Mr. JUSTICE WILSON delivered the opinion of the court:

On April 2, 1935, an election was held for the seven offices of assistant supervisors of the town of Joliet, in Will county. The candidates of the Democratic party were James G. Meade, Frank C. Butala, Edwin F. McHugh, Louis Moroni, Francis M. Johnson, Michael M. Kowalski and Michael Chizmarik. The Republican candidates were John T. Vavrek, David O. Carter, John J. Rogers, Fred Hedstrom, Antone D. Wicevic, Fred W. Smith and William Penrose. The official canvass of the returns from the fifty-three election precincts disclosed that five Democrats, Meade, Moroni, McHugh, Johnson and Butala, and two Republicans, Vavrek and Carter, were elected. They qualified and assumed the duties of their offices. The appellants, Rogers, (now deceased,) Hedstrom, Wicevic, Smith and Penrose, filed a petition in the county court of Will county to contest the election. Eight respondents answered the petition. After hearing evidence and a re-count of all the ballots, the court found, among other things, that unauthorized persons had been afforded access to the ballots of the twenty-ninth precinct; that the seal on the sack containing these ballots had been tampered with, and that the returns made by the officers of election were unimpeached, and, as evidence of the result of the election, should prevail over the ballots. The differences on the re-count from the canvass in the other precincts were inconsequential and did not affect the result. The votes of each of the fourteen candidates on the re-count of the twenty-ninth precinct were subtracted from the total received on the recount of all the precincts, and the votes received by each candidate in the twenty-ninth precinct, as shown by the official returns, added thereto. The court found that Meade, Moroni, McHugh, Vavrek, Johnson, Butala and Carter received the highest number of votes in the order stated and they were declared legally elected to the offices of assistant supervisors. The appellants prosecute this appeal from the part of the judgment finding that the appellees, Meade, Moroni, McHugh, Johnson and Butala, were elected supervisors. A cross-appeal by Chizmarik and Kowalski from the part of the judgment finding Vavrek and Carter to have been elected has been abandoned.

The controlling question presented for decision is whether the trial court erred in excluding the re-count of the ballots of the twenty-ninth precinct and substituting therefor the returns of the election officials. If the ballots were not properly preserved and the returns are used there will be no change in the personnel shown to have been elected both by the official canvass and the election contest. On the other hand, if the ballots were preserved in substantial compliance with the provisions of the Ballot law the re-count should prevail over the returns, and, in consequence, Rogers and Hedstrom would be among the seven elected to the offices of assistant supervisors instead of Johnson and Butala. It is conceded that the ballots from the twenty-ninth precinct were not tampered with prior to their receipt by the town clerk, the lawful custodian, nor after their delivery by him to the county clerk. The inquiry is thus narrowed to a consideration of whether these ballots were properly preserved during the intervening period.

The town clerk, Martin P. Gleason, received the ballots and returns from the twenty-ninth precinct early in the morning of April 3, the day succeeding the election. The ballots were delivered to him in a closed canvas sack, securely fastened by string and sealed with wax. This sack, together with those from the other precincts, was deposited in a vault in the clerk's office and remained there until April 22, when the county clerk obtained the ballots pursuant to the impounding order entered in this proceeding. The office of the town clerk was located on the second floor of the Young building, in Joliet. It could be reached by a stairway and an elevator. The room had no windows facing the street. The town of Joliet rented a portion of the second floor, and it was used not only by the office of the town clerk but also by the offices of the supervisor and assessor. The office of an insurance agency adjoining that of the town clerk occupied the rest of this floor. There was a door between the two offices. Similarly, there were doors between the offices of the clerk and the assessor, and also the assessor and supervisor. Admittance to the town clerk's office could thus be gained from the offices of other town officials and likewise from the quarters of the insurance agency. Furthermore, all tenants of the Young building were supplied with keys to the front door on the main floor.

The vault in the town clerk's office in which the ballots were placed was also used as a store-room for all township offices. It had an iron door with a combination safe lock, to which Gleason and Francis Dougherty, his assistant, had the combination. This door had been installed by the owner of the building, who did not cause its combination to be changed. Gleason recognized the probability that other persons were familiar with the combination but asserted that they would encounter difficulty in opening the vault. The town clerk, called as a witness by the appellants, testified that the vault door remained open most of the time during business hours. Both Gleason and Dougherty testified that in addition to themselves the only persons admitted to the vault between April 2 and 22 were employees of town offices. In particular, they stated that only three, namely, Jacob Brunskol, Marie Halweg and Ruth Oetter actually entered the vault. Of these the first named was assistant supervisor and the other two were employed in the supervisor's office. Each testified that although their duties required them to go into the clerk's vault from time to time they did not in any manner interfere with the ballots. Ruth Oetter added that on the day after the election she observed sacks of ballots in Gleason's office outside the vault. Gleason further testified that either Dougherty, Brunskol or himself was in the office when the door to the vault was open and that the three of them never left the office at the same time. He admitted, however, that he was absent from the city for three days during the period in controversy, namely, on April 13, 14 and 15, and that he did not know what happened in his office on those days, two of which were week days. Although the witness had previously named three persons as those who actually went into the vault, he added, upon re-direct examination, that numerous persons came into his office every day and went into the vault.

From the testimony of John A. Zelko, who testified in behalf of the appellees, it appears that persons having no official business in the vault were afforded easy access to it during the period in controversy. In particular, the witness testified that he was in the town clerk's office each of the six week days immediately after April 2; that the door to the vault was open, and that he saw six or eight persons enter it every day, six of whom he named. Zelko also stated that he and Gleason were present at certain conferences held in the vault between April 2 and 22. The witness observed no attempts at tampering with the ballots. Three of the persons named by Zelko fully corroborated his testimony. One of these witnesses testified that Gleason was with him a part of the time he was in the vault. The parties stipulated that two others would testify to the same effect. The sixth person whom Zelko named was Emmett McGuire, highway commissioner. Zelko stated that he had a conference with McGuire in the vault concerning certain work between the second and twenty-second days of April. Gleason also testified that McGuire was in his office on two occasions during this period, adding, however, that the latter was not in Joliet between the fourth and the sixteenth. McGuire said that he maintained one of the highway commissioner's offices in the town clerk's office from April 1 to 24, and that his duties required him to go into the vault, where he kept some of his records. He claimed, however, that he left the city on the day of the town election and did not return until April 22.

Additional facts and circumstances appearing from the evidence merit scrutiny. On the night of the election the ballots voted in the twenty-ninth precinct were placed in a sack provided by the town clerk for that purpose. Robert Richardson, one of the judges of election, applied sealing wax on the sack, the open end of which was first folded over and sewed. He then impressed the seal furnished by the town clerk for precinct 29 upon the sealing wax. When the sack was produced upon the trial it appeared, upon inspection, that the upper portion of the wax was considerably thinner than the lower part and that the left corner of the upper part had been broken. The wax constituting the upper layer was a much lighter shade of red than the lower layer. In other words, there were then two layers of sealing wax of different hues on the sack. The imprint of the seal appeared only over the lower layer of wax. Examination of the ballot sack at the trial disclosed that the upper layer of additional wax covered at least two-thirds of the impression formed by the seal on the original wax. In short, the impression of the seal was plainly visible on the lower layer of dark sealing wax and not discernible on the upper layer of light-colored wax. The ballot sack has been certified to us and we have examined it. We observe, as the trial court found, that evidence of an impression of a seal upon the upper layer of wax is wanting.

The testimony of the appellants' witnesses failed to account for the strange condition of the seal. Richardson testified that the sewing on the sack on election night and at the trial was exactly the same; that the only difference he could detect in the condition of the seal was that a part of it was broken, and that the visible portion of the impression of the seal was affixed by him on the night of April 2. He also stated that he could see an impression of the seal on the upper layer of wax. According to his testimony he did not observe different shades of coloring in the sealing wax when the seal was impressed. Roland Pierson, another judge of election, testified that he paid scant attention to the process of sealing the sack and that he would not say for sure that it was the same seal. Upon examining the sack he admitted that the top layer of wax was lighter than the lower layer and that he could see no indentation of the wax. He could not say whether it was the same wax. John T. Elens, the third judge, testified that the seal placed on the sack the night of the election was of the same character and in about the same position as the one on it at the trial. He was, however, unable to say whether it was the same seal. The town clerk stated that he did not examine the seals on the ballot sacks on April 22 when he turned them over to the county clerk. He nevertheless said that they were in the same condition as when he and Dougherty deposited them in the vault.

To obtain a reversal of the judgment the appellants contend that the integrity of the ballots in the twenty-ninth precinct was preserved, that there was a substantial compliance with the method of preservation prescribed by the Ballot law, and that they should therefore have been admitted as superior evidence of the result of the election in that particular precinct. The law applicable to the factual situation presented is well settled. The returns of the judges and clerks are prima facie evidence of the result of the election, but the ballots are the original evidence of the votes cast and in case of a contest are better evidence of the result if they have been preserved in the manner prescribed by the statute. ( Talbott v. Thompson, 350 Ill. 86; Graham v. Peters, 248 id. 50.) Conversely, if the evidence discloses that the ballots were exposed to the reach of unauthorized persons, and the returns are not discredited, the ballots will not be regarded as better evidence of the result of the election. ( Talbott v. Thompson, supra; Bolton v. Whalen, 350 Ill. 50.) The question of whether the ballots have been properly preserved is in each action necessarily one of fact, to be determined by the evidence. ( Bullman v. Cooper, 362 Ill. 469; Sibley v. Staiger, 347 id. 288.) The contestant is the moving party, and the burden rests upon him to prove that the ballots have been kept intact. It is not incumbent on the defendant to show that they have been changed, and unless the contestant shows that they have been kept in such a way that there was no reasonable opportunity for tampering with them they cannot overcome the returns. ( Talbott v. Thompson, supra; Bolton v. Whalen, supra; Alexander v. Shaw, 344 Ill. 389; Graham v. Peters, supra.) Nor is it necessary that actual tampering with the ballots be shown, but it is enough to invalidate them as evidence that the opportunity for unlawful interference of unauthorized persons existed. Talbott v. Thompson, supra; Alexander v. Shaw, supra; Rottner v. Buchner, 260 Ill. 475.

Application of the above cardinal principles to the facts in the present case necessarily leads to the conclusion that the ballots of the twenty-ninth precinct lacked probative force. It suffices to observe that during the twenty days the ballots were kept in the vault of the town clerk numerous persons, irrespective of whether engaged upon official business, gained ready admittance to the vault. Reasonable, if not abundant, opportunity for tampering with the ballots existed, and the evidence tends to show actual interference with them. In particular, the condition of the seal on the ballot sack was not the same as when the sack was sealed on the evening of election day. The appellants have failed to sustain their burden of establishing that the ballots were in the same condition when offered in evidence as when they were cast. Although a proceeding to contest an election is statutory, the finding of the trial judge, who observes the demeanor of the witnesses while testifying, will not be disturbed on appeal unless it is palpably against the manifest weight of the evidence. ( Bullman v. Cooper, supra; Baker v. Hinrichs, 359 Ill. 138.) The finding of the trial judge that the ballots of the twenty-ninth precinct, when produced in court, were not in the condition in which the judges of election counted them was not palpably against the manifest weight of the evidence.

The appellants insist, however, that the returns of the twenty-ninth precinct are discredited owing to irregularities of the judges and clerks of election in counting and tallying votes. The three tally-sheets were introduced in evidence by the appellants as petitioners' exhibits 4, 5 and 6. Erasures of tallies in the squares opposite the name of Hedstrom appeared on exhibits 4 and 5. Excluding the erased tallies, exhibits 4 and 5 showed that 284 votes were tabulated for him, and this figure appeared in the final column as the number of votes which he received. There appeared 374 tallies opposite his name on exhibit 6, but the total number of votes received was extended as 284. There appeared 377 tallies in the squares opposite Rogers' name on exhibits 4 and 6 and 277 on exhibit 5. The total number of votes recorded for him on each tally-sheet was 277. Similar irregularities in the tallying were found after the names of Smith and Carter, the latter being one of the successful candidates. Although 280 votes were tallied on exhibit 5 for Carter, and this figure was carried out in the final column as the number of votes he received, the tally-sheet also showed that a few votes tallied in the square marked 380 had been erased. There were no tallies in the squares between those marked 280 and 380. When summoned before the canvassing board at the town clerk's office on April 3 to explain these discrepancies, Richardson, Elens and Pierson placed the following certification on exhibit 5: "We the undersigned judges of the April 2d, 1935, election agree that the tally-sheet of Mrs. Mary DeSignor is correct." Mrs. DeSignor was present when her tally-sheet was adopted as the correct one. The number of votes tallied and extended for Hedstrom, Rogers, Smith and Carter on this sheet is the same in each instance. The judges of election testified that the tabulation of votes on the tally-sheet which they certified was correct. These three witnesses were vouched for and called by the appellants, who attacked the returns. From their testimony it appears that split ballots were tallied as they were counted, and that after tallying those ballots, tallies representing straight party votes were added in the squares opposite the names of the respective candidates. In placing the tabulation of straight ballots upon the tally-sheets, however, the clerks started at the last square in which the votes were to be marked and proceeded backwards. The clerks erroneously commenced tabulating too far over in the line of squares. They discovered their mistake on the night of the election and made the erasures previously mentioned in attempts to correct some of these errors. We have examined the tally-sheets and are of the opinion that the trial court properly found that the returns were not discredited. Irregularities in the conduct of an election, such as those described, which do not proceed from fraud or misbehavior and have no influence upon the result of the election, will not impair the returns. ( Bolton v. Whalen, supra; West v. Sloan, 238 Ill. 330.) Until the contrary is shown, election judges are presumed to have discharged their duty in canvassing the votes and in making returns. ( Clarke v. Bettenhausen, 296 Ill. 373.) There was no proof of want of good faith or of fraudulent conduct on the part of the election officials of the twenty-ninth precinct. On the other hand, the testimony of the three judges that the ballots were accurately counted stands unimpeached. The returns must, in consequence, prevail over the ballots, which were exposed to unlawful interference and apparently subjected to it.

The judgment of the county court is affirmed.

Judgment affirmed.