Jenks, J.:
This application was made against the board of inspectors, the county clerk and the city clerk. None of the other election officials was included. The relator’s affidavit stated that the certified statements of the canvass do not agree with the tally sheets or with the ballot clerk’s return sheet, the said return sheet does not agree with the tally sheets or with the certified statements, and certain excerpts are stated to show irregularities and lack of harmony therein, and to show other defects and irregularities. The opposing affidavits raised certain issues of fact not necessary now to detail. The rule is entirely familiar that the question as to the right to the writ is determined upon the assumption that the averments of the opposing affidavits are true. (People ex rel. Corrigan v. Mayor, 149 N. Y. 215.) But as the Special Term had before it the inspectors’ return and statement of canvass, the original official statement of the result, the inspectors’ return of the first election district, the ballot return sheet and the tally sheets, I think that, aside from the issues raised by the opposing affidavits, and solely upon the facts as shown by these official records, a writ of peremptory mandamus might issue. The question then is whether, under the circumstances, the present writ requires any change or any amendment.
The law (Laws of 1896, chap. 909, § 110, subd. 1, as amd. by Laws of 1898, chap. 335) requires that the board of inspectors shall commence the canvass by comparing the poll books with the registers used on election day, as to the number of electors voting at the election, correcting any mistakes therein, and by counting the ballots found in the ballot boxes without unfolding them, save to ascertain that each is single, and by comparing the number in the ballot box found by this count with the number shown by the poll books and by the ballot clerks’ statement to have been deposited therein. The intention of the law that the check thus afforded by the poll books and by *591the ballot clerks’ return shall prevail over the number of ballots in the box when opened is shown by the provisions that if the count show the latter number to be in excess of the number shown by such return and by such boobs, such excess shall be destroyed; that a ballot in the wrong box may be placed in the right box and counted, provided the number thereby increased does not exceed the number shown by the poll books and the ballot clerks’ return, and that if two.or more ballots be folded together they shall be destroyed if the whole number exceeds the number shown by the poll books and the ballot clerks’ return. The theory, then, of the statute is that before the ballots are canvassed, the number to be canvassed shall agree with the number shown by-the poll books and by the ballot clerks’ return. While the law provides for the destruction of an excess, it does not provide, nor do I see how it could well do so, for a shortage. It would not do to increase the number by putting ballots in the box which had not been voted by the electors. Moreover, the chance of the abstraction of ballots from the box is remote, and not to be compared with the opportunity of putting unlawful ballots therein by enfolding them with other ballots, or by other stealth. As the law thus directs that, so far as is practicable before the canvass, the number to be canvassed has been checked and regulated by these two safeguards, it assumes that this will have been done. Then and thereafter the canvass follows, and the result thereof is recorded on tally sheets furnished in prescribed form. This tally sheet is described by Bartlett, J., in Matter of Stewart (155 N. Y. 545, 551) as “ This contemporaneous, self-proving record, * * * it being the original entry of the casting and canvassing of a vote.” It is so subdivided in columns as to present, with reference to every candidate, the disposition of every vote that is subject to the canvass. The statute (§ 84) provides that at the extreme right of such sheet there shall be a column headed “ Total number of ballots accounted for,” in which shall be enteredx>pposite each office the sum of the total vote cast for all candidates for the office, together with the number of ballots not wholly blank, on which no vote was counted for that office, the total number of wholly blank, and the total number of void ballots, and the votes cast., if any, for candidates for such office whose names are not printed upon the ballot. The statute also provides that “ such sum must equal the number of ballots voted, as *592shown by, the ballot clerks’ return of ballots, and if it does not, there hás been a mistake in the count, and the ballots must be recounted for such office.” The tally sheet in the record before us shows us that the sum in the final column is 482. Item 5 of the ballot, clerks’ return reads: “ The number of sets of official ballots actually voted were 484”’ There is a variance between the tally, sheet and its criterion, arid such variance the statute' declares is a mistake which must be followed by a recount. As Ingraham, J., says, in Matter of Stewart (24 App. Div. 201, 210; affd., 155 N. Y. 545): “ Upon the completion of that count, the ascertainment of the total vote by the -poll clerks upon the tally sheet, the verificatiori of the tally sheet by the inspectors, and the announcement by the inspectors of the result, the canvass is completed;' What happens after relates, to the duties of the election officers in. certifying that result thus announced from the tally sheet as thus kept and prepared, and is merely ministerial. The result .thus announced must comply with the tally sheet as kept by the poll clerks, the correctness-of which has been Verified by the inspectors, and until the correctness of the tally sheet is ascertained,.the statute explicitly provides-that the count is not complete; the -votes have not yet' beéri canvassed. When the record of the count upon the tally sheet has: been completed by the ascertainment of the total vote cast for each officer, and the correctness of the additions ascertained,- the canvass is complete.” How the “recount” would not seem to refer to the> reckoning made by the inspectors when they first opened the ballot boxes, because .the law does not deem that reckoning a count until it agrees or is made to agree with (the poll books- and) the very ballot clerics1 return which, is declared by the statute to establish a, mistake in the count. In other words, such return had already, and before the .procedure detailed- in the tally sheets, been used both to establish the accuracy of the preliminary count made by the inspectors and as the standard of the ballots subject to canvass. And as. the only count prescribed after such inspectors’ preliminary count is the count shown by the tally sheets, I think that the provision for a recount means a recount of the ballots which were canvassed and recorded on the tally sheets, on the theory that. the mistake is due to the fact, that the tally sheets have not set forth in the various columns all of the. ballots subject to canvass; The terms *593“count” and “canvass” are often used interchangeably in the statute, and I think the word thus used may be held to cover whatever steps may be necessary to establish a correct count and, if necessary, an allotment, according to the disposition made by the voter, as evidenced by his action or non-action upon the face of the ballot, of all ballots theretofore, determined by the inspectors as lawfully voted, and thus to cover a recanvass, or, rather, a canvass, if necessary, inasmuch as the law recognizes the result first obtained as a “ mistake.” But I think that the term “ recount ” does not require that in the first instance there should be a recanvass in the sense that the votes actually distributed to the various columns of the tally sheets by the “canvass” already made should be redistributed therein. For the “sum” that presents the “mistake” is that which is designed and required to account for all ballots. And, therefore, if ballots already “ canvassed ” and set forth in the tally sheets be taken from one column and placed in another, any gain in one column is offset by the relative loss in the other. If candidate A gain a vote in his column from candidate B, B loses that vote in his column. Or, if B gain a ballot scheduled in the blank column, the blank column loses that vote. The “ sum ” that is a “ mistake ” is the account for every ballot, and a récanvass which would result in a redistribution of the same number of ballots is simply a variation of the component parts of that sum. Three and three are six. If we take a unit from one three and add it to the other three, the sum of the resultant two and four is still six. And the mistake presented in this proceeding is, so to speak, in the six, not in the subdivision thereof into three and three instead of four and two. I am led, then, to the conclusion that the first step to be taken is to reckon again the number of ballots with reference to the tally sheets; i. e., to ascertain whether the sum total thereof, 482, represents every ballot that could have been canvassed. If the number represented in item 5 of the ballot clerks’ return was, at the time the box was opened, 484, then the inspectors in the preliminary count of the unfolded ballots either did or did not carry out the intent of the law by making the number in the box, counted with a view to the subsequent canvass, conform to that return and to the poll books. We have no record of what was done. We are not *594informed that it was impossible. If they did this, and the . ballot clerks’ return and the poll books then showed 484 ballots, and that preliminary count agreed therewith, then all of the ballots were not canvassed, for the tally sheets accounted for but 482 ballots. If they did not, then they neglected a safeguard provided by the election machinery. If another reckoning of the ballots with reference to the tally sheet shows that there were but 482 ballots that can be accounted for, then, of course, no more than 482 could have been or can now be canvassed, and that has been done. If the new reckoning shows 484, then there was not a canvass in the eye of the law, and there must be a canvass in order that all of the ballots shall be accounted "for on the tally sheets.
. But I cannot see how the recount or the recanvass can be ordered in these proceedings as they now stand. The poll clerks are not before the court. The law (§§ 84, 103, subd. 3, 110, subd. 3) requires the poll clerks to make and to complete the tally sheets, to enter the count therein, to tally the split ballots, to add together the votes and the ballots wholly blank and void, together with the ballots not wholly blank on which no votes were counted for any candidate for the particular office, and to enter in the final, column this very sum whose variance from.the ballot clerks’ return is declared by the statute to be a “.mistake.” They had no notice of this proceeding, they had not appeared, nor do I find even an affidavit from one of them in this record. (People ex rel. Hasbrouck v. Supervisors, 135 N. Y. 522, 532.) If they were before the court, then there is no question but that, they and the election inspectors might be required to make a recount which would include, as I have said, a new canvass if the number of ballots required it. (Baird v. Supervisors, 138 N. Y. 95.)
If, however, there was, as shown by the preliminary count by the inspectors, 484 ballots to be canvassed, and yet an accurate re-reckoning shows that there were but 482 ballots to be found, then 2 ballots have disappeared, and I know of no remedy for that loss. But, notwithstanding the statement in the ballot clerks’ return numbered . 5, there may have been but 482 ballots found, by the inspectors in the box when they opened it and counted them. Mr. Ripple, an inspector, deposes-that the last voter on election day was Mr. Qualey,. who received the last ballot given out, and that the number of the *595ballot was 491, and that a deduction from that number of the ballots returned as spoiled by the voters would show the actual number of votes. If that number was 9,. it would be 482 ; if 7, it would be 484. Item number 3 in the ballots clerks’ return shows the number was 7. The affidavit of Inspector Patterson shows it to have been 9. The poll books are not in the record, but examination of the order made by Marean, J., in the Brush case, now sub judioe, shows a recital, that the number of voters shown by the poll books is 482. How, at the outset, the poll books, a contemporaneous record of the voters, must agree or must be made to agree with the registers, another contemporaneous record of the voters (§ 110, subd. 1), and, of course, the law assumes that the ballot clerks’ return of this item will agree with the poll books, inasmuch as the statute declares the first count made by the inspectors of the ballots must, so far as is possible, agree with the poll books and the ballot clerks’ return. The ballot clerks’ return, as made up, is not so final or so necessarily perfect as to be beyond possible error, and if the error be there placed, of course, there being no return in the eye of the law, a proper return can be required. It is made the criterion of a mistake, it is true, but it presumably agrees with the poll books and with the registers before it is thus constituted. In order to test the worth of the ballot clerks’ return, it may be pertinent to state that as to every item it is not a contemporaneous record in the sense that the poll books or the registers or the tally sheets are a contemporaneous record of the voters. It is made immediately at the close of the polls, but not until the destroyal of the spoiled and mutilated ballots which have been placed in the box devoted to that purpose. The form of return is prescribed by law. (Laws of 1896, chap. 909, §§ 84, 103, subd. 2.) The first item thereof may be determined from the fact that the ballots were placed in their charge; the second, they ascertained from the fact that they withheld the canceled ballots, disposed of them, made a record thereof, and were compelled to compare it with the number of such ballots taken and destroyed by them from the box for mutilated ballots at the close of the polls; the third from the memorandum made thereof by them at the time; the fourth from the fact that they received the original ballots before the opening of the polls, and they send back the unused ballots; the sixth by an addition of the previous items; the *596Seventh from the boxes of the. detached stubs, which stubs they are required to return; the eighth because they retain the unused ballots, and the ninth by addition of the last two items. But as to the fifth item, “ The number of sets of official ballots actually voted,” I .cannot find that the ballot clerks have any contemporaneous record or memoranda made by themselves. That item may be an arithmetical result found by the ballot clerks, or it. may be taken from the contemporaneous records of the poll books and the registers. The poll books and the registers are contemporaneous records of every elector who votes. (§ 103, subds. 1, 3.) The law provides that at the close of the polls, the poll clerks and the inspectors shall compare the .poll books with the registers as to the number of electors voting, and correct any mistakes found therein. (§§ 103, subd. 3, 110, subd. 1.) If this be done, these .two contemporaneous records must agree, and they may. furnish the information to the ballot clerks. Thus, the first check, that of the poll books upon the number , of ballots left in the ballot boxes for canvass, is found in contemporaneous records of the. voters verified by comparison with similar records, to wit, the -registers. But the second check, that of the ballot clerks’ return, so ' far as I can see, is not, in respect to the fifth item, a record of such character. If, so far as this item is concerned, the registers and the poll books being in accord, show the number of voters as 482, and if the re-reckoning show that the number of votes accounted for is 482, is there not, strong reason to believe that the ballot clerks’ return presents an error ? We have no means of knowing how the ballot clerks made up this particular item, or whether they correctly entered ,it.. If they did not take it from the poll books and the registers, it may be that they took it from their data. E. g., looking at the items in their return, it may be that they determined item 5, “ The number of sets of official ballots actually voted were 484,” by subtracting the number returned by them unused, 516, from 1,000, the number originally received by them. But they also account for 7 ballots returned by voters and destroyed by the ballot clerks. If there were 7 ballots destroyed, and 516 returned out of 1,000, how could there have been 484 ballots actually voted ? There may be a mistake in the 1,000, or in the 7, or in the 516. The learned counsel for the appellants, with much acumen and ability, argues that the variance is due to error in the ballot clerks’ return. *597But even if this theory were demonstration, and if the error presented were but clerical, the first step in the procedure is to require at the hands of the election officials a compliance with the statutory duties cast upon them, which the official records before us presumptively show that they did not make. If this be required and be done, and a variance still appear, then it may be pertinent both to reach the error, where it exists, and to right it. Whether the court can do this in the first instance is a question not now necessary to determine. The important requirement now is to have the election officials comply with the statutory duties cast upon them.
In conclusion, the entry of the sum in the tally sheets was, so far as the record shows, simply the entry of the mistake. In the words of Ingraham, J. (supra), “ until the correctness of the tally sheet is ascertained, the statute explicitly provides that the count is not complete ; the votes have not yet been canvassed.” The law required •“ a recount.” It does not appear that this was done, and, therefore, the court may and should order it to be done. And the recount required by the mistake on the tally sheets implies that there shall be a count, whether merely a re-reckoning or, if necessary, a recanvass prescribed by law. That the poll clerks as well as the inspectors are necessary officials in such procedure, I have heretofore ■attempted to show. And the court may order them to be brought in. (13 Ency. Pl. & Pr. 664.) I think that the order should be modified by directing that the proceedings be held open, with leave to apply to the Special Term for an order to bring in the poll clerks, and then for a writ directing the election officials, to wit, the inspectors and the poll clerks, to meet at such time and place as the Special Term may direct, and for the production then and there of the ballots and such records as the court may deem proper, and then and there to recount the vote, by proceeding in such manner as the court shall, in its writ, specify, conforming it therein, in every respect, so far as possible, to the procedure prescribed and required by the statute to be done on the day of election, in view of the provision of section 84 thereof, that “ such sum must equal the number of ballots voted, as shown by the ballot clerks’ return of ballots, and if it does not, there has been a mistake,” and by requiring that the officials shall first re-reckon the number of votes to be accounted for, upon the basis determined by the inspectors at the commence*598ment' of the canvass, as provided for in section 84, and that if such ré-reckoning shall disclose a; number in excess of the number now stated in-the tally sheet under the column headed “Total number of ballots accounted for,” tó wit, 482, then they shall proceed to recanvass all of the ballots of the electors found in the said ballot box, in conformity to the provisions of the Election Law relative to the count,.canvass.and method of' counting, and the making of the official records thereof, but. all only as to the office here in controversy, together with such further provisions as to a return to the said writ as the court shall deem appropriate:
All concurred.
Order modified in accordance with the opinion of Jenks, J., and ■as modified affirmed, without costs. . . . ..