections that an exhibit may contain inaccuracies, ambiguities, or omissions go to the weight and not the admissibility of the evidence. See, e.g., Matador Drilling Company, Inc. v. Post, 662 F.2d 1190, 1199 (5th Cir.1981). Of course, the source of the information about or the circumstances surrounding the preparation of a document may indicate that the document is insufficiently trustworthy to be admissible. See Fed.R.Evid. 803(6). In the case before us, various testimony *695indicated that the animal technicians strove to measure and record the body weights accurately and generally did so. We doubt whether inaccuracies apparent on the face of a document (as opposed to testimony about the unreliability of the procedure used to prepare it) should affect its admissibility as a business record, but in any event our review of the evidence persuades us that there was no abuse of discretion in the trial court’s decision to admit the body weight books.
Defendants also contend that the admission of two exhibits violated their Sixth Amendment confrontational rights. The Confrontation Clause of the Sixth Amendment to the Constitution provides that in criminal prosecutions, “the accused shall enjoy the right ... to be confronted with the witnesses against him____” The confrontation clause and the federal rules limiting the use of hearsay declarations at trial protect similar values, but the conclusion that a statement is admissible under one of the exceptions to the hearsay rule does not eliminate the need to evaluate whether admission of the statement contravenes the defendant’s right of confrontation. See Mattes v. Gagnon, 700 F.2d 1096, 1101 (7th Cir.1983); United States v. Perez, 658 F.2d 654, 660 (9th Cir.1981).
Without attempting to set forth a definitive test for evaluating hearsay testimony under the confrontation clause, the Supreme Court identified the general limitations on the use of hearsay in Ohio v. Roberts, 448 U.S. 56, 100 S.Ct. 2531, 65 L.Ed.2d 597 (1980). First, the Sixth Amendment establishes a “rule of necessity,” so that in the usual case, “the prosecution must either produce, or demonstrate the unavailability of, the declarant whose statement it wishes to use against the defendant.” Id. at 65, 100 S.Ct. at 2538 (citations omitted). Second, once a witness is shown to be unavailable, it must also be demonstrated that the hearsay is trustworthy; it must bear sufficient “indicia of reliability.” 7 Id. at 65-66, 100 S.Ct. at 2538-2539. The rule of necessity is not absolute, however. The government need not produce the hearsay declarant or demonstrate his unavailability where “the utility of trial confrontation is remote.” Id. at 65, n. 7, 100 S.Ct. at 2538, n. 7. Moreover (under the “Dutton exception”), at least where the evidence is neither “crucial” to the presentation, nor “devastating” to the defendant, the evidence may be admissible even though the necessity requirement is not met. See Dutton v. Evans, 400 U.S. 74, 87-89, 91 S.Ct. 210, 219, 27 L.Ed.2d 213 (1970); United States v. McClintock, 748 F.2d 1278, 1292 (9th Cir.1984). The district court relied largely on this “Dutton'exception” in concluding that the confrontation clause was not violated by the admission of evidence here. See United States v. Keplinger, 572 F.Supp. 1068, 1071 (N.D.Ill. 1983).
Defendants complain that admission of two exhibits violated their rights to confrontation. The first exhibit, II-E-19, is the so-called “Kennedy memo.” This is a memorandum by Gerald Kennedy from which it arguably can be inferred that after the final sacrifice of rats in the TCC study, Kennedy took tissues from rats in the late-started “research” study and added them to tissues from the original study for histopathological examination by Dr. Gordon. The memo also indicates different numbers for some animals than those given in a study report. The government argues that admission of this memorandum presents no confrontation problems at all, since the memo has no significance as hearsay, i.e. as a statement offered to prove the truth of the matter asserted, but was only significant because the memo was passed around and discussed at a meeting attended by defendants and so is relevant to show defendants’ knowledge that animals had been commingled at the final sacrifice. Although we agree that the primary impor*696tance of the memo was to show defendants’ knowledge, we cannot accept the government’s argument that the memo presents no hearsay problems whatever. Our review of the record indicates that the Kennedy memo was offered and received as a business record, without limitation on the use to which the evidence could be put. Therefore, the jury could have viewed the memo as direct evidence a) that Gerald Kennedy had given Dr.' Gordon tissues from the late-started research animals and/or b) that Kennedy had changed the numbers of some animals. From these propositions the jury could have drawn the inference, suggested by the government, that Kennedy had changed the numbers in order to conceal the commingling of the late-started animals. Thus we cannot conclude that the hearsay aspect of the exhibit had no significance.
On the other hand, we are persuaded that the evidence was properly admitted under the Dutton rationale. Even though the government did not demonstrate the unavailability of the hearsay declarant, Gerald Kennedy, admission of the memo did not violate the confrontation clause if it was not “crucial” to the government or “devastating” to the defense, and if the utility of cross-examination of Kennedy was remote. The evidence was clear, and the defendants conceded, that Kennedy had indeed given Dr. Gordon slides of tissues from late-started animals. There was various other evidence indicating that numbers for some animals had been changed, including a handwritten reference in another exhibit to the renumbering in Kennedy’s memo, and further testimony that several (unsuccessful) attempts had been made to match numbers of animals indicated in raw data to numbers used on a summary table. Thus, taken for its truth value, the memo was cumulative; it could only have helped to establish two propositions that were essentially indisputable, and so was hardly “crucial” to the government’s case. Moreover, we also find it relevant that the memo was not accusatory in any fashion, nor was there any sort of direct statement that numbers had been changed in order to conceal commingling.
For similar reasons, we believe that the potential utility of cross-examination of Gerald Kennedy was remote. Although defendants acknowledge that in most instances cross-examination of the authors of business records would be pointless, they argue that cross-examination would have been fruitful here because the memo was ambiguous. In light of the other evidence presented we are not certain that the memo can be considered ambiguous, but in any event cross-examination of Kennedy would not appear to have been useful in clarifying any important question. The prime import of the memo is that its circulation at a meeting attended by defendants arguably should have put them on notice, if they did not already know, that commingling might have occurred. Kennedy did not attend that meeting; testimony by him as to his intentions in writing the memo and his perceptions of its meaning would not have been relevant to the issue of what defendants understood or should have understood from examining it. Moreover, since defendants conceded that Kennedy had given Gordon tissues from late-started animals, and since the memo shows on its face that some renumbering occurred, there is little possibility that defendants could have gained much benefit from cross-examining Kennedy about these matters. Defendants apparently would not have attempted to prove that these events did not occur, but only wanted to challenge the prosecution’s “sinister interpretation” of them. But, as noted, Kennedy’s intentions or beliefs (for example, about whether he changed the numbers in order to conceal the inclusion of tissues from late-started rats, or for some other innocent reason) would have been irrelevant to defendants’ knowledge.
Defendants also argue that admission of the body weight books violated defendants’ confrontational rights. The government used the body weight books primarily to support the inference that commingling had occurred. In particular, the government contended that 1) a dash in one weight interval followed by a weight in the *697next weight interval or 2) substantial weight fluctuations between weighing intervals meant that the rat in question had died and had been replaced by a late-started rat. Defendants argue that cross-examination with respect to the makers of the weight books was vital, since the meaning of a dash or weight fluctuation is ambiguous and may have meant something other than a death and replacement. But the government did produce six witnesses, each of whom actually made entries in the weight books or supervised the maintenance of the books. Two of these witnesses were technicians in charge of running the TCC study for most of its duration. The defense had the opportunity to cross-examine each of these witnesses, and indeed elicited from several an acknowledgement that weight fluctuations might have been attributable to factors that were known to other technicians other than animal death and substitution. Possibly these other technicians might have meant something other than a death when they used a dash. Thus, defendants did have the opportunity to cross-examine many if not all of the authors of the documents.
But defendants argue this was not enough. Without naming other technicians who might have recorded weights in the books, defendants contend that the government’s failure to produce all contributors to the body weight books violated their rights. In the absence of the opportunity to cross-examine every technician who contributed to the exhibits, defendants say, they could not show that the innocent “possibilities” suggested on cross-examination were “actualities.” We agree with the government that the potential utility of cross-examining such additional witnesses is negligible. First, the possibility that some additional technicians would testify that they had a significantly different understanding of the meaning of dashes or weight fluctuations than had those who already testified seems remote, if not wholly illusory. Second, even if a few technicians would have ascribed a different, “innocent,” meaning to the use of dashes or to weight fluctuations, this could at most establish that not all of the dashes and fluctuations meant deaths and replacements; such evidence could not prove that commingling did not occur, given the testimony of the technicians who took the stand. Therefore we believe the body weight books fall within the Dutton exception to the confrontation rule, which otherwise would require a demonstration that all the makers of the record be produced as witnesses or shown to be unavailable.
III.
Defendants also contend that their convictions on Count I should be reversed, because the jury might have found them guilty on the basis of omissions from the final TCC report that may not properly be used to support a mail fraud conviction. In that report, defendants represented that T-l was a no-effect level in the testis. As previously noted, a report by Dr. Ribelin, an independent consultant, had concluded that T-l had no significant effect on the testis, but that degenerative changes in the epididymis indicated that a no-effect level had not been reached with respect to the epididymal tissue. Although aware of these conclusions, defendants did not include or refer to them in the final TCC report. The government maintained that this omission, among others, was fraudulent and the district court allowed this evidence to go to the jury on Count I.8
We find defendants’ argument here wholly unpersuasive. It requires no extended discussion of authority to demonstrate that omissions or concealment of material information can constitute fraud or concealment of material information can constitute fraud cognizable under the mail fraud statute, without proof of a duty to disclose the information pursuant to a specific statute or regulation. See, e.g., Unit *698 ed States v. Lindsey, 736 F.2d 433, 436-37 (7th Cir.1984); United States v. Margiotta, 688 F.2d 108, 123-24 (2d Cir.1982), cert. denied, 461 U.S. 913, 103 S.Ct. 1891, 77 L.Ed.2d 282 (1983). See also United States v. Black, 684 F.2d 481, 484 (7th Cir.), cert. denied, 459 U.S. 1043, 103 S.Ct. 463, 74 L.Ed.2d 613 (1982); United States v. Shelton, 669 F.2d 446, 456-58 (7th Cir.), cert. denied, 456 U.S. 934, 102 S.Ct. 1989, 72 L.Ed.2d 454 (1982). Similarly, “[wjhile the existence of a fiduciary duty is relevant and an ingredient in some/mail fraud prosecutions, ... it is not an essential in all such eases.” United States v. Allen, 554 F.2d 398, 410 (10th Cir.), cert. denied, 434 U.S. 836, 98 S.Ct. 124, 54 L.Ed.2d 97 (1977) (citations and footnote omitted).9 Obviously, we do not imply that all or even most instances of non-disclosure of information that someone might find relevant come within the purview of the mail fraud statute; nevertheless, under some circumstances concealment of material information is fraudulent. The scheme to defraud “is not to be measured by technicalities. Rather, the measure of fraud is its departure from moral uprightness, fundamental honesty, fair play and candid dealings in the general life of members in society.” United States v. Lindsey, 736 F.2d at 436. Therefore fraud can be effected not only by deceitful statements but also by statements of half-truths or concealment of material facts. See United States v. Townley, 665 F.2d 579, 585 (5th Cir.), cert. denied, 456 U.S. 1010, 102 S.Ct. 2305, 73 L.Ed.2d 1307 (1982); United States v. Allen, 554 F.2d at 410.
We have already discussed in connection with the sufficiency of the evidence defendants’ arguments that the omission of Ribelin’s conclusions from the final report was not fraudulent; and we see no need to belabor the matter. Moreover, our review of the testimony has convinced us that the issues involving omissions were properly submitted to the jury. Defendants’ heavy emphasis on the physiological differences between the testis and the epididymis does not advance their argument that the concealment of Ribelin’s findings was not fraudulent. First, there was some evidence to support the inference that there was no significant difference between the two entities for purposes of evaluating the effect of TCC. At least, Dr. Ribelin regarded his findings on the epididymis to be within the area of relevance to the question whether TCC affected the testis. More important, let us assume that the testis and epididymis are wholly distinct organs. This assumption would only imply that the statement that T-l was a no-effect level in the testis was not, given Dr. Ribelin’s conclusions, a bald misrepresentation. This state of affairs does not imply, however, that it was not fraudulent for defendants to have omitted any mention of Ribelin’s conclusions from the report. There was ample testimony at trial to the effect that the purpose of the TCC study was to determine whether there was a level at which TCC could be administered without causing ill effects. There was evidence that, while the testes were selected as the target organs, any information about significant effects in other areas would be relevant and important to the central question whether TCC could be safely used in humans. Therefore, there was evidence that Dr. Ribelin’s findings should have been included.10 Of course, defendants presented tes*699timony to the opposite effect. But this only means that there was a real issue as to whether defendants’ failure to mention Dr. Ribelin’s conclusions was fraudulent, which we believe was properly submitted to the jury for resolution. Again, we do not imply that every omission from a scientific report of potentially important information could form the basis of a mail fraud conviction; in this case, however, where the government presented strong evidence of a scheme to defraud consisting of a wide variety of deceptive actions, we believe the omission could properly be considered as part of that scheme.11
IV.
Prior to trial, all defendants moved to suppress certain testimony and materials which they alleged were obtained in violation of their attorney-client privilege. Merrill Thompson and Fred Current were two attorneys hired by Nalco, Inc., IBT’s parent corporation, to represent IBT as a corporation. It was undisputed that an attorney-client relationship existed between IBT and Thompson and Current, and that IBT waived its corporate attorney-client privilege as to all matters relating to this case. Defendants contended, however, that an attorney-client relationship also existed with respect to each of them individually, so that communications made by them to Thompson and Current were protected despite IBT’s waiver of its privilege. The trial court held an extensive six-day hearing on this matter, at which testimony from the two attorneys, as well as Keplinger and Plank,12 among others, was received. Thereafter the court entered a written opinion which included detailed factual findings and concluded that defendants had failed to establish the existence of an attorney-client relationship between themselves individually and the corporate attorneys. Defendants now argue that the court’s factual findings were either unsupported by the evidence, or legally immaterial.
We apply the clearly erroneous standard in reviewing the trial court’s findings of fact. The Supreme Court has recently had occasion to clarify the meaning of that standard in Anderson v. City of Bessemer City, - U.S. -, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985). The Court explained:
If the district court’s account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently. Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.
When findings are based on determinations regarding the credibility of witnesses, Rule 52 demands even greater deference to the trial court’s findings____ [w]hen a trial judge’s finding is based on his decision to credit the testimony of one of two or more witnesses, each of whom has told a coherent and facially plausible story that is not contradicted by extrinsic evidence, that finding, if not internally inconsistent, can virtually never be clear error.
Id., 105 S.Ct. at 1512-13 (citations omitted). These principles apply with special force in this case, since the testimony of the attorneys and that of the defendants directly conflicted in several important particulars, and the trial court was careful to note that it found both Thompson and Current to be “highly credible witnesses.”
*700This court has adopted a number of well-settled principles governing the creation and scope of the attorney-client privilege. In general:
(1) Where legal advice of any kind is sought (2) from a professional legal adviser in his capacity as such, (3) the communications relating to that purpose, (4) made in confidence (5) by the client, (6) are at his instance permanently protected (7) from disclosure by himself or by the legal adviser, (8) except the protection be waived.
Radiant Burners, Inc. v. American Gas Association, 320 F.2d 314, 319 (7th Cir.), cert. denied, 375 U.S. 929, 84 S.Ct. 330, 11 L.Ed.2d 262 (1963) (quoting 8 J. Wigmore, Evidence § 2285 (Rev. ed. 1961)). The district court here found that defendants had failed to demonstrate the existence of several necessary requirements for creation of individual attorney-client relationships between themselves and the IBT corporate attorneys. In particular, the court found that no express agreement for individual representation was made, nor did any of the defendants ever ask the attorneys directly or indirectly to represent him individually. In addition, the court found that the circumstances were not such as to support a finding of any implied individual representation. The defendants never sought individual legal advice or asked questions relating to personal representation, and, quite importantly, never made any confidential statements of an incriminating nature to the attorneys,13 from which it could be inferred that the defendants believed that the attorneys were representing them individually. Neither attorney believed they represented the defendants individually, nor did they believe the defendants thought they enjoyed such representation. Indeed, the district court stated that credible testimony indicates that the relationship between the attorneys and the defendants actually was somewhat adversarial, and by December 1976, had become confrontational. Finally, the court found that communications made by the defendants were not in confidence, both because the information communicated was intended to be disclosed in IBT’s responses to the FDA’s inquiries and because most of defendants’ statements to attorneys were made in the presence of third parties.
We believe that the district court’s findings were fully supported by the evidence. Defendants do not dispute the attorneys’ testimony that defendants never explicitly sought individual legal advice, or asked about individual representation, nor did the attorneys ever indicate to defendants their belief that such a relationship existed. (To the contrary Thompson testified that he once told Plank explicitly that he was not Plank’s lawyer.) Defendants merely argue in various ways that they subjectively believed that the attorneys represented them individually, on the basis of such circumstances as Thompson’s appearance with Keplinger at a meeting with the FDA.14
*701Defendants cite Westinghouse Electric Corporation v. Kerr-McGee Corporation, 580 F.2d 1311, 1319 (7th Cir.), cert. denied, 439 U.S. 955, 99 S.Ct. 353, 58 L.Ed.2d 346 (1978) for the proposition that the scope of the attorney-client relationship “ ‘hinges upon the client’s belief that he is consulting a lawyer in that capacity and his manifested intention to seek professional legal advice.’ ” Id. at 1319 (quoting McCormick on Evidence § 88 at 179 (2d ed. 1972)). But while we do not quibble with that statement, we do not believe it can be deemed equivalent to defendants’ implicit argument that an individual’s mere subjective belief that he is represented individually will always be sufficient to demonstrate that such a relationship existed for the purpose of the attorney-client privilege. First, the quoted language refers to the client’s “manifested intention” to seek legal advice. As already noted, there is little or no evidence in this case that defendants either sought legal advice on an individual basis or manifested in any way their belief that they were being represented individually. Moreover, at least in the absence of any relatively clear indication by the potential client to the attorney that he believed he was being individually represented, we think no individual attorney-client relationship can be inferred without some finding that the potential client’s subjective belief is minimally reasonable. For example, in Westinghouse, the attorneys had specifically represented that the information disclosed to them by the individual clients would be kept confidential. See 580 F.2d at 1314, 1321. See also In re Grand Jury Proceedings (Jackier), 434 F.Supp. 648, 650 (E.D.Mich.1977) (corporate officer may invoke attorney-client privilege individually only if he “makes it clear when he is consulting the company lawyer that he personally is consulting the lawyer, and the lawyer sees fit to accept and give communication knowing the possible conflicts that could arise....”), affd, 570 F.2d 562 (6th Cir.1978). In sum, the district court’s findings were not clearly erroneous and its conclusion that defendants could not invoke the attorney-client privilege to suppress the attorneys’ testimony was correct.
Defendants also contend that the attorneys should have been prevented from testifying on the basis of the “joint defense” doctrine, which generally allows a defendant to assert the attorney-client privilege to protect his statements made in confidence not to his own lawyer, but to an attorney for a co-defendant for a common purpose related to the defense of both. See, e.g., United States v. McPartlin, 595 F.2d 1321, 1336-37 (7th Cir.), cert. denied, 444 U.S. 833, 100 S.Ct. 65, 62 L.Ed.2d 43 (1979). Thus defendants argue that (largely unspecified) communications they made to Thompson and Current while those attorneys represented IBT may not be disclosed because they were made in the course of cooperating about a matter of common interest. We have serious doubts whether IBT and the individual defendants really shared a “common purpose” in the sense necessary to support application of the joint defense doctrine. Regardless, however, defendants’ argument is defeated by the trial court’s finding that the communications made by the defendants to the attorneys were not in confidence. That finding is not clearly erroneous, and so the joint-defense doctrine, even if otherwise applicable, does not prevent their disclosure. See, e.g., United States v. McPartlin, 595 F.2d at 1336.
V.
At trial, neither the government nor the defendants called Gerald Kennedy as a witness. Defendants argue that the district court erred in refusing to give a “missing witness” instruction based on the fact that the government did not call Kennedy, or in allowing defendants to comment on Kennedy’s absence in closing argument. “ ‘The rule ... is that, if a party has it peculiarly within his power to produce witnesses whose testimony would elucidate the transaction, the fact that he does not do it creates the presumption that the testimony, if produced, would be unfavorable.’ ” United States v. Mahone, 537 F.2d 922, *702926 (7th Cir.), cert. denied, 429 U.S. 1025, 97 S.Ct. 646, 50 L.Ed.2d 627 (1976) (quoting Graves v. United States, 150 U.S. 118, 121, 14 S.Ct. 40, 41, 37 L.Ed. 1021 (1893)). We assume for the sake of argument that Kennedy's testimony would have “elucidated the transactions” at issue, since as a member of IBT’s “core group,” he was involved in many relevant aspects of the TCC study. It is also undisputed that Kennedy was physically available to both sides. Thus the only issue we must address is whether, in some pragmatic sense, Kennedy was peculiarly within the government’s power to produce. See United States v. Mahone, 537 F.2d at 926.
A party can show that an absent witness is effectively within the other party’s control, by establishing, for example, that the witness is biased in favor of the opposing party or has some other reason to testify solely on that party’s behalf. See United States v. Mahone, 537 F.2d at 926-27. Defendants make several arguments to support the conclusion that Kennedy was peculiarly within the government’s control. First is the bizarre contention that because the government immunized Kennedy from prosecution in order to obtain his testimony before the grand jury, he was effectively unavailable to defendants. In the usual case, a defendant argues the precise opposite — that a witness is within the government’s control because the government has refused to immunize him, and it can be presumed that he would refuse to testify on grounds of self-incrimination. See, e.g., United States v. Flomenhoft, 714 F.2d 708, 713-14 (7th Cir.1983), cert. denied, 465 U.S. 1068, 104 S.Ct. 1420, 79 L.Ed.2d 745 (1984). But defendants reason that because Kennedy’s immunity grant was conditioned upon his giving truthful testimony, had Kennedy testified for the defense, and contrary to his prior grand jury testimony, he would have subjected himself to perjury charges. On that rationale, any witness who testified favorably toward the government at grand jury proceedings could later be considered unavailable because he might be unwilling to perjure himself by changing his story in favor of the defense. This argument is quite unpersuasive.
Defendants offer a variety of other reasons why Kennedy should be considered within the government’s control, none of which have merit. Kennedy is not an informant, nor is there any evidence that he would have been “vindicated by a conviction of the defendants].” See United States v. Mahone, 537 F.2d at 927. To the contrary, Kennedy could be expected to prefer rather strongly a not-guilty verdict, given his close involvement with the TCC study and his status as an unindicted co-conspirator.15 Kennedy’s cooperation with the government prior to his grand jury testimony also is insufficient to demonstrate that he was effectively unavailable to the defense. Even where a witness entirely refuses to discuss a case with the defense, a missing witness instruction may be appropriately denied. See United States v. Grizaffi, 471 F.2d 69, 74 (7th Cir.1972), cert. denied, 411 U.S. 964, 93 S.Ct. 2141, 36 L.Ed.2d 684 (1973). Yet in this case, Kennedy met with defense counsel on several occasions, with no government representative present. In sum, there is no indication that Kennedy was “peculiarly within the government’s control” and therefore the district court properly refused a missing witness instruction.16
For the same reasons, it was not error for the district court to preclude defendants from arguing the missing witness inference to the jury in closing argument. In a close case, it may be appropriate to allow a party to comment in argument on the significance of a witness’ absence, even where there is insufficient evidence of una*703vailability to warrant a missing witness instruction. See Chicago College of Osteopathic Medicine v. George A. Fuller Company, 719 F.2d 1335, 1352-54 (7th Cir. 1983); United States v. Mahone, 537 F.2d at 927-28. But where the witness appears to be equally available or unavailable to both sides, we think it is well within the district court’s discretion to refuse to allow such argument. See, e.g., Chicago College of Osteopathic Medicine v. George A. Fuller Company, 719 F.2d at 1352-53.17 In some such cases, both sides presumably would argue that the missing person’s testimony would have been unfavorable to the other party, which at least in a complex case like this one would allow the jury to speculate about the meaning of a great deal of non-evidence. We see no constructive purpose to be served by such a procedure, and conclude the district court did not abuse its discretion in preventing commentary on Kennedy’s absence in closing argument.
VI.
Defendants also contend that the district court erred in allowing Dr. Gordon to testify for the government despite the fact that prior to trial he had undergone hypnosis in part to enhance his memory of events about which he testified. Between April 1979 and August 1980, while a criminal investigation of IBT was underway, the government interviewed Dr. Gordon at various times. In the fall of 1980, Dr. Gordon’s attorney in the investigation suggested that Dr. Gordon seek help to alleviate stress he was experiencing in part due to his inability to recall certain relevant events. Dr. Gordon eventually saw Dr. John Adams, a psychologist at Northwestern Medical School who had experience in enhancing memory with hypnosis. Dr. Gordon saw Dr. Adams over five sessions; at four of the sessions Dr. Gordon underwent hypnosis. These sessions had the dual goals of relieving Dr. Gordon’s stress and enhancing his memory. The government apparently was not aware that Dr. Gordon was seeing Dr. Adams.
Defendants contended at trial that Dr. Gordon’s post-hypnotic testimony was so unreliable that it should be excluded, as the result of dangers that may be associated with post-hypnotic testimony generally,18 and their view that such dangers were present in this case. Defendants presented an expert witness, Dr. Bernard Diamond, who testified to this effect. The court also heard testimony by Dr. Adams, Dr. Gordon and Dr. Gordon’s attorney, which concerned, among other things, Dr. Gordon’s recollections both prior to and after the hypnotic sessions and Dr. Adams’ knowledge of the underlying events that were of concern to Dr. Gordon. In addition to this extensive voir dire, the court reviewed a significant amount of ease authority and professional literature concerning the reliability of post-hypnotic testimony. The court subsequently ruled that Dr. Gordon *704was competent to testify despite his experience with hypnosis.
Defendants have made very broad arguments in support of their contention that this court should adopt a rule prohibiting post-hypnotic testimony per se (regardless of the circumstances involved) or, at the very least, adopt a comprehensive set of standards that would govern the admissibility of post-hypnotic testimony in all cases (and would, of course, operate to exclude the testimony admitted here). We decline to adopt a per se rule that would exclude, under any circumstances, testimony of persons who previously underwent hypnosis, concerning a subject which might have been involved in the hypnotic sessions. Defendants have simply offered no reason for us to seriously consider doing so. Moreover, while we are well aware that the reliability of post-hypnotic testimony poses a recurring and serious question, we see no need to pronounce general standards governing admissibility. Even accepting, for the purposes of this case, defendants’ proposed standards, it is clear that the district court did not abuse its discretion in admitting Dr. Gordon’s testimony.
Defendants suggest that, in order to demonstrate that post-hypnotic testimony “is free from the distorting effects inherent in hypnosis,” the government should have demonstrated either a) that adequate safeguards were followed before and during hypnosis to minimize hypnotic distortions of memory, or b) that Gordon’s trial testimony reflected only statements clearly demonstrated to have been made prior to hypnosis. Wright’s Reply Br. at 7. But what defendants largely ignore is that the district court made several explicit factual findings that show that the government met the requirements defendants propose. Most important, the court found that the attempt to enhance Dr. Gordon’s memory was, quite simply, unsuccessful; no additional items of memory were induced. If this finding is correct, then by definition Dr. Gordon’s trial testimony must only have concerned events about which he had knowledge and memory independent of the hypnosis. This finding alone distinguishes this case from others in which there is no doubt that the witness’s memory was enhanced and the court must decide whether that “memory” is sufficiently reliable to justify admission. See, e.g., United States v. Valdez, 722 F.2d at 1196. Moreover, the district court made other findings that buttress its conclusion that the dangers associated with post-hypnotic testimony were minimized here. First, the court found that the prosecution had not initiated or participated in the hypnosis, thus minimizing the chance that it could have deliberately or inadvertently suggested new “memories” to Dr. Gordon. To similar effect, the court found that Dr. Adams was not given any. detailed knowledge about the underlying facts of the case, and so he did not have enough information to allow him to suggest significant new memories. Thus, according to the district court’s findings, there was no apparent source from which Dr. Gordon could have acquired new memories about the relevant events, nor was there any evidence that Dr. Gordon had in fact acquired such memories. Based on these findings, the court concluded that any prejudice caused by Dr. Gordon’s testimony did not substantially outweigh the probative value of that testimony, and so admitted it.
But defendants contend that the hypnotic sessions must have enhanced Dr. Gordon’s memory, because only after those sessions did Dr. Gordon “remember” that he believed he had been pressured by defendants in a meeting in February 1976 to change his opinion about whether T-l was a no-effect level. We disagree. First, generally, both Dr. Gordon and Dr. Adams specifically testified that Dr. Gordon did not have any new memories as the result of the hypnotic sessions. Second, notes of government interviews of Dr. Gordon prior to his hypnosis appear to refer specifically to his feeling about being pressured into changing his views about whether T-l was an effect level. Defendants make various arguments in an attempt to persuade us that these references have other meanings, which we need not detail here. We have *705reviewed the testimony as well as the interview notes and are persuaded that the district court’s finding that Dr. Gordon’s memory was not enhanced by the hypnosis was correct (or, at least, not clearly erroneous).
Based on these facts, we conclude that the district court did not abuse its discretion in admitting Dr. Gordon’s testimony. We emphasize that our holding is a narrow one, and rests on the unique facts of this case: a case in which Gordon himself decided to resort to a hypnotist, the hypnotist had no contact with the prosecution and had not been told what the prosecution’s theory was, no additional recollec-
tions came from the hypnosis and Gordon testified to no more than what he had told government agents prior to the hypnosis. Given these and other factors, we believe the dangers of post-hypnotic testimony were sufficiently remote that Dr. Gordon’s testimony was properly admitted.
VII.
For the foregoing reasons,19 the judgments of conviction are Affirmed.