The Appellant Williamson appeals from a judgment of conviction for violating 18 U.S.C.A. 656 based on a jury verdict of guilty as to all 13 counts charged. A sentence of two years on each count was imposed, four of which are consecutive making an aggregate sentence of eight years. Reversal is sought on two grounds. The first, insufficiency of the evidence was properly preserved by appropriate motions for judgments of acquittal, F.R.Crim.P. 29(a). The second, failure of the Court to charge on accomplice's testimony, is pressed as plain error there having been neither exception nor affirmative request for a proper charge. F.R.Crim.P. 52(b). We overrule the first, but for reasons discussed, we hold as to the second that this is that exceptional case in which an error not preserved must yet be noticed since on this record a fair trial required that the jury be properly instructed on the evaluation and use of accomplice testimony. We therefore reverse.

Williams was the assistant cashier of the Brevard State Bank, an F.D.I.C. insured bank. On the Government's theory and evidence, he was accused of purposefully advancing the Bank's credit to a local automobile dealer, Mike Mackin, on the basis of automobile installment loan contracts known by both of them to be spurious. The indictment, by separate counts as to each such loan contracts charged all of the offenses in the statutory language set forth in 18 U.S.C.A. 656 subjecting to criminal penalties a bank officer or employee who 'embezzles, abstracts, purloins or willfully misapplies any of the moneys, funds or credits' of the bank.1 But after such generality, each count specified that the misuse of bank funds was accomplished by (1) a spurious loan to a specified borrower-purchaser, (2) the issuance of the bank's check to the dealer and (3) a conversion of bank funds by Williamson 'to his own use.'2

It aids an understanding of this case to recognize at the outset that the Government's case was made entirely by the accomplice Mike Mackin. If his testimony is disregarded, there is simply no case. For example, the abundant evidence, if credited, that the particular contracts were not signed by the individuals therein named and hence were spurious ,3 does not implicate Williamson, and certainly not sufficiently to withstand the severe burden of proof beyond a reasonable doubt. What, and all, that implicates Williamson under proper criminal standards, is the testimony of Mackin. This is not the case, therefore, in which the accomplice supplies significant testimony to overcome psychologically weak evidence by a categorical confession of guilt plus the inevitable finger pointed at the accused. All of this makes it important therefore to consider the Government's case inescapably advanced through the story of the accomplice and to weigh it carefully to determine whether the testimony of the accomplice was both so critical and also so suspect that the jury needed the law's traditional caveat as to its evaluation and use.

In so doing, we must keep carefully in mind the procedures followed regularly or frequently by the bank personnel including Williamson. These procedures were normal commercial actions if innocently done but were, of course, significant steps in the consummation of a crime once criminal purpose is assumed or proved. The bank was interested in developing a substantial volume in the financing of installment loans or credits. It had a number of automobile dealer customers whose installment sales paper was financed through the bank. It also apparently had a considerable volume in direct personal loans by the bank to individuals for the purpose of buying appliances, equipment, residential improvements and the like. Williamson was in charge of this department. Included in the automobile dealers was Mike Mackin.

In the Mackin transactions, the bank did not purport to make a loan to the automobile purchaser. Rather, transactions took the form of an assignment by the dealer Mackin of the installment contract to the bank.4 The bank handled his paper principally in one of two ways. Under one procedure, the automobile purchaser would supply credit data5 on a loan application form supplied by and identified with the Brevard Bank. This application form, signed by the prospective buyer, would be brought to the bank. The bank would make a credit check. Sometimes the bank would notify Mackin whereupon he would deliver the original installment sales contract6 signed by the buyer and assigned by Mackin. On other occasions Mackin, in anticipation of ultimate credit approval, would deliver the original signed and assigned sales contract at the same time the application form was submitted. In either event, on the receipt of the assigned sales contract, the loan would be consummated, Mackin would receive the funds by check or credit, and the automobile would be delivered to the purchaser.

A second procedure, frequently followed, had this one substantial difference. Instead of there being a loan application signed by the prospective buyer-borrower, Mackin would simply call the bank by telephone, advise them of the prospective sale and, working from a blank form at his end, give to the clerk or Williamson the credit information called for by the form. Working with an unsigned application, the bank would then make a credit check. If the application was approved, Mackin would be notified and on the receipt by the bank of a properly signed sales contract assigned by Mackin, the bank would advance its credit. Under this procedure, the automobile purchaser would sign only the installment sales contract. More important, the handwriting or handprinting of the loan application would inescapably be that of Williamson or one of his clerical assistants.

It rounds out the picture to state that on final consummation, the bank took the following steps. From the gross amount of the credit to be advanced, it first deducted any insurance or other closing costs. It then deducted a stated percentage amount which was deposited to Mackin's dealer reserve account. This was a cumulative account to protect the bank against losses from defaults on the part of any of the borrower-purchasers. The net balance was then made available to Mackin in one of two ways. The most usual one was for his general account to be credited with the net proceeds thereby subjecting them to his unlimited checking and withdrawals. Occasionally, but apparently very infrequently, the bank would issue its check to Mackin.7

We come then to Mackin's part in this situation and as a witness on the trial. He was a sportscar dealer selling Triumph sportcars. The Brevard Bank had been floor-planning the acquisition of his cars from the manufacturer-distributor. The bank had also been presumably handling much of his sales contract paper. In any event, there were a number of transactions between Mackin and the bank involving the extension of large credits. According to Mackin's version, Williamson became greatly disturbed about a $13,000 overdraft in Mackin's account. Mackin fixed this as April 1961, but it is quite plain that this had to be in 1960 since all of the count transactions precipitated, so he said, by this incident took place prior to January 23, 1961 (see note 3, supra). The overdrafts were, on Mackin's story, represented by specific checks which Williamson had in a little black box. In the course of this discussion, Mackin insisted he had no other course of credit to make good the losses. Mackin testified that at this point Williamson suggested a fantastic scheme of Mackin taking a lonesome boat ride from which he apparently would never return thus setting in train recovery from a large double indemnity accident policy which Mackin would then take out. As Mackin tells it, when this seemed not to be too plausible, the two of them hit upon this scheme of spurious installment sales contracts. Under it, Williamson or Mackin, or both, would select, as needed, the names of automobile purchase prospects whose credit presumably would pass muster at the bank. With the name selected, Mackin would then initiate the routine of loan applications, bank credit inquiry, bank approval, delivery of signed installment sales contracts with Mackin's assignment followed by the bank crediting his account for the net proceeds. Mackin's testimony concerning two other incidents in this whole matter warrants some specific mention. Mackin stated that at one time Williamson called to tell him that Frisbee, executive vice president of the bank, was very upset and for him to come to the bank for a conference which he did. At the bank, he and Williamson conferred with Frisbee. According to Mackin, Frisbee stated that he had information which raised many questions about the November 21, 1960, Skinfill contract, see count 23, note 3, supra. Mackin says that Frisbee finally came out and asked Mackin whether the contract was spurious, and that Mackin admitted that it was. Frisbee asked if there were any others, and, according to Mackin, Williamson then volunteered that there were two or three others which Mackin named. Frisbee then demanded that Mackin pay up the Skinfill contract within 24 hours (which was done). Mackin further testified that despite this ultimatum, Frisbee in effect told him not to worry, that he should be patient and 'just trust your banker.' The other aspect involves the alleged payment of $1,500 and $1,000 in cash to Williamson. Mackin testified that Williamson said that he needed money for personal reasons and demanded that Mackin make funds available. Mackin pinpointed the payoff on two occasions while seated in Williamson's car in front of the bank. Documentary evidence in the form of two checks payable to cash and dated about January 25 and 27 enabled him to fix the time as being within a day or two of such dates.

Of course it is not our function to resolve the credibility issue committed by our system to the jury. And the comments which we make are not to be construed as credibility choices. But with a proper regard to our limited function, we think that a number of circumstances cast such a considerable cloud on the reliability of Mackin's testimony that the jury needed to be specifically cautioned about its use.

There is, first, Mackin's statement as to the incidents which, on his theory, precipitated this whole scheme. He fixed this as April 1961. Obviously this was in error. That error alone is not significant. But the other evidence including that coming from Williamson necessarily fixes it at March or April 1960. Assuming that the jury credited this story, it means that because of the $13,000 overdraft, Williamson and Mackin then and there agreed to set up and operate this scheme of spurious installment sales contracts. Accepting that story, the uncontradicted fact remains that nothing was done until, at the very earliest, August 10, 1960, when the Eleanor Morgan Count 21 transaction was consummated (see note 3, supra).8

Next, Mackin was confused and contradictory about the $13,000 overdraft, the state of his account, what it ought to have been, and what it was. Recognizing, as we must, that the jury did not have to credit Williamson's denial of any statement to Mackin about a $13,000 overdraft, the remainder of the Government's case unmistakenly repudiated Mackin's claim. Frisbee, a Government witness, was positive that there was no such shortage, and further that no such shortage could have been concealed for that length of time. Had there been insufficient-fund checks outstanding, the amount thereof would have been reflected by other appropriate bank entries. He, in effect, corroborated Williamson that the so-called 'shortage' was not in funds to meet outstanding checks. Rather, it was a shortage of funds to meet a draft which would shortly be presented for the payment of a truckload of new cars.

Likewise, there was sharp disagreement between Frisbee, the chief Government witness, and Mackin as to the Skinfill conference in November 1960. Frisbee denied that Mackin or Williamson stated that there were other spurious contracts. According to Frisbee the inquiry was addressed to Mackin alone, and Mackin denied there were other spurious contracts. Frisbee denied also that Williamson volunteered that there were two or three others. In this Frisbee is powerfully corroborated by simple time. Mackin testified that Williamson volunteered that there were several others including one of Grady Henderson. But this conference took place in November 1960 nearly two months prior to the first and only Henderson contract. (See count 3, transaction January 27, 1961, note 3, supra.)

But this is not all. On Mackin's theory, his troubles began in April-May 1960. As pointed out, it was six months thereafter until the first transaction (count 21--Eleanor Morgan) occurred, followed three months later by Skinfill in November (count 23). By this time Frisbee was on to the deal, so much so that he gave Mackin the 24-hour ultimatum which was met. But all of the count transactions took place long thereafter, the first on January 5, 1961, (count 17) and the last on January 27, 1961, (count 3). Moreover, the use of those names suggests a strong probability that they were selected not by Williamson but by Mackin. For example, Henderson (count 3) was Mackin's father-in-law. Yancey (count 5) was a former customer of Mackin, Braum (count 11) had seriously negotiated for a part ownership in Mackin's business, Robinson (count 15) had been Mackin's sales manager for 14 months, King (count 17) had worked for Mackin as parts manager, and Mitchell (count 19) had also negotiated for a 50% Ownership of Mackin's business. Each of these witnesses acknowledged, as Mackin had also, that their relationship with Mackin had afforded him specific detailed data of the kind called for on the loan application blanks. And this was true of the earlier Skinfill transaction (count 23, November 1960) as Skinfill had previously been associated with Mackin at Cape Canaveral.

Mackin's testimony also showed that he did not need an inside man to work this fraud. The first transaction here, Eleanor Morgan (count 21, August 1960), was accomplished by forging her name on the sales contract without her knowledge. Although Mackin was married to Henderson's daughter, he had dates with Eleanor Morgan, a girl friend. He had forged her name to a similar contract without her knowledge at the Cocoa Beach State Bank a month earlier in July 1960.

Finally, there was substantial evidence implicating Mackin in direct efforts to subvert the administration of justice in the trial of this very case. It was uncontradicted that on the eve of trial, and late in the evening of a day on which Williamson's then-counsel had had a conference with Mackin and Mackin's counsel, a telephone call was made by Mackin to the attorney. The attorney took the bold and unusual course of testifying. He related under oath that Mackin called stating in effect that he had some documents he had not shown to the FBI and that he would supply them for $5,000. It was also uncontradicted that on the morning following this conversation, Williamson's counsel notified the United States Attorney of this incident and subsequently was interviewed by FBI agents. Mackin and his counsel refuted this, but the story hardly did credit to either. According to Mackin, he called at his own counsel's suggestion to offer the services of his lawyer for a $5,000 fee. Mackin, on cross examination, had perhaps an understandably indifferent concern that he might have been a tool for unethical solicitation of legal business, but the lawyer could do little more than deny that this was solicitation.

Of course recitation of these circumstances, these actions and the mannerisms of Mackin discloses nothing revolutionary in a criminal trial where out of necessity, enforcement of criminal laws depends frequently on testimony of law breakers and participants in the very crimes under scrutiny. But criminals can, and frequently do, tell a straightforward consistent story which has the ring of truth all the more so because it is a confession of the soul. But law is in step with human nature when it ordinarily recognizes that testimony from this course has to be scrutinized most carefully. That results normally in the giving of a traditional instruction. Undoubtedly had a request been made, the Court would have given such an instruction here. Had the Court declined to do so in the face of a request, we would think it quite certain that, whatever might be the trial court's discretion under circumstances less spectacular,9 it would have been error to decline it here. Phelps v. United States, 5 Cir., 1958, 252 F.2d 49; Freed v. United States, 1920, 49 App.D.C. 392, 266 F. 1012; cf. United States v. Scoblick, 3 Cir., 1955, 225 F.2d 779.

Besides the obvious weaknesses in the testimony of Mackin, an instruction of this kind seems especially needed here. In the first place, the law permits conviction in the federal court on the uncorroborated testimony of an accomplice.10 That means that it can be and frequently is of crucial importance and it most certainly was here. Evidence of that kind--and of that importance--is worthy of special attention by the trial judge as he translates the case into understandable terms for jury resolution. Especially is it so in this case. Williamson could not, indeed did not, deny doing these acts. For example, he freely admitted that the handprinting on the loan applications, identified as his by an FBI expert, was obviously his. This action on his part was quite sufficient to make him an accountable participant in a crime if these actions were done by him in carrying out the scheme. But were these acts taken to effectuate a scheme? Or were they innocent, though perhaps incautious, acts of a loan officer handling papers in the regular frequent routine of the bank?11 The answer to that question turned entirely on belief or nonbelief of Mackin. Similarly, if he received the $1,500 and $1,000 cash payoffs, it is spectacular proof beyond the legal requirements of the crime to show willful participation. But did he get the money? There is, of course, some corroboration in the form of two checks payable to cash. But did the cash go to Williamson? This all depended on accepting Mackin's word. And even there, he was vague as to whether or not the funds given to Williamson really came from these cashed checks.

On the intrinsic facts, credibility made the case a close one. And Judge Hand, stating that 'It is usually desirable to give' the warning as to accomplice testimony, recognized, as we do here, that 'in close cases it may turn the scale * * *.' United States v. Becker, 2 Cir., 62 F.2d 1007, 1009.12

In the final analysis, it is not the parties who determine the charge the judge gives to the jury. The obligation rests squarely on the shoulders of the trial judge. Of course the system of time-tested rules of procedure can rightfully expect competent counsel to request appropriate charges or object to affirmative errors or significant omissions. But there are occasions, and this Court recognizes them year by year, in which the trial court's erroneous action has such immediate and significant consequence that it must be noticed as plain error.13 We think the omission of the charge on accomplice testimony was plain error, and the only way to eradicate it is to grant a new trial.

This result is appropriate for an additional reason since the charge of the court seems to us to have been way too summary in advising the jury as to the nature of the crime and the acts required to be established to return a verdict of guilty. Without intimating that this would measure up to plain error, this critical resolution of the Mackin credibility--and hence the question of guilt or innocence--was placed on the jury with practically no guidance as to the elements of the crime. The charge was a very brief one. At an early point, the Judge read the statute down to the point of the penalty. This, like the indictment, was a statement of all of the offenses (a) through (d), see note 1, supra, not for example, only that of (d) misapplication of funds. None of these terms was ever thereafter defined. Nor was any instruction given concerning what was required to establish any of these elements beyond a reasonable doubt. Indeed, the Judge, confining his instructions entirely to misapplying bank funds,14 treated the case as though the allegations that the defendant did '(a) embezzle, (b) abstract, (c) purloin' (see note 2, supra) did not exist. Did the jury find him guilty for misapplying funds? For embezzlement? For 'abstracting' funds? Or for purloining? It would be inappropriate for us now to indicate what the distinctions are between these crimes. Some acts may constitute each and all of them. Other actions may be one or more, but not all.15 But the jury should not be permitted to speculate on what crime is being submitted to them to resolve or what the contents of each such crime might be.

There were also subsidiary elements such as 'conversion of money to his own use,' and 'intention,' or 'willful misapplication' which called for definition. Again, we need not indicate what the correct standards are, but terms such as 'willful' and 'intention' ordinarily would call for jury instruction.16

Our discussion of the importance of Mackin's testimony makes it plain that there was sufficient evidence to warrant jury submission. It was not error, therefore, to deny Williamson's timely motions for judgment of acquittal. F.R.Crim.P. 29(a). This conclusion assumes, of course, that there was no fatal variance between the proof of the means used to accomplish the crime and that set forth in the indictment. The motions carefully pointed out that since the evidence does not show a (1) 'loan' to the indicated borrower and that there was (2) 'issued said bank's check' but rather, the purchase of third party installment contracts and issuance of a credit to Mackin's account, these elements of the crime were not proved. We are of the view that, whatever might be the legal attributes of these transactions for usury or other purposes,17 the evidence adequately shows a loan. The loan application was on the bank's own identified form as was the installment contract. The bank's processing was more than a credit inquiry in anticipation of a purchase of installment contract under assignment. The bank treated the automobile purchaser as its debtor by establishing a ledger account which identified the purchaser as the 'maker' of the note, etc. And, of course, there was no real difference in operative effect between the bank writing out its check payable to Mackin (or others) and giving him usable credit in his general account.

The Grand Jury through its indictment carefully defined each transaction by the name of the spurious purchaser, the date, and the amount of the credits. Whether it was a 'loan' to the automobile purchaser by the bank or affording the buyer the same effect thereof through the transfer and assignment of the installment sales contract, it was the extension of the bank's credit for that amount on each such occasion. Giving full heed, as we must, to the recently declared decisions of the Supreme Court, Stirone v. United States, 1960, 361 U.S. 212, 80 S. Ct. 270, 4 L. Ed. 2d 252; Russell v. United States, 1962, 369 U.S. 749, 82 S. Ct. 1038, 8 L. Ed. 2d 240,18 the differences are insubstantial, and the variation neither decisive nor prejudicial.

The case is therefore reversed and remanded for a new trial and other and other and further consistent proceedings.

Reversed and remanded.