Mervyn Thompson was shot by a man who had identified himself as “Mr. Palmer.” Thompson described his assailant to the police, and some weeks later was hypnotized in the hope that he would recall more details of the shooting. At a show-up a few days after the hypnosis session, Thompson identified his assailant as a man named David Harker. Harker was convicted of assault with intent to murder by a Maryland jury, and later filed a petition for habeas corpus, under 28 U.S.C. § 2254.
We affirm the district court’s denial of Harker’s petition. Testimony by a witness who has previously been hypnotized must meet criteria of independence and reliability, but such testimony does not per se violate either the sixth amendment right to confrontation or the fourteenth amendment right to due process. In addition, the testimony given by Harker’s fellow inmate did not constitute the admission of an uncoun-seled confession, because the inmate was not a government agent.
I.
On the evening of December 7, 1980, a man who called himself “Mr. Palmer” telephoned Mervyn Thompson to say that Palmer’s son had cut his hand playing on an old truck on Thompson’s property. Palmer asked Thompson to meet him at the truck, to see if any damage had been done to it. At approximately 10:00 p.m., Thompson’s daughter-in-law Barbara drove him down the driveway to the place where the truck and Palmer’s car were parked. Next to the truck was a pole with a hanging light, which had been turned on. Barbara’s headlights were on, as were the headlights of Palmer’s car. Thompson was carrying a six volt flashlight, which he used to inspect the truck with Palmer. They found no recent damage; Thompson signalled Barbara to leave, and she did so.
Thompson had begun to walk back to the house when Palmer called him, saying he had dropped a glove. Thompson looked for a glove, but did not see one. Palmer jumped in his car and Thompson saw that *439he had a shotgun. Palmer shot Thompson, mangling his left arm so badly that it later had to be amputated. After a few shots, Thompson remembered that he was carrying a revolver and fired back. Palmer drove off. Thompson returned to his house for help.
Thompson described Palmer’s car as a red, four-door Honda Accord. According to the testimony of Sergeant Van Horn, Thompson described his assailant as a white male, aged thirty-five to forty-five, about six feet, two inches tall, medium build, medium brown hair, wearing a flannel shirt, tan trousers and glasses. Thompson also worked with police to develop a composite, which was later published in a Harford County newspaper.
Two weeks after the assault, a Maryland state trooper spotted a red Honda Accord driven by a man who resembled the composite. The officer recorded the license plate number and determined that the car was registered to David Harker. At some point between the assault and the hypnosis session, Thompson also looked at a photo array that did not contain a picture of Harker. Thompson did not select any of the photographs as a picture of his assailant.
On January 30, 1981, Thompson consented to be hypnotized by James Roby of the Montgomery Police Department. During that session, Roby told Thompson that he would be able to visualize the events of December 7, and Roby asked Thompson to describe what he was seeing. Roby attempted to record the hypnosis session both on video tape and audio tape. Unfortunately, the lens cap was never removed from the video camera, and Roby succeeded only in producing an audio recording.
A few days after the hypnosis session, police showed Thompson ten photographs. Thompson selected a photo of Harker, and said that it resembled his assailant. The next day, police learned that Harker would be at the Baltimore County Circuit Court within the week. Thompson and his daughter-in-law Barbara went to the courthouse with two police officers, and sat on benches in the hall. Sergeant Van Horn estimated that there were hundreds of people milling around. Thompson saw a man in a camel’s hair coat, and said that it looked like his assailant. The man walked into a courtroom. When he came out, and Thompson was able to get a better look at him, Thompson positively identified the man as his assailant. Later that day, police arrested the man Thompson had identified, David Harker.
Thompson’s photographic, courthouse, and courtroom identifications of Harker were admitted at trial over objection that they were impermissibly tainted by the hypnosis session. Harker was convicted of assault with intent to murder, and sentenced to thirty years. The Maryland Court of Special Appeals affirmed the conviction, Harker v. State, 55 Md.App. 460, 463 A.2d 288 (1983); the Maryland Court of Appeals denied Harker’s petition for a writ of certiorari. Harker then petitioned for a writ of habeas corpus, contending that the trial judge erred in admitting Thompson’s identifications, and in permitting a jailhouse informant to testify. The district court found that the Maryland courts had offered Harker a full and fair opportunity to litigate his contentions, and that the record disclosed no violation of Harker’s constitutional rights. We affirm the decision of the district court that Harker is not entitled to federal habeas relief.
II.
Much of the legal debate over the use of post-hypnotic testimony has its origin in the scientific dialogue on the functioning of human memory. The theory that the memory holds a certain snapshot of the crime has now come into question. Instead, “[mjany memory theorists believe ... that what appears to be recall is actually a constructive process in which information received after an event is integrated by the mind into the memory representation of that event.” United States v. Valdez, 722 F.2d 1196, 1200 (5th Cir.1984). This view of the malleability of memory has caused courts to treat the refreshment of witness *440recollection through hypnosis with some caution.
In cases where hypnosis is used, it is often the victim of the crime who is hypnotized. See, e.g., People v. Shirley, 31 Cal.3d 18, 181 Cal.Rptr. 243, 641 P.2d 775, cert. denied, 459 U.S. 860, 103 S.Ct. 133, 74 L.Ed.2d 114 (1982); State v. Hurd, 86 N.J. 525, 432 A.2d 86 (1981); Harding v. State, 5 Md.App. 230, 246 A.2d 302 (1968), cert. denied, 395 U.S. 949, 89 S.Ct. 2030, 23 L.Ed.2d 468 (1969); United States v. Narciso, 446 F.Supp. 252 (E.D.Mich.1977). Hypnosis, it has been suggested, can help people to remember things that they would not otherwise remember. Dr. Martin Orne, a psychiatrist and frequent expert witness, believes that “hypnosis may be useful in some instances to help bring back forgotten memories following an accident or a crime.” Orne, The Use and Misuse of Hypnosis in Court, 27 Int. J. Clinical & Experimental Hypnosis 311, 317-18, quoted in State v. Hurd, 432 F.2d at 93.
For example, in Chowchilla, California, twenty-six children were kidnapped from a school bus, and the bus driver was able to recall, under hypnosis, a license plate number which was instrumental in the capture of the kidnappers. Note, Excluding Hypnotically Induced Testimony on the “Hearsay Rationale, ” 20 Val.U.L.Rev. 619, 619 n. 6 (1986). Often police attempt to use hypnosis as a trigger to the recollection of the visage of a criminal suspect. See Clay v. Vose, 771 F.2d 1 (1st Cir.1985).
The dilemma, of course, is that the very process that yields investigative fruits may sow testimonial dangers. We recognize that hypnosis has drawbacks, specifically, the potential for suggestibility, confabulation, and hardening of memory, and the attendant danger of misidentification. "While many of the same problems inhere in eyewitness testimony generally, they can be compounded by the technique of hypnosis, a powerful tool, yet one imperfectly understood. Many psychologists believe that a hypnotized person is more likely to be led by suggestions made by a hypnotist or questioner. See Diamond, Inherent Problems in the Use of Pretrial Hypnosis on a Prospective Witness, 68 Calif.L.Rev. 313, 333 (1980); United States v. Valdez, 722 F.2d at 1201. “The attitude, demeanor, and expectations of the hypnotist, his tone of voice, and his body language may all communicate suggestive messages to the subject.” Diamond, 68 Calif.L.Rev. at 333. Moreover, the relationship between hypnotist and subject may be both intense and intimate. Subconsciously, the subject may give the response he or she thinks the hypnotist wants to hear. People v. Shirley, 181 Cal.Rptr. at 269, 641 P.2d at 801. The result may be that the subject’s description of the event departs from reality in order to conform to the questioner’s expectations.
A related problem, called “confabulation,” occurs when the subject fabricates missing details, often using parts of other real memories that are unrelated to the situation the subject is trying to remember. “People want to give an answer, to be helpful, and many will do this at the risk of being incorrect. People want to see crime solved and justice done, and this desire may motivate them to volunteer more than is warranted by their meager memory.” E. Loftus, Eyewitness Testimony at 109 (1979), quoted in People v. Shirley, 181 Cal.Rptr. at 270, 641 P.2d at 801.
The problems of suggestibility and confabulation are compounded by a third phenomenon: memory hardening. After the hypnosis session has ended, the subject may not remember being hypnotized. Diamond, 68 Calif.L.Rev. at 334. Moreover, the subject may not be able to distinguish between events recalled before hypnosis, and those recalled during hypnosis. State v. Hurd, 432 A.2d at 93. “Once a witness makes a recitation under hypnosis, his confidence in that supposed memory — whether genuine or invented — is greatly strengthened. The witness then may have an unshakeable subjective conviction that gives his account on the witness stand the imprimatur of absolute confidence.” United States v. Valdez, 722 F.2d at 1202 (footnote omitted). The worst fear then, is that an *441eyewitness will inaccurately reconstruct the memory of the crime in question, either by suggestion or confabulation, and will then become convinced of the absolute accuracy of the reconstruction through memory hardening.
The purpose of hypnosis, of course, is to improve recall without distorting it. In a sense, hypnotizing an eyewitness is similar to cleaning a valuable painting. A picture covered with grime may have little value for anyone. However, in cleaning the picture, hoping to restore its full distinction; one risks changing the original work.
III.
The legal framework for dealing with the problem of hypnosis must closely parallel the medical debate. The dangers of distortion in witness recollection have led courts to agree upon two basic propositions. First, no witness should ever be permitted to testify while under hypnosis. See State v. Collins, 296 Md. 670, 464 A.2d 1028, 1034 (1983). Second, neither the government nor the defendant should be permitted to introduce statements made under the influence of hypnosis for the truth of the matter asserted. In both cases, courts have considered that statements made under hypnosis are too unreliable. People v. Shirley, 181 Cal.Rptr. at 251, 641 P.2d at 783.
At the same time, the possible utility of hypnosis in the refreshment of memory has persuaded the federal courts and a substantial number of state courts that the testimony of a previously hypnotized witness should not be subject to a rule of per se exclusion at trial. See, e.g., Clay v. Vose, 771 F.2d 1 (1st Cir.1985); United States v. Valdez, 722 F.2d 1196 (5th Cir. 1984); United States v. Awkard, 597 F.2d 667 (9th Cir.), cert. denied, 444 U.S. 885, 969, 100 S.Ct. 179, 460, 62 L.Ed.2d 116, 383 (1979); State v. Hurd, 432 A.2d 86.1 To formulate a per se rule would foreclose whatever investigative benefits may be derived from the responsible use of this technique. The inquiry rather is a more balanced one, and addresses whether the ineourt testimony of a previously hypnotized witness has a basis which is in fact independent of hypnotic influence.
Harker contends that the testimony of Thompson was the product of hypnotic influence and accordingly violated both his fourteenth amendment due process rights and his sixth amendment right to confront the witnesses against him. For the reasons that follow, we reject these two contentions.
A.
It is clear that Harker’s sixth amendment right to confrontation was not violated in any conventional sense. Thompson was present as a witness, and was subject to extensive cross-examination, which explored the circumstances of the criminal encounter by the truck. The defense probed the events the night of the assault, the specifics of the assailant’s appearance, his pattern of speech and accent, and when Thompson first heard the name David Harker. Any inconsistencies in Thompson’s testimony, any flaws in his demeanor, any limitations on his observation, were there for counsel to explore and the jury to evaluate. See Clay v. Vose, 771 F.2d at 4.
Moreover, the jury was fully aware that Thompson had been hypnotized. Lieutenant James Roby, the hypnotist, testified at *442trial and also submitted to cross-examination. Defense counsel properly inquired into Roby’s position in law enforcement, his training in hypnosis, the circumstances of the hypnosis, the possibility of suggestiveness, and the formulation of the questions Roby asked of Thompson. For example, counsel elicited the fact that after Thompson had described his assailant, Roby asked, “Do you think you would recognize his face?” This extensive probing of the hypnotic session itself conforms to the general view of federal courts that prior use of hypnosis on a witness goes to the weight to be given his testimony rather than its admissibility. See United States v. Awkard, 597 F.2d 667 (9th Cir.1979); Clay v. Vose, 771 F.2d 1 (1st Cir.1985).
Expert witnesses on the subject of investigative hypnosis also testified for both sides. The state called Dr. Daniel Stern, who said that in eighty-two percent of the cases in which he had been involved, hypnosis had led to additional information. Stern described hypnosis as a specialized interview technique and discussed the necessary safeguards. The defense cross-examined Stern on the problems with hypnosis, and on compliance with customary safeguards in the hypnosis of Thompson. Dr. Dennis Harrison testified for the defense on the pitfalls of hypnosis, its potential for inaccuracy, and the level of acceptance of investigative hypnosis in the scientific community. Harrison specifically criticized Thompson’s hypnosis session because of problems with recording equipment, the interruption for changing the tape, and the use of a police officer as the hypnotist. Thus the jury had substantial information in the form of expert testimony to aid it in evaluating the effect of the hypnotic session in this case. See Clay v. Vose, 771 F.2d at 4.
Harker nonetheless suggests that hypnosis so programs and fixes the perceptions of a witness that effective cross-examina-becomes impossible. To begin with, the circumstances of this case belie that assertion. The Maryland courts found specifically “that the hypnotic session would not impair defense counsel’s ability to cross-examine the victim because the manner in which Thompson made the photographic and confrontational identification (not positive as to the photograph; hesitant at the confrontation) indicated that he Retained critical judgment, unimpaired by the hypnosis.” Harker v. State, 463 A.2d at 294.
Nor do we think a per se rule of exclusion for post-hypnosis identification testimony on sixth amendment grounds is justified. Suggestibility and memory hardening are not unique to hypnosis; they are potential problems whenever an eyewitness makes an identification. “Without underestimating the seriousness of the problems associated with hypnosis, it should be recognized that psychological research concerning the reliability of ordinary eyewitnesses reveals similar shortcomings.” State v. Hurd, 432 A.2d at 94. These shortcomings occur for several reasons. Any witness who repeats a story a number of times is likely to become increasingly confident in its accuracy. Williams and Hammelmann, Identification Parades, Part I, 1963 Crim.L.Rev. 479, 482, quoted in United States v. Wade, 388 U.S. 218, 229, 87 S.Ct. 1926, 1933, 18 L.Ed.2d 1149 (1967). No one likes to admit error. Moreover, the line-up, photo or hypnotic identification is more recent in time than the witnessing of the crime. “Regardless of how the initial misidentification comes about, the witness thereafter is apt to retain in his memory the image of the photograph rather than of the person actually seen, reducing the trustworthiness of subsequent lineup or courtroom identification.” Simmons v. United States, 390 U.S. 377, 383-84, 88 S.Ct. 967, 971, 19 L.Ed.2d 1247 (1968).
Despite this widespread recognition of the problem, the Supreme Court has never suggested that memory hardening, a danger in all identifications, poses a per se violation of the sixth amendment right. See Simmons v. United States, 390 U.S. at 384, 88 S.Ct. at 971. If that were the case, a vast amount of critical eyewitness testimony would ipso facto be excluded. The answer lies in apprising the jury of the *443uses and dangers of hypnosis, in permitting full exploration of the hypnotic event, and in cross-examining the subjection the opportunity for pre-hypnotic observation. These steps were taken in this case, and Harker accordingly suffered no violation of his sixth amendment right to confrontation.
B.
We next consider whether Thompson’s identification of Harker was so unreliable that it violated Harker’s fourteenth amendment right to due process. Harker contends that the hypnosis session was imper-missibly suggestive and tainted the subsequent photo array and show-up. We affirm the decision of the district court that Harker’s right to due process was not violated and that neither pre-trial procedures nor Thompson’s courtroom testimony gave rise to “a very substantial likelihood of irreparable misidentification.” Simmons v. United States, 390 U.S. at 384, 88 S.Ct. at 971.
The testimony of a previously hypnotized witness may contravene due process in one two ways. First, the in-court identification may lack a basis which is independent from the hypnotic session, and second, the procedures used during the hypnosis session may themselves be defective. The exclusion of evidence from the jury is, however, a drastic sanction, one that is limited to identification testimony which is manifestly suspect. “Short of that point, such evidence is for the jury to weigh ... for evidence with some element of untrustwor-thiness is customary grist for the jury mill.” Manson v. Brathwaite, 432 U.S. 98, 116, 97 S.Ct. 2243, 2254, 53 L.Ed.2d 140 (1977).
The Supreme Court has previously established the factors to be considered in assessing the quality of an independent basis for the identification. There is no violation of due process if
under the ‘totality of the circumstances’ the identification was reliable even though the confrontation procedure was suggestive____ [T]he factors to be considered in evaluating the likelihood of misidentification include the opportunity of the witness to view the criminal at the time of the crime, the witness’ degree of attention, the accuracy of the witness’ prior description of the criminal, the level of certainty demonstrated by the witness at the confrontation, and the length of time between the crime and the confrontation.
Neil v. Biggers, 409 U.S. 188, 199, 93 S.Ct. 375, 382, 34 L.Ed.2d 401 (1972). See also Manson v. Brathwaite, 432 U.S. at 114, 97 S.Ct. at 2253; Willis v. Garrison, 624 F.2d 491 (4th Cir.1980).
Thompson had a good opportunity to see his assailant. The lights along the driveway were lit, and the headlights of the assailant’s car were on. Thompson estimated that he spent between five and eight minutes with his assailant. Before hypnosis, Thompson and the police produced a composite. Deputy Jordan Watts testified that he had shown the composite to David Harker’s father, and that the father had said it closely resembled his son. Under hypnosis, Thompson gave a description of his assailant that closely matched the description he had given to police shortly after the shooting; he did not change his story during or after hypnosis. The consistency in Thompson’s identifications before, during, and after hypnosis provided ample basis for permitting Thompson to testify before the jury.
This clear independent basis for Thompson’s identification of Hark'eFcompensates for the deficienciSs in the hypnosis session itself. We agree with the Maryland Court of Special Appeals that “it would obviously have been better to have a video tape of the hypnosis session,” Harker v. State, 463 A.2d at 295, enabling the accused to challenge the techniques used by the hypnotist and the “court to determine what information or suggestions the witness may have received during the session and what recall was first elicited through hypnosis.” State v. Hurd, 432 A.2d at 97. Although the hypnotist attempted to record the session on both audio and video tape, the lens cap was never removed from the video camera. *444The audio tape of the session was, however, available with some small inaudible gaps.
The person who hypnotized Thompson, Lieutenant James Roby, was a trained law enforcement hypnotist and a member of the Montgomery County Police Department. The shooting occurred in Harford County, and Roby did not participate in the investigation except for the hypnosis session. Thus there was much less temptation for the hypnotist to ask questions which would confirm a desired result. Roby, in fact, testified that he knew very little about the case prior to the hypnosis session; he had not heard the name David Harker, nor had he ever seen a photograph or sketch of Harker. Perhaps for that reason, no photographs or sketches of the assailant were shown Thompson during the session. It has been said that “inaccuracy in hypnotic recall increases with the number and detail of questions” and that “the least-distorted recollections are provided by free narration,” United States v. Valdez, 722 F.2d at 1203 (citing Orne, The Use and Misuse of Hypnosis in Court, 27 Int’l J. Clinical and Experimental Hypnosis 311, 319-321 (1979)). The Maryland trial court found that the session in this case did not exhibit an excess of direction or guidance. See Harker v. State, 463 A.2d at 294.
Finally, Harker’s rights were not violated by the photo array or the show-up. Harker argues that of the ten photographs shown to Thompson, only his showed a subject wearing a plaid flannel shirt over another shirt. During the hypnosis session, Thompson had described his assailant as wearing a plaid flannel shirt over a colored shirt with a collar. The picture of Harker showed him wearing a flannel shirt over a white tee shirt. Another photograph in the array also showed a subject wearing a flannel or velour shirt over a dark shirt with a collar. The large number of photographs shown Thompson further militates against impermissible suggestiveness. See Simmons v. United States, 390 U.S. at 385, 88 S.Ct. at 971. Moreover, facial features, rather than clothing, appear to have been more important to the identification in this case, and police often lack a variety of photographs of the same suspect from which to select.
The testimony on the show-up at the courthouse reveals no due process violation. Police did not tell Thompson that his assailant would be at the courthouse. Thompson was not sure of the identification at first, but was positive after a second look at Harker. The show-up was not so suggestive as to produce any substantial likelihood of misidentification. See Willis v. Garrison, 624 F.2d 491.
IV.
Harker also objects to the use of the testimony of one Larry Eley, a fellow inmate, who testified that Harker had told him about shooting Thompson. Harker contends first that Eley was an incompetent witness and secondly that Eley was a government agent and extracted an un-counseled confession. We find no merit in either contention.
Pointing to post-conviction testimony from a single psychiatrist, Harker moved for a new trial, arguing that the state should have known that Eley was mentally unstable, and that the use of his testimony was therefore fundamentally unfair. This argument fails for several reasons, the most compelling of which is the trial court’s conclusion that exclusion of the evidence would not have changed the outcome of the trial. The evidence did not tend to establish affirmatively the innocence of the accused, and there remained substantial evidence that established his guilt.
Harker next argues that Eley was a government agent who extracted an uncounseled confession. In United States v. Henry, 447 U.S. 264, 100 S.Ct. 2183, 65 L.Ed.2d 115 (1980), the Supreme Court held that the use of testimony by a fellow inmate of the accused who was a paid informant acting under instructions of the government violated the sixth amendment right to counsel. However, “the mere *445presence of a jailhouse informant who had been instructed to overhear conversations and to engage a criminal defendant in some conversations would not necessarily be unconstitutional.” Henry, 447 U.S. at 276, 100 S.Ct. at 2190 (Powell, J. concurring).
Eley was not paid, nor was he acting under the instructions or solicitations of the government. He responded to a general request for information, and no promises were made to him. Since Eley was not a government agent, there was no violation of Harker’s right to counsel. “[T]he Sixth Amendment is not violated whenever — by luck or happenstance — the State obtains incriminating statements from the accused after the right to counsel has attached.” Maine v. Moulton, — U.S.-, 106 S.Ct. 477, 487, 88 L.Ed.2d 481 (1985). See also Kuhlmann v. Wilson, — U.S. -, 106 S.Ct. 2616, 91 L.Ed.2d 364 (1986).
For the foregoing reasons, the decision of the district court is
AFFIRMED.
. Some courts have held that the use of hypnosis affects credibility, but not admissibility. Others have permitted post-hypnotic testimony only if certain procedural safeguards are present. Still others have found post-hypnotic testimony inadmissible per se. Ruffra, Hypnotically Induced Testimony: Should It Be Admitted?, 19 Crim.L.Bull. 293, 293-94 (1983). For a thorough discussion of the development of the case law, see People v. Shirley, 31 Cal.3d 18, 181 Cal.Rptr. 243, 641 P.2d 775 (1982).
When Harker was tried, Maryland permitted the introduction of post-hypnotic testimony provided that the judge found it reliable. Polk v. State, 48 Md.App. 382, 427 A.2d 1041 (1981). Currently, under Maryland law, hypnotically-enhanced testimony is not admissible, although a witness would be permitted to testify in accord with statements made prior to hypnosis. State v. Collins, 296 Md. 670, 464 A.2d 1028 (1983). Our only inquiry in a federal habeas proceeding is, of course, a constitutional one.