Dey v. Scully

MEMORANDUM AND ORDER BLOCK, District Judge: Petitioner Robert Dey, a/k/a Robert Connyer (“Dey”), brings this petition for a writ of habeas corpus pursuant to 28 U.S.C. § 2254, challenging his 1989 convictions on one count of criminal sale of a controlled substance in the third degree (crack cocaine) and one count of criminal possession of a controlled substance in the seventh degree (heroin), following a jury trial in New York State Supreme Court, Queens County. 1 Dey argues (1) that the exclusion of his arrest photograph deprived him of a fair trial and due process with respect to his conviction for sale of crack cocaine, and (2) that the evidence was insufficient to support his conviction for possession of heroin. Dey’s petition is granted with respect to his conviction for sale of crack cocaine and denied with respect to his conviction for possession of heroin. I. BACKGROUND A. The Arrest and Identification On the evening of January 6, 1988, members of the New York City Police Department’s (“NYPD”) Queens Special Anti-Crack Unit conducted a “buy and bust” operation in the vicinity of 170th Street and Liberty Avenue in Queens, New York. In a buy and bust operation, an undercover police officer (“undercover”) is given pre-recorded “buy money” to purchase narcotics from a drug seller at á particular location. Once a purchase is completed, the undercover radios a physical description of the seller to a nearby backup team, which then locates and arrests the seller. Sometime shortly after 8 p.m., one of the two undercovers involved in the buy and bust operation that evening (shield number 22602) approached an individual, who later introduced himself as Robert, and asked him if “he was working.” (Tr. at 183.) 2 Although Robert answered “no,” he added that he “could show [the undercover] where to get something.” (Id.) The undercover and Robert then walked south on 170th Street toward Liberty Avenue and engaged in a brief conversation about the cost and color of the caps of the vials of crack cocaine to be purchased. After about a minute and a half, they arrived at a residential house located at 170-06 Liberty Avenue and Robert went inside. When he returned to the street he handed the undercover two “red cap” vials of crack cocaine. In exchange, the undercover handed Robert a ten-dollar bill from the pre-recorded buy money. The undercover then left the area and headed to his ear. Approximately one minute after making the exchange, the undercover reached his car and radioed the backup team, reported these events, and gave a physical description of Robert. No contemporaneous notes of the description radioed by the undercover exist; however, the undercover did prepare a report upon returning to the precinct. In that *961 report, apparently created between 10:00 p.m. and 12:30 a.m., the undercover recorded the description of the second individual who sold to him that night as “J[ohn] D[oe]Robert-m[ale]/b[lack]/bm’gundy jacket, blue jeans, black/white sneakers, dark-skinned[,] mustache/goatee.” 3 (Ex. A to Decl. of Carol Gette in Support of Pet.) Another report, called a DD5, prepared on January 10, states that the description was “M[ale]/B[lack], Red Jacket, bl jeans, blk & wt sneakers, darked skined (sic).” (Ex. C to Decl. of Carol Gette in Support of Pet.) Upon receiving this information, the backup team proceeded to the location where the sale had occurred, but did not see anyone matching the description. After a search of the surrounding area, which lasted approximately two minutes, two members of the backup team, Police Officer Steven Curasi (“Curasi”) and Sergeant James McCool (“McCool”), observed an individual who they believed matched the description at the corner of 170th Street and 93rd Avenue, approximately two blocks from the site of the sale. The individual the officers observed was Dey. The officers placed Dey under arrest at approximately 8:10 p.m. and conducted a quick pat and frisk search before handcuffing him. During the course of this search, Curasi recovered one glass crack pipe, but nothing else — no pre-recorded buy money and no drugs. A police van was then summoned to the arrest location and Dey was placed, with his hands handcuffed behind his back, on the floor of the rear left side in the back of the van. Because Dey was not the first individual arrested during the course of the buy and bust operation, there was at least one other person in the van at that time. However, there was nothing on the floor where Dey was placed. In addition, the van was searched before the evening’s operation began and was found to be empty. After Dey was placed in the van, the buy and bust operation continued at several other locations. By the time it ended and the van, the undercovers, and the backup team returned to the 113th precinct (“precinct”) at 10:00 p.m., seven individuals, including Dey, had been arrested and placed in the van. 4 Although the arrestees could move around in the back of the van, when they arrived at the precinct Dey was seated in the same area in which he was initially placed. As Curasi lifted him in order to help remove him from the van he noticed seven glassine envelopes directly under Dey. It was later discovered that these glassine envelopes contained heroin. Upon entering the precinct, Dey and the other arrestees were placed in a holding pen. At 12:30 a.m., Dey was removed from the holding pen for a confirmatory identification by the undercover. Confirmatory identifications proceeded in the following way: the undercover officer stood with the arresting officer on one side of a doorway that contained a viewing glass; the individuals who were arrested as a result of that undercover’s buys were then brought before the viewing glass, individually, in the order of their arrest; the undercover then simply confirmed or denied whether the individual was the person from whom he had bought drugs at a particular place and time. Accordingly, undercover 22602, whose buys led to two arrests that evening, stood with Curasi on one side of the doorway while each of the arrestees who had allegedly sold the undercover drugs were placed before him for identification. As the second person arrested as a result of the undercover’s buys, Dey was the second person placed before him for identification. The undercover identified him as the individual who had sold him crack cocaine in front of 170-06 Liberty Avenue. *962 B. The State Court Proceedings 1. The Hearing At a “Mapp-Wade” hearing held on September 28, 1988, Curasi, the only witness called, testified about the circumstances leading to Dey’s arrest and the manner in which the glassine envelopes of heroin were discovered. He testified that the description radioed by the undercover was “a male black with a goatee, approximately 5’10, 5’11, red jacket, blue jeans and black and white sneakers.” (M-W Hr’g at 4.) 5 He stated that the individual he and McCool arrested matched “the exact description that was radioed to us,” and identified Dey, in-court, as the individual they had arrested. (Id.) He added that at the confirmatory identification, the undercover pointed to Dey and said “that’s him, you know, wearing whatever,” and he explained what transpired at the location. (M-W Hr’g at 24.) The court concluded that Curasi’s testimony did not “indicate any departure from good or accepted police procedure or any violation of the defendant’s Constitutional rights,” and that the “identification procedure and testimony were legal and valid.” (M-W Hr’g at 28.) The court also denied a motion to suppress the seized evidence. 2. The Trial At trial, the primary issue was the identification of Dey — specifically, whether the backup team had arrested the right individual based on the physical description radioed by the undercover. The undercover made an in-court identification of Dey as the “Robert” who had sold him drugs in front of 170-06 Liberty Avenue, 6 and testified that the description he had radioed was a “male black, dark skin, he had a mustache and goatee, he had on a burg[u]ndy jacket, blue jeans and black and white sneakers.” (Tr. at 188, 209.) He testified that at the confirmatory identification Dey “looked the same way.” (Tr. at 190.) On redirect, he elaborated that he based his identification at the precinct on the individual’s “appearance,” meaning, “[t]he way [he] saw him at the time of the buy. The description of what he had on. His facial hair. His skin color. The color of his clothing.” (Tr. at 217-218.) However, he admitted that he did not write down the description of the individual from whom he bought drugs until after he returned to the precinct that night, and thus, the recorded description was based on his memory. Curasi’s testimony at trial mirrored his testimony at the MappfWade hearing. He stated that the description was a “male black, approximately five-ten, red jacket, blue jeans, black and white sneakers, beard, mustache,” and that the individual he and McCool arrested “was, in fact, wearing those articles of clothing and did match the general physical description given by the undercover officer.” (Tr. at 87-88.) He identified Dey, in court, as that individual. He also stated that Dey at the confirmatory identification “was dressed in the same article of clothing that he was arrested in,” a “red burgundy type jacket.” (Tr. at 137.) McCool testified that the description was: “Male black, dark skinned, mustache and goatee with a burg[u]ndy jacket, and, I believe, jeans.” (Tr. at 37.) He stated that the individual he and Curasi arrested “matched the description” and, specifically, that the individual was wearing a burgundy jacket. (Tr. at 38.) He also identified Dey, in court, as the individual he and Curasi arrested. He stated that when the undercover performed the confirmatory identification, Dey was wearing a “burg[u]ndy jacket the whole time.” (Tr. at 69.) Karen Connyer (“Connyer”), Dey’s wife, provided a description of the clothing Dey was wearing on the night of his arrest that contradicted the police description in every respect. She testified that he was wearing “a blue hooded jacket that’s called a bench warmer and it comes down past the hips. It had an attached hood. He was wearing another blue hooded sweatshirt under that, a pair of maroon or wine colored corduroy *963 pants and green gym shoes.” (Tr. at 222.) She testified that she and Dey were walking home that evening from a friend’s house on 171st Street, with their dog, when a passing vehicle backfired and the dog broke loose. She stated that she gave chase to the dog, but did not catch up with it until she was almost at home. According to Connyer, she did not have any contact with her husband until shortly after midnight when he called and told her that he had been arrested. On three separate occasions, both before and after Connyer’s testimony, the defense attempted to move Dey’s official arrest photograph into evidence. This photograph, dated January 8,1988, two days after the arrest, depicts Dey wearing a blue hooded jacket and blue hooded sweatshirt, and thus is corroborative of Connyer’s testimony. It also shows Dey to be a black male with a goatee and a large belly. Despite Dey’s considerable girth, the jacket in the picture fits him quite well. The court refused to admit the arrest photograph, stating that “to offer a photograph taken two days after an arrest to tiy to prove the point of what the defendant allegedly was wearing would be an unfair inference for the jury to conclude and draw.” (Tr. at 30.) The court later elaborated on its ruling as follows: [A]s I stated to you earlier when I ruled on this subject matter, I said that so long as you have that kind of interval of time from the time of his arrest to the time of this photograph, two days, or even if you wanted to call it only one day or one-and-a-half days, the fact of the matter is a great many events may occur. Visits, change of clothing, any number of possibilities. (Tr. at 102.) Despite the court’s refusal to admit the arrest photograph, the issue of the identification remained hotly contested. In fact, the first issue addressed by the Assistant District Attorney (“ADA”) in his closing was the veracity of Connyer’s testimony about the clothing Dey wore on the night he was arrested: “The truth or the accuracy of that story that she had given you must be viewed in light of the fact that she is, in fact, an interested witness.” (Tr. at 260.) In order to reinforce this point, the ADA requested that the court include an interested witness charge as part of its instructions. That request was granted. The jury was charged on the morning of March 8, 1989. At 4:30 p.m. that afternoon, the jury submitted a note stating that it was “hopelessly deadlocked” on Counts One (sale of crack cocaine) and Two (possession of heroin). (Tr. at 320.) In response, the court gave the jury a modified Allen charge, encouraging it to continue deliberating and stating that “if it’s at all possible, it’s desirable that a jury should reach a verdict one way or the other.” (Id.) The jury returned a verdict the next day, convicting Dey on each of the three counts in the indictment. The court sentenced Dey to concurrent terms of four and a half to nine years on the criminal sale count and to one year on each of the criminal possession counts. 3. The Appeals Dey’s conviction was subsequently affirmed by the Appellate Division. People v. Dey, 184 A.D.2d 779, 585 N.Y.S.2d 485 (2d Dep’t 1992). That court briefly addressed each of the issues presented in the instant petition, concluding that (1) it was not error to refuse to admit the arrest photograph into evidence to establish that Dey had been the victim of mistaken identification because “[Dey] was confined to a holding cell with numerous other prisoners,” and thus, the photograph was “inherently unreliable as proof of his appearance at the time of the drug transaction”; and (2) viewing the evidence in the light most favorable to the prosecution, it was legally sufficient to establish Dey’s guilt on the heroin possession count beyond a reasonable doubt. Thereafter, the New York Court of Appeals denied Dey’s application for leave to appeal. People v. Connyer, 80 N.Y.2d 928, 603 N.E.2d 960, 589 N.Y.S.2d 855 (1992); People v. Dey, 80 N.Y.2d 928, 603 N.E.2d 961, 589 N.Y.S.2d 856 (1992). C. The Habeas Petition and Hearing Under 28 U.S.C. § 2254(d), federal courts conducting habeas proceedings must *964 give a “presumption of correctness” to “a determination after a hearing on the merits of a factual issue, made by a State court of competent jurisdiction,” unless the conditions for one of seven listed exceptions are met or unless the state court findings are not “fairly supported” by the record as a whole. See Ventura v. Meachum, 957 F.2d 1048, 1054 (2d Cir.1992); Campaneria v. Reid, 891 F.2d 1014, 1019 (2d Cir.1989). “Deference need not be given when ‘the material facts were not adequately developed at the State court hearing,’ ” Pagan v. Keane, 984 F.2d 61, 64 (2d Cir.1993) (quoting 28 U.S.C. § 2254(d)(3)), or when “the factfinding procedure employed by the State court was not adequate to afford a full and fair hearing.” 28 U.S.C. § 2254(d)(2). Nor does § 2254(d) require deference to state court findings “on mixed issues of fact and law such as whether a plea agreement was entered into voluntarily ... or whether there has been ineffective assistance of counsel.” Ventura, 957 F.2d at 1055; see also Tyson v. Trigg, 883 F.Supp. 1213, 1218 (S.D.Ind.1994) (petitioner is correct in his contention that the “decision to exclude [the proffered] evidence raises a mixed question of law and fact and that no deference is due the state courts’ conclusions on this matter.”), aff'd, 50 F.3d 436 (7th Cir.1995), cert. denied, — U.S. —, 116 S.Ct. 697, 133 L.Ed.2d 655 (1996). 7 Although at one time “district courts were required to hold evidentiary hearings in habeas cases whenever ‘for any reason not attributable to the inexcusable neglect of petitioner, evidence crucial to the adequate consideration of the constitutional claim was not developed at the state hearing,’” Pagan, 984 F.2d at 63-64 (quoting Townsend v. Sain, 372 U.S. 293, 317, 83 S.Ct. 745, 759, 9 L.Ed.2d 770 (1963)), a hearing is now required only when “the petitioner can establish cause for his failure to develop an adequate factual record below and prejudice resulting from that failure; it no longer suffices to show merely that he did not deliberately bypass the opportunity to present facts to the state forum.” Pagan, 984 F.2d at 64; see Keeney v. Tamayo-Reyes, 504 U.S. 1, 112 S.Ct. 1715, 118 L.Ed.2d 318 (1992). Despite this recent limitation on the situations in which evidentiary hearings are mandatory, federal “ ‘district courts ... still possess the discretion ... to hold hearings even where they are not mandatory.’ ” Pagan, 984 F.2d at 64 (quoting Keeney, 504 U.S. at 23, 112 S.Ct. at 1727). “Factors relevant to the District Court’s discretionary determination include the existence of a factual dispute, ... the strength of the proffered evidence, ... the thoroughness of the prior proceedings ... and the nature of the state court determination.” Id. Since the state court failed to hold any evidentiary hearing regarding the arrest photograph, and instead, made a conclusory determination based solely on the length of time it was taken after Dey’s arrest, the Court was not required to defer to the state court’s finding. Accordingly, in the exercise of its discretion, the Court ordered an evidentiary hearing in order to elucidate the facts surrounding the arrest photograph and the processing of Dey after his arrest. 1. Typical Booking Procedures Many of the specific facts of Dey’s processing have been lost in the course of time due to the disappearance of documentary evidence and the limits of human memory. Nonetheless, as a result of the hearing, the Court is able to discern the likely booking process, which concluded with the taking of the arrest photograph and Dey’s arraignment on January 8,1988. *965 a. The 113th Precinct Curasi’s testimony revealed that the following procedures were in effect at the precinct at the time of Dey’s arrest. After an individual was arrested, he would be brought to the precinct, presented before the precinct desk to be “logged into” the “command log,” and fingerprinted. (Hr’g at 71.) 8 An arrest report would be completed and a Polaroid photograph, which would not be the official arrest photograph, may or may not be taken during this initial processing. 9 The arrestee would then be moved to another area where he would be extensively searched. On many occasions, this search would consist of a full “strip” search. After the search was completed, the arrestee would put whatever clothing that was removed during the search back on and be placed into the precinct’s holding cell. There is some uncertainty concerning the exact number of arrestees that could be held in the holding cell. According to Curasi, it held as many as ten to twenty arrestees at one time. McCool testified that it could hold as many as twenty-five to thirty arrestees. More importantly, McCool revealed that the precinct holding cell was not one large area, but rather, consisted of several individual cells, each of which held no more than three arrestees: [MR. DUMBACH]: Do you recall how many persons the holding cell in the 113th Precinct could hold in January of 1988? [McCOOL]: Depending if you put more than one person in an individual cell. [MR. DUMBACH]: Yes. Let’s assume that’s the case. [McCOOL]: If you could put three people in there, I would say probably 25 to 30. (Hr’g at 106.) Sometimes the arrestees were handcuffed while in the cells and other times they were' not. Contrary to the trial court’s conjectures, however, arrestees were not allowed to receive visitors who might bring a change of clothing into the holding cells. Nor were they allowed to shower or clean up or even wash. Moreover, throughout this initial processing, the arrestees were under constant observation by police officers: THE COURT: In the holding cell these people are under observation[ ] all the time? [CURASI]: Yes. (Hr’g at 64.) The arresting officer would remain at the station house with his arrestee until the initial processing was complete and the arrestee was removed, in handcuffs, to Queens Central Booking (“Central Booking”). Typically, this transfer would occur within three hours after the initial arrest. It is undisputed, however, that Dey was at the precinct until at least 12:30 a.m., the time he was brought before the undercover for the confirmatory identification, and probably did not arrive at Central Booking until approximately 3:30 a.m. See infra Part I.C.2. b. Queens Central Booking Sergeant Michael Goff (“Goff’), who was assigned to the Queens Court section in January of 1988, described the processing of new arrestees at Central Booking. An arrestee *966 would arrive with a handwritten “on-line” booking sheet and completed fingerprint cards. Upon entering Central Booking, two one by one inch Polaroid pictures would immediately be taken. One Polaroid was attached to the “movement slip” — a document that listed the arresting officer, the arrestee, and the charge — which stayed with the arrest ee wherever he was moved so that officers were aware of his identity. The other was attached to the arrest package, which was sent to the District Attorney’s Office to be placed in its file once the criminal history sheet (“rap sheet”) was returned. When an arrestee was brought to court for arraignment, the movement slip was brought with him, but was destroyed if he was released. If he was remanded to the Department of Corrections (“Corrections”), it was transferred with his property to Corrections, where it was used until Corrections generated its own paperwork. At that point, the movement slip and attached Polaroid were destroyed. Thus, the Polaroid affixed to the movement slip was almost certainly destroyed shortly after Dey’s arrest. Although the Polaroid that was part of the arrest package should still be in existence, the District Attorney’s Office has represented that it is unable to locate its file. After an arrestee was photographed, he was put into a temporary holding cell in the front of Central Booking while the police officer generated the arrestee’s arrest number and dropped off the arrestee’s fingerprint cards. As opposed to the temporary holding cells, the main holding cells were located in the back of Central Booking. If these were empty, the arrestee would eventually be moved from the temporary holding cells to the main holding cells, where he would wait for the official arrest photograph to be taken, and his arraignment. However, if the main holding cells were full, the arrestee would remain temporarily in the front cell until he and other arrestees from his precinct could be returned to the precinct, where they would await the completion of their paperwork. Once the paperwork for a majority of the arrestees returned to a given precinct was completed, they would be returned to Central Booking for their official arrest photograph and arraignment. 10 Because Dey’s name cannot be found on the prisoner rosters of the male cells at Central Booking between January 6th and January 8th, it is likely that he did not remain at Central Booking after his initial transfer from the precinct, and that he was returned to the precinct while his paperwork was being completed. 11 The arrest photograph was taken by an official NYPD photographer in an office in the rear of Central Booking. In the photograph, an arrestee “usually ha[s] the clothing [he] w[as] arrested in.” (Hr’g at 50.) According to Walter Taylor, an NYPD Senior Photographer, “[everybody comes in street clothes. They don’t — usually they don’t allow people to change clothes when they get arrested. They don’t bring in clothes or allow them to bring in other garments.” (Hr’g at 51.) Taylor stated that “[n]ormally, from my experience working in the mug room, [they wear] the same clothes that they are arrested in. They don’t change clothes down there.” (Id.) Goff explained the effort that is made to take the picture of the arrestee in the clothing he wore at the time he was brought into the precinct: That little Polaroid picture that was taken when you are brought in is on the movement slip. When they are photographing *967 and removing you to court the movement slip stays with you. They take them and they will call the prisoner’s name out of the cell and direct him into the photo room. They look at the picture and they’ll say to them — at least they’re supposed to — they will say that’s not what you were wearing when you came in, where is your coat or you had a hat on or you had whatever, where is it? If they have it and it is back in the cell they tell them to go get it. (Hr’g at 121.) Such care is taken because the arrest photograph serves to memorialize the appearance of the arrestee at the time he entered the precinct, including his clothing: THE COURT: The process is to try as best as possible to have the photographs, be it the Polaroid or official photograph, represent what the prisoner was wearing at the time he was brought into the precinct? [GOFF]: Yes, sir. THE COURT: Do you have some system where you try to see that that happens consistently? [GOFF]: The escort officer who is assigned to bring them out and photograph them is supposed to review the Polaroid picture to see as much as he can see on the Polaroid picture that that is what the individual was wearing when he goes in. THE COURT: If the individual is not dressed the same, what happens then? [GOFF]: They would question him as to where those clothes went, if he had taken the jacket off or sweatshirt. Sometimes they’ve taken a coat off and left it in the cell. THE COURT: In that situation do they direct the prisoner to put on his original clothing? [GOFF]: Yes. THE COURT: The reason for that being what? [GOFF]: Because they want to take the picture of him the way he was brought in. In the case of a robbery suspect they even take a long photo of him as the official photo so they have his complete attire. THE COURT: They consider it important to try to have the picture represent the actual appearance of a prisoner as he’s brought in to be booked, so to speak? [GOFF]: Yes. (Hr’g at 121-22.) The integrity of the arrest photograph was confirmed by NYPD Police Officer Richard Barber (“Barber”), who has been assigned to Central Booking for the past ten years and has been a cell attendant in the facility: [BARBER]: When you call them out to have a photo taken, when they come out sometimes you have to tell them bring what you had with you and they will go back inside. They will put on clothes they had taken off. And at least several occasions, especially during the colder months, when I take the photo, because I have to ID when he comes out to make sure it’s the right person, I saw a guy that stood out with a park[a] and actually he’s coming out and has a light jacket on. I say where’s your jacket and he says I swapped it with another guy. THE COURT: It happens? [BARBER]: It happens. THE COURT: You take the picture of the person with different clothing on? [BARBER]: No. I — if I pick it out I’ll make them change it back. THE COURT: You really want to try to get that person photographed in the clothing in which he walked into Central Booking? [BARBER]: Correct. *968 THE COURT: That’s what the policy and practice is, exactly the way he came in? [BARBER]: Yes. THE COURT: There is a good and sufficient reason for that, I take it, or is that— [BARBER]: For identifying purposes. THE COURT: You consider that it is important because there could be issues of identification that may have some relevancy and it is good police practice to see that that happens, correct? [BARBER]: Correct. (Hr’g at 143^44.) 2. The Timing of Dey’s Transfers and Arrest Photo As previously mentioned, Dey was arrested at 8:10 p.m. on January 6, brought to the precinct at 10:00 p.m., and identified by the undercover at 12:30 a.m. The exact time of his arrival at Central Booking cannot be determined. However, it is undisputed that the police at Central Booking received his initial arrest paperwork and created his arrest number in the police “on-line” booking system at 3:39 a.m. Because arrestees arrive at Central Booking with their paperwork in “95 percent of the cases,” it is most likely that Dey arrived there at about that time. (Hr’g at 130.) 12 As Goff stated: “If the body and the paperwork arrive in Central Booking at the same time, we take it. If [the precinct] send[s] the paperwork down without the body because he is being debriefed or held for a lineup or he’s going to the hospital, they are required ... to take the picture and send it with the paperwork ... so there’s always a picture on these forms when they arrive.” (Hr’g at 134.) 13 Since the police procedures should ensure that Dey’s appearance in the arrest photograph, taken on January 8th, mirrors his appearance upon entering Central Booking, his appearance in the arrest photograph can be traced back to 3:39 a.m on January 7th. Based on the probable time the Polaroid was taken at Central Booking (3:39 a.m.), the time of the showup at the precinct (12:30 a.m.), and the drive time from the precinct to Central Booking (a half-hour), it appears Dey had a three hour window of opportunity during which he could have changed his coat while in the holding cell at the precinct if he was, in fact, wearing a red or burgundy coat at the time he was identified by the undercover. 3. Connyer’s Testimony At the habeas hearing, Connyer testified consistently with her trial testimony. She stated that Dey, on the night of his arrest, “was wearing green Puma gym shoes, ankle length, maroon corduroy pants and a blue three-quarter jacket with a hood.” (Hr’g at 19.) She characterized the shade of blue as “between a royal bright blue and a powder blue,” and related that they had purchased the jacket at “Robbins” on 14th Street in 1987. (Hr’g at 20-21.) She further stated that she remembered he was wearing a blue jacket that night because “[i]t was the only one that covered his body that much.” (Hr’g at 20.) She explained that: He had two other jackets that were summer, spring-like jackets. The plaid two-tone ones, blue and black, and the other one was blue and it has the pockets in the front. This one was a three-quarter coat, it zippered up- the front and snapped. It may or may not have had a string on the bottom to tighten, but I’m not sure, and the hood was an attached hood and the snaps also closed up the front and tied. (Hr’g at 20.) Many of the details of her testimony at trial and at the habeas hearing *969 are corroborated by the arrest photograph which, she testified, she never saw. II. DISCUSSION A. The Failure to Admit the Arrest Photograph Dey’s petition requires the Court to determine whether the trial court violated a constitutional right of Dey by excluding his arrest photograph from evidence. In making this determination the Court engages in a two part analysis, examining 1) whether the exclusion was error under state law, and 2) whether the error amounted to the denial of the constitutional right to a fundamentally fair trial. 1. Was it error to exclude the arrest photograph? Under New York law, “[ejvidence is relevant if it has ‘any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.’ ” People v. Fagan, 215 A.D.2d 686, 628 N.Y.S.2d 118, 119 (2d Dep’t 1995) (quoting People v. Davis, 43 N.Y.2d 17, 27, 371 N.E.2d 456, 400 N.Y.S.2d 735 (1977)); People v. Lewis, 69 N.Y.2d 321, 325, 506 N.E.2d 915, 916, 514 N.Y.S.2d 205, 207 (1987) (“Evidence is relevant if it has any tendency in reason to prove any material fact.”) (citations and internal quotations omitted). “All relevant evidence is admissible unless its admission violates some exclusionary rule.” Lewis, 69 N.Y.2d at 325, 506 N.E.2d at 916, 514 N.Y.S.2d at 207; People v. Buie, 86 N.Y.2d 501, 515, 658 N.E.2d 192, 200, 634 N.Y.S.2d 415, 423 (1995) (Simons, J., concurring) (“Generally, all relevant, i.e., probative, evidence is admissible unless there is some legal reason to exclude it[.]”). “[R]ulings on relevance and admissibility rest within the trial court’s discretion.” People v. Barnes, 109 A.D.2d 179, 184, 491 N.Y.S.2d 864, 868 (4th Dep’t 1985). The state trial court refused to admit Dey’s arrest photograph into evidence because, presumably, it was not taken at or about the time of the arrest. Based on the evidence adduced at the habeas hearing and New York easelaw on the relevance and admissibility of arrest photographs, this Court concludes that the trial court abused its discretion and committed error. See People v. Husband, 135 A.D.2d 406, 407, 522 N.Y.S.2d 132, 133 (1st Dep’t 1987) (trial court committed reversible error by refusing to permit defense witness to give testimony clearly relevant to the central issue of the trial); People v. Figueroa, 80 A.D.2d 520, 522, 436 N.Y.S.2d 1, 3 (1st Dep’t 1981) (conviction for unlawful possession of a gun reversed where defendant claimed he had obtained gun innocently when he helped to break up a fight and trial court excluded testimony of defense witness that she was with defendant “until a short time” before the quarrel and she saw no signs that he had a gun). Although the arrest photograph is not conclusive on the issue of Dey’s clothing on the night in question, because it was taken on January 8th and not January 6th, it is certainly relevant in light of the testimony adduced at the habeas hearing detailing the care taken to ensure that an arrest photograph is an accurate representation of a defendant’s appearance at the time of his arrest. Dey’s opportunity, such as it was, to exchange clothing with a fellow prisoner does not undermine the relevance of the arrest photograph or render it inherently unreliable. See Barnes, 109 A.D.2d at 183, 491 N.Y.S.2d at 868 (in order for evidence to be relevant, “[a]ll that is necessary is ‘that it tend to convince that the fact sought to be established is so [or not so]. That it is equivocal * * * does not render it inadmissible.’ ”) (citations omitted). Indeed, since Dey was handcuffed behind his back when he was being transported, and since it is highly unlikely that he changed clothing at Cental Booking (because of the Polaroid .typically taken upon arrival), his only realistic opportunity to change clothing was at the 113th precinct during the three horn* period between the confirmatory identification and Dey’s transport to Central Booking. For that to have happened, the following scenario would have had to unfold: 1) one or two other arrestees were placed in the individual, three-person cell with Dey; 2) one of them was of considerable girth; 3) he was wearing a blue jacket large enough to fit Dey; 4) he *970 was willing to exchange the jacket with Dey; and 5) the exchange was made without arousing the suspicion of any of the guards. This unlikely scenario reinforces the probative value of the arrest photograph. 14 In summarily dismissing the relevance and reliability of Dey’s arrest photograph, both the trial court and the Appellate Division ignored the numerous cases in which arrest photographs have been admitted and deemed relevant on the issue of identification. Arrest photographs have consistently been admitted after cross-examination in order to corroborate a witness’ identification testimony. See, e.g., People v. Walker, 216 A.D.2d 425, 628 N.Y.S.2d 951, 952 (2d Dep’t 1995) (holding that arrest photograph which “confirmed the complainant’s testimony that the robber was wearing a dark grey jacket at the time of attack,” was admissible), appeal dismissed, 87 N.Y.2d 926, 664 N.E.2d 519, 641 N.Y.S.2d 608 (1996); People v. Rios, 156 A.D.2d 397, 548 N.Y.S.2d 348, 349 (2d Dep’t 1989) (holding that arrest photograph, showing the defendant wearing a red sweatshirt which the complainant had described, was properly admitted after defense counsel “opened the door” during his cross examination of the complaint on the issue of the defendant’s attire, and noting that “[i]t is well established that arrest photographs may be admitted to establish a defendant’s appearance at the time of the crime”), appeal denied, 75 N.Y.2d 923, 554 N.E.2d 79, 555 N.Y.S.2d 42 (1990). Moreover, they have also been admitted as evidence-in-chief. People v. Jones, 187 A.D.2d 730, 731, 590 N.Y.S.2d 296, 296 (2d Dep’t 1992) (holding that arrest photograph which, according to the undercover officer, depicted the defendant wearing the same red hooded jacket that he wore when he sold the drugs to the officer, was admissible “to establish the defendant’s appearance at the time of arrest”), appeal denied, 81 N.Y.2d 790, 610 N.E.2d 410, 594 N.Y.S.2d 737 (1993); People v. Santana, 162 A.D.2d 191, 192, 556 N.Y.S.2d 316 (1st Dep’t 1990) (holding that arrest photograph which, according to the undercover officer, depicted the defendant wearing the same sheepskin coat he wore when he sold him the drugs, did not constitute improper bolstering because “photographs are admissible to establish a defendant’s appearance on the date of the crime”); People v. Peters, 135 A.D.2d 841, 522 N.Y.S.2d 944, 945 (2d Dep’t 1987) (holding that arrest photograph, which was introduced into evidence after the arresting officer testified that it was a fair and accurate representation of the defendant as he appeared at the time of his arrest, was properly admitted to establish the defendant’s appearance on the date of the crime). 15 2. Was the error of a constitutional dimension? The Court’s conclusion that the trial court erred does not end the inquiry. The Court must also determine that the error was of constitutional dimension because “federal habeas corpus review of state erimi *971 nal proceedings under 28 U.S.C. § 2254 is limited to errors of constitutional magnitude.” Alvarez v. Scully, 838 F.Supp. 1000, 1005 (S.D.N.Y.1993); see also Ciak v. United States, 59 F.3d 296, 301 (2d Cir.1995) (“[P]risoners seeking habeas relief must prove that constitutional violations occurred at trial[.]”); Brown v. Doe, 2 F.3d 1236, 1249 (2d Cir.1993) (“The habeas remedy is not ours to grant unless a petitioner has established a constitutional violation[.]”). An “erroneous evidentiary ruling[ ] do[es] not automatically rise to the level of constitutional error.” Rosario v. Kuhlman, 839 F.2d 918, 923 (2d Cir.1988); Taylor v. Curry, 708 F.2d 886, 891 (2d Cir.1983). Thus, the court will grant habeas relief as a result of such an error “only where the petitioner can show that [it] deprived [him] of a fundamentally fair trial.” Taylor, 708 F.2d at 891; see Alvarez, 833 F.Supp. at 1005 (“Discretionary state court evidentiary rulings normally do not rise to a constitutional level so as to be cognizable in a federal habeas corpus proceeding absent a showing that ... the error so infected the proceedings as to have rendered a petitioner’s trial fundamentally unfair.”). The concept of fundamental fairness has been described as “elusive.” Collins v. Scully, 755 F.2d 16, 18 (2d Cir.1985) (quoting Taylor, 708 F.2d at 891). In making the determination of whether a defendant’s right to a fundamentally fair trial has been violated, courts have been directed to examine “the materiality of the excluded evidence.” Rosario, 839 F.2d at 925. Such inquiry requires courts to evaluate “what the impact of the excluded evidence would have been if it had been admitted,” Collins, 755 F.2d at 18, and more precisely, whether it “could have created a reasonable doubt that did not otherwise exist[.]” Rosario, 839 F.2d at 927. 16 As the Supreme Court has explained: The proper standard of materiality must reflect our overriding concern with the justice of the finding' of guilt. Such a finding is permissible only if supported by evidence establishing guilt beyond reasonable doubt. It necessarily follows that if the omitted evidence creates a reasonable doubt that did not otherwise exist, constitutional error has been committed. This means that the omission must be evaluated in the context of the entire record. If there is no reasonable doubt about guilt whether or not the additional evidence is considered, there is no justification for a new trial. ■ On the other hand, if the verdict is already of questionable validity, additional evidence of relatively minor importance might be sufficient to create a reasonable doubt. United States v. Agurs, 427 U.S. 97, 112-13, 96 S.Ct. 2392, 2401-02, 49 L.Ed.2d 342 (1976) (footnotes omitted). 17 Despite respondent’s contention that the evidence against Dey was “overwhelming,” this Court finds that it was not. Rather, it was almost entirely based on a brief description and subsequent confirmatory identification by a single undercover officer. 18 Al *972 though both Curasi and McCool testified that Dey was the individual they arrested in response to the undercover’s description, and that he matched the description radioed by the undercover, neither officer actually observed Dey selling any drugs or engaging in any suspicious behavior. Moreover, Dey neither possessed any pre-recorded buy money, nor any vials of crack cocaine, the drug sold to the undercover. In short, there was no tangible evidence against him on the charge of criminal sale of crack cocaine. In evaluating the excluded evidence in context, it should be noted that the arrest photograph would not have been merely cumulative of Connyer’s testimony. As Dey’s wife, Connyer was marked as an interested witness in respondent’s closing and the court’s charge, and thus, her testimony was significantly undermined. The jury should have been allowed to examine the arrest photograph not only because it corroborated Connyer’s detailed testimony regarding the blue jacket and blue sweatshirt, but also because it showed that the jacket fit. After considering the arrest photograph, the jury may well have concluded that Connyer had truthfully testified that Dey was wearing maroon pants and a blue jacket and that, perhaps, the officers had inadvertently inverted the colors of these garments. If the arrest photograph had been admitted, it would have directly and strongly supported the possibility that Dey was misidentified. See Lyons, 99 F.3d at 504 (“Evidence which directly and strongly supports the possibility that a criminal defendant was misidentified certainly raises the specter of a reasonable doubt.”). The Second Circuit recently noted that “eyewitness testimony is often highly inaccurate” and that the “issue of misidentification is absolutely fundamental to a criminal trial.” Lyons v. Johnson, 99 F.3d 499, 504 (2d Cir.1996). In affirming a district court’s grant of habeas in a case that “seem[ed] solidly built upon the identification of three eyewitnesses,” the court stated that: There can be no reasonable doubt that inaccurate eyewitness testimony may be one of the most prejudicial features of a criminal trial____ Because the intrinsic unreliability of eyewitness identifications is so often compounded with the distorting effects of the natural suggestion to the witness that the person on trial is the guilty one, it is no exaggeration to venture that the influence of improper suggestion upon identifying witnesses probably accounts for more miscarriages of justice than any other single factor — perhaps it is responsible for more such errors than all other factors combined. Id. at 504 (citations and internal quotations omitted). Accordingly, even a case built on several eyewitness identifications may warrant habeas relief if evidence undermining confidence in those identifications was improperly excluded. Based on its review of the entire state record and the evidence at the habeas hearing, the Court concludes that the arrest photograph, if received in evidence, could have created a reasonable doubt that did not otherwise exist, and thus, Dey was deprived of a fundamentally fair trial. This determination is reinforced by the fact that the jury had been “hopelessly deadlocked” and only reached a verdict after receiving an Allen-type charge. See Mason v. Scully, 16 F.3d 38, 45 (2d Cir.1994) (noting that jury returned guilty verdict only after receiving an Allen charge, in support of its finding that error warranted grant of habeas writ). 3. Should the Court apply harmless error analysis? In Chapman v. California, 386 U.S. 18, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967), the *973 Supreme Court rejected the argument that the Constitution requires a blanket rule of automatic reversal in the case of constitutional error, concluding instead “that there may be some constitutional errors which in the setting of a particular case are so unimportant and insignificant that they may, consistent with the Federal Constitution, be deemed harmless.” Id. at 22, 87 S.Ct. at 827. Accordingly, the Court held that the appropriate question for determining whether an error is harmless is “whether the error ‘contributed[d] to the verdict,’” Peck v. United States, 102 F.3d 1319, 1320 (2d Cir.1996) (Newman, C.J., concurring in dissolving in banc court) (quoting Chapman, 386 U.S. at 24, 87 S.Ct. at 828), and stated, with respect to the degree of certainty required, “that before a federal constitutional error can be held harmless, the court must be able to declare a belief that it was harmless beyond a reasonable doubt.” Chapman, 386 U.S. at 24, 87 S.Ct. at 828. Subsequently, however, the Court clarified in Brecht v. Abrahamson, 507 U.S. 619, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993), that the Chapman standard should be applied only on direet review and that a more deferential standard, whether the error “had a substantial or injurious effect or influence in determining the jury’s verdict,” was the appropriate standard for habeas review. Brecht, 507 U.S. at 637, 113 S.Ct. at 1722. The Court further refined the harmless error analysis for habeas cases in O’Neal v. McAninch, 513 U.S. 432, 115 S.Ct. 992, 130 L.Ed.2d 947 (1995), adding, with respect to the degree of certainty required: “When a federal judge in a habeas proceeding is in grave doubt about whether a trial error of federal law had ‘substantial and injurious effect or influence in determining the jury’s verdict,’ that error is not harmless. And, the petitioner must win.” Id. at -, 115 S.Ct. at 994. 19 According to the Supreme Court, whether harmless error analysis should be applied depends on the type of constitutional error that occurred in the trial court. In Arizona v. Fulminante, 499 U.S. 279, 306-12, 111 S.Ct. 1246, 1262-66, 113 L.Ed.2d 302 (1991), a sharply divided Supreme Court 20 distinguished between two classes of constitutional errors: the vast majority, denominated “trial errors,” which are subject to harmless error review, and a very limited class of errors called “structural,” which require automatic reversal regardless *974 whether the error had any appreciable effect on the outcome of the trial. Yarborough v. Keane, 101 F.3d 894, 896 (2d Cir.1996). Structural errors, such- as the deprivation of the right to counsel, lie at one end of the spectrum of constitutional errors and “require[ ] automatic reversal of the conviction because they infect the entire trial process.” Brecht v. Abrahamson, 507 U.S. 619, 629-30, 113 S.Ct. 1710, 1717, 123 L.Ed.2d 353 (1993); see also Fulminante, 499 U.S. at 309-10, 111 S.Ct. at 1264-65 (listing other structural errors, such as trial before a judge who was not impartial, unlawful exclusion of members of the defendant’s race from a grand jury, the right to self-representation at trial, and the right to public trial). “Errors are properly categorized as structural only if they so fundamentally undermine the fairness or the validity of the trial that they require voiding its result regardless of identifiable prejudice.” Yarborough, 101 F.3d at 897. Accordingly, harmless error analysis is not applied to structural errors. 21 See California v. Roy, — U.S. —, —, 117 S.Ct. 337, 338, 136 L.Ed.2d 266 (1996) (“ ‘[Structural defects in the constitution of the trial mechanism ... defy analysis by ‘harmless-error’ standards])]”’) (quoting Brecht, 507 U.S. at 629, 113 S.Ct. at 1717). Trial error, on the .other hand, such as the prosecutor’s use of a defendant’s post-arrest, post-Miranda silence for impeachment purposes, see, e.g., Brecht, 507 U.S. at 635, 113 S.Ct. at 1720-21, is amenable to harmless-error analysis “because it ‘may ... be quantitatively assessed in the context of other evidence presented in order to determine [the effect it had on the trial].’ ” Brecht, 507 U.S. at 629, 113 S.Ct. at 1717 (quoting Fulminante, 499 U.S. at 280, 111 S.Ct. at 1249); see also Fulminante, 499 U.S. at 306, 111 S.Ct. at 1262-63 (listing constitutional errors to which harmless error analysis can be applied such as use of a confession obtained in violation of Massiah and admission of evidence obtained in violation of the Fourth Amendment). Because the erroneous exclusion of evidence may “be quantitatively assessed in the context of other evidence presented,” it may appear at first blush that the type of error presented by Dey’s petition requires harmless error analysis. However, when the Supreme Court divided constitutional error into “structural” and “trial” error in Fulminante, it was considering in that case error which was intrinsically of constitutional significance — the admission of a coerced confession. It was not considering the type of error presented here, the erroneous exclusion of evidence which only attains constitutional significance when considered in the context of the entire trial in order to ascertain whether it amounted to a deprivation of a fundamentally fair trial. Harmless error analysis is simply inapplicable to error that only attains constitutional significance when considered in the context of the entire trial because such analysis inheres in the initial finding that the error was constitutionally significant. A determination that such error was not harmless, after having already concluded that it denied the defendant a fundamentally fair trial, would be tautological. An analysis of the standards used to determine, on the one hand, whether an erroneous exclusion of evidence was constitutionally significant, and, on the other hand, whether an intrinsically constitutional trial error was harmless, illustrates why harmless *975 error analysis is here unnecessary. Since the fundamental fairness analysis — -whether the evidence creates a reasonable doubt that did not otherwise exist — is more stringent than the harmless error standard — whether the error had a substantial and injurious effect or influence in determining the jury’s verdict — a deprivation of fundamental fairness can never be harmless. See Agurs, 427 U.S. at 111-12, 96 S.Ct. at 2401 (holding that the constitutional standard of materiality, which is used to determine fundamental fairness, must impose a higher burden on the defendant than the Brecht/Kotteakos harmless error standard, which is used on habeas review). 22 As one district court correctly noted: In eases such as this one, where the alleged constitutional error is the deprivation of the fundamental right to a fair trial, the finding of an error that deprives the defendant of his constitutional right to a fair trial cannot by definition be harmless. Accordingly, the court’s broader inquiry into the constitutional dimensions of an erroneous exclusion of evidence subsumes the harmless error analysis and obviates the need for any additional inquiry[.] Flanders v. Meachum, 824 F.Supp. 290, 299 n. 8 (D.Conn.1993), rev’d on other grounds, 13 F.3d 600 (2d Cir.1994). Although the Supreme Court has never addressed the conceptual irrationality of applying harmless error analysis to the erroneous exclusion of evidence that results in the constitutional deprivation of a fair trial, it has addressed this type of issue in the context of exculpatory evidence improperly withheld by a prosecutor. Under United States v. Bagley, 473 U.S. 667, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985), that type of evidence must also be evaluated in the context of the entire trial and will amount to a constitutional violation “only if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.” Bagley, 473 U.S. at 682, 105 S.Ct. at 3383; see supra note 17. Harmless error analysis is inappropriate in this context because, as the Supreme Court recently explained: [Ojnee a reviewing court applying Bagley has found constitutional error there is no need for further harmless-error review. Assuming arguendo that a harmless error inquiry were to apply, a Bagley error could not be treated as harmless, since a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceedings would have been different, necessarily entails the conclusion that the suppression must have had substantial and injurious effect or influence in determining the jury’s verdict. Kyles v. Whitley, 514 U.S. 419, -, 115 S.Ct. 1555, 1566, 131 L.Ed.2d 490 (1995) (internal citations and quotations omitted). See also Gilday v. Callahan, 59 F.3d 257, 268 (1st Cir.1995) (“[T]he Bagley materiality standard necessarily requires a court to find an impact on the jury verdict sufficiently substantial to satisfy the Brecht harmless error test.”); cf. Hill v. Lockhart, 28 F.3d 832, 839 (8th Cir.1994) (holding that harmless error analysis is inapplicable once a determination has been made that the claimed ineffective assistance of counsel was constitutionally significant) (“[I]t is unnecessary to add a separate layer of harmless-error analysis to an evaluation of whether a petitioner in a habeas case has presented a constitutionally significant claim for ineffective assistance of counsel.”). Courts analyzing the type of error in this case — the erroneous exclusion of evidence— have often improperly commingled harmless *976 error analysis with the initial finding of constitutional error. See, e.g., Lyons, 99 F.3d at 502 n. 5 (noting that “[t]he instant case, involving error in evidentiary rulings is clearly ‘trial error’ of the sort governed by the Brecht/Kotteakos [harmless error] standard”); Vargas v. Walker, No. CV-94-3571, 1996 WL 118555, at * 5 n. 1 (E.D.N.Y. Feb. 26, 1996) (noting that “[o]ther judges have applied the ‘harmless error’ standard in determining whether or not an erroneous evidentiary ruling amounts to a constitutional violation,” but concluding that “[i]t makes no difference which standard is applied in the present case because the outcome is the same under both”); Thomas v. Scully, 854 F.Supp. 944, 955 (E.D.N.Y.1994) (applying Brecht harmless error standard in order to determine whether an erroneous evidentiary ruling, warranted issuance of a habeas writ) (“[I]n order for a writ of habeas corpus to issue as a result of an erroneous evidentiary ruling, the petitioner must show that the evidentiary ruling had [a] substantial and injurious effect or influence in determining the jury’s verdict.”) (citations and internal quotations omitted). But see Flanders, 824 F.Supp. at 299 n. 8 (“A court’s determination of whether or not the erroneous exclusion of evidence is an error of constitutional magnitude is not to be confused with the narrower inquiry of whether a constitutional error is harmless[.]”). Other courts, while announcing the proper standard for determining whether an erroneous evidentiary ruling rises to a constitutional error, have nonetheless needlessly applied harmless error analysis. See, e.g., Laboy v. Demskie, 947 F.Supp. 733 (S.D.N.Y.1996) (applying Brecht/Kotteakos harmless error standard after finding that erroneously excluded evidence “would not have created a reasonable doubt that did not otherwise exist”); Perfetto v. Hoke, 898 F.Supp. 105, 116 (E.D.N.Y.1995) (finding that excluded evidence would have created “no reasonable doubt that did not otherwise exist,” and that “[i]or the same reasons, any constitutional error in excluding the reports did not have a ‘substantial and injurious effect or influence in determining the jury’s verdict.’ ”) (citation omitted); Headley v. Tilgham, 860 F.Supp. 956, 962 (D.Conn.1994) (“If a reviewing court finds that an evidentiary ruling deprived the defendant of a fair trial, the court must then determine whether that trial error was harmless.”), rev’d on other grounds, 53 F.3d 472 (2d Cir.1995). For the foregoing reasons, the Court need not engage in a redundant harmless error analysis. Accordingly, the Court’s prior determination that the exclusion of the arrest photograph deprived Dey of a fundamentally fair trial warrants granting De/s petition for a writ of habeas corpus on his conviction for criminal sale of crack cocaine. B. The Insufficiency of the Evidence of Heroin Possession A habeas petitioner raising a sufficiency challenge bears a heavy burden. The petition will be granted only if the federal court finds that “no rational trier of fact could have found proof of guilt beyond a reasonable doubt.” Jackson v. Virginia, 443 U.S. 307, 324, 99 S.Ct. 2781, 2791-92, 61 L.Ed.2d 560 (1979). In making this determination, “all possible inferences that may be drawn from the evidence must be construed in the prosecution’s favor.” Maldonado v. Scully, 86 F.3d 32, 35 (2d Cir.1996). Moreover, “[g]uilt beyond a reasonable doubt may be established entirely by circumstantial evidence, and this evidence must not be reviewed piecemeal, but rather as a whole.” Id. Applying these principles to the present case, the Court concludes that the evidence was sufficient to support petitioner’s conviction for possession of heroin. Section 220.03 of the New York Penal Law provides that a person is guilty of criminal possession of a controlled substance in the seventh degree when he “knowingly and unlawfully possesses a controlled substance.” N.Y. Penal Law § 220.03. Possession is defined as having “physical possession or otherwise to exercise dominion or control over tangible property.” N.Y. Penal Law § 10.00(8). Dominion and control can be established when the controlled substance is within a person’s control and reach, and is therefore available to that person. People v. Lemmons, 40 N.Y.2d 505, 509-10, 354 N.E.2d *977 836, 839, 387 N.Y.S.2d 97, 100 (1976); People v. Lynch, 116 A.D.2d 56, 61, 500 N.Y.S.2d 236, 239 (1st Dep’t 1986) (dominion and control may be inferred from the defendant’s dominion and control over the area where the contraband was found); see also Martinez v. Reynolds, 888 F.Supp. 459, 463 (E.D.N.Y.1995) (“Constructive possession under New York law requires a showing that the defendant ‘exercised dominion or control over the property by a sufficient level of control over the area in which the contraband is found[.]’”) (quoting People v. Manini, 79 N.Y.2d 561, 573, 594 N.E.2d 563, 584 N.Y.S.2d 282 (1992)). Construing the evidence in the light most favorable to the prosecution, the record in this case reveals ample evidence which could “reasonably support a finding of guilt beyond a reasonable doubt.” Jackson, 443 U.S. at 318, 99 S.Ct. at 2788-89. A detailed search of the van was conducted prior to its departure on the evening in question, including the area where the prisoners were detained, and no heroin was found. At the time Curasi placed Dey into the van, there was nothing on the floor in the area where Dey was seated. However, while Curasi was helping Dey out of the van after it arrived at the precinct, he noticed seven glassine envelopes of heroin directly beneath Dey. Although Dey was searched before he first entered the van, that search was a brief street frisk for weapons, and not an extensive search which would necessarily reveal all secreted contraband. The fact that six other prisoners were in the back of the van does not preclude a finding of constructive possession. People v. Torres, 68 N.Y.2d 677, 679, 496 N.E.2d 684, 685, 505 N.Y.S.2d 595, 596 (1986); People v. Diaz, 41 A.D.2d 382, 384, 343 N.Y.S.2d 474, 477 (1st Dep’t 1973), aff'd, 34 N.Y.2d 689, 312 N.E.2d 478, 356 N.Y.S.2d 295 (1974). Accordingly, the jury could reasonably conclude that Dey constructively possessed the heroin found directly beneath him. III. CONCLUSION The Court concludes that Dey’s petition should be granted with respect to his conviction for sale of crack cocaine and denied with respect to his conviction for possession of heroin. Accordingly, the case is remanded for a new trial on the charge of criminal sale of a controlled substance in the third degree. Footnotes [1] Dey was also convicted of one count of criminal possession of a controlled substance in the seventh degree (crack cocaine), based on residue in a glass crack pipe that was found in his possession. He does not challenge that conviction in this petition. [2] References to "Tr." are to the page numbers of the transcript of the state trial, which commenced on March 2, 1989. [3] The Court believes the report was prepared after the undercover returned to the precinct at 10:00 p.m., but prior to the 12:30 a.m. post-arrest identification, because the markings on the form designating the time of the post-arrest identification are made in pencil, while the rest of the report is typewritten. Moreover, the date of the post-arrest identification is typed as January 6, suggesting that it was filled-out prior to the identification, which occurred at 12:30 a.m. on January-7. [4] The Court notes that McCool’s testimony conflicts with Curasi's in this respect. McCool testified that they arrived at the 113th Precinct at "10 of 10.” (Tr. 58.) This discrepancy is, however, not material. [5] References to "M-W Hr’g" are to the page numbers of the transcript of the Mapp-Wade hearing. [6] As defense counsel noted on the record, however, Dey was the only black male in the courtroom except for the Assistant District Attorney and members of the jury. (Tr. at 189.) [7] On April 24, 1996, President Clinton signed into law the Antiterrorism and Effective Death Penalty Act of 1996 (“the Act"), Pub.L. No. 104-132 (1996), which, inter alia, imposes new restrictions on habeas corpus petitions and modifies § 2254. Dey’s petition was filed on June 14, 1993, and thus, a question of retroactive application arises. This question, however, was answered in the negative by the Second Circuit in Boria v. Keane, 90 F.3d 36 (2d Cir.1996). In Boria, the court held that the Act does not apply to pending non-capital habeas petitions because it provides no indication that Congress intended it to apply to such cases and because of the presumption " 'deeply rooted’ in our jurisprudence that absent some clear signal from Congress, a statute will not apply retroactively.” Id. at 38 (quoting Landgraf v. USI Film Products, 511 U.S. 244, -, 114 S.Ct. 1483, 1497, 128 L.Ed.2d 229 (1994)). [8] References to "Hr’g" are to the page numbers of the transcript of the habeas hearing held by this Court on February 22, 1996. [9] Curasi testified that a photograph might have been taken within an hour of an arrestee’s arrival at the precinct: [CURASI]: I have on occasion taken photos. There have been times when I have. THE COURT: Normal procedure is— [CURASI]: A photo would be taken. THE COURT: A photo is taken right away, right? [CURASI]: Yes. THE COURT: Is it taken while they are in the holding cell or while they are being fingerprinted, or is it taken sometime after they are fingerprinted? [CURASI]: In the course of the processing they would be pulled aside and photographed. THE COURT: Sort of fingerprinted and the photo is taken? [CURASI]: Maybe initially. It could be at the end of the process. THE COURT: It is within an hour or so of the arrest? [CURASI]: That would be fair. (Hr’g at 63-64.) If any such photograph ever existed, it is no longer available. [10] Occasionally, the official arrest photograph would be taken before an arrestee was returned to his precinct to await the completion of his paperwork. In such a case, once his paperwork was completed he would be brought straight to court for arraignment. [11] Goff testified that the only reason Dey would not be on the Central Booking prisoner roster was "because he never was in the cells or was only temporarily held in the main Central Booking holding area, otherwise he would have been required to be put on [it]. The only reason I could think is because when he was brought in the cells were full and he was shipped out.” (Hr'g at 119.) Confirmation of Dey’s transfer back to the precinct, however, cannot be obtained because the prisoner log sheet of the 113th Precinct is missing. This is not surprising. As Goff noted, "[t]he precincts are also required to maintain [prisoner logs]. However, [Central Booking] keep[s] [its logs] for a lot longer than the precincts probably keep theirs.” (Id.) [12] Respondent concedes the likelihood of this fact in his brief, stating that "[p]etitioner most likely arrived [at Central Booking] at 3:39 a.m. on January 7[J" (Resp’t Supplemental Mem. of Law at 10.) [13] There is, however, no way to know whether the Polaroid was taken upon Dey’s arrival at Central Booking at 3:39 a.m. or at an earlier time at the precinct. Of course, if the Polaroid was taken at the precinct, Dey would have had even less time in which to trade the red or burgundy coat, which he was purportedly wearing at the time of the sale, for the blue coat, which he was wearing in the arrest photograph. [14] The Court’s determination that the exclusion was erroneous would remain the same under federal law. The arrest photograph would have aided the jury in making its ultimate determination of Dey’s guilt or innocence, Fed.R.Evid. 401, and its probative value "posed no realistic threat of confusing the issues before the jury, let alone one which presented a danger of 'substantially outweigh[ing]’ the evidence's probative value." United States v. Blum, 62 F.3d 63, 68 (2d Cir.1995). Moreover, " 'given the importance of the ... testimony to the defense, whatever confusion ... that may have resulted from its admission would have to have been overwhelming to satisfy Rule 403’s balancing test.' " Id. (quoting United States v. Harvey, 547 F.2d 720, 723 (2d Cir.1976)). See United States v. Taylor, 92 F.3d 1313, 1331 (2d Cir.1996) (district court did not abuse discretion in excluding tape of conversation which was neither "critical to ’ [the] defense," nor would have had “a substantial effect on the jury’s verdict.”) (citations omitted). [15] Although the arrest photographs in these corroboration and direct evidence cases were submitted by the prosecution and not the defense, the relevance of an arrest photograph cannot plausibly be predicated upon the party that offers it into evidence. In fact, the trial court recognized this logic in stating that it would not allow the arrest photograph into evidence even "if the People were attempting to introduce it as supposed evidence against your client for the very same reasons that [it's] not allowing it now[.]” (Tr. at 28.) Of course, an erroneous ruling does not become a correct one just because the court would malte the same error against the other party. [16] The Court notes some confusion about the precise wording of this standard. Prior to Rosario, the Second Circuit stated that "the test for determining whether the ruling denied the defendant a fair trial is whether it would have created 'a reasonable doubt that did not otherwise ex-istí.]' " Collins, 755 F.2d at 18 (quoting United States v. Agurs, 427 U.S. 97, 112, 96 S.Ct. 2392, 2402, 49 L.Ed.2d 342 (1976)) (emphasis added). The Court’s ruling, however, would remain the same under either version of the standard. [17] The constitutional standard of materiality has been explained similarly with respect to determining whether information not disclosed to the defense creates constitutional error: ”[M]ateriality exists 'only if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.’ ” United States v. Gambino, 59 F.3d 353, 365 (2d Cir.1995) (quoting United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 3383, 87 L.Ed.2d 481 (1985)). "A ‘reasonable probability’ is 'a probability sufficient to undermine confidence in the outcome’ of the case, (citation omitted); hence, the undisclosed evidence will be deemed material only if it 'could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict.' " United States v. Payne, 63 F.3d 1200, 1208 (2d Cir.1995) (quoting Kyles v. Whitley, — U.S. —, —, 115 S.Ct. 1555, 1566, 131 L.Ed.2d 490 (1995)). [18] In fact, respondent essentially conceded this point at oral argument: THE COURT: [B]asically, it seems to me that the ID by the undercover agent is all impor *972 taut here and, you know, certainly there is nothing else real separate and apart from that that would warrant or justify the finding of this defendant guilty; do you agree with that? [ADA]: Yes. ****** THE COURT: It does seem to me the ID by the undercover agent is all important. [ADA]: Of course it is. (Hr'g at 99-100.) [19] Despite Brecht, some federal courts have held that the Chapman harmless error standard should still be applied, even on habeas review, if no harmless error analysis was performed on direct review. See Starr v. Lockhart, 23 F.3d 1280, 1292 (8th Cir.1994) (applying the Chapman harmless error standard on habeas review where state courts had not found constitutional error on direct review, and thus, had not performed harmless error analysis); Orndorff v. Lockhart, 998 F.2d 1426, 1430 (8th Cir.1993) (same); Lyons v. Johnson, 912 F.Supp. 679 (S.D.N.Y.) (same), aff'd on other grounds, 99 F.3d 499 (2d Cir.1996). But see Davis v. Executive Director of Dep’t of Corrections, 100 F.3d 750, 778 (10th Cir.1996) (Brecht standard applies in all federal habeas proceedings); Sherman v. Smith, 89 F.3d 1134, 1140 (4th Cir.1996); Tyson v. Trigg, 50 F.3d 436, 446-47 (7th Cir.1995) (same); Horsley v. Alabama, 45 F.3d 1486, 1495 n. 11 (11th Cir.) (same), cert. denied, - U.S. -, 116 S.Ct. 410, 133 L.Ed.2d 328 (1995). The Second Circuit has expressly declined to resolve this issue. Lyons, 99 F.3d at 503 ("Due to the magnitude of the trial court's error ... we need not decide whether Brecht can be limited as [the district court] suggests or whether Chapman or Brecht is the appropriate standard to apply in this case.”); see also Hanna v. Riveland 87 F.3d 1034, 1038 n. 2 (9th Cir.1996) (refusing to decide the issue because "even under Brecht’s less stringent standard of review, the error cannot be deemed harmless.”). This issue is not before the Court because, as explained more fully below, harmless error analysis is inapplicable when the claim of constitutional error is based on the denial of a fundamentally fair trial. [20] See Fulminante, 499 U.S. at 290-91, 111 S.Ct. at 1254-55 (5-4 decision) (White, J., dissenting) ("The majority ... draw[s] a meaningless dichotomy between ‘trial errors’ and ‘structural defects’ in the trial process ... [which] fails [because] our jurisprudence on harmless error has not classified so neatly the errors at issue.”); Brecht, 507 U.S. at 648 n. *, 113 S.Ct. at 1727 n. * (White, J., dissenting) (“As I explained in Fulminante, I have serious doubt regarding the effort to classify in systematic fashion constitutional violations as either ‘trial errors' — that are subject to harmless error analysis — or ‘structural defects’ — that are not.”); United States v. Wiles, 102 F.3d 1043, 1069 n. 5 (10th Cir.1996) ("The distinction between trial and structural error was bom in controversy in the 1991 Fulminante decision.”). [21] In Yarborough, the Second Circuit recently explained that harmless error analysis can, at times, be applied to rights that have previously been categorized as "structural,” stating: We do not understand Fulminante’s list of examples of violations that have been held exempt from harmless error review to mean that any violation of the same constitutional right is a ‘structural defect,’ regardless whether the error is significant or trivial. Nor does the fact that the Supreme Court has applied harmless error analysis to one level of violation of a particular right necessarily mean that even the most egregious violations of that right would also require demonstrated prejudice. Unless the Supreme Court has held otherwise, errors of a quality that undermines the structural integrity and fairness of the proceeding might be deemed structural notwithstanding that less significant violations of the same constitutional right have been subjected to harmless error analysis. To determine whether an error is properly categorized as structural, we must look not only at the right violated, but also at the particular nature, context, and significance of the violation. Yarborough., 101 F.3d at 897. [22] In Kotteakos v. United States, 328 U.S. 750, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946), the Supreme Court interpreted the predecessor to the current federal harmless error statute, 28 U.S.C. 2111, which directs courts in both civil and criminal cases to enter judgment, although error occurred at the trial level, if that error did not "affect the substantial rights of the parties.” The Court held that in determining whether remand is warranted, a reviewing court should ask whether “the error had substantial and injurious effect or influence in determining the jury’s verdict.” Kotteakos, 328 U.S. at 776, 66 S.Ct. at 1253. Thereafter, the Court, in Brecht, imposed the Kotteakos standard on federal district courts’ collateral review of state court convictions. See Brecht, 507 U.S. at 638, 113 S.Ct. at 1722 (”[W]e hold that the Kotteakos harmless-error standard applies in determining whether habeas relief must be granted because of constitutional error of the trial type.”).