This case is before us for the second time. The facts which are pertinent to this appeal remain the same:

"During a period spanning more than five years the Pittsburgh area was plagued by a series of bank robberies all thought to have been committed by the same `Commuter Bandit.' The exploits of this unknown robber and his ability to avoid detection aroused considerable public attention. The arrest of a suspect, William Zeiler, on June 23, 1967 was attended by massive publicity. For several days newspapers and television stations prominently displayed pictures of the arrest, comments by the chief FBI agent in Pittsburgh, and photographs of Zeiler in comparison with earlier composite sketches of the `Commuter Bandit.'"

United States v. Zeiler, 427 F.2d 1305, 1306 (3d Cir. 1970). Although defendant was arrested for only one robbery, the FBI began showing his photograph to the victims of the fifteen or more other robberies thought to have been committed by the "Commuter Bandit." The resulting identifications ultimately led to defendant's being charged in two indictments with having perpetrated a total of eleven bank robberies between 1963 and 1967. In 1968, he was tried separately on each indictment and was convicted of having committed three of those robberies.

On appeal, this Court held that, once an accused is taken into custody, he is entitled under the Sixth Amendment to have counsel present whenever law enforcement authorities confront witnesses with a series of photographs for identification. We reversed all three convictions and remanded the cases for new trials on the ground that it was constitutional error for the Government to have elicited testimony from its witnesses at both trials that they had previously identified Zeiler from photographs exhibited in the absence of counsel. We then proceeded to consider whether the witnesses' courtroom identifications of Zeiler would be admissible on remand. With respect to the one robbery conviction resulting from the second of Zeiler's two trials (involving Criminal Action No. 67-187), we examined the photographs which had been shown to the witnesses, found them unduly suggestive, and held that, under the circumstances, they had so influenced the witnesses as to render such witnesses incompetent to make an in-court identification of Zeiler at his retrial. Since the record of the first trial, which had involved Criminal No. 67-186 and resulted in two convictions, did not indicate what photographs were shown to the identifying witnesses in the absence of Zeiler's counsel, we remanded for a pretrial hearing in order to afford the Government an opportunity to prove "by clear and convincing evidence" that its witnesses were not so influenced by the prior photographic confrontations as to be incompetent to make in-court identifications.

On remand, a hearing was held in Criminal No. 67-186, as a result of which the district court ruled on the basis of our previous opinion that all of the Government's witnesses who had previously been shown photographs of Zeiler outside his attorney's presence were incompetent to testify on retrial. The Government now appeals that portion of the district court's order which suppresses the testimony of four eyewitnesses to one of the robberies charged in Criminal No. 67-186, namely, that of the Bloomfield Office of Pittsburgh National Bank (hereinafter referred to as "the Bloomfield robbery"). Pursuant to 18 U.S.C. § 3731, the Government has certified that the appeal is not taken for the purpose of delay and that the suppressed evidence constitutes substantial proof of the charges pending against the defendant.1

As we ruled in our previous opinion in this case, the witnesses' in-court identifications are admissible only if the Government establishes by clear and convincing evidence that they have an origin independent of any photographic identification made outside the presence of counsel. In ordering a hearing on this matter in Criminal No. 67-186, we issued the following guidelines:

"[I]f on remand the conduct of the photographic identifications is found to have been significantly less suggestive in this case than the other, that fact, together with all other circumstances that support or militate against the claim of each witness that his in-court identification has not been influenced by the photographic identification, must be considered by the trial judge in deciding whether the prosecution had carried its heavy burden of proving that the witness was not influenced and thus made incompetent for subsequent in-court identification." 427 F.2d at 1309.

We have carefully examined both the transcript of the suppression hearing and the actual photographs which were shown to the witnesses in question and disagree with the district court's conclusion that there is "no significant difference" between the photographic identification procedure used in the FBI's investigation of the Bloomfield robbery and that which we condemned in our previous opinion. We further believe that the Government established by clear and convincing evidence that the eyewitness identifications did have an independent source.

The following factors, previously mentioned by the Supreme Court in lineup cases, have an equally important bearing in this type of case upon the true basis of each witness' in-court identification: (1) the manner in which the pretrial identification was conducted; (2) the witness' prior opportunity to observe the alleged criminal act; (3) the existence of any discrepancies between the defendant's actual description and any description given by the witness before the photographic identification; (4) any previous identification by the witness of some other person; (5) any previous identification of the defendant himself; (6) failure to identify the defendant on a prior occasion; and (7) the lapse of time between the alleged act and the out-of-court identification. United States v. Wade, 388 U.S. 218, 241, 87 S. Ct. 1926, 18 L. Ed. 2d 1149 (1967).

With respect to the first criterion, none of the specific elements of suggestiveness which we found in the photographic array shown to the witnesses at the second trial are present here. Whereas three of the eight photographs in the previous group were of Zeiler, the defendant is only pictured once in the present group of six photographs. All the photographs in the previous group were police "mug shots" except those of Zeiler, which were ordinary snapshots and might easily have reminded the viewers that the "Commuter Bandit" had only recently been arrested, while the present group includes two other pictures besides that of Zeiler which are not "mug shots." Whereas in the previous group only Zeiler was pictured wearing glasses, as had the actual robber, neither Zeiler nor any of the other individuals in the present group is shown wearing glasses. Furthermore, the identification procedure itself shows no indication of improper suggestion. Each witness was interviewed separately outside the presence of the others; the photographs were arranged in a random order with Zeiler's pictures somewhere in the middle so as not to stand out; and the investigating FBI agent neither suggested that any particular person in the pictures was under suspicion nor commented that the witnesses had picked the "right" man. In contrast with our conclusion concerning the previous group of photographs, therefore, we find nothing to suggest that the attention of each of the witnesses in question was irreparably "focused on Zeiler's picture as identifying the suspect with whom the investigators were concerned." 427 F.2d at 1308.

Furthermore, the Government elicited substantial evidence of the other criteria which tend to show that the in-court identifications at the first trial were of independent origin. It established, for example, that all four witnesses to the Bloomfield robbery had the opportunity to observe the robber unmasked in a well-lighted room with no obstructions to their view for estimated periods ranging from two to seven minutes and at estimated distances ranging from three to six feet. The witnesses gave descriptions of the robber ranging from 5'11" to 6'2" in height and 210 to 225 pounds in weight, while Zeiler described himself as being 6'1" tall and weighing 235 pounds at the time of his arrest. Although the witnesses had been shown other sets of photographs previously, none had identified any other person prior to the photographic display in question, nor had any of them failed to identify the defendant on any other occasion. Whereas we found the witnesses at Zeiler's second trial to be particularly susceptible to suggestion because the robbery they had observed occurred three and a half years before the photographic identifications were made, the Bloomfield robbery was relatively fresh in the minds of the present witnesses since only six weeks had elapsed before they were shown the photographs.

Finally, in attempting to prove that each witness' in-court identification of the defendant had a source independent of the photographic identification, the Government produced evidence of the fifth type listed above which showed that all of the witnesses had positively identified Zeiler on the basis of newspaper and television photographs of his arrest several days before they were questioned by the FBI. The district judge, however, with little or no specific evidence upon which to rely, found as a fact that the FBI had deliberately initiated publicity in the news media — apparently with the intention of influencing later eyewitness identifications — and had thereby created a sufficient degree of suggestiveness to render the witnesses in this case incompetent to identify Zeiler at trial. A hearing on this specific issue had been held at an earlier stage of the case by another district judge, who concluded that the arrest publicity was "not designed by the authorities as a pre-lineup identification technique" and was not even planned or initiated by the Government. United States v. Zeiler, 278 F. Supp. 112, 114 (W.D.Pa.1968). Nevertheless, the court below stated:

"They [the witnesses] had all seen the publicity before they saw the photographs. And the publicity they did see, he was in the custody of the officers and the newspaper accounts said, `Commuter bandit job suspect held.' So the suggestiveness had to be present. * * *"

In our previous opinion, we found it unnecessary to decide "whether publicity alone would require reversal" in the absence of a suggestive photographic identification procedure. 427 F.2d at 1308 n.4. Nor do we decide this issue now. Even assuming that the Government was in fact responsible for the news media's publicity concerning Zeiler's arrest as the "Commuter Bandit,"2 we do not believe that such publicity was "so impermissibly suggestive as to give rise to a very substantial likelihood of irreparable misidentification." Simmons v. United States, 390 U.S. 377, 384, 88 S. Ct. 967, 971, 19 L. Ed. 2d 1247 (1968). There is no evidence that defendant was ever connected in the press with the Bloomfield robbery. Indeed, he was not even indicted for that robbery until more than a month after his initial arrest. Since, as the Government's witnesses testified, they obviously could not know whether they had been robbed by the "Commuter Bandit" or by some other person until they saw Zeiler's picture on television and in the newspapers, we do not find the witnesses' in-court identifications to have been tainted by the publicity concerning the arrest of the "Commuter Bandit" for one completely unrelated bank robbery.

We therefore conclude that the Government has proved by clear and convincing evidence that the four witnesses to the Bloomfield robbery are competent to make courtroom identifications of the defendant upon his being retried in the district court.

The order of the district court suppressing the testimony of Judith Scully, Frank J. Perry, Charles Edward Bruder, and Marian A. Cunnings is reversed, and the cases are remanded for new trials consistent with this opinion.