The

and as a result of the recount Vare lost 14, Wilson gained 8, and the other senatorial candidates gained 5 votes in the aggregate. ballot box contained 35 ballots which could not lawfully have been placed there, and the list of voters contained the names of five 'repeaters," as well as 18 names not on the register.

Forty-eighth ward, tenth division.-In the tenth division of the fortyeighth ward Vare lost exactly 10 votes as a result of the recount and Wilson gained 8. The ballot box contained 10 specimen ballots (which are printed on pink paper). These ballots were all marked stright Republican by the same individual and had apparently been counted by the election officers as valid votes; but there could have been no excuse for the use of them, since 243 unused official ballots were returned by the election officers.

Forty-eighth ward, eleventh division. In the eleventh division of the forty-eighth ward, although apparently proper tallies appeared on the tally sheets, Vare lost 39 votes and Wilson gained 33 when the ballots were recounted. The ballot box contained 24 ballots bearing evidence of fraudulent marking, as heretofore described and set forth in detail in Mr. Melcher's report. The voters' list included tne names of 20 "repeaters, as well as 26 names not on the registration books. Thirty-nine of the names in the voters' list appeared in alphabetical order, and the number of names marked as voted in the voting check list fell short of the ballots in the box by 67.

ELECTION OFFICERS

It has been stated that the election in each division is conducted by a board of five election officers—a judge, a majority and minority inspector, and two clerks. The judge and majority and minority inspectors who conducted the polls at this election were those who were chosen by the voters, to serve for two years, at an election in November, 1925. There were in that year 1,492 election districts in Philadelphia, and a total of 4,253 election officers were chosen therein. Of these, 2,839, or 67 per cent, were sponsored by the Republican organization, the sponsorship being evidenced by the signatures of one or both of the.division leaders to the nominating petition, by virtue of which the name of the particular candidate appeared on the ballot at the primary preceding the election of 1925.

The records themselves thus show that in the case of at least twothirds of the men elected to serve at the 1926 elections the Republican organization was definitely responsible for their choice. As a matter of fact, the actual responsibility of the organization was undoubtedly even more extensive than this, since it is a matter of common knowledge that the Republican division leaders are the persons who usually circulate the petitions to nominate judges and inspectors of election.

Legality of election boards. For the purpose of insuring a bipartisan election board, the constitution of Pennsylvania provides that "each elector shall have the right to vote for the judge and one inspector;" two inspectors (the two receiving the highest number of votes) being elected in each division. (Art. 8, sec. 14.) The clerks are merely appointed by the inspectors. Each election board should therefore contain one minority representative and a clerk in sympathy with him.

Furthermore, when members of an election board are absent on election day, there are elaborate provisions of law for so completing the board that minority representation will be insured. (Act of 1839, P. L. 522, secs. 16 and 17.)

Comparison of the records of the election held on November 2, 1926, with the returns of the 1925 election, shows, however, that only 25 per cent of the persons who signed these records as judges and inspectors were the duly elected officers. And a comparison of the 1926 election records with the registration books of that year shows that in 609 divisions the whole election board comprised none but enrolled Republicans.

The validity of the elections in those districts is open to serious question, quite apart from the existence of any fraudulent practices

PITTSBURGH

REGISTRATION

Registration books from 31 divisions of the city of Pittsburgh were submitted to experts for an examination similar to that made of signatures in the registration books of Philadelphia. These books were selected from divisions in the second, third, fourth, fifth, ninth, tenth, eleventh, twenty-first, and twenty-seventh wards of Pittsburgh.

In the 31 registers examined 142 signatures were found attributable to the persons who made the other entries in them and 48 signatures had been falsified by other persons. Time did not permit of a more

extensive examination.

PREPARATION OF BALLOTS FOR ELECTION

The law governing the preparation of ballots for election is the same in Allegheny County as in Philadelphia, and the county commissioners are similarly responsible and are required to keep a record like that obtained from the county commissioners of Philadelphia. However, no record of this sort was produced in response to the Senate's subpoena, and the explanation was given that the records of ballots issued had been destroyed. The material necessary in order to make a proper check of the disposition of ballots issued in Allegheny County has therefore not been available.

ELECTION

The same examination of election records in Pittsburgh has been made as that which was made in Philadelphia.

The comparison of votes originally returned for senatorial candidates with the results of the recount of those votes shows the existence of less fraudulent counting than in Philadelphia. One hundred and sixty-seven of the 689 election districts of Pittsburgh show a correct count of the senatorial vote by the election officers, and the average chance of a Pittsburgh voter to have his vote counted correctly was therefore more than twice as great as that of a Philadelphia voter. With 689 divisions in the city, Vare gained 184 votes in 90 divisions and lost 1.381 in 320 divisions, a net loss of 1,197 votes. Wilson gained 1,287 votes in 327 divisions and lost 164 in 83 divisions, a net

gain of 1,123. Other candidates for Senator gained 30 votes in the entire city. As already stated, it must be borne in mind that wherever in this report reference is made to a number of votes received, counted, gained, or lost by a particular candidate or candidates, the figure given represents what appears on the face of the ballots, and therefore often includes ballots protested on behalf of one or the other of the parties to the contest.

The tally sheets in Pittsburgh also disclose a situation quite different from that existing in Philadelphia. In only five divisions in the whole city were the tally sheets lacking in tallies.

The number of names recorded as voted in Pittsburgh, although not registered, and also the number of "repeaters" shown by the lists of voters appear to have been inconsiderable.

A considerable number of fraudulent ballots were cast, however. One way in which this is made evident is by comparing the total number of names recorded as voting in the voting check list and the total number of names recorded in the list of voters for each division with the highest vote returned as cast for any office in the division and with the total number of ballots in the ballot boxes. In 150 divisions the highest vote for any office exceeds the number of names written in the list of voters (after the deduction of repeated names) and in 170 divisions it exceeds the number of names checked in the voting check list as having voted. These divisions are scattered through every ward in the city.

The total number of ballots in the boxes of 106 divisions exceeds the number of names written in the lists of voters (after the deduction of repeated names), and in 147 divisions there were more ballots in the boxes than there were names checked in the voting check lists. In 7 divisions the number of ballots cast actually exceeded the number of registered voters in the divisions, and in 4 other divisions the election officers returned more votes than there were registered voters.

Since the record of ballots issued to the respective divisions was not available, it has not been possible to determine whether or not any fictitious ballots put in an appearance at the polling places. The unused ballots and stubs were available, however, and the number of stubs in each division has been compared with the number of detached unused ballots, and with the number of spoiled and cancelled ballots, in order to ascertain whether or not the ballot boxes contain ballots which could not be accounted for on the basis of the number of stubs returned. This comparison disclosed that the boxes of 85 divisions contained a total of 2,154 ballots which could not be accounted for on this basis. Unused ballots which had been detached from the stubs and so were ready to be handed to a voter for marking, or to be intermingled with the voted ballots when the box was opened, were found in 249 election districts, more than a third of the total in Pittsburgh. The number of detached unused ballots in these divisions amounted to 1,580.

The same examination of ballots in the boxes which was made for Philadelphia was repeated for Pittsburgh, and this examination disclosed that ballot-box stuffing had occurred on the same scale. The total number of ballots found in the boxes in such condition that they could not have been lawfully inserted there was 331. A large proportion of these ballots were totally lacking in creases and the remain

der, although appearing to have been folded, could not possibly have been passed through the slot even though folded in accordance with the creases which remain. The only way that any of these 331 ballots could have been inserted in the box was at a time when the lid of the ballot box was open, either before the polls were opened or after they were closed. These ballots were distributed among the boxes of 33 divisions.

In arriving at these figures care has been taken to include only ballots regarding the condition of which there could be no reasonable question. As in Philadelphia, a very large proportion of the stuffed ballots were found together in groups when the ballot boxes were opened, and an equally large proportion bear indications of having been marked by the same person.

No examination of Pittsburgh ballots has been made by handwriting experts. In each of 32 divisions in Pittsburgh the ballot box disclosed five or more ballots on each of which a cross was either erased or scored out, and another cross inserted if necessary to the advantage of Mr. Vare. The total of ballots so tampered with was 354.

In Pittsburgh, as in Philadelphia, attempts were often made to account for the deposit of fraudulent ballots in the boxes by inserting false names in the lists of voters. In 33 divisions, sections of the lists of voters, which purported to be a current record of the names of voters in the order of their appearance to vote, contain from 10 to 71 names, all listed in alphabetical order. The total of names in alphabetical groups is 876.

Illustrative of the conditions found in Pittsburgh are the following divisions:

Third ward, second division.-In the third ward, second division, the registers contained 17 forgeries by one or more of the registrars, and three by other persons. Vare lost 5 and Wilson gained 7 in the recount, a total difference of 12 votes as compared with the original returns. There were 189 ballots in the box, but only 188 stubs, and every ballot contained a purported valid vote for Senator. The voters' list contained the names of four "repeaters" and 44 names, all told, were listed in alphabetical order. But the total number of names in the voters' list, after deducting the "repeaters," was only 183, and the total registration in the division was only 187-two less than the senatorial vote.

Third ward, sixth division.-In the third ward, sixth division, the election board's return for United States Senator was: Vare, 126; Wilson, 0. The recount showed 128 ballots in the box, all marked for Vare and none for Wilson or any other candidate. Only 126 stubs were returned to account for these ballots. Every ballot but two was found folded separately, indicating that they had never been opened, much less counted, by the election officers. The two additional ballots were found folded together. The voters' list contained the names of two people who were not registered and the names of seven "repeaters.' Fifty-five of the names on the list were arranged in alphabetical order. The net number of names in the voters' list (deducting the "repeaters") was only 119.

Third ward.-Conditions throughout the whole third ward seem to have been as bad as in the two divisions indicated. The total of

S. Rept. 1858, 70-2--4

ballots found in the boxes which could not have been lawfully deposited there was 91, distributed through six divisions, and no less than 262 names were found arranged in alphabetical order in the voters' lists of this ward.

Fourth ward.-Conditions in the fourth ward were little better. Here the most manifest fraud was in tampering with ballots after they had apparently been lawfully voted by electors, and in the insertion of names in the voters' lists in alphabetical order. A total of 63 ballots containing erasures and alterations was found in this ward, and a total of 141 names was entered in alphabetical order in the voters' lists.

No recount could be made of the senatorial vote in five divisions of this ward, since in three of them the ballots were not delivered to the Sergeant at Arms of the Senate, and in the other two only ballots dealing with the primary election were so delivered. Particular attention, however, may be given to the first and fourth divisions.

Fourth ward, first division. In the first division, although Wilson was originally credited by the election officers with four votes, the recount showed that not single one of the ballots in the box had been marked for any candidate but Vare, though Vare's true vote was 18 less than that which had been credited to him. But the ballots in the box numbered 25 less than the stubs returned from this division, from which it may reasonably be inferred that these ballots were destroyed by the election officers. Eight ballots were found in the box totally lacking in creases; and to conceal the deposit of these and other ballots, 39 names were entered in alphabetical order in the list of voters. It may be noted in passing that the total number of names in the list of voters in this division was 23 more than the number of ballots in the box. This, of course, would tend to explain, at least in part, the disappearance of the 25 ballots unaccounted for. Fourth ward, fourth division.-In the fourth division of the fourth ward, although Vare was credited with 100 votes and Wilson with 5 votes on the election officers' return, the box was found to contain only 23 ballots altogether, 18 of which were marked for Vare, 4 for Wilson, and 1 for Kane. Here, upon the basis of stubs returned, 82 ballots were not accounted for, and the number of names written in the list of voters exceeded the number of ballots in the box by exactly the same amount. Of the names in the list of voters, 19 were arranged in alphabetical order.

An examination of the signatures to the tally sheets and return sheets certified to by the election officers of this division discloses that all five names appear to have been written by the same person, as well as the signatures to the oaths of office required by law.

Fifth ward, first division. In the fifth ward, first division, with 119 voters registered, the election officers' return showed that all but two had voted, and the 117 ballots in the box were marked; 114 for Vare and 3 for Wilson. The board filed no list of voters in the prothonotary's office, but in the voting check list they recorded only 93 as voting. The number of ballots in the box (and also the vote for Senator), hence, was 24 greater than the number of voters recorded, and 30 uncreased ballots were among those found in the box.

Fifth ward, twelfth division. In the fifth ward, twelfth division, 224 net stubs were returned by the election board, but only 212 ballots were found in the box. The voters' list numbered 213 names