delivered the opinion of the court.

So far as the proposition that a juror cannot testify to sustain his verdict is concerned, the modern and better doctrine, that we have heretofore recently announced, is that he may testify in support of his verdict. An examination of the authorities will disclose practical unanimity in the modern holdings on this point. There is, however, a fatal defect in this record of conviction, which is this: The state failed to meet the prima facie showing made by the defendant that there had been opportunity to communicate improperly with the jury. Mrs. Jane Wilkinson and Mrs. Adaline Dunn testified, on the motion for a new trial, that they “saw a person not a member of the jury and not a bailiff in conversation with the jury, at least one member thereof.” J. L. Jones testified that “on the evening of the 12th of April the jury was completed in the case of State v. Nunnery, and placed in charge of two bailiffs; that they were kept in the courtroom and one of the small jury rooms during the night of the 12th of April; that the prisoner, Nunnery, was kept in the other small room, and the guards were in the large courtroom, where the jury was. And affiant knows that on the night of the 12th of April seventeen men slept or stayed in the same room, twelve of whom composed the jury trying the defendant, the other five being two bailiffs to the jury and the three guards of the prisoner, Nunnery; that one of said guards that stayed in the room with the jury was a son-in-law of the deceased’s brother, the said brother and the father-in-law of the said guard taking a very prominent and active part in the prosecution of defendant. Affiant further says that he knows that some of said jury had communication and conversation with some of the *546guards of said prisoner, but the extent and character of which he does not know. He does know that the guards, or some of them, talked with the jury, or some of them, but he does not know the extent or character of the conversation. He further says that there was opportunity to carry on almost any kind of conversation desired; that said jurors separated, so that some of them remained in the small jury room, and others came out where the guards were, and held conversation with them, at least to some extent.” The defendant, Nunnery, testified that “he saw some person, unknown to him, and who was standing in the dark under a cedar tree within a few feet of the courthouse door, approach some one of the jury, and have some words with such juror, as they passed by the cedar tree; that he does not know who the party was nor which of the jury he spoke to; that there were parties in the courtroom the night before the jury rendered their verdict, while the jury were in the room, and had some talk with some member or members of said jury; that the night before the verdict was i*endered he was under guard in one of the jury rooms opening into the courtroom, where the jury was, and that some of the guards with him left the room he was in, entered the jury room, and remained there some length of time. The door was open a part of the time.”

The learned judge held that a prima facie case was made as to opportunity for improper communication with the jury, and that the state would have to rebut the showing, or a new trial would be granted. He then continued the case until the next term of the court. At that term the state introduced only three of the twelve jurors, with the two bailiffs who had the jury in charge, the circuit clerk, and certain other witnesses. But it is plain that the showing made by the state does not sufficiently meet the case made by the defendant. The juror McGee, by his testimony, practically confirms the testimony for the defendant as to the number of parties who were in the courtroom, where the jury were during their deliberations; indeed, this particular *547fact, as to the guards of the prisoner, and other persons besides the guards, being in the courtroom while the jury was deliberating, is practically conceded — at any rate, indisputably established. And the rebuttal testimony introduced by the state, given the fullest possible effect, fails to show that the testimony of J. L. Jones as to the conversations had with the jury is not true. All the jury manifestly should have been examined under the peculiar facts of this case. It was easy to have done this, and overthrow — if the rest of the jury had so testified — the prima facie case made by the defendant. The three jurors that did testify do not pretend to say more than that they were not talked to. Non constat but some one or more of the other nine jurors were talked to, as' Jones testified. On the fundamental proposition that the purity of the verdict of the jury must be inflexibly maintained, this court has spoken with but one voice from the beginning until now. Chief Justice Smith, in Pope & Jacobs v. State, 36 Miss., 136, has stated the rule succinctly and clearly, as follows: “And where facts are established which show that improper influence might have been brought to bear upon the jury, and there is no opposing testimony which negatives the presumption thus created, according to the settled rule of this court, the verdict will be deemed vicious. Hare’s case, 4 How., 192; McCann’s case, 9 Smed. & M., 469; Organ’s case, 26 Miss., 78; Ned & Taylor’s case, 33 Miss., 370. The rule applicable in all cases of this character, and which has been uniformly recognized in this court, is laid down in McCann’s case, above quoted, as follows: ‘The evil to be guarded against is improper influence; and where an exposure to such influence is shown, and it is not shown that it failed of effect, then the presumption is against the purity of the verdict.’ ”

It is perfectly manifest, from the careful and painstaking examination which we have given to the testimony introduced by the defendant and for the state on the motion for a new trial, that the state failed to meet the burden imposed.

The judgment is reversed, and the cause remanded.