OPINION
HARBISON, Justice.Appellants filed two petitions for a referendum election on proposed amendments to *953the charter of the Metropolitan Government of Nashville and Davidson County, Tennessee (Metro). The petitions were filed on May 16, 198$, but their legal sufficiency was challenged by local officials on the ground that they did not contain signatures of enough qualified voters to satisfy the charter provisions. The Chancellor held that the number of signatures on the petitions was sufficient, and Metro appealed.
While that appeal was pending Metro officials began checking the signatures on the petitions for authenticity. This apparently had not been done prior to the chancery hearing even though that hearing was not held until late June, 1983, several weeks after the petitions were filed. The Chancellor’s order holding the number of signatures to be adequate also provided that the proposed amendments should be placed on the ballot for an August 4, 1983 general Metropolitan election, provided that the signatures appearing on the petitions were valid and were those of qualified voters.
Checking samples of the signatures on the petitions against official registration records by computers, election officials concluded that large numbers of them were invalid — a sufficient percentage to disqualify the petitions. Accordingly they “de-cer-tified” the petitions, over objection of appellants.
This Court then dismissed Metro’s appeal as being premature. The cause was remanded to the Chancellor to hear the exceptions of appellants to the methods used in checking the signatures on the petitions. After an evidentiary hearing held on July 26, 1983, the Chancellor concluded that the procedures used by Metro officials were reasonable and not arbitrary. He upheld their determination that enough signatures on the petitions were invalid to disqualify them. Appellants seek review of that decision, and appellees renew the issue raised in their original appeal as to the number of signatures required by the charter. Because the latter question would be disposi-tive if appellees are correct, we will first discuss that issue.
(1) The Charter Requirements
The Metropolitan charter, § 19.01, requires that a petition for a referendum on a proposed amendment be signed by “ten (10) percent of the number of the registered voters of Nashville-Davidson County voting in the preceding general election.”
The issue is whether this reference is to a preceding Metropolitan general election (regularly held in August) or the previous state general election, which occurred in November, 1982. If the August 1982 or August 1979 Metropolitan elections are meant, facially the petitions contain a sufficient number of signatures. If the reference is to the state general election held in November 1982 (to which no Metropolitan offices were subject), the number is insufficient.1
The Chancellor held that since the subject involved is the amendment of the Metropolitan charter, the intent of the Charter Commissioners was to refer to the number of votes cast in a Metropolitan election rather than to the number in a state or national election. We agree. The charter, § 15.01, provides for Metropolitan general elections and refers to them as such. We think that the reference in § 19.01 under consideration here clearly is to municipal elections. The judgment of the Chancellor with respect to that question is affirmed.
(2) Verification of Signatures
Because of the large number of signatures to be checked against the permanent voter registration rolls, the Metropolitan Clerk decided to employ electronic data processing for this purpose. Obviously other methods had been used on previous occasions.
We quote from the Chancellor’s opinion:
*954“Employing the assistance of the Department of Data Processing and Computer Services, most of the names and addresses representing the signatures and addresses on the two petitions were keyed in to a library file on the computer. The permanent voter registration file is already contained in the computer. A program was run comparing the names in the library file with the names in the permanent voter registration file, and the computer reported how many matches were obtained.
“From Petition # 1 which contained 13,005 signatures, Data Processing keyed in 9,898 signatures to be compared with the permanent voter registration file. Of these, the computer reported that only 2,925 names matched names on the permanent voter registration file. The Metropolitan Clerk concluded that there were not sufficient signatures of registered voters on Petition # 1 to equal the required 9,473, since even if all 3,107 of the unchecked signatures were valid, the total of valid signatures would be only 6,023.
“From Petition # 2 which contained 12,127 signatures, Data Processing keyed in 7,156 signatures to be compared with the permanent voter registration file. Of these, the computer reported that only 2,041 names matched names on the permanent voter registration file. The Metropolitan Clerk concluded that there were not sufficient signatures of registered voters on Petition # 2 to equal the required 9,473, since even if all 4,971 of the unchecked signatures were valid, the total of valid signatures would be only 7,012.
“Under the method used by Data Processing, the addresses were not checked. Illegible names were not checked. David R. Johnson, Assistant Director of Data Processing, testified that the computer would report a match only if the name keyed in corresponded exactly, letter by letter, with a name ,|ts listed on the permanent voter registration file. If a middle initial was omitted or if a middle name was spelled out in full when the name on the permanent voter registration file had a middle initial, the name would not be reported as a match by the computer. Any variation whatsoever would result in the signature being rejected as not that of a qualified registered voter.”
Although concluding that this method of verifying signatures was “not ideal under all circumstances,” the Chancellor held that under the facts of this case it was “not improper.” He held that a public official must be allowed discretion in the performance of a duty involving the exercise of judgment,2 and he noted that the task of ascertaining the validity of more than 25,-000 handwritten signatures was arduous. He stated that the Metropolitan Clerk could not be required “to exhaust every imaginable avenue” in an attempt to verify signatures of persons signing the petitions differently from their official registration.
There can be no doubt as to these general observations. Further, it appears to be well settled that officers receiving petitions such as those involved here are required to exercise quasi-judicial functions in passing upon their sufficiency. Their decisions ordinarily cannot be reviewed in mandamus proceedings like these, provided that the officials have not acted in bad faith or arbitrarily. See generally 26 Am. Jur.2d, Elections § 200 (1966); 29 C.J.S., Elections § 69 (1965).
Nevertheless it must be conceded, as recognized by the Chancellor, that the methods used in the present case were very stringent. They undoubtedly resulted in the invalidation of numerous signatures of *955qualified voters — indeed, the disqualifications seem to have been almost on a wholesale basis. We conclude that these methods contravene the public policy of the state as declared by the General Assembly and must, therefore, be deemed arbitrary.
The state Election Code, Title 2 T.C.A., governs “[a]ll elections for public office, for candidacy for public office and, ... questions submitted to the people .... ” T.C.A. § 2-1-103. Apparently some of its provisions were not called to the attention of the Metropolitan electio'n officials, nor were they referred to by the Chancellor.
Particularly pertinent here are the provisions of T.C.A. § 2-l-107(a) and (b) (1982 Supp.). The first subsection requires that every person signing a petition for nomination of a candidate or for a referendum shall include a residence address. Failure to include at least a street address operates to disqualify the signature. Apparently in the present case street addresses were not even checked, although these are criteria for valid signatures.
The second subdivision is even more pertinent. It requires that signatures on petitions conform to those on permanent registration cards with respect to script or printing. However, it expressly states that failure to comply with this requirement “shall not operate to disqualify any nominating signature or candidate’s signature.” T.C.A. § 2-l-107(b).
By Tenn. Public Acts 1981, ch. 478, the following significant amendment was added to this provision:
“Provided, further, a person’s regular signature shall be accepted just as his legal signature would be accepted. For example, for the purposes of this section ‘Joe Public’ shall be accepted just as ‘Joseph Q. Public’ would be accepted.”
The methods used by Metro officials in this case failed to take into consideration the foregoing statutes, and in some instances positively contravened them. The results obtained from the methods used are not in conformity with law and cannot be approved.
Of course, these petitions must be verified by appropriate means, either by revised data processing or by other methods used in the past. Petitions of this sort are not without precedent. While local officials have a certain amount of discretion, they obviously must conform to state statutory requirements.
The decision of the Chancellor as to the second issue is reversed. The cause is remanded to the trial court for further proceedings consistent herewith and for determination as to when these referenda shall be submitted to the voters if the petitions are found to contain sufficient signatures of qualified voters. All costs incident to proceedings in this Court are taxed to appel-lees.
FONES, C.J., and COOPER, BROCK and DROWOTA, JJ., concur.. One petition contained 13,005 signatures and the other 12,127. In the November state election 139,523 votes were cast. In a Metropolitan election held in August 1982 there were 94,730 votes. In the last previous mayoral election held in August 1979 about 89,000 votes were cast.
. Citing Bradley v. State ex rel. Haggard, 222 Tenn. 535, 438 S.W.2d 738, 740 (1969), a reapportionment case.