Hunter and Gambrill were convicted of rape, robbery, assault with a dangerous weapon and unauthorized use of a motor vehicle. We reverse the judgments because of impermissible suggestion in the identification of the subjects and order a new tral.

At about 10:30 P.M. on October 1, 1968, a twenty-two year old female nursing student (hereinafter referred to as Mary) and her male companion (hereinafter referred to as John) drove in John's Chevrolet to a well-lighted picnic area in Rock Creek Park off of Beach Drive just north of the Military Road overpass in the District of Columbia. Around 12:30 A.M., when all other cars had left the area, two young men (subsequently identified as appellants Hunter and Gambrill), both wearing handkerchief-type masks covering their lower face below their eyes,1 approached the car, one on each side, and with the taller man threatening John and Mary with a small calibre pistol (.32 or .38),2 forced them out of the car. The assailants then forced their two victims to walk to a dimly lighted sandy beach area several feet below the level of the picnic area and some twenty-five feet away. There they tied John's hands, made him lie on the ground and took his watch and wallet containing money, credit cards and his driver's license. The shorter of the two men (subsequently identified as Gambrill) then sat on John while the taller man (subsequently identified as Hunter) unsuccessfully attempted to rape Mary. The two men then traded places and the shorter one (Gambrill) raped her. Thereafter, the two men tied up Mary, took her wrist watch and gagged her, apparently with the handkerchief the shorter of the two had been using as a mask. The assailants then left in John's car.

Mary and John freed themselves and Mary went to a hospital while John reported the crimes to the police, went with them to the scene of the crime and gave the police the following description of the attackers:

The taller individual [was] about six feet tall and approximately 180 pounds, Negro, medium complexion. * * * wearing a mask that covered a portion of his face below his eyes and covered down in a V-type as if a hanky or something had been folded over and attempted to tie around his face [sic]. He was wearing a light-colored jacket and dark-colored trousers.

The second individual [was] shorter than the first individual; being about five-ten in height, between around 150 pounds, medium build, medium complexion. He was a Negro, also wearing a light-colored jacket and dark trousers. He had a similar type mask on and also covered his face to similar extent of the first individual [sic].

The next morning Mary gave the following description to the police:

[O]ne seemed to be taller and larger and he had short hair, and one was smaller and he seemed to be thinner and had more hair; had bushier hair.

The next night, about twenty-four hours after the attack, Hunter and Gambrill were arrested in a Plymouth exhibiting stolen license plates. Hunter was driving the car and Gambrill was in the front passenger's seat. As the police flagged the car to the curb, they observed Gambrill lean forward as if to place something on the floor. When the car was searched following the arrest, the license plates from the stolen Plymouth were found under the front seat where Gambrill had been riding and three of John's credit cards and his driver's license were also found under the front seat floor mat on the passenger side. John's Chevrolet was found the next day less than a block from where Hunter lived and a fingerprint identified as Hunter's was found on the left front vent window.

On October 3, 1968, Hunter and Gambrill were charged with the instant offenses.

On October 8, 1968, the police staged a seven-man lineup which included Hunter and Gambrill. It was viewed separately by John and Mary and both testified they were told prior to viewing the lineup that two persons arrested as suspects in the crimes would be in the lineup.3 However, John was unable to identify any person in the lineup as a person he had seen commit the crimes.4 Mary picked out Hunter and a person other than Gambrill as the two men who most "reminded" her of her assailants.

Hunter and Gambrill were each indicted for rape, robbery, assault with a dangerous weapon and unauthorized use of a vehicle and were convicted on all counts. These appeals followed.

* Appellants' principal contention on this appeal is that they were denied due process of law when the trial court permitted Mary to identify both of them in court. A central point in this contention arises from the fact that at a conference held on January 23, 1969 (six months before trial) in the office of the Assistant United States Attorney, an unidentified police officer showed Mary two colored pictures of Hunter and Gambrill which had been taken on the night of their arrest. One picture showed side views, from the top of their hips up, of both Hunter and Gambrill, and the other showed similar length front views of the two men. From the testimony, it appears that, at the conference in question, Mary and an officer of the Metropolitan Police became engaged in a discussion concerning the relative heights of her assailants. The officer stated that the two were of the same height while Mary believed that one was taller than the other. To settle the controversy, the officer pulled from his pocket the two pictures in question and showed them to Mary while the Assistant United States Attorney was otherwise engaged. After both Mary and the officer looked at the pictures, they agreed that the two men were of different heights. Thus Mary's recollection was proved correct.

Appellants did not discover that Mary had been shown these two pictures until she was being cross-examined at trial and after a pretrial suppression hearing had been held. When the information concerning the pictures came to light, a second suppression hearing was held out of the presence of the jury. The court decided that there was an independent source for Mary's identification and thus refused to grant the mistrial requested by appellants on the ground that Mary had made an in-court identification of both Hunter and Gambrill.

In considering whether this was a proper decision with respect to the in-court identification, we do not question the testimony that the United States Attorney was not aware of, and played no part in, the showing of the pictures to Mary. We are concerned only with whether the surrounding facts support a conclusion that Mary should have been permitted to make an identification of appellants in court on June 19, 1969. To resolve that question we look to the totality of the surrounding circumstances5 and follow the principles and reasoning of Simmons v. United States, 390 U.S. 377, 88 S. Ct. 967, 19 L. Ed. 2d 1247 (1968).

While Simmons involved the showing of photographs prior to arrest and indictment, and here the photographs were shown after appellants had been charged and indicted, the following remarks and reasoning of the Court on the use and possible misuse of photographs in identification are fully applicable to the facts here:

[a] witness may have obtained only a brief glimpse of a criminal, or may have seen him under poor conditions. Even if the police follow the most correct photographic identification procedures and show him the pictures of a number of individuals without indicating whom they suspect, there is some danger that the witness may make an incorrect identification. This danger will be increased if the police display to the witness only the picture of a single individual who generally resembles the person he saw. * * * The chance of misidentification is also heightened if the police indicate to the witness that they have other evidence that one of the persons pictured committed the crime. Regardless of how the initial misidentification comes about, the witness is thereafter apt to retain in his memory the image of the photograph rather than of the person actually seen. * * *6

After pointing out the merits of photographic identification and some of the pitfalls the Court referred to United States v. Wade, 388 U.S. 218, 87 S. Ct. 1926, 18 L. Ed. 2d 1149 (1967) and Gilbert v. California, 388 U.S. 263, 87 S. Ct. 1951, 18 L. Ed. 2d 1178 (1967) as departing from the prior rule that the manner of extra-judicial identification affects only the weight of such evidence and not its admissibility. Simmons then prescribed a general standard for appellate courts to follow with respect to the use of photographs in connection with identification of suspects:

[W]e hold that each case must be considered on its own facts, and that convictions based on eyewitness identification at trial following a pretrial identification by photograph will be set aside on that ground only if the photographic identification procedure was so impermissibly suggestive as to give rise to a very substantial likelihood of irreparable misidentification. This standard accords with our resolution of a similar issue in Stovall v. Denno, 388 U.S. 293, 301-302, 87 S. Ct. 1967, 1972-1973, [18 L. Ed. 2d 1199,] and with decisions of other courts on the question of identification by photograph.

Simmons v. United States, 390 U.S. at 384, 88 S.Ct. at 971. (Emphasis added). It is particularly difficult to determine which circumstances involve a "very substantial likelihood of irreparable misidentification," for it is apparent that not every in-court identification which follows an impermissibly suggestive photographic identification will be excluded. A combined reading of several cases, however, leads us to the conclusion that an impermissibly suggestive photographic display should be considered to give rise to a very substantial likelihood of irreparable misidentification if the Government is unable to show, by clear and convincing evidence, that a subsequent in-court identification is based on a source independent of the photographic display.7

Thus interpreted, two inquiries are relevant in determining whether, under Simmons, Mary should have been permitted to make an in-court identification of appellants: (1) whether the photographic display itself was impermissibly suggestive and (2) if so, whether there is nevertheless an independent source for the in-court identification. The object is to obtain a reliable identification.

The answer to the first question must be in the affirmative, for there could hardly be a more suggestive display than the one here used. Mary was shown two color pictures of the two appellants standing side-by-side. In one, both were facing the camera and in the other side views of both men were shown. At the time she viewed the photographs, Mary knew that the persons there depicted had been arrested, indicted and were awaiting trial for the crimes which occurred involving her on October 2, 1968.8 Indeed the very purpose for which she was present in the Assistant U. S. Attorney's office was to prepare for that trial. Moreover, her prior description of the assailants was none too detailed and she could only be said to have identified one of the men before in person. In that identification she was "not sure." Under such circumstances, the photographs were suggestive to an impermissible degree.9

Our conclusion in this respect is not altered by Mary's testimony that she did not "identify" the appellants when she was shown their pictures on this occasion. In this first step on the two part inquiry discussed above, the focus is on the photographic display itself and not on the reaction of the witness to it.

We next turn to the second part of the inquiry, i. e., did the Government show, by clear and convincing evidence, that Mary's in-court identification had a source independent of the suggestive photographs. Clearly this second part of the inquiry is dictated both by the language of Simmons and by common sense, for as the Court of Appeals for the Fifth Circuit has explained

if a teller in a bank is held up by a person he knows well, a picture spread, regardless of its suggestiveness, is unlikely to affect the teller's identification. But if the witness caught only a fleeting glimpse of an unknown fleeing felon, the likelihood of misidentification is substantially increased by a suggestive picture spread. Between these two extremes, the determination must be made upon the facts in the particular case.10

The trial court found Mary's statements that she was basing her in-court identification on her observations on the night of the crime to be very convincing. Her statements that indicated she possessed an independent recollection were very positive and her candor was impressive.11 However, statements by witnesses to the effect that they do have an independent source for their identification cannot by themselves be taken conclusively to supply the "clear and convincing" evidence which the Government must produce to meet its burden.12 What is required is that the court consider and evaluate such testimony with all the other circumstances of the case. Many of the other factors which must be considered were outlined in United States v. Wade, 388 U.S. 218, 241, 87 S. Ct. 1926, 1940, 18 L. Ed. 2d 1149 (1967), where the Court said that among the relevant factors were

the prior opportunity to observe the alleged criminal act, the existence of any discrepancy between any pre-lineup description and the defendant's actual description, any identification prior to lineup of another person, the identification by picture of the defendant prior to the lineup, failure to identify the defendant on a prior occasion, and the lapse of time between the alleged act and the lineup identification.13

Testimony concerning the first of these factors, Mary's prior opportunity to observe, while indicating that the lighting conditions were not the best, did indicate that they were adequate to enable an observer to distinguish faces and other physical features.14 The assailants spent a total of about twenty minutes in the company of John and Mary and each was in very close proximity to Mary for between three and ten minutes. It is not clear, however, whether either John or Mary ever saw the attackers without their masks, or, if they did, for how long and at what proximity.

It is also clear that Mary did have an ample opportunity to observe some features of her assailants but the faces of both men were obscured during a substantial portion of the encounter by handkerchief-type masks, and when Gambrill allegedly removed his mask so he could use it to gag Mary, she was tied up and they were both in the more dimly lighted area by the beach. The description that Mary gave of her assailants was limited to "taller and larger and he had short hair" and the other "one was smaller and he seemed to be thinner and had more hair; had bushier hair."15 To the extent that her observations were contemporaneously expressed they focused on their hair and general physical makeup. Her recollection of other distinctive features of her assailants was limited. She could not recall what the men were wearing on the night of the crime,16 did not remember whether Hunter had a mask on17 and was unable to recall whether or not he had a mustache.18 The testimony was unclear as to whether either assailant was wearing his mask during the period of about ten minutes while each tried to rape her and had his face about five inches from hers.19 She did indicate that her in-court identification of Gambrill was based on her recollection of his hair and eyes alone,20 but did not indicate with any precision what was distinctive about those features other than her description of his hair as "more hair, had bushier hair." Understandably, she was, in her own words, "terrified" and "hysterical" on the night of the crime.21

Of course she did pick out Hunter from the seven-man lineup she viewed on October 8, 1968, five days after the crime; but at the lineup she failed to identify Gambrill and instead picked a third person. As to Hunter and the third person she testified at trial:

A I told them that I didn't * * feel sure of any of the ones, but of all the people there, these two reminded me, most, of the men.

Q Did you use any particular words to describe this?

A No.

Q And what was the statement you made to the police?

A I told them I wasn't sure of my identification of the men but I thought that those two were most like the men. (Emphasis added.)

I went to a lineup and were told that the suspects were there but I could not identify them.22

She had been "scared"23 and "terrified" at the lineup because of certain events that were occurring at that time and had even been afraid to go up as close as the officers thought she should.24 However, three months later in the U. S. Attorney's Office she stated she was still unable to identify Hunter and Gambrill from the two colored pictures taken the day after the crime which she was then shown. She studied the pictures for about two minutes25 and, in reply to the query of the police officer who showed them to her, stated that she "wasn't sure * * * I said I couldn't be sure" from those pictures that the two men were her attackers.26 Actually the pictures have some dark shadows particularly around the face and head.

Several additional factors merit consideration. In the first place, John, who had given a detailed description to the police27 and who had looked at the attackers at the time of the crime with the intention of remembering any distinguishing characteristics they might have had,28 was never able to identify Hunter and Gambrill as the assailants. There are thus no other eyewitness identifications by which the reliability of Mary's identification can be gauged. Secondly, between the time that she identified the appellants at the pretrial suppression hearing held at the start of the trial and the time that she identified them at trial, Mary was shown the photographs in question by the Assistant U. S. Attorney29 and also saw a photograph of a 10-man lineup in which both Hunter and Gambrill had appeared just prior to their appearance in the 7-man lineup which she viewed.30 The positive character of her in-court identification must therefore be viewed as the possible result of the influence of improper suggestion in light of the fact that just before she made it, she had seen several pictures of appellants.

A final point to be considered is Mary's candor on the witness stand. Though she stated repeatedly that she was basing her in-court identification on her observations on the night of the crime, she did indicate that appellants looked the same in the questioned pictures as they did in the courtroom,31 that she "might [have been] looking" to identify two men sitting at the table with defense counsel as her attackers32 and that she felt that seeing the two pictures and being told that these were the two suspects "probably" had some element of suggestion.33

After considering the above testimony, and hearing arguments of counsel, the trial court concluded that the Government had established, by clear and convincing evidence, that Mary's proposed in-court identification of appellants rested on an independent source. The conclusions of the trial court in this respect are entitled to a great deal of respect, for it is the trial court alone which has the opportunity to assess the credibility of witnesses and therefore the inherent reliability of testimony.34 Moreover, absent a constitutional issue, the weight to be given to the testimony of witnesses is entirely for the trier of fact to determine35 and here appellants were additionally convicted by a jury which heard the circumstances surrounding Mary's identification presented with skill, determination and vigor by able counsel. The viability of both these principles is confirmed by the great infrequency with which determinations that an independent source exists are reversed on appeal.36

However well established these principles may be, they cannot derogate from the equally well-established principle that, when a constitutional issue is present, our own view of the entire record may impel us to conclude that the evidence concerning the existence of an independent source is insufficient to satisfy the burden which rests upon the Government.37 This is one such case, for we are unable to conclude in the circumstances here present that the identification procedures which were used with respect to both accused were not so impermissibly suggestive as to give rise to a very substantial likelihood of irreparable misidentification.

To summarize the facts surrounding the identification: Following the offenses both victims viewed both Hunter and Gambrill at the regular 7-man police lineup on October 8, 1968. The victims knew that two men suspected by the police were in the lineup but only Mary was able to pick out anyone and she only selected Hunter as one who was "most like," but she "wasn't sure." Mary also picked out a third person but neither victim selected Gambrill. Then, on January 28, 1969, Mary was shown the two colored pictures of both suspects by the unidentified police officer. She observed the two colored pictures for less than two minutes but "couldn't be sure these were the men." Next, on June 17, 1969, during the pretrial suppression hearing Mary observed appellants in court where they were seated at counsel table and were apparently the only two Negroes there.38 Mary testified that at that time she recognized Hunter as the one she had identified at the lineup and also stated "I recognize both of them now."39 In doing so, she also admitted she did not pick out Gambrill from the lineup and stated she did not remember seeing him at that time.40 Then, before Mary testified at the actual trial on June 19, 1969, the U. S. Attorney also showed her the two colored pictures again and she saw a photograph of a lineup which included both suspects.

It must be recognized that any exposure of an eyewitness to the suspects or their pictures is suggestive. Even the lineup carried its own suggestivity by the mere presence of the seven individuals on the stage and being told that two arrested suspects are in the lineup carries another form of suggestiveness. A witness to such lineup might subsequently remember any face in a lineup even though he was unable to identify it in the first instance. When all this is coupled with the fact that two men in the array meet the physical description and the victim is subsequently shown two pictures of those suspects and told that they are the men charged with the crime, suggestiveness is compounded. To this is also added that suggestiveness which inheres in the eyewitness observing appellants as the only Negroes seated at counsel table at the suppression hearing when she is asked to identify them. Finally the victim positively identified both suspects at the trial. So a witness who at the lineup held immediately after the crime did not "feel sure of any of the ones" (she was not absolutely sure), and who selected one person other than Gambrill, and who testified before the grand jury that she "could not identify"41 the suspects at the lineup, finally testifies at the trial eight and one half months later that she did "recognize both of them now." Thus the final result is that she positively identified Gambrill whom she had been unable to identify previously and her identification of Hunter changed from "not sure" to positive. Something happened between the lineup and the trial with respect to her ability to identify both men.

As with many witnesses, her desire to be absolutely honest may have caused her to be overcautious and inexact in articulating the degree of certainty that she possessed as to her first identification of Hunter, but her complete inability to select Gambrill at the first lineup held within six days of the crime is a circumstance that compels us, in view of the intervening exhibition of the two single photographs, and the 10-man lineup photograph and the strong suggestiveness that comes from seeing two Negroes seated at counsel table, both of whom she had viewed at the lineup, to conclude that her identification of appellants cannot stand.42

In addition, we are unable to accept the Government's argument that Mary's identification was harmless error. While there was other evidence implicating appellants in the crime,43 we cannot say that, beyond a reasonable doubt, Mary's positive identification of appellants as her attackers did not contribute to the verdict of the jury.44

This result is to be regretted, particularly because of the strong case against Hunter and the adequate case against Gambrill but we have no recourse under the controlling principles laid down in Simmons when constitutional error of this magnitude exists and can be said to be a factor which contributed substantially to the guilty verdict. Certainly the positive identification by the victim of both appellants meets this standard. Even though the showing of the two photographs to Mary was what might be termed an accidental act by the police officer, and one for which we assess no blame against him or the U. S. Attorney, we cannot conclude after a detailed reading and study of the entire transcript that such showing, together with other suggestive acts which occurred, did not affect the reliability of the victim's identification of appellants. Something happened in the eight months' interval between the first lineup and the trial that changed the ability of the victim to make a visual identification of both her attackers with greater positiveness than she did within a few days of the crime and we do not find in the record that clear and convincing evidence that is required before we can conclude that the change was not caused by the numerous suggestive factors which were proved to exist, the most suggestive of which was the display of the two photographs on January 28th.

The case is accordingly reversed and remanded for a new trial which must proceed without an in-court identification by Mary of Hunter and Gambrill: she may however testify that she did select Hunter, and failed to select Gambrill, at the lineup on October 8th, and as to all circumstances relevant to her identification.

Though what we have said thus far disposes of the case, at least two other issues might again arise on a subsequent trial and should therefore be treated at this time.45 Before and during trial, both Hunter and Gambrill asked for a severance. Gambrill asked for a severance before trial because he contended that the evidence against Hunter was far greater than the evidence against himself and that he would be prejudiced in a joint trial by a "spillover" of evidence. During the trial, he asked for a severance after Hunter's alibi witnesses began to testify. Characterizing their testimony as "inherent[ly] incredible," he argued through counsel that, since the alibi involved both Hunter and himself, he was being prejudiced by its use. The court declined to sever the defendants but did instruct the jury, at the request of counsel for Gambrill, that Hunter's witnesses were offered on behalf of Hunter alone. Shortly thereafter, Hunter requested a severance when the trial court refused to allow one of his alibi witnesses to testify after objection to such testimony was made by counsel for Gambrill.

The decision to grant a severance of defendants properly joined for trial is one over which the trial court possesses great discretion and exercise of that discretion will be reversed on appeal only when it is shown to have been clearly abused.46 The general rule is that defendants charged with jointly committing a criminal offense are to be jointly tried.47 While a disparate quantum of evidence against each of two codefendants may conceivably require a severance under some circumstances, the evidence against Hunter was not so "far more damaging" than the evidence against Gambrill that a severance was required.48 We thus conclude that the failure of the trial court to grant either of Gambrill's motions for severance was not an abuse of its discretion.

Hunter's motion presents somewhat more complex problems. He called two alibi witnesses, his brother and his brother's female friend. The brother testified first, and his testimony placed Hunter and Gambrill together in Hyattsville, Maryland, on the night of the crime. After the U. S. Attorney had finished his cross-examination, counsel for Gambrill, at a bench conference, moved for a severance, claiming that Gambrill was not with Hunter in Hyattsville on the night of the crime, that the alibi presented by the brother was "inherently incredible" and that Gambrill was being prejudiced by its use. The trial court refused to sever the two, but, as stated above, did instruct the jury that Hunter's witnesses were offered on behalf of Hunter alone. Counsel for Hunter voiced no objection to the instruction at the time it was given or at any subsequent time during the trial.

Following the testimony of Hunter's second alibi witness, his counsel requested that he be allowed to call Hunter's father as a witness for the defense, stating that the father had just indicated to him that he could corroborate some of the testimony given by Hunter's two witnesses. Counsel for Gambrill objected on the grounds that the father had been present in the courtroom during the entire trial, and the court sustained his objection. Specifically because of this ruling, counsel for Hunter moved for a mistrial and for severance. Both motions were denied.

On appeal, Hunter contends that he should have been granted the requested severance (1) because the instruction given by the court at Gambrill's request amounted to a repudiation of the alibi presented by Hunter's witnesses which severely prejudiced Hunter's defense and (2) because Gambrill's objection precluded Hunter's use of an allegedly valuable witness.

We need not decide whether Hunter was entitled to a severance on ground (2).49 His counsel now knows that the father has information useful to the defense. Consequently, the father can and should be excluded from the courtroom during retrial until it is time for him to testify.

Ground (1) is not so easily handled. The jury could certainly have construed the instruction given by the trial court as a statement by Gambrill that he was not at Hunter's house at the times mentioned by the witnesses or, more generally, that he knew the alibi presented by the witnesses was untrue in other particulars.50 The jury could then have taken the next step and, choosing to believe Gambrill, concluded that Hunter's witnesses were lying and used this conclusion as added evidence of his guilt.

The risk that one may be caught in a lie and that the lie may be used by the jury as evidence of guilt is a risk taken by all those who believe that mendacity is a preferable alternative to jail.51 If Gambrill had taken the stand, and had testified that he was not present during the times described by Hunter's witnesses, we might have serious doubts whether the trial court's denial of a subsequent motion to sever made by Hunter would have been reversible error.52 That being the case, one may well question whether a more equivocal repudiation, such as the one here embodied in the instruction, should be grounds for a severance.

To understand the crux of the problem before us, however, it is necessary to shift the focus somewhat. The central problem is that Gambrill did not take the stand and was not subject to cross-examination concerning either his whereabouts on the night of the crime or his reasons for refusing to subscribe to the alibi presented by Hunter's witnesses. In substance, then, the claim here made by Hunter is that the instruction given by the court amounted to an out-of-court statement by Gambrill which Hunter had no chance to explore but which disputed evidence offered on his behalf. Use of such a statement, in his view, cannot be made in a joint trial.

In evaluating this claim, we note first of all that counsel for Hunter did not object to the instruction either when it was given following the testimony of the first alibi witness or when it was given in the early part of the court's charge to the jury. Nor did he ever request a severance on the ground that the instruction was prejudicial to the interests of his client.53 In addition, he did not object when informed that the U. S. Attorney intended to ask the second alibi witness whether she knew that Gambrill had said that he was not present at the times and places she had testified he was.54 Moreover, neither counsel for Hunter or Gambrill nor the U. S. Attorney referred to Gambrill's repudiation in closing argument.

In this posture, it is a very close question whether it was plain error for the trial court not to sever the defendants. Moreover, to hold that it was reversible error under these circumstances for the court not to have granted a severance sua sponte would place in the hands of joint defendants the power to obtain severance through fabrication. To so hold would also require us to find prejudice in the instruction where the very competent and able counsel trying the case apparently found none at the time.55

We need not decide whether the failure to sever was reversible error. The case must be retried because of the erroneous admission of Mary's in-court identifications. However, despite the foregoing difficulties, we conclude that, under the circumstances of this case, if it appears to the court that the same situation will reoccur, separate trials should be afforded appellants on remand. There is no other way to insure that the jury can consider the case against Hunter uninfluenced by the untested suggestion that his witnesses were lying when they described his whereabouts on the night of the crime. It may well be that they were lying and it is certain that either they were lying or Gambrill was. Nevertheless, we cannot countenance a trial in which the jury is given the opportunity to make such a crucial determination on the basis of "testimony" which is insulated from the revealing probes of cross-examination.56 In our view, to allow such a trial would be to allow a serious erosion of a defendant's right57 to present a defense to the charges of which he stands accused.

Two further considerations strengthen our conclusion that separate trials should be afforded on remand if, prior to trial, an adequate showing is made that the same situation will reoccur. First of all, with the excision of Mary's in-court identification, we believe that the evidence against Hunter tends to become "far more damaging" than the evidence against Gambrill.58 In essence, the evidence against Hunter now is that he matches the description given by John on the night of the crime, that he was picked from a lineup by Mary shortly after the crime occurred, that John's car was found several blocks from his house several days after the crime, that his fingerprint was found on one of the car's windows and that he was arrested in company with Gambrill and found to be in close proximity to, if not in possession of, John's credit cards twenty-four hours after the crime took place. The evidence against Gambrill is that he matches John's description and was arrested in company with Hunter and found to be in close proximity to, if not in possession of John's credit cards twenty-four hours after the offense.

We do not intimate that the evidence against Gambrill is insufficient to support a conviction should the jury return one, nor are we of the view that the disparate quanta of evidence now present in the case would necessarily justify a severance in all cases. We do conclude, however, that the difference in the amount of evidence against each defendant, when considered in light of the other elements present in the case, is a factor which supports our conclusion that, because of that fact and others, severance should be granted on remand.

Secondly, the principal reason for the rule permitting the joinder of offenses is economy of time and efficiency in the judicial process.59 Relevant to our decision, then, is the consideration that if the demonstrated antagonism continues between these two defendants insofar as their alibi is concerned, it is quite likely to contribute to difficulties at trial and on appeal that would be counter productive to the normal reasons for trying the cases jointly. Once more, this is not a factor which, by itself, would necessarily justify a severance, but it, too, is a factor which, in the present context, supports the conclusion we here reach.60

Finally, Gambrill argues that discovery of the credit cards in the car when he was arrested was insufficient corroboration of his participation in the rape to permit the case against him to go to the jury. We recently discussed at length the standards governing corroboration in "sex offenses" and see no necessity for another extended discussion at this time. Suffice it to say that the necessity for corroboration of the identity of the offender is less pressing than is the need that there be corroboration of the fact that an offense occurred, and in both cases the evidence which will supply the requisite corroboration may be far less than would, by itself, sustain a conviction.61 With these principles in mind, we conclude that the trial court was correct when it determined that the credit cards provided enough corroboration to allow the jury to consider the case against Gambrill.

We have carefully considered the other points raised by appellants but, convinced that the circumstances which gave rise to them are unlikely to occur again, decline further comment. The case is therefore remanded to the District Court for a new trial at which Mary will not be permitted to identify appellants.

So ordered.