The official canvass gave Stentz 18,192 votes, Drolet 17,989, and Stentz was certified as the duly elected State’s Attorney of Kankakee County by a margin of 203 votes. In the court recount, the trial court rejected the entire vote in the second precinct in Pembroke Township, 88 votes in another precinct and 13 votes in a third precinct. The summarized result was 17,942.064 votes for Drolet and 17,769.936 votes for Stentz. Drolet was adjudged to be the duly elected State’s Attorney and it is that judgment we now review.
In Pembroke No. 2, the official canvass gave Stentz 323 votes and Drolet 36 votes, all of which the trial court rejected with the finding that the many statutory violations constituted weighty evidence of fraud. No actual fraud is charged or established. The judges, with a single exception, were wholly inexperienced and untrained. The township had been divided into three precincts by the county board in September before the election. Due to the shortness of time and the fact that precinct No. 2 was a rural precinct, the precinct binders, as well as the *204voter registration, cards, leave much to be desired as to completeness and accuracy and contributed to the confusion in this precinct.
The judges took no voter applications for the first 2% hours. The election was then halted and an assistant State’s Attorney brought a new ballot box, sealed the old one, and instructed the judges to keep the ballots cast up to this time separate. There is a dispute as to whether this instruction meant only for the day or forever. In any event, the ballots in both boxes were dumped together, counted, and there is no hypothesis upon which it can be ascertained what votes were cast without applications. The range of the testimony is from 35-50 to 175 in number. During the day word was sent out to those who had voted without applications and some returned to the polls and filed applications. How many is a matter of dispute and the number is not ascertainable. Whether any voted the second time was not known. The judges relied on memory and were of the opinion that no one voted twice. The judges seldom, if ever during the day, compared the signature on the application with any signatures in the voters’ registration binder. Seven persons known to one of the judges to reside outside the precinct applied for, received, marked and cast ballots. Their names were in the precinct binder, but their applications were endorsed by one of the judges as residing out of the precinct. An unknown number of persons received, marked and cast ballots upon no further evidence than the presentation of a voter’s registration card issued by a former county clerk before the division of the precinct, or the location of the names on a 1963 binder precinct list compiled and prepared prior to the division into three precincts. Some thirteen persons obtained, marked, and cast ballots by making affidavits, one being on forms supplied as a part of the furnished election supplies and twelve on letter-sized paper without a supporting affidavit as required by the statute. Some thirty-*205six people voted whose applications bore names neither in the precinct binders nor on the printed 1964 voters’ list, nor were they in possession of a valid registration card, nor did they make affidavit. In this state of the record, it is impossible to segregate the valid from the invalid ballots with any degree of mathematical certainty.
The choice between disenfranchising an entire precinct and judicially exculpating the election official where no actual fraud is either alleged or established is not an easy one. It is one thing to nullify an entire election where the same matters may be considered in another election on another day. It is one thing to be able to segregate illegal ballots and adjust the results accordingly on a pro rata basis. It is quite another matter to be unable to segregate illegal ballots and adjust the results accordingly on a pro rata basis. It is quite another matter to be unable to segregate or mathematically compute the number of ballots illegally cast. Where fraud was present and the number of illegal ballots were unascertainable, the votes of an entire precinct were rejected. Emery v. Hennessy, 331 Ill 296, 162 NE 835; Lehman v. Hill, 414 Ill 173, 111 NE2d 120. The distinction between ascertainable and unascertainable alleged votes was pointed out in Thornton v. Gardner, 30 Ill2d 234, 195 NE2d 723. In Lehman, the Supreme Court uses this language:
“. . . That election is free and equal where all of the qualified electors in the precinct are carefully distinguished from the unqualified and are protected in the right to deposit their ballots in safety and unprejudiced by fraud. That election is not free and equal where the true electors are not separated from the false, where the ballot is not deposited in safety, or where it is supplanted by fraud. People v. Hoffman, 116 Ill 587, 5 NE 596, 8 NE 788, 56 Am Rep 793. . . .”
*206Morandi v. Heiman, 23 Ill2d 365, 178 NE2d 314, recognizes that provisions designed for the information and guidance of the election officials should be liberally-construed, particularly after the election has taken place, and that the policy of directory construction may be applied to avoid disenfranchisement of the innocent voter. It is, however, emphatically stated at page 373 of that opinion:
“. . . Such policy, however, has always been subject to exceptions where the statute specifically provides that a thing shall be done in the manner indicated, or where the irregularity has rendered doubtful the evidence from which the result is to be declared (People ex rel. Comerford v. Miller, 314 Ill 474, 478, 145 NE 685), or would tend strongly to overthrow the safe conduct of elections. People ex rel. Agnew v. Graham, 267 Ill 426, 441, 442, 108 NE 699. . . .”
In Morandi, the court rejected 739 uninitialled absentee ballots cast for one candidate and 403 uninitialled absentee ballots counted for another. No fraud was charged or proved. The results for a county office were reversed by such holding. An offer of proof was made suggesting that the folds in the absentee ballots readily identified each absentee ballot as absentee ballots have 8 folds and ballots in person have only 6. The Supreme Court rejected this proof and in so doing said:
. . To attempt to identify a validly cast ballot from an illegal one upon such fortuitous circumstance, would be a dangerous rule to permit to creep into the conduct of our elections. (Cf., Griffin v. Rausa, 2 Ill2d 421, 118 NE2d 249.) It would be a dangerous rule to establish that election judges may disregard the plain provisions of the statutes and thereby defeat the intention of the law to prevent actual fraud from being committed and to disarm *207the constituted authority of the efficient means provided by statute for detecting such fraud. Cf., Laird v. Williams, 281 Ill 233, 239, 118 NE 73; Harvey v. Sullivan, 406 Ill 472, 475, 477, 94 NE2d 424. . . .”
“. . . It is clear that no discretion is reposed in the voter by the language of the Permanent Registration Act as to the necessity for his registration in the manner required by the Act. It was the evident intention of the General Assembly to put teeth into the Permanent Registration Act and to prevent frauds arising out of illegal or indifferent registration. . . .”
It is thus clear that not only the Legislature but the Supreme Court has indicated the avowed purpose of the Voter Registration Act. The provision of the Election Code, Ill Rev Stats c 46, is a legislative expression of how and in what manner an election should be conducted. On the record before us we deal not with a single failure to comply with the rules governing elections, but with multiple violations. Inadvertent noncompliance with some legislative directions for the conduct of elections need not vitiate the vote in a given precinct. Where, as here, the violations are so numerous and so gross that, an accurate calculation of the valid and invalid votes is impossible, we reach the end of a dead-end street. We cannot go further. We cannot determine the number of legal votes cast for either party. We cannot apportion. The gross disregard for numerous simple and plain statutory requirements should not, under these circumstances, be judicially condoned. We therefore conclude that the election in that precinct is abortive. It is better that the voters in one precinct in one election be disenfranchised than that a judicial key to fraud and chicanery be molded or that the tempta*208tion to ignore the legislative directives for the assurance of a free and equal election be encouraged. The trial court correctly excluded the votes in Pembroke No. 2, and surcharged the final for each party accordingly. When this is done, Stentz has 17,896 votes left and Drolet has 17,953. Drolet’s majority is 84 at this point.
It is urged that there should be an apportionment in Pembroke No. 2 according to the legal votes cast as held in Neff v. George, 364 Ill 306, 4 NE2d 388, and Smoda v. Gallagher, 412 Ill 271, 106 NE2d 181. This same principle was recognized in Thornton v. Gardner, 30 Ill2d 234, 196 NE2d 723. However, in Thornton the number of illegal votes was readily ascertained. In Lehman v. Hill, 414 Ill 173, 111 NE2d 120, the Supreme Court recognized the apportionment rule of Neff and Smoda but refused to apportion for the reason that “here the particular votes claimed to be illegal could not be segregated, proved and exactly computed.” Here a like situation exists. It is impossible to apportion when the number of votes to be apportioned is impossible of accurate computation and determination. In our judgment the apportionment rule does not apply where the number of legal and illegal votes rest upon pure guess, conjecture and surmise.
In the second precinct of the first ward in Kankakee, the trial court rejected 88 ballots and apportioned them on the basis of 85.9% charged to Stentz and 14.1% charged to Drolet. The defendant Stentz does not question or argue on this appeal the apportioned deduction of 10 votes on the stated percentage. Accordingly, if any objections were available, they are now waived and we thus deduct 8.592 votes from Stentz, leaving him with 17,860.41 and deduct 1.41 votes from Drolet leaving him with 17,951.59 votes. At this point, the summarized result leaves Drolet with 91.18 votes in the lead. We need go no further to affirm the result in the trial court. The trial court deducted a total of 422.064 from Stentz and *209as noted herein 323 plus 8.59 or 331.59 were properly deducted. If we were to restore to Stentz without deciding the points raised as to them, the remaining 90.474 votes deducted by the trial court, Drolet would still win. However, in restoring to Stentz the remaining ballots would likewise require a like action for Drolet and the restoration to him of 9.795 votes. Thus, the final result would be Stentz 17,950.884 and Drolet 17,961.385 or a majority for Drolet of 10.501. In addition, we should observe that of the 90.474 votes we have tentatively restored to Stentz without deciding the objections, four in the second precinct of the seventh ward were rejected because of distinguishing marks, the action of the trial court as to them is not questioned and therefore they should not be counted for Stentz. So treated, Drolet’s majority is 14.501. Other points raised in this appeal relate to the applications for ballots, improper affidavits for ballots, applications to which the judges’ initials or names were signed by others, and a problem of charged improper assistance to other voters. It would unduly extend this opinion to discuss in detail all of these points and, as already noted, it would not affect the result. Thus, the judgment of the trial court should be and it is hereby affirmed.
CORYN, P. J. and ALLOY, J., concur.