The prosecuting attorney of Greene county at the March term, 1911, of the criminal court of said county, filed an information charging the defendant'with murder in the first degree in having ' shot his wife to death with a shotgun at their home .on the morning of January 24,1911. Upon a trial defendant was convicted of murder in the first degree and his punishment assessed at imprisonment in the penitentiary for life.
The defendant with his family, consisting of his wife and two small children, the elder under the age of five years, occupied a frame dwelling on the county road near the town of Bassville, Greene county: The dwelling consisted of three rooms downstairs and one *113partitioned off upstairs, two other upstairs rooms being simply floored but otherwise unfinished and separated from each other only by open studding. A brick flue ran up through the center of the building and a narrow inclosed stairway from the first to the second story ran alongside the flue. Defendant was very poor, of humble station, ignorant, and of a low order of mentality. His vocation consisted in doing day’s labor at whatever he could turn his hand to. Like many of his class found in certain country districts he was addicted to hunting such small game as rabbits and partridges, and in consequence of this habit was possessed of a shotgun.
About four or five o’clock of the morning of January 24,1911, persons living in the same neighborhood as defendant were aroused by his hallooing, as he ran up and down the road, that his house was on fire, and that everything was burning up; the neighbors responded to his call and upon arriving at the scene found the upper part of the building in flames to such an extent that it was impossible to get upstairs. Several persons went into the lower rooms and carried out different articles of household furniture. It appears that the defendant, after having aroused the neighbors, ran back to his home where those arriving found him apparently in great distress, standing near or leaning on a hack or old vehicle in which were his little children covered up with bedclothes, the shotgun also being in the hack; upon being asked where his wife was, defendant said in a hysterical manner he knew, or he was sure, she was in the burning building .; after the flue fell such portions of her body as had not been entirely consumed by the fire were seen lying on the ruins of the flue. Her limbs were almost entirely burned off, as well as her head, the trunk of her body being very badly burned and charred. While the house was burning, frequent reports were heard as of *114gunshots which were explained, to have been the discharge of cartridges which defendant said he had kept to the number of twelve or fifteen hundred, for the purpose of sale, as well as for use in his shotgun; these cartridges were said to have been on the steps of the stairway leading to the upper rooms and in a large earthen churn which stood immediately under the upper part or head of the stairway. Among other testimony introduced on behalf of the State, it was shown that the defendant the afternoon before the fire had gone out hunting and had killed two or three rabbits which he had taken to a country store near by and sold, and with the proceeds had bought a small quantity of coal oil; that he also bought at the same time, with cash his father had paid him for a day’s wages, a 25-pound sack of flour, a can of baking powder, and some candy for the children. A witness named Russell Bass assisted the defendant in carrying the goods purchased part of the way towards his home, when he was met by his wife and children, his wife talcing the coal oil can, one of the children the baking powder can, and the smaller one the candy. When they reached their home the wife at once filled two lamps with coal oil, and a lantern which was used by the husband to give light at the wood pile while he cut up wood for the kitchen stove to enable her to cook supper. The filling of the lamps and the lantern took, defendant says, nearly all the oil he had bought. The testimony as to what occurred after defendant met his wife and children as he was on his way home from the store, and after he reached home, is confined wholly to his statements. He says it was his purpose the next day to' go to work for his father who lived in the neighborhood, and he and his wife talked the matter over after they had eaten supper, and agreed that it was best for her and •the children to go with him and stay at his father’s while he was employed there; that they retired early, .all sleeping in one bed; that he was awakened some*115time between twelve and three o’clock by the barking of neighbors’ dogs, when he arose and placed a stick of wood on the fire and went back to bed; he went to sleep at once and heard nothing more until sometime between four and five o’clock in the morning, when he was awakened by his wife calling him to get up, that the house was on fire. He arose hastily, put on a portion of his clothing, and he and his wife carried the children out and put them in the old hack, which stood •some distance from the building, taking out some bedclothes to wrap them in. They then drew some water from the well with which to put out the fire, and worked together in their efforts to accomplish this and thought they had gotten it under control when the defendant went upstairs and discovered it had broken out afresh; he and his wife then drew more water from the well, put it in a tub, and carried it upstairs, and threw it on the fire; finding their efforts unavailing and that it was dangerous to remain longer upstairs en account of the fire and suffocating smoke, the defendant called to his wife and told her they had better get out of the building, and for her to follow him. That the steam from the water and smoke from the fire had become sot dense he could not see his wife, but knew she was near him when he called to her. That he ran to a window and either raised it or kicked it out, he doesn’t remember which, and let himself down to the ground, it being his intention to pull the hack up under the window so that his wife might jump out on the feather-bed which they had carried out and placed in the hack. That the fall to the ground rendered him partially unconscious for a few minutes, and vdien he had come to himself he ran to the neighbors to give the alarm, not knowing whether his “wife had escaped from the building or not. His testimony as to what he did thereafter is in a measure corroborated by the neighbors who were aroused by his cries for help. He says he first went to the house of a neighbor named Davis, and *116appealed to Mm to come and help Mm, and then ran to the house of an uncle, Sampson Bass, and hallooed that Ms house was on fire and to come and help him put it out; others heard his cries and rushed to the scene; he then ran back to the house and found Ms little children crying and that he was so exhausted and overcome he could do nothing except stay with the children and lean on the hack for support. The witnesses all testified that his cries for help were those of one in distress, and that he evinced great excitement. When asked by Davis where his wife was, defendant said she was dead, and afterwards that she was carrying things out of the house. His inability to express himself clearly and consistently, even in a cool and collected condition, is evident from his testimony as found in the record. When the neighbors who had responded to his call reached the house, they found the upper part of it in flames, and after they had taken out such articles of household furniture as they were able to carry, the brick flue fell and the wife’s body was discovered on the ruins, with her limbs as well as her head partially burned up and the trunk of her body very much burned and charred. While the fire was in progress there was an almost continuous discharge of the cartridges which sounded, as the witnesses say, like a sham battle.
The theory of the State was that the defendant had taken the coal oil which he had bought the night before at B'assville, and after shooting his wife, had, distributed the oil over the interior of the house and furniture and had then set fire to the building with the intention of destroying the evidences of his crime. The purchase of the coal oil is the only circumstance in support of this theory. There was no odor of coal oil evident to any of the witnesses, and the can in which he had brought it home was found intact after thé fire except that the spouts were melted off, giving a color of truth to the defendant’s statement that his *117wife had emptied the can in filling two lamps and a lantern the evening before. The lantern was found almost filled with oil and was thus taken out of the burning building. There is no testimony as to the lamps.
About ten months before the wife’s death defendant came to one T. M. Hunter, a blacksmith, living in Springfield, and said that he and his wife had parted and that the wife was then in Springfield and that defendant had gone up to see the children and that she would not let him see them, but that he would get a gun and would see them; the blacksmith warned him not to do anything of that kind; and that defendant did nothing further than to make the remark; that a reconciliation was effected soon thereafter between the parties, and there was no subsequent differences between them. This fact was introduced by the State as showing a motive for the crime. There was evidence that the wife was several months advanced in pregnancy at the time of her death.
In further corroboration of the testimony of the defendant that his wife and he had endeavored to put out the fire, witnesses stated that upon going to the well while the house was burning they found the well rope and the well curb wet as though water had recently been drawn therefrom.
The remains of the wife were buried the next day after the fire, but were subsequently exhumed on two different occasions, the first about six days after the burial; upon the second exhumation the heart and others organs were removed and examined. These, oh account of having been subjected to great heat, had undergone marked physical changes, rendering it difficult for the physicians who made the autopsy to reach any definite conclusion as to the cause of the wife’s ■death or to indicate anything except that some shot were found in the cavity of the trunk and a few, as testified to by one of the witnesses, in the lower part *118of the pericardium, or the investing membrane of the, heart, there also being what appeared to be wounds in the heart. Tbe State’s contention is that these were due to a gunshot fired from a shotgun in the hands of the defendant; that of the defense being that, the shot came from exploding cartridges, and that the' wife’s death was due to suffocation or other causés. Iban having been shot by the defendant. There was. no evidence as to the condition of the shotgun which was found in the hack, where defendant said he had left it the night before the fire; whether the gun was-loaded or unloaded, or having been loaded had recently been discharged, does not appear; in fact, there is little or no evidence in regard to the gun, except its location. There was much evidence pro and con as to-the condition of the blood found in the heart and large vessels connecting therewith, in an attempt to show whether the shot found in the body entered it before or after death and as to whether the death was caused by a gunshot or suffocation, and as to the force of shot from cartridges when caused to explode by heat. All of this testimony, expert and otherwise, upon impartial analysis is found to be very unsatisfactory as affording reasons for a conclusion, one way or the other, as to the cause of the wife’s death.
Defendant in and out of court related substantially the same story he had told from the beginning. Discharged after two preliminary examinations, he went to Arkansas, not for the purpose it appears of flight, or to evade arrest, but acting upon the advice of relatives who desired to get him away from the scene of his trouble; he returned when he learned that an information had been filed against him.
Other facts not hereinbefore stated will, if necessary to an elucidation of the case, be set forth in their order in the opinion.
The court instructed the jury as to the formal requisites necessary to authorize the finding of defend*119ant guilty as charged; the technical words used in an indictment for murder in the first degree being defined according to approved definitions. The necessary presumption following the use of a deadly weapon was' elaborately explained and what was necessary for the jury to find to constitute the defendant guilty of murder in the first degree; the effect and the weight to be given the testimony of physicians and experts was explained in an instruction which has often received the approval of this court; that the presumption arising from statements made by the defendant against or for himself, together with the caution to be observed by the jury in considering such statements; the distinction between direct and circumstantial evidence was clearly made, and the manner in which each was to be considered was correctly defined; this was followed by an explanation of the weight to be given circumstantial evidence; also as to the manner in which the testimony of the defendant was to be considered generally on account of his being the defendant and on trial. The jury were further told that they were the judges of the credibility of the witnesses, and the weight to be given their testimony, and as to what they might take into consideration in weighing the testimony of such witnesses. The presumption of innocence of the defendant was reiterated and that it devolved on the State to prove his guilt beyond a reasonable doubt, and the latter was defined in an instruction. No substantial objection can be urged to the forms of the instructions given so far as they attempted to direct the jury as to the law if there was evidence sufficient to authorize a verdict of guilt. Twelve instructions asked by the defendant were refused by the court. They were in the main substantially the' same as the instructions given, but were expressed in different words, and the* defendant suffered no prejudice by reason of their refusal.
*120Experiments ■ Cartridges and Coai o¡i. I. In view of the importance of this case, involving as it does the' life imprisonment of the defendant, the manner pursued by thé State in the examination of witnesses who had experimented in the explosion of shells, deserves serious consideration. This examination consisted in putting a number of persons on the stand and examining them as to the effect of exploding loaded shells by heat, evidently for the purpose of refuting the theory it was anticipated the defendant would advance that the shots found in the wife’s body were from shells stored on the stairway and in an earthen churn under the stairway, and which were exploded by the heat when the dwelling was burning.
The general rule is that a non-expert witness will not be permitted to testify to the results of experiments made out of court, but that a witness who is an expert and has made experiments under conditions and circumstances as nearly similar as possible to those in the concrete case, may be permitted to state the result of his experiments made out of court. [Riggs v. Railroad, 216 Mo. 304; Underhill, Cr. Ev. (2 Ed.), sec. 233, p. 422.] Evidence based on experiments, however, should be received with the greatest caution. The cautions to be observed are that unless the experiments are shown to have been made under essentially the same conditions as in the concrete case, the tendency is to confuse and mislead rather than enlighten the jury. [2 Whar. Cr. Ev. (10 Ed.), sec. 783a, p. 1541; Daniels v. Stock, 130 Pac. 1. c. 1034.] Confining ourselves, therefore, for tbe time being, to the rules announced, the important matter to be first considered is whether or not the witnesses were experts, and second, as to whether or not the experiments were made under similar conditions and circumstances as when the shells were exploded in the burning building.
*121The first witness examined by the State, on this subject, was one D. A. Wiseman, a farmer living in the same neighborhood as the defendant, and who it seems on his own initiative determined to test the explosive force of shells. He says he bought a- box 'of shotgun shells, but doesn’t know whether they were the kind that were in the defendant’s house or not; that he tried them and that they wouldn’t shoot anything. This seems to have been the extent of his experience in the testing of shells.
D. J. Holland, who had for about two years been at work for the Remington Arms Company and the Union Metallic Cartridge Company, defines himself as a “shooter and salesman” or a demonstrator in rifles and shotguns, and states that he had thrown shells into a fire or where a fire had been built around them, and the only things that exploded were the primers; that the explosion would make no more noise than the burning of the powder; that if shells were put into a stone vessel and exploded from heat and fire around it, there would not be much, if any, penetration of the shot therefrom; that there is nothing to hold the shells after they start to burn; that you might take the wads off of a shell and hold it in your hand and when you released the wads it would loosen the binding string and the shot would have no penetrating force. Wit-, ness does not think that shells packed in a stone churn and exploded by the heat from a burning building would have any more penetrating force than if openly thrown into the fire.
Arthur Kilham stated that he was a “shooter and salesman” for the Dupont Powder Company; that about all he has ever done since he was eleven years old was to test shotgun shells, and that at different times he has placed shells in stoves and bonfires and places like that, and when so placed they would burn like wood; that a fire burning around them would cause no more explosion than if they were thrown into an *122open fire; that a shot’s force in a shell is due to its confinement.
Fred Wingo stated that he was in the general mercantile business and that he had made some experiments in exploding shotgun shells unconfined in a fire; that he had used up two boxes of shells, one of black powder and one of smokeless, by throwing them into a fire and by building a fire around them and some of the exploded shells flew out and struck some of the men that were standing around; that other experiments were made by setting the shells on their bases and raking fire over them, and that the shots therefrom did not penetrate a cardboard.
In none of these cases was it shown that the shells used were identical.or reasonably similar to those in the concrete case. No attempt was made to show, a proximate similarity of the one case to the other, either as to the character of the shells, whether they were of metal or paper, or the quantity of the powder, or the nature of the shot. The witnesses, other than the representatives of the Remington Arms Company and the Dupont Powder Company, show themselves by their statements to have had no experience whatever in testing the penetrating force of shot from exploded shells other than in the experiments of which they gave testimony. It is apparent that the “great caution” referred to by the eminent text-writers above and which has the support of many cases, was not observed in the selection of the experts or the admission of the testimony in regard to the experiments made by the witnesses and others afterwards examined along like lines. It would profit nothing to cite and discuss the large number of cases found in the reports of the courts of last resort in the United States and elsewhere, showing in many instances the utter unreliability, especially on account of dissimilar conditions, of testimony of this character. While it is true that evidence in regard to *123experimental tests rests largely in the discretion of the courts, we have found no case where they have exercised their discretion to permit the admission of testimony based upon such uncertain and dissimilar experiments as in the case at bar. There are other objections to this character of testimony than those stated. If sanctioned, it permits the State to select its experts and make its experiments to sustain any theory it may adopt, in the absence of the defendant or his counsel, and while it is held that this goes to the weight rather than the admissibility of the testimony (Moore v. State, 96 Tenn. 209) it is a dangerous power even when lodged in the fairest hands, and should be exercised with great care. Such evidence is at best an inference based upon an inference, and not upon a fact from which it is sought to have a jury find a verdict of guilty. We cannot-give our sanction to the admission of this character of testimony, especially under the facts in the case before us. A jury should not be left to decide a principal fact by making remote inferences from other facts having no connection with those in dispute.
As was said in effect in Jim v. State, 4 Hump. (Tenn.) 288: If the life of an individual is at stake upon an indictment for murder, the court should not permit a verdict to stand which has been obtained not by calm, deliberate examination of the proof, but by uncertain experiments, from which the witnesses have deduced. certain opinions but who may have no actual knowledge of the facts. And in State v. Sanders, 68 Mo. 202, one of the few cases to be found in our reports discussing this question, defendant’s counsel told the jury in his argument in a felony case, in regard to certain footprints, that the jury might try for themselves whether such worn-out boots as witnesses for the prosecution had described, would make such tracks as they had testified to, and upon its being shown that members of the jury without leave made the experi*124ment out of court, it was held to authorize a reversal of the judgment of conviction and the granting of a new trial, notwithstanding the fact that the experiment had been suggested by counsel for the defense. This subject is discussed incidentally in the Hyde Case, 234 Mo. 200, in regard to the testimony of a nurse, who,, after having testified as to her familiarity with cyanide of potash, was further permitted to smell a physician’s fingers which had been moistened with this drug and state if she recognized the odor. This as an experiment, was held to be prejudicial error.
Evidence of a similar character to that introduced by the State in regard to the shells was offered and admitted over defendant’s objections, in regard to the explosive effect of coal oil, if confined in a can and left in a building being consumed by fire; this testimony is subject to the same objections as that in regard to the shells, and to the additional objection that there was no evidence upon which to base the assumptions made in framing the hypothetical question put by the State in-regard to the explosive effect of coal oil. The only evidence in regard to the coal oil was that the evening before the fire the defendant had-brought home some coal oil which had been emptied by the wife out of the can into lamps and a lantern, leaving little, if any, in the can, and the lantern was found, as shown by uncontradicted testimony, .to have been full of oil and the lamps were not accounted for, having evidently been destroyed in the fire.
It furthermore appeared that there was no evidence either to the sense of smell or sight of any coal oil having been distributed or poured over articles in the rooms, which were entered and furniture, etc., taken therefrom by the neighbors after the upper part of the building was rendered inaccessible by the fire. The theory, therefore, that the defendant had poured oil over the furniture, etc., and then set fire to the building to destroy the evidence of his crime is based upon *125nothing except assumption. Its presentation to the jury was, in our opinion, prejudicial.
Motive. II. The prosecution in its effort to show a motive for the crime, introduced evidence of a temporary estrangement between the defendant and his wife almost ten months before the wife’s death and that- defendant had said to a witness named Hunter, during the estrangement, that unless his wife consented for him to see the children he (defendant) would “get a gun and go and see them.” The witness advised against this step and nothing more came of it; the parties soon thereafter became reconciled and lived happily together up to the time of the wife’s death.
This occurrence does not, in our opinion, tend to establish a motive for the murder, and the introduction of testimony in regard thereto on account of the remoteness of the event and the circumstances following it has little if any probative force and, as was aptly -said by Faris, J., in State v. Concelia, 250 Mo. 411, “"Where facts exist, as they do here, if there is no motive shown, no guilt can, with legal certainty, be attributed to the defendant.”
Sufficient •to Convict. III. There is a graver question involved in this case than the admission of prejudicial testimony, and that is whether there was upon the whole, sufficient evidence upon which to base a verdict of guilty.
The respondent contends that the -testimony introduced by the State established a motive on the part of the defendant for the ldlling of his wife; that the evidence establishing such motive, with all the other facts and circumstances, justified the court in .submitting the case to the jury, and when so submitted the fact of the defendant’s guilt or innocence was for the jury alone and its verdict should not be disturbed unless prejudicial error was committed during *126the trial. Oil the part of the defense it is contended that there was no snch proof of the corpus delicti as the law requires and that there was an utter absence of motive for the alleged crime and that all the facts and circumstances, if conceded to be true, did not authorize the submission of the case to the jury, and that the verdict and judgment were based upon conjecture and suspicion.
Appei"ate°f court. It is a rule often announced that this court will not in a criminal case weigh the evidence, that this is for the jury. The rule as thus stated is too general. What is really, meant is that where a rec- or(l discloses that there is a conflict between material testimony, the court will not undertake to decide which is of the greater probative force, because this is the exclusive province of the jury; but where the court finds upon an analysis of the testimony, which it must of necessity make in every case, that there is an absence of substantial evidence to convict, then the question becomes exclusively one for the court, which it must decide, as any other assignment of error. [State v. Donnington, 246 Mo. 1. c. 354; State v. Gordon, 199 Mo. 561; State v. Francis, 199 Mo. 671; State v. DeWitt, 191 Mo. 1. c. 58; State v. Lockhart, 188 Mo. 427; State v. Swisher, 186 Mo. 1; State v. Scott, 177 Mo. 1. c. 673; State v. Crabtree, 170 Mo. 642.]
Corpus Delicti. As preliminary to an examination of the evidence, it is proper to consider the proof of the corpus delicti; it involves two things: first, the criminal act, and, second, the agency of the'accused in its commission. [State v. Crabtree, 170 Mo. 1. c. 650; State v. Dickson, 78 Mo. 1. c. 447.] To state the matter concretely as applicable to the case at bar, it is requisite, first, that the deceased should be shown to have died from the effects of a wound; and, second, that the wound was unlawfully inflicted by the person charged. The corpus delicti, therefore, consists *127not merely of the objective crime, bnt also of the agency of the accused in the crime. In a homicide case the corpus delicti cannot be said to be established until it has been proved that the death was not caused by natural causes or by accident. The foregoing is the summary of much reasoning on this subject by many courts of last resort, and is to be found in other and more elaborate language in Dr. Wharton’s treatise on Homicide (3 Ed.), sec. 587.
The charred trunk of the wife’s body in this case, with the limbs and head partially destroyed by fire, was taken from the ruins of the burned building as soon as the fire would permit. As to the identification of the remains there seems to be no question.
Physicians who were examined as experts stated hesitatingly that there appeared to be a wound in the posterior portion of the trunk not perceptible on the outside, probably due to the charred condition of the body, and also wounds in the upper part of the heart, and a small number of shot in the pericardium, and about a spoonful of shot in the lower part of the abdominal cavity mixed with partly clotted and decomposed blood and portions of the viscera. This is the gist of the testimony that the wife’s death was due to a criminal act.
The agency of the accused in the crime is confined to the fact that he had a shotgun, and to the absence of evidence that the wife was shot by some one else, and the circumstance that he ran away from the burning building to alarm the neighbors and left his wife to be consumed by the flames.
If we proceed upon the assumption that the shot from the wife’s body was evidence of a criminal act, in order to connect the defendant therewith, there should be proof of some circumstances to indicate a motive for the act. “Motive,” says Ferriss, J., in State v. Hyde, 234 Mo. 1. c. 226, “in murder, is the impulse or purpose that induces.the murderer to kill *128Ms victim.” More briefly put, we would say that in homicide, motive is the mainspring wMch impels the murderer to commit the crime. It is well established that the presence of motive may be shown after proof of the corpus delicti to explain the character of the killing and to strengthen the probability that the accused was impelled thereby rather than by other causes, but where the corpus delicti has not been fully proved it is, in our opinion, equally permissible to show a lack of motive, as tending to lessen the force of whatever proof may have been made of the corpus delicti. Although the shot found in the wife’s body might justify the presumption that her death was caused by a ■criminal act, still if no express or direct motive can be attributed to the defendant for- the murder, and there is no other evidence to show his participation in the .same than the ownership of a shotgun, then such a train of incriminatory circumstances should be shown as to afford a reasonable ground upon which to base a motive for the commission of the crime by him. We are not unmindful of the rule as announced in State v. Dunn, 179 Mo. 95, and other similar cases, that it is not necessary for the State in a murder case to prove a motive for the crime, but this rule has no application where evidence of motive becomes necessary to a proof of the corpus delicti; and, it will be found in each of the cases where the rule was applied, that there was a complete proof of the body of the crime.
We find since making the foregoing -deduction limiting the application of the rule in regard to the proof of motive, a better and simpler statement of same by Underhill and Clark, 12 Cyc. 149, title £ £ Criminal Law, ’ ’ where it is said that “the existence or non-existence •of motive is immaterial, where the guilt of the accused is clearly established.” From which the converse' proposition must follow that where the guilt of the .accused is not so established, proof of motive becomes *129material. This conclusion finds support in State v. Crabtree, 170 Mo. 1. c. 651.
Whatever light may be thrown upon defendant’s life must be drawn from the testimony; from it motive for this crime, if it exists, must be found. His life was but that of the “short and simple annals of the poor,” always limited in area, meagre in events and filled with minor circumstances'of much sameness. Unlike those with ampler means and wider horizons, and as a consequence having more duties and cares to occupy their time and withdraw their minds from their home circles, defendant lived, even when at toil, very near to his wife and children. On the morrow which followed the fire, he was to go to his father’s who lived a few miles away, for a few days’ labor and his wife and children were to accompany him. When he worked for others in the neighborhood it had been their custom to accompany him, as was stated by a Mrs. Hollman for whom he labored at different times; and wdien engaged in the daily toil incident to the every day life of a poor man the wife would often assist him. The night before the fire he cut up some wood for the kitchen stove by the light of a lantern and his wife helped him to store it away. She was, therefore, a help to him instead of a burden, and he could not have desired to rid himself of her on this account. They had two small children which it is shown they both loved; their care necessarily fell most upon the mother and even selfishness would have prompted him to not destroy her on this account. His habits, poverty and the traits of his character delineated by those who knew him from boyhood, render the idea of “another woman in the ease” an absurdity. .For him to have killed her because she was pregnant and likely soon to be confined argues him to be a monster, and is in view of all of the other evidence in regard to him beyond even violent conjecture. WRen he was returning *130from the store the evening before the murder, she met him and helped him carry home the articles he had purchased; they talked pleasantly about the immediate affairs of their future the night before her death. The testimony of neighbors who usually in country districts know all that is going on in the vicinity, is to the effect that kindly relations existed between the defendant and his wife; he treated her, they say, “as men ordinarily do their wives.” He was possessed of no vicious habits; of an inert nature and of not more than mediocre mental ability he was naturally not of an aggressive or domineering disposition, and although a ne’er-do-well, his only vice, if it can be called such, was shiftlessness. There is nothing in this analysis of his life and character which, of course, is based on the testimony, to indicate that he would commit murder, and it is unreasonable to an unbiased mind that he would have murdered his wife under the circumstances contended for by the State. His conduct in leaving his wife in the burning building and running away to alarm the neighborhood may not unfairly be attributed to his lack of moral fibre or.physical courage, and the abject fear which overcame him when he discovered the fire could not be extinguished. The testimony of distinterested witnesses is that the curb and well rope were wet when they reached the premises, thus corroborating the contention of the defendant that he and his wife had been drawing water from the well with which to extinguish the fire. If they were trying to extinguish the fire, does it not follow that he did not start it for the purpose of destroying evidences of crime?
The testimony of all of the witnesses is that the house was burning from the top downwards or in other words that the fire started, as defendant states, in the second story, probably near the flue. If as contended by the State, the fire was started by the defendant after having shot his wife, to destroy the evidence of his *131crime, and it evidently was not started until sometime between four and five o ’clock in the morning, wbat was the wife, if not helping to extinguish the fire, doing in the second story at this hour, where defendant, if at all, must have shot her; and if she was not shot while in the second story, why would he shoot her while downstairs, and there is no evidence physical or otherwise that he did, then carry the body upstairs and start the fire there, when as every school boy knows, whether he has any knowledge of physics or not, that a fire started from below or at the bottom of any inflammable material burns more rapidly and completely than when ignited from the top. If it is possible to conceive that defendant killed his wife, and then carried her body upstairs, is it not most reasonable to conclude that having removed his children, he would have set fire to the building in one of the lower rooms where combustion would have been rapid and complete, and if his alarm was a mere ruse, the building would have been so completely enveloped with flames when the neighbors arrived that all evidence of his crime would have been obliterated; instead, however, the lower story was still so accessible or free from fire when the neighbors came that they were enabled to carry many articles of furniture and bedding therefrom, and if there was evidence of the murder the record makes no mention of same.
Much as defendant may be censured for having failed to do what the man of ordinary manhood would have deemed his highest duty, namely, to save his wife from the flames, his failure to do so is not a circumstance sufficient to indicate that he is guilty of murder. The State’s ease, therefore, as it seems to us, outside of the facts we have enumerated, is based wholly upon conjecture bom and nurtured in that abhorrence normally felt for a .creature who becomes so possessed with fear that he will abandon his wife probably to a horrible death in an attempt to save himself. Taking *132into consideration the delicate condition of the wife at the time of the fire, and the terrible physical effort she must have made to assist him in extinguishing it, the conclusion is not unreasonable that when they reached the second floor of their dwelling and had ■emptied the tub of water, when as he says the smoke became so dense he could not see her, she may have been suffocated or rendered unconscious, and, therefore unable to move before her body was burned; the :shot in'her body may have been from the cartridges, ■of which there were many on the stairstep^ and three ■or four gallons of same in the stone churn underneath where her body was evidently lying before the flue fell, u,nd which were discharged by the heat with such rapidity that those present likened the reports to a sham battle.
There is nothing to indicate that the reports heard by the witnesses during the fire could have been produced by anything else than the exploding cartridges. If exploded by a force sufficient to make the reports described, the shot or the shells, if metal, dependent upon whether they rested upon the primer or shot ends of the cartridges, must have been driven somewhere, and as the wife’s body was immediately over the stone churn in which there were hundreds of shells, many of them metal, it is not unreasonable that the body was penetrated by some of the shot therefrom. This reasoning involves two assumptions: the first is that the reports indicated confinement of the cartridges in a manner entirely different from that testified to by the witnesses of the State in their experiments, and, the second, that when so confined and exploded they would be productive of sufficient force to carry the shell or shot a considerable distance; if not, there would have been no reports — the testimony of the only expert witnesses introduced by the State was that unconfined cartridges make no reports, therefore, confined car*133tridges do. This reasoning, however, is only made in passing, so to speak, because although the wife’s body may have been penetrated by shot other than from the shells exploded by the heat, there is nothing, absolutely nothing, which can be dignified with the title of evidence, to connect the defendant therewith.
A fair but not infallible test as to the sufficiency of facts necessary to convict in a case based wholly on circumstantial evidence may be made by comparing such evidence with that of other cases possessing similar features and involving kindred crimes; but such a course is, in our opinion, not necessary here; the facts are few and simple, and the circumstances are all within the range of average human experience. Contrast or comparison with other cases would, as a consequence, tend more to confuse than enlighten.
We have, therefore, refrained from discussing the facts by contrast or comparison in such cases as State v. Francis, 199 Mo. 671; State v. Gordon, 199 Mo. 561; State v. Nesenhener, 164 Mo. 461; and State v. Crabtree, 170 Mo. 642, in which the evidence was wholly circumstantial, and as we think much stronger than in the case at bar, yet the court found it in each case insufficient to sustain the conviction.
While the reasoning leading to the conclusion we have reached in this case is largely analogous, it is based upon as broad a foundation of fact as that upon which the jury found the defendant guilty. We do not have to rely, however, upon analogous reasoning or weigh it with that to the contrary, to determine the guilt or innocence of defendant; we are authorized and it is our duty to confine ourselves to the actual facts to determine the defendant’s connection with the crime; giving all of these their full force, irrespective of any other cause that may have occasioned the wife’s death, or in other words, weighing them at their full worth as evidence, it is our opinion they are not suffi*134ciently ample and convincing to base a verdict thereon of guilty, and we so hold.
The judgment should, therefore, be reversed, and the defendant discharged. It is so ordered.
All con■cur; Faris and Brown, JJ., in result.