Anderson, J.,
delivered the opinion of the conrt.
The appellant, Swinton Permenter, was convicted of mnrder and sentenced to hang, and appeals to this court. He was convicted on evidence wholly circumstantial, made up of many different facts and circumstances. There is grave doubt, from the record in this case of appellant’s guilt. The giving of the second instruction for the state is assigned as error. That instruction is as follows: “The court charges the jury that circumstantial evidence has been received in every age of the common law, and may arise so high in the scale of belief as to generate full and complete conviction in the minds of the jury of defendant’s guilt; and when it does arise so high in the scale of belief as to generate full conviction in the minds of the jury of defendant’s guilt beyond a reasonable doubt, then they are authorized to act upon it, and convict the defendant of the crime charged. ’ ’
This instruction was clearly erroneous. It is substantially the same instruction which was condemned by the court in Williams v. State, 95 Miss. 671, 49 South. 513. The court said of the instruction in that case: “It is elementary law that a conviction may be had on circumstantial evidence alone, when by it guilt is proven beyond a reasonable doubt; but it is also elementary that, before such evidence can be said to prove guilt beyond a reasonable doubt, it must exclude every other reasonable hypothesis than that of guilt.” The fatal defect in the instruction is that it authorizes the jury to convict on circumstantial evidence which shows guilt beyond a reasonable doubt, without going further and informing the jury that the evidence must be so strong as to exclude every other reasonable hypothesis than that of guilt; in other words, explaining what it takes to show guilt where the evidence is circumstantial. This addition to the instruction is made necessary by the inherent difference in direct and circumstantial evidence. That difference is well stated in Haywood v. State, 90 Miss. 461, 43 South. *462614, as follows: ‘ ‘ Circumstantial evidence is a different kind of evidence, wholly different from evidence consisting of the direct and positive testimony of eyewitnesses. In the one case, if the jury believe the testimony of the witnesses, the fact of the killing is established by their direct statements that they saw the party killed. In the other case, if the jury believe the party was killed beyond all reasonable doubt, they do so believe it from a chain of circumstances, the absence of any one link in which chain destroys the value of all other circumstances, no matter how absolutely proven. There has never been a better illustration of the weakness of circumstantial evidence than that which tells us that ‘it is no stronger than the weakest link in the chain. ’ This essential difference in the very nature of the two kinds of testimony, circumstantial and direct, is such and so clear that- it is 'not the law that circumstantial evidence is as good as any other kind of evidence. True enough, if, as stated, the circumstantial evidence excludes every other reasonable hypothesis than that of guilt beyond all reasonable doubt, the jury are just as much bound to convict as if the guilt had been shown by the direct evidence of eyewitnesses; but this does not alter the fact that the inherent nature of the two kinds of evidence is different, nor the other fact that the only thing which invests mere circumstances with force of proof is the absolute exclusion of every other reasonable hypothesis than that of guilt. . . „ The clear difference between circumstantial evidence and direct evidence is pointed out by Prof. Wigmore in the first volume of his work on Evidence (section 25), in a quotation from Wills on Circumstantial Evidence, cited with approval by him as follows: ‘ The different writers, ancient and modern, on the subject of evidence, have concurred in treating circumstantial evidence as inferior in cogency and effect to direct evidence, a conclusion which seems to follow necessarily from the very nature of the different kinds of evidence.’ ”
*463Proof of guilt hy circumstantial evidence consists in the proof of facts from which guilt is inferred. The fact of guilt is arrived at hy process of reasoning and deduction from the proven facts. Where direct evidence is relied on to establish guilt, only one step is taken, namely, the facts are proven, which must show guilt beyond a reasonable doubt. But where circumstantial evidence is relied on, another step must be taken. In addition to proving the facts beyond a reasonable doubt, those facts must be such as to exclude, beyond a reasonable doubt, -every other hypothesis than that of guilt. In other words, where circumstantial evidence is relied on, proof of the facts beyond a reasonable doifbt of itself proves nothing, unless the inference deducible from the facts so proven excludes beyond a reasonable doubt every other hypothesis than that of giult. The fault of the instruction is that it authorizes the jury to convict if the facts are proven beyond a reasonable doubt, wholly disregarding whether the inference of guilt drawn therefrom excludes beyond a reasonable doubt every other hypothesis. Under the facts of this case, an instruction properly defining circumstantial evidence would have been peculiarly applicable. On the other hand, the instruction under consideration, in view of the facts of the case, was in a marked degree inapplicable and misleading. The error in it is neither cured by any other single instruction, nor all of the instructions taken together. We refrain from further comment on the evidence, as there may be another trial of the ease. Suffice it to say, the instruction in question was calculated to materially influence the jury in rendering their verdict.
The other errors alleged are not well founded.
Reversed and remanded.