I dissent.
While I agree with the holding that the ordinance in question is subject to referendum proceedings, I dissent from the remainder of the discussion and particularly from the conclusions reached in the majority opinion. The opinion concludes; (1) that the trial court properly upheld the clerk’s determination that 94 signatures on the referendum petition were invalid because, although the clerk’s conclusion was arguably incorrect, the similarities of the signatures were not so great or the dissimilarities so minor as to render the clerk’s rejection unreasonable or arbitrary; and (2) that this is an appeal on a judgment roll and therefore the sufficiency of the evidence to support the findings is not open to review.
As we shall see, these conclusions are incorrect since the law clearly declares that the burden of proof was upon the clerk to establish the signatures were not genuine and it affirmatively appears on the face of the record that the trial court failed to require the clerk to meet this burden. That the evidence may have been sufficient under the trial court’s inappropriate standard cannot justify this court in upholding the judgment.
The initiative and referendum are powers reserved to the people, not powers granted to them. (Cal. Const., art. IV, § 1.) For this reason statutes dealing with the referendum should be afforded the same liberal construction accorded to election statutes so that citizens may, whenever possible, be protected in the exercise of their reserved power. (Ley v. Dominguez (1931) 212 Cal. 587, 593 [299 P. 713].) As was said in McFadden v. Jordan (1948) 32 Cal.2d 330, 332 [196 P.2d 787], “The right of initiative *459is precious to the people and is one which the courts are zealous to preserve to the fullest tenable measure of spirit as well as letter.” The clerk must perform his duties in such manner as will, whenever possible, protect rather than defeat the right of the people to exercise their referendary power. (Ley v. Dominguez (1931) supra, 212 Cal. at p. 602.)
Consistent with this long-prevailing policy the Elections Code establishes a rebuttable presumption that the signatures on initiative and referendum petitions are genuine. Section 3514 of the Elections Code provides that verified initiative and referendum petitions “shall be prima facie evidence that the signatures thereon are genuine and that the persons signing the same are qualified electors. Unless and until it be otherwise proven upon official investigation, it shall be presumed that the petition presented contains the signatures of the requisite number of qualified electors.”1 (Evid. Code, § 602.) The presumption imposes upon the party attacking the genuineness of the signatures on a referendum petition, here the clerk and the interveners, the burden of proving that the signatures are not genuine. (Evid. Code, § § 605, 606.) In other jurisdictions, too, the burden of proof is upon the person or agency attacking the validity of signatures on election petitions. (See, e.g., In re Initiative Petition No. 260, State Question No. 377 (Okla. 1956) 298 P.2d 753, 757; Woodward v. Barbur (1911) 59 Ore. 70 [116 P. 101, 103-104]; State v. Black (1964) 149 W.Va. 124 [139 S.E.2d 166, 171]; see 42 Am.Jur.2d 704; 5 McQuillin, Municipal Corporations (3d ed.) p. 235.)
The majority opinion, in holding that the clerk has met his burden by a bare assertion of the invalidity of the signatures and a showing that his conclusion was not unreasonable or arbitrary, eviscerates the presumption of validity which should be accorded to the signatures. It is not surprising that no authority is cited in the opinion for this startling inversion of legislative intent, for the cases have stated or held not only that the burden of demonstrating the invalidity of signatures on a petition is upon the person protesting their genuineness but also that it is for the trial court rather than the clerk to determine whether the signatures are genuine. Thus it has been said that the question whether fraud has been committed is not for the clerk but for a court of equity in appropriate proceedings. (Ley v. Dominguez (1931) supra, 212 Cal. 587, 602; Williams v. Gill (1924) 65 Cal.App. 129, 132 [223 P. 559].)
In recognition of this rule a number of cases have held that the trial court must take evidence and make findings of fact upon the question whether signatures on a petition are genuine. (See, e.g., Ratio v. Board of Trustees *460(1925) 75 Cal.App. 724, 727-728 [243 P. 466]; City of Watts v. Superior Court (1918) 36 Cal.App. 692, 698 [173 P. 183].) Authorities in other jurisdictions are in accord. (State v. Carter (1914) 257 Mo. 52 [165 S.W. 773, 780-781]; In re Initiative Petition No. 23 (1912) 35 Okla. 49 [127 P. 862, 864-865]; State v. Olcott (1912) 62 Ore. 277 [125 P. 303]; State v. Garner (1962) 147 W.Va. 293 [128 S.E.2d 185, 190]; see also State v. Rupert (1918) 99 Ohio St. 17 [122 N.E. 39, 40]; Hindman v. Boyd (1906) 42 Wash. 17 [84 P. 609, 613].). None of these cases indicate that the court must defer to the judgment of the official charged with examining the petition unless his acts are unreasonable or arbitrary. Indeed, reported decisions which recite evidence indicate clearly that the persons who protested the signatures were compelled to introduce evidence to prove their assertion of invalidity.2
The wisdom of these rules is graphically illustrated by the circumstances of the present case. The clerk initially disqualified the 94 signatures in issue on grounds other than the genuineness of the signatures. When the trial court held that the grounds of disqualification were improper, the clerk then stated that he would have refused to certify the signatures as valid in any event because a comparison of each signature on the petition with the registration affidavit of the voter indicated to him that the signer of the petition was not the voter in question. Although the judge disagreed with this conclusion, he indicated he had no power to substitute his opinion for that of the clerk since the similarities were not so great or the dissimilarities so minor as to render the clerk’s determination unreasonable or arbitrary.
No evidence aside from the opinion of the clerk or his deputies was introduced at the trial to support his conclusion. No attempt was made to contact the signatories to determine if they had in fact signed the petition for the purpose of disproving their signatures at the trial. In other words, the clerk’s burden of proof was held by the trial court to be satisfied by his ipse dixit statement that the signatures were not genuine.
Justification for requiring the clerk to prove the invalidity of the signatures is supported by the unlikelihood under the circumstances of the present case that the signatures were forged. The documents contained in the record show that the signatures in question appear at random on numerous petitions circulated by a large number of persons. Each circulator attached an affidavit asserting.upon information and belief that the signa*461tures were genuine and that each signer was a voter in the county. (Elec. Code, §§ 3701, 3754.) I have examined a sampling of signatures which the clerk held to be invalid after comparing them with the signatures on the affidavits of registration. It is evident from an inexpert comparison of these signatures why the trial court expressed the personal opinion that the signatures on the petitions were genuine. In view of the very close resemblance of these petitioners’ signatures on the petition with their affidavits of registration, the likelihood of forgery appears minimal. It would have to be assumed that some persons examined certain affidavits of registration at random to ascertain the appearance of the signatures of certain voters and that they then signed these names to various petitions also chosen at random. Or, in the alternative, we would be required to speculate that 94 persons who were familiar with the signatures of these voters, in their enthusiasm for the proposal set forth in the referendum petition, forged the names of the voters on the petition.* *3
It must be remembered that the Elections Code makes it a felony punishable by up to 14 years in prison to subscribe or cause one to subscribe the name of another to a referendum petition (Elec. Code, § 29221) and that it is a crime for the circulator to make a false affidavit attesting to the genuineness of a signature on a petition (Elec. Code, § 29218) or to file a petition with a false signature (Elec. Code, § 29216).4 While I do not imply that forgeries never occur, the unlikelihood of one risking serious penal sanctions supplies ample justification, if justification is necessary, for the legislative policy of placing the burden of proving that a signature is not genuine on the person who protests its validity.
As we have seen, the clerk is enjoined to perform his duties in such a manner as will, whenever possible, protect rather than defeat the right of the people to exercise their referendary power. (Ley v. Dominguez (1931) supra, 212 Cal. 587, 602.) He must consider that the signing of a petition is not accomplished with the same formalities as the attestation of a will. It was said in Chester v. Hall (1921) 55 Cal.App. 611, 617 [204 P. 237], “ ‘Those who circulate the petition will necessarily be drawn from the *462ranks of volunteers or those who, for a very small consideration, call attention to their fellow-citizens to the measure proposed, and solicit their interest therein. Necessarily, even with the best safeguards that can be thrown around the circulation of petitions, where such a large number of names are required, inaccuracies and technical departure from prescribed forms are certain to occur every time a petition is circulated. The people who sign the petitions often, if not generally, lack both convenience and the best writing materials to distinctly, legibly, and permanently attach their names thereto. All of these things are proper to be noted and taken into consideration in the administration of this law. It can be made effective or defeated by the officers charged with its administration, and it is our duty to sustain it, rather than destroy, if it can be accomplished within the law.’ ” (Italics added.)
The trial court’s error in the present case was compounded by its refusal to permit petitioners to produce evidence that the signatures were genuine. Leaving aside for the moment the question whether petitioners should have been barred from introducing such evidence because they did not promptly assert their desire to do so,* ***5 the court did not reject the offer of proof because it was tardy. The record establishes that petitioners attempted to introduce affidavits of the signatories attesting to the genuineness of the signatures and also offered to introduce testimonial evidence to the same effect. In its memorandum opinion the trial court indicated that the rejection of this offer was based on the court’s conclusion that such evidence was irrelevant in view of the clerk’s determination that the signatures were forgeries.6 A clearer violation of a voter’s right to exercise the power of referendum reserved to him under the Constitution is difficult to conceive.
One other matter should be noted. The clerk contends that if the trial court must go beyond a determination of whether he acted reasonably, courts will be burdened with an incredibly complex and time-consuming task since a referendum petition could contain hundreds of thousands of *463signatures. This argument is without substantial merit. The number of signatures disqualified by the clerk may indeed be large but only a small proportion of these would be disqualified because of an allegedly forged signature.
Under existing law, as the majority opinion admits (ante, p. 456), the clerk’s determination as to the sufficiency of a petition is not conclusive and may be challenged and reviewed by a court in a mandamus proceeding. The instant case is probably a fairly typical illustration of proceedings when the clerk’s factual determination is challenged. Here the clerk initially disqualified 944 signatures out of 6,719. Of these, petitioners conceded that 474 were properly disallowed and the clerk conceded that he had erred as to 42 signatures which he should have allowed. A considerable body of case law has evolved in California on the question whether various defects in referendum petitions justify the clerk’s refusal to validate signatures thereon. For example, in Ley v. Dominguez (1931) supra, 212 Cal. 587, 597, it was held that the failure by a signer to insert a precinct number after his name did not invalidate his signature and in Chester v. Hall (1921) supra, 55 Cal.App. 611, 618-619, it was held that the absence of a date after the signature did not disqualify the signature. The clerk restored over 200 signatures after the standards set forth in these and other cases were applied by the court.
This left only 233 signatures in question and of these 108 were invalidated because of asserted duplication of names and the absence of a circulator’s affidavit. Only 125 were rejected because the signatures were assertedly not genuine.7 Thus, ultimately, over 98 percent of the signatures on the petition were either found to be sufficient before trial or were restored during trial by the concession of the parties or the application of established rules.
In these 125 individual instances, we do not face difficult evidentiary situations in which the signers of a document are deceased or otherwise unavailable and the genuineness of signatures must be proved by others. Here the voters who signed the petitions are alive and well, and anxious to affirm the exercise of their referendum rights by sworn testimony that each in fact personally signed the petitions in question, but they are restrained from doing so only because a county official is of the opinion from a comparison of signatures with the affidavits of registration that the signatures were forged. This incomprehensible result can only chill the faith of registered voters in the rationality of the democratic process.
*464The United States Supreme Court has been increasingly solicitous of protecting the right of franchise against interference by the states. (See, e.g., Louisiana v. United States (1965) 380 U.S. 145 [13 L.Ed.2d 709, 85 S.Ct. 817]; Harman v. Forssenius (1964) 380 U.S. 528 [14 L.Ed.2d 50, 85 S.Ct. 1177]; Reynolds v. Sims (1964) 377 U.S. 533 [12 L.Ed.2d 506, 84 S.Ct. 1362]; Wesberry v. Sanders (1964) 376 U.S. 1 [11 L.Ed.2d 481, 84 S.Ct. 526]; Gray v. Sanders (1963) 372 U.S. 368 [9 L.Ed.2d 821, 83 S.Ct. 801].) The holding of the majority opinion is inconsistent with the underlying rationale of these decisions. It exalts form—the opinion of a public servant—over substance—the acts of electors.
I would reverse the judgment.
Peters, J., and Sullivan, J., concurred.
Appellants’ petition for a rehearing was denied May 21, 1970. Draper, J.,* sat in place of Tobriner, J., who deemed himself disqualified. Peters, J., Mosk, J., and Sullivan, J., were of the opinion that the petition should be granted.
The presumption was previously contained in article IV, section 1, of the California Constitution, repealed on November 8, 1966.
In State v. Olcott, supra, the evidence of handwriting experts was introduced to prove invalidity and in In re Initiative Petition No. 23, supra, the official who refused to certify the petition attempted to contact the persons whose allegedly forged signatures appeared thereon.
We would also be required to attribute to those who allegedly falsified the signatures an uncommon degree of carelessness. Initially, all these signatures failed to meet with the clerk’s approval for some reason other than authenticity. Presumably, the persons who took the trouble to forge the names failed in some respect to follow the requirements for a proper signature as to every one of the asserted forgeries.
In other jurisdictions where there is no statutory presumption of genuineness of the signatures on a petition the presumption is held to arise as a matter of law from the prohibitions against forgery and the requirement of verification. (State v. Carter, supra, 165 S.W. 773, 780; Tyler v. Secretary of State (1962) 229 Md. 397 [184 A.2d 101, 103-104]; State v. Olcott, supra, 125 P. 303, 307; Woodward v. Barbur, supra, 116 P. 101, 103-104; Rousso v. Meyers (1964) 64 Wn.2d 53 [390 P.2d 557, 560].)
After some days of trial the court announced that it ruled in favor of petitioners on the 94 signatures in question and that borderline cases had been resolved in their favor. The court stated in its memorandum opinion that this conclusion was intended to apply not to handwriting comparison but to other reasons initially advanced by the clerk for disqualifying the signatures, such as the failure to add an address, the use of initials, or some other technical defect.
The court stated in its memorandum opinion with regard to the petitioners’ claim that they should have been permitted to introduce evidence regarding the genuineness of the signatures, “. . . it is difficult to envision what further examination could have been conducted. In every case the signatures were personally examined and compared by the court. In many cases the court agreed with the Clerk. In all cases dissimilarities were apparent which . . . were given greater significance by the Clerk than by the court. But in every case the reason for the Clerk’s action was apparent from the comparison alone.”
These 125 signatures consist of the 94 discussed above and 31 additional signatures as to which the trial court agreed with the clerk’s determination of dissimilarity.
Assigned by the Chairman of the Judicial Council.