A district court jury found the defendants Coffey and Coward guilty of both illegally conspiring to distribute and distributing Schedule II controlled substances.1 An extensive 1978 investigation of illegal Schedule II drug sales in the Landis, North Carolina area had implicated both pharmacists. Agents of the North Carolina Board of Pharmacy testified that the defendants routinely filled prescriptions ostensibly written by a local physician, Gerald C. Shingleton. The doctor, indicted as a co-conspirator but not tried due to his absence from the state, supplied signed blank prescription pads to the defendants. The patient's name and medication were routinely added after a telephone order from Shingleton's receptionists. Those same receptionists supplied prescription refills in the doctor's absence for a $3.00 fee, gave verbal approval for the pharmacists to make refills, and frequently referred patients to the defendants' stores.

Defendant Coffey not only collected $3.00 prescription fees for Shingleton for called-in refills but also actively promoted such refill procedures, permitting the receptionists to complete the blank forms in his presence. Defendant Coward authorized refills by telephone and referred an undercover agent to another area physician who would readily prescribe large quantities of Schedule II drugs without questions. After Dr. Shingleton left town in August, 1978, Coward paid a receptionist to keep his office open and to send any refills to Coward.

The grand jury indictment charged a conspiracy between Shingleton, Coffey, and Coward. In addition to the single three-party conspiracy count, the indictment charged twenty-five substantive violations of the Controlled Substances Act (21 U.S.C. § 841(a)(1)) by Dr. Shingleton, five violations by Coffey, and seventeen violations by Coward.

The evidence at trial showed unrelated conspiracies between Coffey and Shingleton, and between Coward and Shingleton. The government produced abundant evidence from which a jury could infer the separate conspiracies. At the end of the government's case, however, the district court found that no evidence supported the conspiracy count charging Coward, Coffey, and Shingleton with a single conspiracy. The defendants contend this shows a fatal variance between the indictment and proof at trial. We agree and reverse the lower court's judgment both as to the conspiracy and the drug distribution convictions.

Throughout the trial the district court carefully cautioned the jury as to which items of evidence related to each defendant. The trial, nonetheless, was often confusing for both the attorneys and the jury. The two defendants had similar names. Many witnesses presented details of 51 separate conspiratorial overt acts. Detailed testimony revealed undercover agents' separate contacts with each individual defendant and with Dr. Shingleton, the common perpetrator, with each of them but not jointly with both.

The most significant incident illustrating the confusion occurred during the government's closing argument to the jury. The Assistant United States Attorney commented on an undercover agent's conversation with Coward. Coward had identified another physician who would readily write prescriptions in quantity for Schedule II controlled substances. The government attorney mistakenly named Coffey as the person making that statement to the undercover agent. The attorney later apologized and explained his mistake to the jury. The jury, however, after a day-and-a-half of deliberation, expressed their confusion. They requested the court to inform them as to which one of the defendants had taken part in the incriminating conversation. Coward understandably objected to any clarification while Coffey urged the court to provide the information. The jury's request was refused.

The Supreme Court drew the parameter of error for such variances in conspiracy cases in Berger v. United States, 295 U.S. 78, 55 S. Ct. 629, 79 L. Ed. 1314 (1935), and Kotteakos v. United States, 328 U.S. 750, 66 S. Ct. 1239, 90 L. Ed. 1557 (1946). The Berger opinion held that a variance between the indictment and proof of multiple conspiracies rather than a single conspiracy would not result in reversible error unless it resulted in prejudice to the defendant. The Court said:

It is settled by the great weight of authority that, although an indictment charges a conspiracy involving several persons and the proof establishes the conspiracy against some of them only, the variance is not material.

. . . a variance such as that here dealt with . . . was not prejudicial and hence not fatal.

Id. at 83-84, 55 S.Ct. at 631.

In Kotteakos the Supreme Court held that the allegation of a single conspiracy, but proof of separate conspiracies under the circumstance of that case could, without more, demonstrate prejudice. The defendants in Kotteakos were numerous and the conspiracies complex. This Court in United States v. Goss, 329 F.2d 180 (4th Cir. 1946), in reaching a similar result, stressed the large number of co-conspirator defendants. In the instant case, there were only two defendants and the issues were not as complex as either Kotteakos or Goss. The trial court, moreover, correctly instructed the jury. It is not necessary, however, to determine if the complexity of the trial approached the dimensions of that in the Kotteakos or Goss trial. The jury's deliberations and their failure to understand the connection of incriminating evidence to a particular defendant demonstrated their confusion. This exhibited actual prejudice judged by the Berger criterion.

The trial court was eminently fair in the conduct of the trial and did everything possible to assure that the jury only considered the appropriate evidence against each defendant. This proved to be an impossible chore, however, due both to the subtleties of the evidence and to its emanation from common sources. It is highly probable that the jury's confusion also affected its verdict on the substantive drug distribution charges. As stressed in Goss, the misjoinder of the defendants accentuated the prejudicial effect of the variance. Although the defendant did not move for a Rule 14 severance at the end of the government's case, it is now clear that the defendants should have been tried separately.

The judgment, both as to the conspiracy and drug distribution convictions, is therefore reversed and remanded and, upon remand, the defendants should be granted separate trials.

REVERSED AND REMANDED.