ANDERSON, Justice, for the Court:

On October 17, 1985, a Harrison County jury found the defendant/appellant, James Robert Steele, guilty of capital murder in the death of Christina Renee Sinclair, a twenty-three month old child. See, Sections 97-3-19(2)(f) and 97-5-39(2), Mississippi Code Annotated (Supp. 1988). The next day, after a separate hearing, the jury sentenced Steele to life imprisonment. Steele's Motion for New Trial or J.N.O.V. was overruled on October 31, 1986. In reversing and discharging the defendant, we address two of Steele's seven assignments of error.

I. FACTS

Steele joined the Navy in December 1981, and arrived at the Construction Battalion Center (Seabee Base) in Gulfport on May 7, 1983. Sometime in May or June 1984, he and Kathy Sinclair (Sinclair) met at a church softball game and began dating. Sinclair had a young daughter named Christina Renee Sinclair (Christina).

Because Sinclair was in the process of a divorce, she and Christina lived with her parents, Mary and Charles Red, in Long Beach, Ms. For approximately one month before Christina's injury, Christina spent her days at a day care center, while her mother went to work in the parts department of a car dealership in Gulfport. As Steele's and Sinclair's relationship developed, Steele would sometimes pick Christina up from day care and take her to the beach, where Sinclair would join them when she got off work around 5:30 or 6 p.m. In addition, Steele stated he had stayed with or babysat Christina ten or fifteen times over the course of his four-month relationship with Sinclair. Usually, he babysat when the Reds and Sinclair were going to church or choir practice.

On October 10, 1984, Sinclair picked up Christina at day care around 5:15 p.m. Sinclair fed and bathed Christina but did not wash Christina's hair. By arrangement, Steele arrived at the Red's home around 6:30 p.m. Sinclair was going to join her parents at choir practice, so Steele offered to babysit. When Sinclair left, shortly before 8 p.m., Christina was already dressed for bed. She was sitting on the floor of the den, playing with a new tea set. Steele was going to watch the World Series. The following is Steele's account, which account Steele first gave on the night of the incident, and continued to give throughout the investigation and proceedings.

Christina, who was being toilet-trained let Steele know she had to go to the bathroom. Steele took her to the bathroom, removed her disposable diaper, seated her on her trainer seat and returned to the ballgame. Christina came back to the den a few minutes later and Steele went to find her diapers. They were damp and he could not find anymore, so he put a pair of her panties on her. Christina then said, "rock" and Steele rocked her while watching the game. After she fell asleep, Steele put her in her bed under the covers.

Fifteen or twenty minutes later, Christina came back to the den and wanted to be rocked again. Steele, intent on the game, did not notice when she fell asleep, but put her back in bed sometime later. Three to six minutes later, he thought he heard her whine. He went to check and she was lying on the floor, more of less parallel to the bed, with her head facing the head of the bed, and with her eyes wide open. Steele picked her up and as he put her back in bed, told her to go to sleep. He thought he placed her on her stomach and on top of the covers this second time. Two to five minutes later he thought he heard her again. This time, the noise was unusual and he could not describe it.

He found Christina in bed. However, she was on her back, with her eyes rolled up in her head and she was gagging. Steele put his finger in her mouth to see if she had swallowed her tongue or if she had something lodged in her throat. Finding nothing, he "became really scared" and did not know what to do. He picked her up and rushed out the front door. When he picked her up, she appeared to pass out. He noticed that if he kept her moving or shook her, "she would stay half-way revived". Steele ran to a neighbor's house and let himself in their front door. The man of the house took Christina and the woman called an ambulance. Steele stated he told the man he had to keep her moving or she would stop breathing.

Christina's mother returned from choir practice just after the ambulance arrived. Steele gave Christina to a medic and went inside the Red's home to get Sinclair. He told her Christina had fallen out of bed and was hurt. As they were going back outside the Reds arrived and everyone went to the hospital. Steele's statement was introduced by the state as part of its case-in-chief. At all times, Steele's only explanation for Christina's injury was that she must have fallen out of bed.

X-rays and a CAT scan were performed immediately. The x-rays showed massive fracturing on the right side of the skull. The CAT scan showed the fracturing on the right side of the skull and swelling of the right hemisphere of the brain which was pushing the right hemisphere over into the space occupied by the left hemisphere. Christina was placed in an intensive care unit. Her vital signs were monitored and a pediatric nurse was with her at all times until October 13, 1984, at which time adult intensive care nurses took over, until her death on October 23, 1984.

Kim Hammond, who had graduated from nursing school three and one-half years prior to trial and had been a registered nurse in pediatrics for three years at the time of trial, testified for the state. She testified that on the evening of October 12, 1984, two nights after Christina entered the hospital, she noticed several uniform lines on Christina's ears and a grid-like patterned burn on Christina's left buttock. Both areas of burns, the ears and buttock, were scabbing. The scabs on the ears were dry and could be flaked off with a fingernail. The burn on the left buttock was deeper and the scab would have bled if picked. At trial, Hammond gave her expert opinion that the burns were two to four days old on October 12, 1984. The burns, then, could have been inflicted as early as October 8, 1984, or as late as October 10, 1984.

Kathy Sinclair was in the room and Hammond asked her if she had noticed the marks. Sinclair said she had not seen the marks before and, particularly, had not seen them when she bathed Christina on October 10. Charles Red testified that a nurse showed him the burns and told him that one of the treating physicians, Dr. Reeves, felt that the burns were caused by a hair dryer or a floor furnace. There is no floor furnace in the Red's house.

Both Charles and Mary Red testified that they had seen no marks until they were pointed out at the hospital. Mary Red testified that she and Christina had bathed together on the evening of October 9, 1984, and that she had not noticed anything unusual. Mary Red also stated that she washed and dried Christina's hair that night. The hair dryer worked fine. On cross-examination, Mary stated that on October 9, 1984, the hair dryer would run, but it would "hang up". Both the Reds and Sinclair stated that they had not accidentally or otherwise burned Christina.

Steele testified that he had never bathed or dressed Christina and on October 10, 1984, she was already bathed and dressed for bed when he arrived. He had, however, dried her hair a couple of times, but he stated the last time was two or three months prior to October 10, 1984. Steele stated that he had not seen any marks on Christina and had not burned her or injured her in any way. On cross examination, he stated that he had not and would not have seen the burns on her ears because she had long hair, and that he had not and would not have seen the burns on her buttock when he changed her panties because she was facing him. Steele also testified that he had not burned Christina.

Upon learning the doctor's theory of the cause of the burns, Charles Red went home and conducted experiments with the hair dryers he found in his bathroom and the bathroom that Mary Red and Kathy and Christina used. He testified that he did not mention his experiments to anyone until after Christina was declared legally dead on October 23, 1984. Unfortunately, Red did not conduct his experiment on the hair dryer that had caused the burns because Kathy Sinclair had taken the dryer from the bathroom and put it in a brown paper trashbag in her bedroom.

Sinclair testified that on October 11, 1984, she left the hospital and went home to shower and freshen up. The hairdryer was broken; it would run, but it would not blow hot air. When she returned to the hospital she told her mother about it and her mother told her to use an extra dryer that was in the cabinet in Mr. Red's bathroom. A few days later, she put the hair dryer in the trashbag in her bedroom. Sinclair also stated her father did not tell her about his experiments until after October 23, 1984.

Apparently, on October 25, 1984, Charles Red and one of his sons went to their lawyers to discuss the case. The hair dryer experiments were discussed. When they returned home, Mary Red, Kathy Sinclair and Charles Red were in the living room discussing the case and the hair dryer experiments were again mentioned. Kathy kept trying to interrupt her father and finally told him he had tested the wrong dryer. Later that day, Kathy went with her lawyers to the Long Beach Police Department and turned over the hair dryer.

Roger Morrison, a criminal forensic scientist with the Alabama Department of Forensic Sciences, testified for the state about the hair dryer. He stated that the hair dryer worked when he plugged it in. The blower would jam when the hair dryer was held in some positions. This was due to a broken blower vein, and when the blower vein was removed, the dryer worked "quite well." Morrison stated that the hair dryer had a maximum temperature of 194 degrees Fahrenheit and that even working properly, this temperature would burn someone's skin. If the dryer were pressed against an object, such as skin or cloth, the air flow would be restricted and the temperature would quickly rise to 282 degrees F., at which point a switch in the dryer would shut it off. In a restricted air flow situation, the wire grid itself would go up to between 275 degrees and 300 degrees F. Finally, Morrison stated that wet hair would not burn because the evaporation as the hair is drying would keep the hair cool. There was no evidence that Christina's hair had been burned.

John Kilborn, another scientist with the Alabama Department of Forensic Sciences, testified that the metal grid on the hair dryer contained small samples of skin and connective tissue. There was not enough skin or tissue present to determine whose skin or tissue it was. Finally, Kilborn lifted one partial print from the hair dryer, but did not himself do a comparison. An investigator with the Long Beach Police Department later testified that the print was unidentifiable.

The state's case against Steele necessarily hinged on expert medical testimony because the evidence was entirely circumstantial. Dr. David Reeves, Christina's pediatrician; Dr. Victor Bazzone, a Gulfport neurosurgeon; Dr. Byron Brogdon, a radiologist and a professor at the University of South Alabama Medical School; Dr. Thomas Bennett, Mississippi's state medical examiner; and Dr. Paul McGary, a pathologist with the New Orleans' coroner's office and LSU medical school professor of pathology; testified for the state. Reeves and Bazzone were the only treating physicians to testify for the state. Brogdon and Bennett never saw Christina, but reviewed the entire medical record. McGary examined Christina's brain, which was sent to him after an autopsy was completed in Alabama. He also reviewed some, but not all of the CAT scans performed on Christina while she was in the hospital.

Dr. Harry Albert Danielson, a Gulfport neurosurgeon, and Dr. Richard Lindenberg, a neuropathologist and forensic neuropathologist, testified for the defense. Dr. Danielson examined Christina on October 20, 1984, because Dr. Bazzone was out of town. He also reviewed all of the medical records and looked in on Christina from time to time. Dr. Lindenberg, who received his medical education in Germany just prior to World War II and was invited to the United States in 1947 to continue his neuropathology work for the U.S. Government, was not a treating physician. He based his conclusions in this case on the autopsy report; Dr. McGary's brain notes and pathology report; a slice of the brain; sixteen microscopic tissue sections of the brain; autopsy photographs; the complete hospital records; and Steele's statements.

Christina's bed was twenty-two inches off the floor. The floor was concrete covered with foam rubber padding and shag carpeting, which combined totalled one-half to three-quarters of an inch thick. Christina was two feet, eight and one-half inches tall. If she had been standing on the bed, she would have fallen 54 1/2 inches to the floor. When first examined in the emergency room at Gulfport Memorial Hospital, Reeves believed that Christina had sustained two or three separate skull fractures, the autopsy revealed that it was one fracture. It was uncontroverted that this was not a depressed skull fracture, but rather a linear fracture consistent with the head hitting a large flat surface and not consistent with the head being hit with a small object such as a hammer or blow dryer.

As will be recalled, the October 10, 1984, CAT scan showed severe swelling of the right hemisphere of the brain. This CAT scan also indicated that there was no pressure on the mid-brain, a lower portion of the brain. A CAT scan done on October 11, 1984, showed no change in the swelling and still no pressure on the mid-brain. During the night from October 12 to October 13, 1984, something happened to Christina. Around midnight, there was a sharp rise in her blood pressure and pulse. Prior to this, Christina had fits of restlessness and her pupils responded to light. While she would not respond to verbal stimuli, she did respond to the pain of a pin prick. After the rise in blood pressure, Christina stopped moving, her pupils were dilated and unresponsive to light and she did not respond to any stimuli. On October 13, 1984, a third CAT scan was performed. This CAT scan indicated that the swelling had increased and pressure was being exerted on the mid-brain. The basal cisterns surrounding the mid-brain were obliterated by swollen brain tissue.

All of the state's medical experts were of the opinion, based on a reasonable degree of medical probability, that Christina's fracture and brain injury were inconsistent with Steele's account. They opined that the injury was too severe to have been sustained by a fall from the bed from either a prone or standing position. The fact that the autopsy indicated there was no subdural hemorrhaging did not affect their opinion. Dr. Reeves testified that the fact that there was only one massive fracture did not change his opinion that the injury could not have occurred the way Steele said it did. All of the experts, both those for the defense and the state, testified that Christina died from brain swelling which put pressure on the lower part of her brain, the mid-brain, which is responsible for respiration and heart beat.

Dr. Bazzone, who testified for the state, was of the opinion that Christina could have received a skull fracture by falling from a standing position on the bed and pitching backwards. However, Bazzone stated that while it was possible, it was improbable that a fracture from such a fall would be as severe as that sustained by Christina.

Dr. Danielson, a neurosurgeon and the only treating physician to testify for the defense, stated that Christina's skull fracture and resulting brain injury were consistent with Steele's story, but in his opinion and based on the severity of the fracture, Christina had to have been standing in bed when she fell. With that qualification, he maintained his opinion, based on a reasonable medical probability, that Christina's injury could have been sustained by a fall from the bed.

Dr. Lindenberg testified that, based on his experience, study of the case, and a reasonable medical probability, in his opinion Christina's fracture was consistent with Steele's account. Lindenberg explained that when a person falls backwards and hits the back part of the skull, that area momentarily flattens out and results in a larger impact area. When the back of the head flattens, then the top and side of the skull must bulge. This creates a pulling force that pulls or tears the bone apart. Lindenberg then stated that studies indicate that the fracture which occurs in such a fall, actually starts on the top and below the area on impact and the fracture lines then run back to the point of impact.

Lindenberg also testified that if the fall had been a severe fall, with a lot of power, she would have "contused the anterior part, the forehead, the temporal lobe" and visible bleeding spots or bruises would have shown up in the front part of the brain. The autopsy report indicates that there were no such bruises and there were no bruises present on the CAT scans. Finally, on this point, Lindenberg testified that if someone threw or slammed a child to the floor, the frontal contusions would be apparent.

Furthermore, Lindenberg attributed the brain swelling not to trauma from impact, but to an interference with the blood's flow out of the brain. He noted that the fracture crossed the right sigmoid sinus located at the base of the brain on the right side. These sinuses located on each side of the brain are responsible for getting blood out of the brain. In his opinion, the brain swelling was caused by the occlusion of the right sinus when the fracture crossed it. On cross examination Lindenberg insisted that there would have been frontal contusions if the child had been forcefully thrown to the floor instead of falling out of bed.

The state called Dr. Paul McGary, the New Orleans pathologist, to testify in rebuttal. McGary testified that the fact that one sigmoid sinus might have been occluded or clogged by the fracture at the moment of impact is insignificant and would not have caused swelling because there are two sigmoid sinuses and they both drain from both sides of the brain, not one from one side and the other from the other side. McGary concluded that the brain immediately began to swell due to trauma and continued swelling and ultimately caused Christina's death. During cross examination it developed that McGary had seen the CAT scans from October 10 and October 13, but had not seen the October 11, 1984, CAT scan.

On surrebuttal, Lindenberg reiterated that occlusion of one sigmoid sinus would cause swelling to that side of the brain. He also stated that McGary's conclusions were inaccurate and "sloppy" because McGary had not seen the CAT scan taken on October, 11, 1984.

II.

THE TRIAL COURT ERRED IN DENYING THE DEFENDANT'S MOTION FOR JUDGMENT OF ACQUITTAL NOTWITHSTANDING THE VERDICT.

The corpus delicti in a homicide case consists of (1) the death of a human being, and (2) a criminal agency causing the death. May v. State, 524 So.2d 957, 966 (Miss. 1988) and cases cited therein. The state may prove the crime (corpus delicti) by circumstantial evidence, but where the case is based wholly on circumstantial evidence, the state must prove the defendant's guilt beyond a reasonable doubt and to the exclusion of every reasonable hypothesis consistent with innocence. Leflore v. State, 535 So.2d 68, 70 (Miss. 1988); Montgomery v. State, 515 So.2d 845, 848 (Miss. 1987); and Westbrook v. State, 202 Miss. 426, 32 So.2d 251, 251 (1947).

The defendant's position is that he should have been granted a judgment of acquittal notwithstanding the verdict because the state failed to prove criminal agency beyond a reasonable doubt and to the exclusion of every reasonable hypothesis consistent with innocence. The state rebuts by arguing that the jury was properly instructed on circumstantial evidence, that the circumstantial evidence was sufficient, and, therefore, the defendant was not entitled to a judgment of acquittal notwithstanding the verdict. The standard used in addressing a motion for jnov was clearly set out by this Court in Pharr v. State, 465 So.2d 294 (Miss. 1984):

The motion for judgment of acquittal notwithstanding the verdict tests the legal sufficiency of the evidence supporting the verdict of guilty. It is a renewal of the defendant's request for a peremptory instruction made at the close of all the evidence. It asks the court to hold, as a matter of law, that the verdict may not stand.

Where a defendant has moved for j.n.o.v., the trial court must consider all of the evidence — not just the evidence which supports the state's case — in the light most favorable to the state. May v. State, 460 So.2d 778, 781 (Miss. 1984). The state must be given the benefit of all favorable inferences that may reasonably be drawn from the evidence. Glass v. State, 278 So.2d 384, 386 (Miss. 1973). If the facts and inferences so considered point in favor of the defendant with sufficient force that reasonable men could not have found beyond a reasonable doubt that the defendant was guilty, granting the motion is required. On the other hand, if there is substantial evidence opposed to the motion, that is, evidence of such quality and weight, that, having in mind the beyond a reasonable doubt burden of proof standard, reasonable fair-minded men in the exercise of impartial judgment might reach different conclusions, the motion should be denied and the jury's verdict allowed to stand. May v. State, 460 So.2d 778, 781 (Miss. 1984).

In other words, once the jury has returned a verdict of guilty in a criminal case, we are not at liberty to direct that the defendant be discharged, short of a conclusion on our part that the evidence, taken in the light most favorable to the verdict, no reasonable, hypothetical juror could find beyond a reasonable doubt that the defendant was guilty. [citations omitted].

Pharr, 465 So.2d at 301.

The question before the Court, then, is whether the state met its burden of proving criminal agency, such that a hypothetical, reasonable juror could find Steele guilty beyond a reasonable doubt and to the exclusion of every reasonable hypothesis consistent with innocence. As we have often noted:

"It is fundamental that convictions of crime cannot be sustained on proof which amounts to no more than a possibility or even when it amounts to a probability, but it must rise to the height which will exclude every reasonable doubt; that when in any essential respect the state relies on circumstantial evidence, it must be such as to exclude every other reasonable hypothesis than that the contention of the state is true, and that throughout the burden of proof is on the state. It is our duty here to maintain these principles."

Hester v. State, 463 So.2d 1087, 1093 (Miss. 1985) and Hemphill v. State, 304 So.2d 654, 655 (Miss. 1974) quoting Westbrook, 32 So.2d at 252.

Accepting the evidence in the light most favorable to the state, including all reasonable, favorable inferences, as we must, the circumstances support the state's hypothesis. The circumstances, however, do not exclude the reasonable hypothesis that Christina's injury was an accident, the exact cause of which remains unknown.

"It is always insufficient where assuming all to be proved which the evidence tends to prove, some other hypothesis may still be true, for it is the actual exclusion of every other hypothesis which invests mere circumstances with the force of truth. Whenever, therefore, the evidence leaves it indifferent which of several hypotheses is true, or merely establishes some finite probability in favor of one hypothesis rather than another, such existence cannot amount to proof, however great the probability may be." (Emphasis added).

Hester, 463 So.2d at 1094, quoting Algheri v. State, 25 Miss. 584, 589 (1853).

The problem in this case is that the evidence "merely establishes some finite probability in favor of" the state's hypothesis. The state proved that Christina was burned by a hair dryer on October 8, 9 or 10, 1984. According to the testimony, the defendant probably did it between 8:00 and 9:30 p.m. on October 10, 1984, because the Reds and Sinclair stated that they neither burned Christina, nor noticed any marks before October 12, 1984. According to the nurse's testimony, however, the fact remains that the burns could have been inflicted by someone other than Steele on the 8th or 9th of October.

The state also proved that Christina probably did not sustain her head injury by falling off of her bed from a lying or standing position. After eliciting this testimony from medical experts, the state introduced Steele's tape-recorded statement concerning the events of the night of October 10, 1984. Taken in the light most favorable to the state, Steele's statement does not add anything to the jury's pool of knowledge. Steele's statement indicates that he does not know what occurred in Christina's bedroom.

All of the favorable evidence considered, there is no evidence, satisfying the beyond a reasonable doubt and exclusion of every reasonable hypothesis consistent with innocence burden of proof, linking Steele with Christina's injury. The fact that Steele was alone with Christina for slightly more than an hour and a half does not give rise to a reasonable inference of guilt where it is merely more probable than not that Christina did not fall out of her bed from either one of the two positions about which the state's witnesses testified. For instance, the state did not exclude the possibility that Christina fell while jumping on her bed or climbing on her chest-of-drawers which was located twelve or thirteen inches from the head of her bed. The state's medical experts were not asked whether Christina could have sustained her head injury under those circumstances.

We are loath to set aside the jury's guilty verdict, but must do so in the this case. The evidence was legally insufficient to establish anything more than a probability of guilt and did not "invest mere circumstances with the force of truth." The conviction must, therefore, be reversed and the defendant discharged. We hasten to add that, in this opinion, we do not answer the question of whether medical probabilities can ever support a conviction based on circumstantial evidence of criminal agency. We simply hold that, on the facts of this case, the state's proof of criminal agency was so deficient that no reasonable hypothetical juror could have found, beyond a reasonable doubt and to the exclusion of every reasonable hypothesis consistent with innocence, that Steele killed Christina Sinclair.

In addressing the following assignment of error, we wish to emphasize a recent change in our criminal discovery rules.

III. THE NECESSITY OF DISCLOSING TO THE DEFENSE THE PRIOR STATEMENTS OF STATE WITNESSES

At various times prior to trial and during trial, defense counsel requested copies of prior statements of state witnesses. The trial court ruled that seven of eight prior statements contained no material inconsistencies and refused to turn them over to defense counsel. At the conclusion of the guilt phase, the lower court did supply defense counsel with all of the prior statements. The defendant then objected and moved for a mistrial, which motion was never ruled on, and later assigned this denial of the statements as a ground for new trial. In his bench opinion on the motion for new trial, the trial judge pointed to two examples, and ruled that the inconsistencies were not material, but rather, minor.

We have carefully reviewed the witnesses' prior statements and, while it is unnecessary at this juncture to go into any detail, we are convinced that the defense would have been materially advanced by access to those statements. We would only point out that our recent decisions in Porter v. State (Miss. No. 58111, dec'd February 15, 1989) (not yet reported) and Brock v. State, 530 So.2d 146 (Miss. 1988) make it clear that, under Rule 4.06 of the Uniform Criminal Rules of Circuit Court Practice, as amended by these decisions, all such statements, whether materially inconsistent or not, must, upon request, be turned over to the defense during discovery. The only exception to full disclosure must be based upon the trial court's determination that "`. . . there is a substantial risk to any person of physical harm, intimidation, . . ., resulting from such disclosure, which outweighs any usefulness of the disclosure to defense counsel.'" Porter, slip opinion at 3; quoting Rule 4.06, Unif.Crim.R.Cir.Ct. Prac.

REVERSED AND APPELLANT DISCHARGED.

ROY NOBLE LEE, C.J., HAWKINS and DAN M. LEE, P.JJ., and PRATHER, ROBERTSON, SULLIVAN and BLASS, JJ., concur.

PITTMAN, J., not participating.