William Evans-Smith v. John B. Taylor

OPINION MURNAGHAN, Circuit Judge: Appellant, William Evans-Smith, is serving a twenty-year sentence at Virginia’s Staunton Correctional Center for the 1985 murder of his wife, Barbara Evans-Smith. He was first convicted after a jury trial and *900 sentenced to five years imprisonment in August 1985. That conviction was overturned on appeal to the Virginia Court of Appeals on the ground that much of the evidence used to obtain the conviction was either improperly admitted hearsay evidence or evidence that was more prejudicial than probative. A second trial was set and Evans-Smith moved for a change of venue on account of pretrial publicity; his motion was denied after an extensive hearing. During the course of the second trial, which began on April 17, 1989, nearly four years after the first trial, Evans-Smith filed a Motion to Strike, arguing that the solely circumstantial nature of the evidence presented by the Commonwealth was insufficient to establish guilt. His motion was denied. The jury found Evans-Smith guilty of second-degree murder and recommended the maximum twenty-year sentence. Evans-Smith then moved to set aside the verdict, again arguing that the evidence was insufficient, and also arguing, among other grounds, that the court’s prior denial of the motion to change venue deprived him of a fair trial. He also moved for a new trial based upon alleged juror misconduct. Both motions were denied. Bond was denied, and Evans-Smith was sentenced to twenty years imprisonment. Evans-Smith appealed through the state system. His appeals were denied. He next filed a state habeas petition, raising only the speedy trial issue. The denial of that petition was affirmed by the Virginia Supreme Court. The present action is a § 2254 habeas appeal. In the district court, both the Commonwealth and Evans-Smith moved for summary judgment. After a hearing, the district court granted the Commonwealth’s motion and denied Evans-Smith’s. Evans-Smith appeals the summary judgment entered by the district court on the grounds that the evidence was insufficient to prove his guilt, that his sentence of twenty years following the second trial violated the rule announced in North Carolina v. Pearce, 395 U.S. 711, 89 S.Ct. 2072, 23 L.Ed.2d 656 (1969), and that the jury that convicted him was tainted by pretrial publicity. FACTS In obtaining Evans-Smith’s conviction, the Commonwealth proceeded on the theory that Evans-Smith Mlled or rendered his wife unconscious in the kitchen of their farmhouse, then dragged her body upstairs and staged a rape/robbery scene before leaving for work. Evans-Smith asserted throughout that he is innocent, that the murder was committed by an unidentified third party or parties, and that the Commonwealth failed to investigate properly the possibility that someone other than Evans-Smith was the murderer. The prosecution agrees that the evidence used to convict Evans-Smith was entirely circumstantial. According to Loudoun County Medical Examiner Dr. Hoeker, Barbara Evans-Smith (hereinafter “the victim”) died sometime between 4:00 and 8:00 a.m. on the morning of April 15, 1985. His calculation was based on the degree of rigor mortis (60%) he observed upon his arrival at the crime scene at approximately 1:00 p.m. that afternoon. Dr. Byer, the medical examiner who performed the autopsy, determined that the victim died by strangulation thirty to sixty minutes after eating. The contents of the victim’s stomach were consistent with the breakfast that Evans-Smith stated he had shared with his wife. The victim’s body was found in her second-floor bedroom of the couple’s Loudoun County farmhouse with a pair of pantyhose wrapped three times around her neck and knotted. She was clad in a nightshirt and robe which had been ripped open, leaving her bare from the waist down. While the condition of her clothing suggested the possibility of a sexual assault, Dr. Byer found no seminal fluid or genital abrasions on the victim. There was a small amount of unexplained blood in her vaginal cavity. A pair of white underwear had been stuffed into the victim’s mouth. Dr. Byer stated that the underwear was loosely packed, a fact which could lead to the inference that the underwear had been placed in the mouth after death, since the loose packing would not have served as an effective gag. The Commonwealth urged the jury to consider the underwear as evidence that the murder had been staged as a sexual assault. *901 Dr. Byer identified internal injury and hemorrhaging to the back of the victim’s head which could have been caused by her head hitting the wooden stairs, as she was dragged to the second floor. The doctor testified, however, that “any blunt force trauma” could have caused the injury. He also noted triangular-shaped abrasions on the victim’s neck near her jaw line, which he believed were the victim’s fingernail scratches, created while she was trying to remove the ligature from around her neck. There was no blood on the hands of the victim. The bedroom in which the victim was found looked as if it had been ransacked, with items from the closets and dresser drawers strewn around the room. Evans-Smith’s bedroom, separated from the victim’s bedroom by a shared bathroom, had a similar appearance, but the rest of the house was in order, with the exception of an overturned, downstairs hall table and slightly skewed furniture in the kitchen. • The Commonwealth argued that since only a few pieces of furniture had been moved, and many valuables were left either untouched or “neatly” laid out on the beds upstairs, the crime scene was consistent with a staged crime and not a real robbery. Evans-Smith’s daughter, Leslie Cook, noted in testifying for the prosecution that the front door was ajar when she arrived at the scene at approximately 11:00 a.m., that one of the two dogs was closed up in the laundry room, and that the other was outside. Cook also testified that she observed only one set of tire tracks on the newly raked gravel driveway as she came onto the property. She and another prosecution witness, William Laycock, testified to the presence of another set of tire tracks running across the hayfield toward a second exit on the morning in question. This second set of tracks was not investigated by the police, and the defense argued at trial that their existence implied that someone other than Evans-Smith had been to the house that morning. 1 Special Agent Ritchie of the Virginia State Police, a key participant in the investigation of the crime scene, placed bags over the victim’s hands and feet upon his arrival to preserve any evidence on them during transport, and collected several other items from the crime scene for analysis, including jewelry boxes which had been tampered with and which Ritchie believed could be fingerprinted. 2 No fingerprints were taken from the furniture and the other items that were found out of place on the first floor. Nor was fingerprinting undertaken' along the staircase. Ritchie also observed two parallel scratch marks; which were photographed, on the linoleum floor in the kitchen. Although the Commonwealth attempted to cite these scratch marks as evidence that the victim was killed in the kitchen, then dragged across the floor and upstairs, State Police Special Agent Schultze, testifying for the prosecution, stated that the scratch marks could not have been caused by the soft slippers the victim was apparently wearing at the approximate time of the murder. Her slippers were found in different rooms on the first floor. At oral argument, the Commonwealth asserted that its argument was that the victim’s bare heels had created the marks. . Among the evidence collected by Ritchie were seventy-three individual hair samples from around and on the body of the victim and from the stairway leading to the second floor, tissues removed from the wastebasket in the shared bathroom, the victim’s nightshirt and robe, and various other items. Of the fourteen hair samples taken from the stairway, Forensic Scientist Deanna Dabbs found ten of them to be unidentifiable, three to be consistent with the hair of the victim (one of which had been forcibly removed), and one blond head hair to belong to neither the victim nor to Evans-Smith. Thirty-nine of the fifty-nine hair samples taken from the victim’s body or bedroom were *902 found to be consistent with the victim’s hair, seven were unidentifiable, and seven were consistent with the hair of Evans-Smith. None of the hair attributed to Evans-Smith had been forcibly removed. Of the remaining hair, one brown Caucasian head hair taken from the bag enclosing the victim’s clenched left hand was found to belong to neither the victim nor to Evans-Smith. 3 Five hair samples taken from a pubic combing of the victim were definitely not Evans-Smith’s, and could not be identified as belonging to.the victim. 4 Dabbs analyzed several small blood stains on the victim’s robe and a blood-stained tissue from the shared bathroom’s wastebasket. There was also a band-aid found in the wastebasket with an amount of blood on it too minute to test. Two of seven blood stains on the robe were consistent with the same type as Evans-Smith’s blood. 5 Blood of the same type as Evans-Smith’s was also found on a tissue retrieved from a wastebasket in the shared bathroom. Evans-Smith told the police that he and his wife had cleared limbs, thorny locusts, and other brush from parts of their farm on the Saturday and Sunday morning preceding Monday’s murder. 6 Commonwealth witnesses corroborated this statement. While clearing the brush, Evans-Smith allegedly acquired several scratches on his arm and hands. At about noon on Sunday, after patching up his scratches with large band-aids, Evans-Smith and his wife attended the Oatland Point-to-Point horse race. Upon their return that afternoon, Evans-Smith raked the gravel driveway with a york rake. Evans-Smith stated to investigators that he arose on the morning of the murder at 5:00 a.m., as usual, and drove down the one-quarter mile lane to the mailbox to retrieve the newspaper. 7 While at the mailbox, Evans-Smith claimed, he caught a glimpse of a dark-colored van parked on the road near a neighbor’s driveway, but thought nothing of it. 8 He returned to the house, placed the newspaper on the hood of his wife’s truck— as was their custom — and completed some farm chores. Evans-Smith stated that he returned to the house to breakfast with his *903 wife at approximately 6:00 a.m., and after signing a birthday card for their son-in-law at his wife’s request, kissing his wife goodbye in the kitchen, and watching her wave to him from her upstairs bedroom, 9 he began his one and one-half to two hour commute to American University’s Foreign Area Studies Office in Washington, D.C. 10 According to Evans-Smith, the traffic coming into the city was “extremely slow” that morning. The prosecution put on witnesses who testified that there had been nothing extraordinary, such as accidents, about the traffic that morning. The Commonwealth argued that Evans-Smith complained “bitterly” about the traffic, and that his reaction was inconsistent with the normal state of rush hour traffic. The trial transcript suggests that Evans-Smith did complain, but perhaps not “bitterly.” The testimony of a co-worker was that [Our conversation] was concerned with the traffic that morning, that it was exceptionally bad. The traffic was very often bad coming over the Cabin John Bridge and it was exceptionally bad that morning, yeah. And Mr. Evans-Smith was upset ... in a way that is not abnormal with the traffic. • Two investigators who interviewed Evans-Smith two days after the murder testified that Evans-Smith told them at that time he had not made any stops on his way to work. He later told someone else and testified at the first trial that he had stopped at the university president’s office to drop off some books. The office, however, was closed, and he proceeded on to work. According to prosecution witness Mimi Tandler, who was writing a book on the case, Evans-Smith told her after the first trial that he had stopped at a car wash on the way to work. Evans-Smith stated that he arrived at his office at approximately 8:16 a.m. or “some time before 8:30 a.m.” Evans-Smith’s coworker, Dorothy Lohman, testified that Evans-Smith usually arrived at work between 7:30 or 8:00 a.m., but that he was not at his desk at 8:45 a.m. when she went to his office. She had a meeting with Evans-Smith concerning edits of a manuscript later that morning, however, at which time he appeared completely normal. Prosecution witness and coworker James Rudolph testified that the office arrival time was flexible and that it was normal for people to arrive between 8:00 and 8:30 a.m. He further testified that he had spoken with Evans-Smith in the office parking lot at 8:30 or 8:35 on the morning in question. This conversation was the source of the comments about the traffic. A defense witness stated that Evans-Smith arrived at work on April 15 between 8:00 and 8:30 a.m. Yet another stated that he had spoken with Evans-Smith in the men’s bathroom on the same floor as Evans-Smith’s office at 8:30 a.m. 11 Evans-Smith was contacted by the police at his office and told of the murder that *904 afternoon. Lieutenant Brown and Investigator Merchant testified that Evans-Smith then asked a series of questions about whether his wife had been raped or disfigured, whether she was clothed or not, where she was located, and who found her. Although such questions might reasonably be expected, the prosecution argued that such questions suggested he knew what the murder scene looked like, and, in any event, were odd reactions for someone told of the murder of his wife. Brown testified that as soon as he told Evans-Smith that a burglary had occurred, Evans-Smith gasped, “Oh my God, the van,” and then provided a description of the van, as noted above. The investigators and Evans-Smith then left Washington for Loudoun County. The investigators, who trailed him on the way back, testified that Evans-Smith, upon reaching stop lights, would put his head in his hands or, alternatively, look in his rear view mirror. When they reached the farm in Loudoun County, Evans-Smith was not allowed to visit the crime scene to see his wife’s body. He engaged in a physical tussle with the authorities, during which they first observed several scratches on his arms. 12 The scratches on his arms were the source of much testimony at the trial. The prosecution’s theory was that the scratches were caused by the victim’s fingernails in the course of a violent struggle with Evans-Smith. The Commonwealth based its argument on the testimony of its witness, Dr. Solin, who had examined Evans-Smith’s scratches and had determined that they could have been caused by the victim’s fingernails. He testified that his knowledge of the fingernails was based on Commonwealth photographs. Dr. Solin did not himself examine the victim’s fingernails. Two other prosecution witnesses, who saw only photographs of Evans-Smith’s scratches but had made á careful inspection of the victim’s fingernails, concluded that her fingernails could not possibly have caused the scratches. These two witnesses, Loudoun Medical Examiner George T. Hocker and Virginia Deputy Chief Medical Examiner Dr. James C. Byer, both testified that the victim’s fingernails were too short and dull to have made the scratches. They opined that the scratches were more likely made, as Evans-Smith claimed, by his work clearing thorn bushes. 13 Evans-Smith’s daughter, who testified for the prosecution, stated that she had seen her parents clearing out thorn bushes the day before the murder. Moreover, the only debris found under the victim’s fingernails was identified by the prosecution witnesses as having resulted from her attempts to release the ligature from around her neck. None of the debris was consistent with Evans-Smith’s blood or skin; indeed, there was no blood on the victim’s hands at all. Evans-Smith had stated that the blood found on the tissues and medicated band-aid in his bathroom had come from his and his wife’s having treated his scratches from the thorns. Prosecution witness, Dabbs, testified that the spots on the victim’s robe might have been placed there at different times, one stain having been on the robe as long as eight weeks previously, and the other less than one month previously. The test was not capable of determining whether the blood had been placed on the victim’s robe the day of the murder. *905 On the evening of the murder, after the day’s investigation had concluded, Evans-Smith was asked by the authorities to make a list of any items he believed were missing from the house. Two days later Evans-Smith asked to speak with Investigators Brown and Merchant. During the ensuing interrogation, he mentioned that his wife kept her jewelry and hay money in an aqua jewelry box in her bedroom. Prosecution witness Joe Rogers, a farmer who leased and cultivated the land adjoining Evans-Smith’s farm, testified that Evans-Smith had phoned him ostensibly to discuss that year’s planting. During the course of the conversation, Evans-Smith asked him to keep an eye out for anything suspicious that he might find and report it. Shortly thereafter, an employee of Rogers, while discing the neighboring field, found the aqua box in the field, slightly covered by dirt. There were no fingerprints on the box, but its composition as well as the time it had been exposed to the elements made it unlikely that such fingerprints would still exist. Several other facts bear mention. During the week following the murder, Evans-Smith called his coworker, Rudolph, who was then vacationing in New Mexico. During their conversation, Evans-Smith related that he had been indicted, and he believed'the state had bungled the investigation. He then related to Rudolph his activities on the morning of the murder, telling him of the stop at the president’s office and his late arrival at the office. He seemed concerned, however, that Rudolph remembered his arrival time to be approximately 8:35, because he had already told the police that he thought he had arrived at work at 8:15 a.m. According to Rudolph, Evans-Smith mentioned that “it would be nice if he and I recalled the same time of arrival that morning.” The Commonwealth argued that the conversation was evidence that Evans-Smith was trying to make other people’s stories consistent with his own. Rudolph testified, however, that Evans-Smith did not threaten him, and that “he was not noticeably angry. I’m sure he was upset, everybody was upset.” Rudolph testified that Evans-Smith was not angry with him. In addition, he testified further that his watch often ran five minutes fast, and that he could have seen Evans-Smith at 8:30 rather than at 8:35 a.m. Evans-Smith attempted, with no success, to have the police investigate the possibility that two horse trainers, Robert Rivera and Rudy Ruiz, who were working at a horse barn on adjacent property at the approximate time of the murder, may have been involved in the crime. In support of this theory, Evans-Smith pointed to the fact that Rivera was on a $100,000 bond for drug charges and that he had been ruled off of the Charles Town race track, after drugging horses. There was some suggestion that Rivera’s whereabouts on the morning of the murder were inconsistent with his statements at trial. The Rivera theory suffers from incomplete development in the record. On the basis of the foregoing evidence, essentially all there is, William Evans-Smith was convicted of the second degree murder .of his wife of over forty years. DISCUSSION The standard of review for a claim of insufficient evidence is whether, viewing the evidence in the light most favorable to the prosecution, Goldsmith v. Witkowski, 981 F.2d 697, 701 (4th Cir.1992), cert. denied, — U.S. -, 113 S.Ct. 3020, 125 L.Ed.2d 709 (1993), and according the benefit of all reasonable inferences to the government, United States v. Tresvant, 677 F.2d 1018, 1021 (4th Cir.1982), any rational trier of fact could have found the defendant guilty beyond a reasonable doubt. Jackson v. Virginia, 443 U.S. 307, 319, 99 S.Ct. 2781, 2789, 61 L.Ed.2d 560 (1979). The standard is obviously rigorous. As the Supreme Court recently has reiterated, “the writ of habeas corpus has historically been regarded as an extraordinary remedy.” Brecht v. Abrahamson, — U.S. -, -, 113 S.Ct. 1710, 1719, 123 L.Ed.2d 353 (1993). Nevertheless, however circumscribed, the writ remains a remedy, “a bulwark against convictions that violate fundamental fairness.” Id. (quoting Engle v. Issac, 456 U.S. 107, 102 S.Ct. 1558, 71 L.Ed.2d 783 (1982)) (internal quotation omitted). The very existence of the Jackson test *906 presupposes that juries accurately charged on the elements of a crime and on the strict burden of persuasion to which they must hold the prosecution, nevertheless “may occasionally convict even when it can be said that no rational trier of fact could find guilt beyond a reasonable doubt.” [The test] was adopted to provide an additional safeguard against that possibility, and was to give added assurance that guilt should never be found except on a rationally supportable “state of near certitude.” West v. Wright, 931 F.2d 262, 268 (4th Cir.1991), rev’d on other grounds, — U.S. -, 112 S.Ct. 2482, 120 L.Ed.2d 225 (1992) (quoting Jackson v. Virginia, 443 U.S. at 315, 99 S.Ct. at 2786). The question before us in the instant case, therefore, is whether the wholly circumstantial evidence, taken together, could allow any rational trier of fact to conclude beyond a reasonable doubt that Evans-Smith murdered his wife. In conducting such an inquiry we consider all the evidence presented in the record. The prosecution’s theory of the case was that Evans-Smith, the victim’s husband, for whatever reason but probably in the heat of passion (of which there was no proof), murdered his wife or rendered her unconscious in the kitchen after the two had breakfast. No evidence of motive nor of heat of passion was introduced at trial. Indeed, the only evidence before the jury was that the Evans-Smiths had been seen the day before the murder enjoying one another’s company. The prosecution’s theory was that he then dragged his wife up the stairs, and placed her on the floor of her bedroom. Realizing, according to the prosecution, that he would be a primary suspect, Evans-Smith then supposedly staged a robbery/rape scene and drove to work, perhaps arriving a little later than usual. According to the Commonwealth, Evans-Smith’s statements and behavior after the crime suggested to the prosecution an attempt to cast blame elsewhere, lending further supposed support to the conclusion that he in fact had committed the murder and then had attempted to cover it up. A comprehensive review of the facts has been set out above, but, at the risk of repeating many of them, it seems worthwhile to explore fully the facts on which the Commonwealth particularly relies in its brief. They are quoted at length as follows: 1. When the victim’s daughter arrived at the home that morning, she found one of her parents’ two dogs, significantly the one which she described as more protective of the victim, penned up in the laundry room. The daughter had never seen the dog thus confined before. 14 2. The victim, a 64-year-old woman, was found naked from the waist down, with her nightshirt torn open and a pair of panties placed in her mouth in a manner so as to suggest that they had been used as a gag. She was killed by strangulation with a ligature, consisting of a pair of pantyhose wrapped three times around the neck and knotted. 15 Her room was in disarray, with drawers pulled from dressers and various items from the closet placed on the floor. A button which matched her nightgown was found under the bed next to where she lay[,] establishing that the nightgown had been ripped open in the bedroom. 16 3. The kitchen furniture was out of place, there were drag marks consistent with *907 the victim’s heels being dragged across the kitchen floor, 17 and a table was overturned in a hallway- leading from the kitchen to the stairs to the victim’s room. The autopsy revealed head injuries which could have been caused by the victim’s head hitting the stairs as she was dragged from the kitchen, up the stairs to her room. 18 4. There was no evidence that the victim had been sexually molested in any way. There were no bruises, abrasions or tears on her breasts or genitalia, nor was there any evidence of seminal fluid anywhere on her body, clothing or bedclothes. 19 5. Dr. Hocker, the medical examiner who responded to the scene, testified without contradiction that the victim’s body had been placed where it was found, and that she had not fallen dead there. His opinion on this was not questioned on cross-examination. It was stipulated that the testimony of Dr. Beyer (sic), the medical examiner who performed the autopsy, would be that the panties probably were placed in her mouth after she was strangled. Since there would be no need for a gag after the victim had been killed, both of these observations support the conclusion that the scene was “rigged” to look like something other than what it was. 20 6.Other than the victim’s room, the defendant’s bedroom was the only other *908 room in the house that showed any signs even vaguely consistent with a burglary. Two jewelry boxes had been emptied on the defendant’s bed, but the photographs revealed that the contents had been spread neatly out on the bed, and items of obvious value were left in place. Silverware was left untouched in other rooms. 21 7. There was no evidence that anyone other than the defendant had been in the home on the morning of the killing. There was no blood at the scene other than blood consistent with the victim and the defendant. Blood consistent with the defendant’s blood was found on tissues in a wastebasket in the room where the victim’s body was found. 22 8. When informed of the fact that his 64-year-old wife had been killed, the defendant asked a series of considered objective questions which the jury could infer suggested his awareness of the appearance of the scene of the crime, and were calculated to highlight the staging of the scene to simulate a rape and robbery by an unknown intruder. Specifically, Evans-Smith asked if his wife had been raped, whether she was clothed or not, whether she had been disfigured, where she was, and who had found her. 23 9. The defendant gave contradictory statements as to where his wife was when he last saw her, and gave contradictory statements as to whether he stopped on his way to work, claiming first that he made no stops at all, and later claiming to have stopped at the provost’s office to deliver books, and finally claiming he stopped at a car wash. 24 10. At one point he met with several friends in what the jury could infer was an effort to get them all to agree to his story about delivering books on the morning for the killing. He even went so far as to call one person in New Mexico in what the jury could infer was an effort to get that person to say he had arrived at work on the morning of the murder earlier than he did. 25 *909 11. He gave contradictory statements as to the location of a mysterious van he claimed to have seen when he claimed to have driven to the bottom of his driveway to retrieve the morning newspaper. Several people who had driven the road that morning at about the time the defendant claimed to have seen the van testified that there had been no such van present. The road was a narrow, two lane dirt road on which a van would be readily seen. The defendant claimed initially that he had driven to the bottom of his driveway to get his paper, that that was when he had seen the van. His daughter testified that when she arrived at the house, however, there was only one set of tire marks in the freshly raked driveway. Those tire marks would have been made by the defendant when he left to go to work. 26 12. Dr. Byer .testified that the autopsy revealed the victim died within thirty to sixty minutes of her last intake of food. The defendant told investigators that he ate breakfast with her at about 5:30 to 6:00 a.m. and that he left for work about half an hour later, at his usual time of 6:30 a.m. Coworkers testified that the defendant usually arrived at work between 7:30 and 8:00 a.m., but there was testimony that he was not at work until between 8:30 and 8:45 on the morning of the killing. 27 13.Although the defendant had complained bitterly about traffic that morning when he spoke to others, there was testimony from several law enforcement officers that there were no unusual traffic tie-ups on the defendant’s route that morning. There was no evidence adduced at trial to support the defendant’s claim of unusual traffic congestion. This evidence placed the defendant with the victim during virtually all of the time frame in which the medical examiner fixed her death. 28 At the outset, it is worth noting that circumstantial evidence need not exclude every reasonable hypothesis of innocence. Stamper v. Muncie, 944 F.2d 170, 174 (4th Cir.1991). Rather “circumstances altogether inconclusive, if separately considered, may, by their number and joint operation ... be sufficient to constitute conclusive proof.” Id. (internal quotation marks and citations omitted). Even allowing for this important caveat, we believe that a rational trier of fact could not determine that the evidence supports a finding of guilt beyond a reasonable doubt Instead, what the evidence shows is that Evans-Smith was at his home the morning of the murder, a fact he does not dispute and would be his most customary behavior. In addition, according all reasonable inferences to the benefit of the prosecution, 29 we *910 believe a rational trier of fact possibly could conclude beyond a reasonable doubt that the crime scene was staged. It is the next step — an inference that Evans-Smith was within “a state of near certitude” the perpetrator of this terrible crime — that the evidence simply cannot support. “A determination in federal collateral review that a state court conviction by jury verdict was not supported by constitutionally sufficient evidence is one to be made with special caution and anxiety.” West v. Wright, 931 F.2d at 270. As in West, the instant case requires not only the reversal of a federal district court, but also of the state trial and appellate courts. We conclude, however, that the evidence assessed in its entirety and in the light most favorable to the prosecution is not sufficient to persuade any rational trier of fact of the petitioner’s guilt beyond a reasonable doubt, that Evans-Smith’s conviction on the evidence therefore violated his constitutional right to due process, and that he is entitled to the relief sought in this habeas corpus proceeding. Evans-Smith could no doubt have acted as the prosecution has contended he did. However, the suppositions are derived from acts of his consistent with a husband’s perfectly proper behavior mixed with imaginations of behavior for which his presence has not been proven. To start with the assumption that the crime was committed and then to show that each piece of circumstantial evidence can be explained in a consistent manner is fundamentally different from examining each piece of evidence and finally concluding beyond a reasonable doubt that the defendant was guilty. The prosecution has attempted to accomplish only the first alternative, not the second. As the Supreme Court has long taught “[i]t is the duty of the Government to establish ... guilt beyond a reasonable doubt.” In re Winship, 397 U.S. 358, 362, 90 S.Ct. 1068, 1071, 25 L.Ed.2d 368 (1970) (quoting Leland v. Oregon, 343 U.S. 790, 802-803, 72 S.Ct. 1002, 1009-10, 96 L.Ed. 1302 (Frankfurter, J., dissenting)). Such guilt has not been established, and, as a consequence, Evans-Smith’s right to due process has been violated. We are unable to conclude that any rational trier of fact could have found Evans-Smith guilty beyond a reasonable doubt. 30 We therefore remand the proceeding to the district court with directions to issue the writ vacating Evans-Smith’s conviction. SO ORDERED. Footnotes [1] Under cross-examination, Laycock, a friend of Cook and an early arrival on the crime scene, stated that it would have been unlikely that Evans-Smith, a farmer, would have driven through this particular field, as such a path would have damaged his crops. [2] The results of the fingerprint analysis are not indicated in the record, and neither party makes mention of them. [3] Whether this hair was found in the closed hand or simply in the bag is disputed. Not disputed is that the bagging technique is used to preserve evidence. The person responsible for bagging the victim’s hands testified that the purpose was to preserve evidence and that he did his job properly. While the Commonwealth argued to the jury that the hair probably came from one of the investigators on the murder scene (an admission that their investigative procedures were flawed), no effort was made to match the hair found in the evidence bag with that of any of the investigators on the scene. [4] Hair samples sometimes may be identified as belonging to a specific person, sometimes a specific person may be ruled out as the source of the hair, and sometimes the test simply cannot determine whether the hair belongs to a specific person or not. [5] Dabbs, a prosecution witness, testified that the blood was not only consistent with Evans-Smiths’, but also with that of 58% of the population. [6] Evans-Smith did not testify in the second trial, although he did in the first. [7] Evidence was introduced that it was Evans-Smith's habit to walk down the driveway to pick up the newspaper, but that sometimes he drove. Apparently, the prosecution hoped to show that he had in fact walked, not driven, down the driveway to get the paper. His having walked would make consistent evidence from his daughter that there was but one set of tire tracks on the drive, and call into question Evans-Smith's statements concerning his sighting of the dark-colored van. See infra, text at note 8. The testimonial evidence as to the tire tracks was disputed by the defense, which asserted that the york rake itself necessarily would have created a set of tracks, making it impossible for there to have been only one set. In addition, evidence was submitted to the effect that the driveway was narrow, and that it was common that one car would follow the tracks of another. [8] The prosecution called a number of witnesses who testified that they had driven on the road in question on the same morning and had never seen a dark colored van. None of these witnesses had driven past the point where Evans-Smith said he saw the van within fifteen minutes of either side of the alleged sighting. A witness for the defense testified that she had seen a green van with Louisiana plates speeding through the area. The van had cut her off, and had acted strangely enough that she reported it to the police. Patricia Packard, who lived opposite the entrance to Evans-Smith’s driveway, testified that she remembered hearing and seeing a dark-colored vehicle hurriedly departing the Evans-Smith driveway the morning of the murder between 9:00 and 11:00 a.m. [9] The prosecution presented evidence that Evans-Smith had been inconsistent, telling at least one person that his wife had waved to hirti not from upstairs, but from the kitchen window— both were on the same side of the house. This contradiction typifies the sort of attenuated evidence on which the prosecution relied in asserting that Evans-Smith concocted a story about the circumstances of that morning. [10] No evidence was submitted by the Commonwealth as to the time Evans-Smith's commute customarily took. Nor was any evidence submitted demonstrating that the time Evans-Smith said it took to travel the distance on the day of the murder was implausible. [11] The dispute over the time of Evans-Smith's arrival provides a useful illustration of the quality of the evidence in this case. The prosecution argued that the "inconsistency” in the statements about when he arrived, combined with an arrival time which, according to one witness, was later than usual, suggests that Evans-Smith was otherwise engaged that morning, and that he had lied about his whereabouts. Our task emphatically is not to reweigh the evidence; however, even viewing the evidence in a light most favorable to the prosecution, it is difficult, if not impossible, to determine why his arrival time, which at the most was only somewhat later than usual, indicates anything about Evans-Smith’s activities that morning. Anyone with a daily commute of up to two hours in rush hour traffic could reasonably be expected to arrive late. Such lateness of itself is not proof of a crime beyond a reasonable doubt. Moreover, more witnesses stated that he had arrived at the normal time than stated that he had been late, thus corroborating Evans-Smith's own statements. Indeed, the witness who said that Evans-Smith had been late (i.e., later than 8:45) based her conclusion on her not having seen Evans-Smith in his office, while the witnesses who testified that he was there by about 8:30 actually saw him. [12] The defense argued to the jury that the Commonwealth's refusal to allow Evans-Smith to see the body, combined with its refusal to investigate any alternative theories and its insufficient fingerprinting, suggested that the authorities already considered him a suspect, especially since other individuals had been allowed on the scene. The Commonwealth argued that when Evans-Smith was invited to view the body several days later and after the autopsy, he refused, demonstrating that the tussle in front of his house was not genuine, but merely a performance. Yet, if he had shown indifference in seeing his wife's body at the first possible moment and then sought to see the body several days later, the prosecution would no doubt have instanced the scenario as proof of murder. [13] The Commonwealth attempted to refute this finding by putting on testimonial evidence that Evans-Smith ordinarily wore gloves when he worked on the farm. There was testimony, however, that he did not always wear gloves. Several prosecution witnesses who had seen Evans-Smith at the Point-to-Point race the day before the murder testified that they had not seen any scratches or bandages the day of the race. Two defense witnesses testified that they had in fact seen bandages on his hands. [14] The fact is meant to support an inference that, since Evans-Smith knew the dog’s habits, Evans-Smith would know which of the two dogs he had to lock up to commit the crime. While we can give the prosecution the benefit of the inference that the dog was locked up because it was more protective of the victim on the day of the murder than the other dog, it is unclear why any attacker (Evans-Smith or someone else) would not have noticed the difference in protectiveness of the two dogs, if that difference were as great as the prosecution suggests. [15] All parties assume that the pantyhose belonged to the victim. The Commonwealth asserted that they were particularly convenient to Evans-Smith, since he would have known their location. Another pair of pantyhose was found, in plain view, in a dish on top of the victim's dresser. [16] The inference meant to be drawn here, particularly from the button, is that having been killed in the kitchen and dragged up the stairs, the victim’s gown was tom open after she had been placed dead on the floor of her bedroom. The *907 inference is somewhat remotely meant to support the theory of a staged rape. [17] The statement that the marks were consistent with the victim's being dragged is simply unsupported by the record. Indeed, the record suggests that drag marks could not have been made by the heels being dragged. The prosecution's witness Schultz, a state crime scene technician, clearly stated that the slippers the victim apparently had been wearing could not have caused the drag marks on the linoleum floor. No evidence was adduced supporting the prosecution’s assertion that the victim's bare heels, had she not been wearing slippers, caused the marks. To the extent, therefore, that the Commonwealth's theory is helped by the murder having been committed in the kitchen, and the body having been dragged up the stairs, the marks do not appear to support such an inference. [18] It is somewhat unclear as to whether the prosecution advanced the theory that Evans-Smith killed his wife in the kitchen and then dragged her upstairs, or whether he knocked her out in the kitchen, and then strangled her upstairs. In his closing argument, the Commonwealth’s attorney suggested that he was advancing either theory, "The evidence is, I would submit, that she was either rendered unconscious at the table or strangled at the table and dragged upstairs.” Plainly, the theory that she was "rendered unconscious” raises a difficult problem for the Commonwealth's theory that the victim's head was bruised as she was being dragged up the stairs. The record suggests that the victim’s head suffered a defined internal contusion; even if the contusion represented more than one head injury (an inference not supported by the record), all of the injuries — those caused by being rendered unconscious and those caused by the stairs would have had to occur in the same part of the head. To say the least, this seems unlikely, and subject to reasonable doubt. If the victim's head did not hit the stairs and the skid marks on the kitchen floor were not caused by her slippers, there is no evidence that she was "dragged” upstairs. In order to remain consistent with the Commonwealth's theory, we would have to infer that Evans-Smith, a 64-year-old man in admittedly good shape, knocked out his wife and then carried or dragged her upstairs without doing any further damage to her body or any damage to the house. In that connection, it is worth noting that the victim’s body was carried down the stairs by three medical technicians, one of whom noted that she was heavy. Finally, the bruise on the victim's head could have been caused, according to the medical examiner, by any blow. The idea that the blow was caused by a step was inventively advanced by the prosecution to support the dragging theory. As stated before, there is no evidence that she was dragged, therefore there is no more evidence connecting the blow on the head to a step than there is connecting the blow to any other blunt object. What inference was there for the jury to draw beyond a reasonable doubt? [19] There was at least a little blood in the victim's vagina. Obviously, sexual assaults frequently occur without physical trauma or evidence being left behind. Nevertheless, viewing the evidence in a light most favorable to the prosecution, it is plausible that involved was someone who wanted the crime scene to appear consistent with that of a sexual assault. [20] The prosecution urges us to believe that the crime scene was staged. Even if a rational fact-finder would accept this theory, the Commonwealth's bald assertion that no one but Evans-Smith would have staged the crime scene simply is insufficient ground for us to conclude that a rational fact-finder could connect Evans-Smith beyond a reasonable doubt to the killing. [21] There was testimony that a friend of the victim’s had called the house and let the phone ring three times sometime after 8:00 a.m. The defense suggested that the repeated phone calls could have scared the intruders away before they completed the crime. While we note the defense's theory is not implausible, we must give the benefit of the inference to the prosecution. That inference suggests that someone staged the crime scene. Even if we infer that the "staged” quality of the crime scene suggests a cover-up, there is nothing linking Evans-Smith to the crime, except for the unsubstantiated prosecution's theory that no one else could have had any reason to stage a rape and robbery. [22] While all inferences must be made in favor of the prosecution, leaps of logic should not be. Here, the Commonwealth asserts that the bloody tissues were left behind by Evans-Smith, who had used them to wipe scratches caused by a struggle with his wife, though prosecution evidence was to the effect that the victim could not have caused the scratches. Such careless littering by Evans-Smith would be, at the very least, somewhat inconsistent with the theory that he had killed his wife, and then had attempted to stage a robbery scene. [23] The Commonwealth's inference that such comments suggest guilt would perhaps be supportable if there were anything linking Evans-Smith to the crime. Otherwise, the questions seem to be perfectly reasonable ones to be expected of a bereaved husband, and it is difficult to see how they are consistent with his guilt. [24] These inconsistencies are treated at some length in the facts section. In our view, although they received much attention at trial, the contradictions appear inconsequential, since Evans-Smith appears to have arrived at work at about the usual time. [25] We have searched the record in vain, paying particular attention to the passages cited by the prosecution, for support for the first sentence of the paragraph. The second is mentioned in the facts section. Once again, what is required here, in order to find for the Commonwealth, is a kind of double inference. We must infer that the content of Evans-Smith's phone conversation with someone in New Mexico was meant to make his colleague's story consistent with his own, and then we must infer that the reason he desired consistency was to create an alibi since he had committed the murder. However, the first inference might lead only to the conclusion that whether or not he had committed the murder, it would be in Evans-Smith's interest to present a consistent story to the police. After all, the recipient of the phone call testified that Evans-Smith simply said that it would be "nice" if the two' of them had the same story. He did not appear angry, merely upset. [26] There was testimony tending to establish the presence of the van, and a track was shown across a cultivated field, indicating the presence on the scene of an unknown party or parties. [27] The evidence shows that Evans-Smith could have been with his wife when she was killed. Only one prosecution witness testified that Evans-Smith was regularly at work between 7:30 and 8:00, and only one (the same one) testified that he was not at his desk at 8:45 a.m. [28] The record suggests that Evans-Smith complained about the traffic, but it does not suggest that he did so ‘‘bitterly.” Most of the evidence adduced at trial suggests that Evans-Smith arrived at work around 8:30 a.m. According to the medical examiner, the victim could have died anywhere from 30 to 60 minutes after she finished eating. Evans-Smith stated that he left for work after breakfast at 6:20 or 6:25, which would not have been inconsistent with his arrival time. The time of death would then have been anywhere from 7:00 to 7:30 a.m. Even assuming that he left at 7:00, and that the drive took the least amount of time it ever did, one hour and a half, the medical examiner’s estimate would still allow for the victim's being killed in his absence. We must not re-weigh the evidence, but it is worth illustrating that the victim could easily have been killed after Evans-Smith left. Indeed, the rigor mortis analysis allowed for the possibility that she was killed as late as 8:00 a.m. If she had continued breakfast after her husband's departure, the victim could have been killed at 8:00 a.m. [29] It is essential to remember that Jackson requires that we review a 11 the evidence and then determine whether a rational trier of fact could have found guilt beyond a reasonable doubt. See Jackson. 443 U.S. at 319, 99 S.Ct. at 2789, {‘‘all of the evidence is to be considered in the light most favorable to the prosecution”) (emphasis in original). Favoring the prosecution with all inferences does not mean that we must ignore evidence that is in the record, but which they *910 ignore. So, for example, as mentioned in the facts section, the unidentified hair taken from the evidence bag enclosing the victim’s left hand as well as the unidentified hair from the pubic combing suggest a view of the events contrary to the one provided by the Commonwealth. [30] Having found the evidence insufficient to sustain the conviction, we have no need to reach the other grounds of appeal.