This case presents an appeal from a district court's denial on summary judgment of a writ of habeas corpus under 28 U.S.C. Sec. 2254. The only grounds alleged in support of the request for the writ are that unnecessarily and unduly suggestive pretrial confrontations tainted the two trial identifications of the petitioner, Joseph Kosik, and made those identifications so unreliable that their admission violated Kosik's Fourteenth Amendment right to procedural due process. For the reasons stated below, we reject that contention and will affirm the district court.

* In determining the facts of this case, we are required by 28 U.S.C. Sec. 2254(d) to presume the accuracy of the findings of the state courts.1 Sumner v. Mata, 449 U.S. 539, 544-47, 101 S. Ct. 764, 767-69, 66 L. Ed. 2d 722 (1981); see also Wainwright v. Goode, 464 U.S. 78, 104 S. Ct. 378, 382, 78 L. Ed. 2d 187 (1983); Burns v. Clusen, 798 F.2d 931, 940-41 (7th Cir.1986); Holleman v. Duckworth, 700 F.2d 391, 395 (7th Cir.), cert. denied, 464 U.S. 834, 104 S. Ct. 116, 78 L. Ed. 2d 116 (1983). The Illinois appellate court in the direct appeal of the instant case provided a detailed account of the evidence at trial. People v. Kosik, 110 Ill.App.3d 930, 66 Ill. Dec. 555, 443 N.E.2d 238 (1983). The following factual summary is based on that opinion.2

On July 22, 1979, at about 1:00 a.m. four men were standing on the northwest corner of Division Street and Cicero Avenue in Chicago. A dark-colored car traveling north on Cicero made a left-hand turn onto Division. As it made its turn onto Division gunshots were fired from the passenger's side of the car, hitting one of the four men twice. The shooting victim, Saul James, Jr., was paralyzed as a result.

All four men who had been standing near the street corner testified as eyewitnesses at the trial of Joseph Kosik. All four testified to the facts recited above. Three of them testified that there were two white men in the car.3 Another three also testified that the entire incident took a very short time, their estimates ranging from a few seconds to two minutes.4 The Illinois appellate court found that the area was "well-lighted" and that the entire incident "occurred quickly."

The two primary eyewitnesses, the victim, James, and his cousin, Charles Lemon, also testified more specifically about both the car and its occupants. Both testified that the car was a 1975 or 1976 black Buick Century. One of the other witnesses also offered the testimony that the car was a dark blue or black 1975-77 Buick Regal or Century with a dark vinyl roof and opera windows. Lemon read "JJK2" on the car's license plate, and James saw that the first three letters were "JJK." Lemon also observed a little flower, or flowers, hanging from the rear view mirror. James observed opera windows on the car. Both men testified to having seen the faces or profiles of the driver and the passenger. James described the driver after the incident as having long, dark hair and a mustache, and being in his twenties. At trial he added that the driver's hair was curly and he had a beard. Both James and Lemon testified that the passenger had blond hair and severe acne.

Lemon further testified that after the shooting he went back to the area of the incident and found in the neighborhood a 1976 black Buick Century with the license plate "JJK235," which he identified as the car used in the shooting. At trial he identified pictures of the car. He waited in the area until he saw someone come out of a nearby house. Before the individual approached the car Lemon recognized him as the driver he had seen during the shooting. Later in the day, Lemon again returned to the area, saw the driver again, and followed him around the block, where he saw the driver engaged in conversation with a man Lemon identified as the passenger who had fired the shots. When Chicago police detective Mook met with Lemon about one month after the shooting, Lemon positively identified Kosik from a photo array.

About one-and-a-half months after the shooting, Lemon was present at two lineups containing the same individuals. Both included Kosik. Lemon failed to recognize him. Some time between the shooting and the lineups (and after Kosik's first conversation with Detective Mook), Kosik shaved off his mustache and cut his hair.

James, the victim, was the other witness who identified Kosik as the driver. James was also shown a photo array before the preliminary hearing. At that time he was in the hospital under medication. He selected a picture and stated that "this might be him [the driver]," but at this point did not make a positive identification. The picture was of Kosik. Kosik's picture in this photo array and in the one Lemon viewed had been taken before he cut his hair and shaved off his mustache.

James and Lemon later appeared at a preliminary hearing held about two months after the incident. Both men positively identified Kosik as the driver. When they made their identifications on the record Kosik's case had already been called and Kosik had moved from a full gallery to a position near his attorney. James testified at trial that he had positively recognized Kosik among several others in the gallery before the case was called at the preliminary hearing.5

Kosik testified at trial in his own behalf, alleging that he had been elsewhere at the time of the shooting incident. He testified that the license plate number "JJK235" was for a 1975 black Buick Century registered to him and he identified photographs of the car. These were the same photographs Lemon identified as depicting the car used in the shooting. Kosik further testified that his car had a "poppy" or "donut" "something like a Veteran's Day" symbol hanging from the rear view mirror. Kosik admitted that he cut his hair and shaved off his mustache after his initial conversation with Detective Mook. He said it was his practice to do this periodically.

At trial, Lemon and James both positively identified Kosik as the driver of the car in the shooting. Kosik was convicted on an accountability theory of armed violence and aggravated battery in the shooting of James. On direct appeal the armed violence conviction was vacated, and the case was remanded for the trial court to issue a new sentence on the remaining aggravated battery conviction. Only this latter conviction is before this court in the appeal at bar.

Kosik alleges only two grounds for the writ, both under the same theory of law.6 He asserts that the trial identifications of him by both James and Lemon violated his right to procedural due process because they were so tainted by unconstitutionally suggestive identification procedures used at the preliminary hearing that they gave 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); accord Manson v. Brathwaite, 432 U.S. 98, 106, 97 S. Ct. 2243, 2248-49, 53 L. Ed. 2d 140 (1977); Neil v. Biggers, 409 U.S. 188, 196-97, 93 S. Ct. 375, 381, 34 L. Ed. 2d 401 (1972); United States v. Goodman, 797 F.2d 468, 470 (7th Cir.1986); United States ex rel. Moore v. Illinois, 577 F.2d 411, 413-14 (7th Cir.1978), on remand from 434 U.S. 220, 98 S. Ct. 458, 54 L. Ed. 2d 424 (1977), cert. denied, 440 U.S. 919, 99 S. Ct. 1242, 59 L. Ed. 2d 471 (1979).

The test we must apply here is a well settled, two-part inquiry. Under the first prong, the federal district court below found that the procedures used at the preliminary hearing were unduly suggestive. Under the second prong an identification after unduly suggestive procedures will still be admissible if it is not so unreliable under the totality of the circumstances that it creates a very substantial likelihood of irreparable misidentification. Brathwaite, 432 U.S. at 109-14, 97 S.Ct. at 2250-53; Biggers, 409 U.S. at 198-99, 93 S.Ct. at 382; United States v. Briggs, 700 F.2d 408, 412 (7th Cir.), cert. denied, 462 U.S. 1110, 103 S. Ct. 2463, 77 L. Ed. 2d 1340 (1983) & 461 U.S. 947, 103 S. Ct. 2129, 77 L. Ed. 2d 1307 (1983). In holding that the identifications possess sufficient indicia of reliability to prevent a very substantial likelihood of irreparable misidentification we assume, without deciding, that the pretrial confrontations under the circumstances of this case were unduly suggestive.7

While this second prong requires a determination of reliability upon the totality of the circumstances, the Supreme Court has also identified several specific factors to look to in determining the reliability of the identification. These factors are:

The opportunity of the witness to view the criminal at the time of the crime, the witness' degree of attention, the accuracy of his prior description of the criminal, the level of certainty demonstrated at the confrontation, and the time between the crime and the confrontation. Against these factors is to be weighed the corrupting influence of the suggestive identification itself.

Brathwaite, 432 U.S. at 114, 97 S.Ct. at 2253; accord Biggers, 409 U.S. at 199, 93 S.Ct. at 382; Goodman, 797 F.2d at 470; Moore, 577 F.2d at 414.

It is, of course, not our function in this setting to judge the ultimate accuracy of the identifications; that decision was made by the jury in its role as finder of the facts.8 Our role is the much more limited one of determining if the identification was so unreliable that the defendant's due process right to fair judicial procedures should have precluded an identification at trial. This judgment of minimally acceptable reliability has obtained constitutional stature because of the great evidentiary impact and statistically questionable validity of an eyewitness's identification from the stand of a defendant as the perpetrator of a criminal act.9 See, e.g., United States v. Wade, 388 U.S. 218, 228 & n. 6, 87 S. Ct. 1926, 1933 & n. 6, 18 L. Ed. 2d 1149 (1967) (noting unreliability of eyewitness testimony). Thus we determine here only whether the entire evidence regarding these identifications afforded them sufficient reliability to reach the jury despite any undue suggestiveness at the preliminary hearing.

Before considering individually the specific factors of the test, we note two crucial indicia of reliability pertinent to several of the factors and also directly supportive of the reliability of the identifications. First, both James and Lemon testified consistently with regard to the car's appearance and its occupants' features. The two other witnesses also bolstered the description of the car. Corroborated eyewitness testimony is much less suspect than one individual's visual impression. See, e.g., United States v. Cox, 428 F.2d 683, 686 (7th Cir.1970). And while it is possible that witnesses may collude in preparing testimony, no such suggestion has been made or supported on this appeal.10

Second, the descriptions of the car and its license plate offered by Lemon and James were specific and consistent. Moreover, a third witness testified that the car was a dark blue or black 1975-77 Buick Regal or Century with two doors, opera windows, and a dark vinyl roof. Although the description of the vehicle offered by the fourth witness was less specific, it was also consistent. A car matching these descriptions and the partial license plate numbers was registered to the defendant. We do not by any means suggest that the mere use of a car in a crime will subject that car's owner to the presumption that he was present. But other circumstances exist in this case that substantially support such an inference. Two different witnesses both described a driver with an appearance similar to Kosik's, the incident took place in Kosik's neighborhood, two different witnesses testified to recognizing Kosik during the incident, two different witnesses testified to partial license plates that matched Kosik's, and four witnesses testified to seeing a car with physical characteristics that matched Kosik's car. This evidence substantially corroborates the identification of the car's registered owner, Kosik, as the driver.

We also note at this point that one of our recent cases contains facts that in their fundamentals parallel those before us now. In United States v. Briggs, 700 F.2d 408 (7th Cir.), cert. denied, 462 U.S. 1110, 103 S. Ct. 2463, 77 L. Ed. 2d 1340 (1983) & 461 U.S. 947, 103 S. Ct. 2129, 77 L. Ed. 2d 1307 (1983), we held admissible at trial identifications made by two different witnesses who had both been tentative at best in their earlier photo identifications. The witnesses in Briggs also identified the individual in the courtroom prior to any suggestive proceedings. Id. at 411-14. Petitioner urges us to find that there are additional factors in this case that require a different result. While some of the factors noted are material, our analysis of them requires us to reach the same result we did in Briggs.

Turning to the first of the individual factors of the reliability test, we believe that James and Lemon had an adequate opportunity to view the individual at the crime. James and Lemon had only a short period of time under street lamps at night in which they actually could have viewed the car from which the shots were fired. They almost certainly had somewhat less time to view the driver within the car, as neither testified to looking inside the car as it first started its turn. Moreover, the passenger at some point after the turn was completed may well have partially or entirely blocked the witnesses' view of the driver. Finally, at least one lane of Division apparently separated the car and its occupants from the witnesses on the curb.

These conditions are not ideal for making identifications. But we cannot say that they precluded the possibility of obtaining a clear view of the driver sufficient to form a definite and specific impression of his appearance. The area was well lit and the driver could be viewed through the car's windows for at least a substantial portion of the time that elapsed. For a portion of that time the car was less than half the road's width from the witnesses. These facts establish a physical opportunity to view the individual sufficient later to identify him. E.g., Moore, 577 F.2d at 414. While a limited opportunity to view an individual during a crime certainly provides a lower indicium of reliability than a better opportunity,11 it does not preclude an ultimate conclusion of constitutional reliability where there are other sufficient indicia of reliability--unless, of course, there was no opportunity during the crime to obtain a definite impression of the individual's appearance. This factor is, after all, only a single component of a test based on the totality of the circumstances.

Moreover, in United States v. Cox, 428 F.2d 683 (7th Cir.1970), we found an identification admissible under an opportunity to view roughly analogous to the opportunity in this case. The witness in Cox saw the individual at the crime for a period of several seconds under good visibility at a distance of forty-five to sixty feet. Id. at 685-86. In the case at bar we are required by the findings of the state courts and the federal district court to consider the lighting good, and the distance involved here was somewhat shorter. Even if we assume, arguendo, that the street lamps did not provide visibility equivalent to that in Cox (where the crime occurred during daylight), the shorter distance largely makes up for the poorer light. Finally, Cox does not portend to establish the minimum factual threshold the government must meet to show a sufficient opportunity to view, and we consider the opportunity to view here to be, at the extreme, only slightly poorer than that in Cox.

The degree of attention of the witnesses is the second specific factor in the test for reliability. While opportunity to view focuses on the physical possibility of obtaining a reliable identification, attentiveness requires the court to determine if the witnesses actually took advantage of the opportunity. Attentiveness also requires the court to determine what degree of attentiveness the witnesses displayed. The attentiveness shown by the witnesses here varies. It appears from the record that the two minor witnesses threw themselves to the ground upon hearing the shots and did not look at the car again until it was leaving the scene. Both James and Lemon testified to continuing to look into the car.

Both parties to this action have argued that under gunfire one necessarily is, or is not, attentive. It takes no more than common sense to recognize that neither result necessarily follows. We do not undermine this observation if we accept the expert testimony offered at trial, which suggested that during gunfire people statistically are likely to concentrate their gaze upon the weapon. Here James's and Lemon's identification testimony includes descriptions of the driver, who was for a good portion of the incident directly behind the weapon and the person firing it. Moreover, the simple fact that people are statistically more likely to do something cannot normally, and does not here, undermine the specific testimony that James and Lemon did in fact look into the car. The specific testimony suggests that both James and Lemon were highly attentive: they noticed specifics about both the car (model, year, color, partial license number, flowers on the mirror, opera windows) and its occupants (acne, hair length and color, skin color, facial hair, approximate age).

Kosik relied heavily in the briefs to this court and at oral argument upon the First Circuit's opinion in Velez v. Schmer, 724 F.2d 249 (1st Cir.1984), for the proposition that a witness's high degree of attention to the details of a car's appearance during a brief encounter necessarily detracts from that witness's ability to describe reliably the car's occupants. We do not read the First Circuit to have gone so far.

Kosik's proposition is logically flawed in at least three respects. First, the car in the shooting was at least one lane's width from the curb. Scientific evidence is not required to show (nor does the scientific evidence offered here successfully controvert) that this distance might allow an individual at the curb to observe virtually simultaneously several of the specifics the witnesses in this case have in fact described--including, for example, the flowers hanging from the mirror and the appearances of the car's two occupants. Second, the other details about the car (its make, model, year, license number, and opera windows) were most likely seen either as the car began its turn onto Division or as it sped away. During these periods the witnesses would necessarily have had a less satisfactory opportunity to see into the car and to observe its occupants (a fact we have taken into account in our discussion of the opportunity to view the driver). Thus the observation of the license plate and of the body of the car does not detract from the witnesses' potentially optimal use of the time they did have to observe the two individuals within the car. Finally, it is beyond cavil that individuals differ in their attentiveness to their physical environment. The specificity of the observations of the car made by the witnesses here tends to establish their attentiveness to all the particulars of the scene throughout the incident, including the appearances of the car's occupants.

This analysis is perfectly consistent with Velez, for in Velez, unlike the case at bar, the witnesses to the crime offered virtually no description of the individuals. In fact, the Velez court noted that "[w]e have seen no case with such a total lack of description of the individual." Velez, 724 F.2d at 252. Thus the substantial specific testimony regarding the car's appearance in Velez was in that case starkly contrasted by the almost complete absence of a description of the individual. Here, on the other hand, the attentiveness to the car's appearance complements the substantial specific testimony of multiple witnesses regarding the driver's appearance. Velez does not hold that witnesses who had a short opportunity to view the individual can only describe either the individual or some other object seen during the incident. Such a holding would be ludicrous. Velez simply and accurately notes that where there is an absence of specific testimony regarding one component of a scene and substantial specific testimony regarding another, the implication is that more attention was paid to the component that was more completely described. See Velez, 724 F.2d at 251-52. In this case the car and its passengers were all well described. Thus while we concur in the teaching of Velez, it simply does not apply here.

The third specific factor to consider is the accuracy of the witness's prior description of the individual. This factor helps the court determine if and when the witness developed and expressed a concrete and specific impression of the individual's characteristics firm enough to remain reliable despite the vagaries of time and the pressures of any undue suggestiveness. The prior description of Kosik by James was accurate. Before the suggestive procedures James described accurately Kosik's approximate age, skin color, hair color, and he also noticed Kosik's mustache. At trial he added that the driver's hair was curly and that he had a beard. This description is not exhaustive but it includes several specific factors, many of which were communicated before the assumed suggestive procedures. It bolsters the reliability of the identifications. See, e.g., Biggers, 409 U.S. at 200, 93 S.Ct. at 383; Goodman, 797 F.2d at 470.

Lemon evidently gave detective Mook a description of the driver about one week after the offense, but that description is not in the trial record. Our inability to ascertain the accuracy of this prior description obviously prevents this fact from weighing in favor of the reliability of Lemon's identification. However, it is also important to note that since it is an omission rather than an affirmative inaccuracy it does not cast doubt on other evidence supporting the reliability of the identification. Cf. United States ex rel. Hudson v. Brierton, 699 F.2d 917, 925 (7th Cir.1983) (holding finding of reliability not clearly erroneous despite lack of prior description). Indeed, as the district court cogently noted, Lemon's identification of Kosik from the photo array prior to the suggestive procedures serves to help support the inference that Lemon had obtained a definite impression of the driver's appearance at the crime. While his later failure to identify Kosik in two lineups might well in other circumstances significantly undercut this inference, in this case Kosik by the day of the lineups had cut his hair and shaved off his mustache. The absence in the record of a prior verbal description by Lemon of the driver does not under these circumstances weigh significantly, if at all, against the reliability of Lemon's identification of Kosik at trial.

The fourth specific factor is the level of certainty displayed by the witness in his confrontation with the individual. Here Lemon and James made positive identifications of Kosik at both the preliminary hearing and the trial. Determinations of the reliability suggested by a witness's certainty after the use of suggestive procedures are complicated by the possibility that the certainty may reflect the corrupting effect of the suggestive procedures. See, e.g., Brierton, 699 F.2d at 925; Moore, 577 F.2d at 415. This difficulty has not prevented courts from finding sufficient certainty even when the evidence of certainty comes from confrontations that took place after the invocation of suggestive procedures. See, e.g., Brathwaite, 432 U.S. at 115-16, 97 S.Ct. at 2253-54; Biggers, 409 U.S. at 200-01, 93 S.Ct. at 382-83; Brierton, 699 F.2d at 925. In such cases, there exists sufficient other evidence to establish reliability or to convince the court that the certainty is the result of the identification at the crime. Here we also have testimony of identifications by both witnesses of Kosik made before the suggestive procedures. Lemon identified Kosik coming out of a house later on the same day of the shooting, and also picked out Kosik's photo from an array one month after the proceeding. James identified Kosik in the gallery of the courtroom before the use of the suggestive procedures at the hearing. Both men were certain of these identifications. This testimony tends to establish certainty at confrontations prior to the suggestive procedures, and bolsters the inference that the certainty exhibited at the confrontations after the suggestive procedures was not tainted by them.

Kosik argues that James's failure to positively identify Kosik from the photo array should undercut this certainty. We have, however, already ruled that a previous failure to make a positive identification from a photo array does not necessarily, or even normally, make the later identification less certain. United States v. Briggs, 700 F.2d 408, 413 (7th Cir.), cert. denied, 462 U.S. 1110, 103 S. Ct. 2463, 77 L. Ed. 2d 1340 (1983) & 461 U.S. 947, 103 S. Ct. 2129, 77 L. Ed. 2d 1307 (1983); United States ex rel. Moore v. Illinois, 577 F.2d 411, 413-14 (7th Cir.1978), on remand from 434 U.S. 220, 98 S. Ct. 458, 54 L. Ed. 2d 424 (1977); accord United States v. Black, 412 F.2d 687, 689 (6th Cir.1969), cert. denied, 396 U.S. 1018, 90 S. Ct. 583, 24 L. Ed. 2d 509 (1970). The rationale is unexceptionable; the photographs may be poor or old or from an angle different from the one presented to the witness at the crime. The witness may also not wish to take the serious step of identifying someone as a possible criminal without personally seeing the individual. Accord Moore, 577 F.2d at 415. We see no worthwhile reason to depart from our earlier decisions, and counsel has not presented one. In this case James made a tentative selection of Kosik's photograph from the array. While this is certainly not a positive identification, it is equally certain that it adds an indicium of reliability to the identification of Kosik as the driver.

Petitioner also questions James's identification of Kosik while Kosik was in the gallery section of the courtroom prior to the assumed suggestive procedures at the preliminary hearing. We have previously recognized that such testimony bolsters the reliability of an identification, at least where, as here, there is no evidence that the witness wished to wreak vengeance upon an innocent person, or was coached or otherwise led improperly to attempt such an identification. Briggs, 700 F.2d at 413, cf. Moore, 577 F.2d at 415 & n. 15. While this testimony is far from dispositive, it does not weaken the certainty of James's identification.

Petitioner also argues that Lemon's failure to identify Kosik at the lineups should make us doubt the certainty of his identifications of Kosik. Besides sharing the weaknesses of petitioner's argument as to James's, we note that Kosik had altered his appearance prior to the lineups. Lemon may have been looking only for someone with longer hair and a mustache, which Kosik still had in the photo Lemon had already picked out. At the preliminary hearing, where we presume Kosik still had shorter hair and no mustache, Lemon may have taken into account the possibility that Kosik may have altered his appearance. While this construction of the facts is by no means certain, it is perfectly plausible, and in a case like this our task is only to exclude the highly suspicious and fundamentally questionable. The plausibility of this construction of the evidence is also supported by the district court's specific finding that Lemon was "perhaps the most credible witness." In sum, petitioner's arguments do not convince us that we should significantly question that certainty exhibited by Lemon in his identifications of Kosik.

Under the last specific factor, the time between the shooting incident and the various pretrial identifications was in no case greater than two months, which is not enough by itself to raise serious questions about reliability. See, e.g., Goodman, 797 F.2d at 471 (two months from incident to identification at suggestive pretrial proceeding and five months from incident to identification at trial did not in the circumstances of the case make the identification inadmissible); cf. Biggers, 409 U.S. at 201, 93 S.Ct. at 383 (lapse of seven months between view of individual at crime and first identification of defendant did not make identification inadmissible). Lemon, of course, made a positive photo identification about one month after the incident.

Considering these circumstances in their entirety against the corrupting effect of the assumed unduly suggestive procedures, we conclude that there was not a substantial likelihood of irreparable misidentification.12 We have assumed the undue suggestiveness of the procedures at the preliminary hearing where the prosecution asked for identifications of Kosik in court when he was near his attorney, his case had been called, and the charges against him read. Against this one must set the several incidents prior to the suggestive procedure in which Lemon and James tentatively or positively identified Kosik as the driver and provided specific and consistent evidence supporting those identifications. These are not so weakened by the failures of Lemon to recognize Kosik in the lineups or the failure of James to be positive in picking Kosik out of the photo array that the trial identifications should have been suppressed. The significant corroborative testimony regarding the car further supports the reliability of the identifications, as does the consistent testimony regarding the appearance of the passengers within the car. This testimony could not have been tainted by the suggestive procedures. While James's testimony regarding the length of the incident and Lemon's testimony regarding his search for the car and identification of Kosik immediately after the incident might both represent fabrications that cast doubt on the rest of their testimony, there is no significant evidence suggesting that is the case, and we conclude that the jury had the right to determine for itself these subsidiary matters and the ultimate accuracy of the trial identifications.13 There was no constitutional error in admitting the trial identifications.

The judgment of the trial court denying petitioner-appellant Joseph Kosik a writ of habeas corpus is AFFIRMED.