Deadlocks, Vacancies and Scandals, 1866 to 1912: What Went Wrong with Legislative Election

Lesson 02 ended with the Act of July 25, 1866 and Daniel Clark's prediction that a faction blocking a Senate election to the detriment of state business would not happen "once in a hundred years." This lesson is the record of the next forty-six. It leans on George H. Haynes, whose 1906 book The Election of Senators compiled the deadlock and bribery data that the Senate Historical Office still uses, and on the Senate's own contested-election files. Every figure below was checked against one or both.

The deadlock count

Haynes defined a deadlock carefully, because the word could be stretched. He counted only a contest "so bitter and unyielding that its only issue could be, either the preventing of any election, or the choice of a senator who would win his high office not because of any preeminent qualifications, but because he chanced to be found possessed of such qualities that the hostile and disappointed factions in the joint assembly could be reorganized under his banner and led to victory." He measured each in calendar days "from the date on which the two houses balloted separately till the date of the final vote," because the 1866 Act required at least one vote on each of those days. His data came from Appleton's Annual Cyclopedia and the New York Tribune Almanac, checked against state legislative journals in the Massachusetts State Library and by correspondence with the secretaries of state.

The table that follows is his "Record of Deadlocks," 1891 to 1905, reduced to one line per state. The full table has 46 rows, two of them double deadlocks in which both of a state's seats were open (Delaware 1901, Nebraska 1901) and two of them double elections at the end of a long fight (Delaware 1903). Where the Internet Archive's scan is unreadable, the cell says so rather than guessing.

State	Deadlocks 1891-1905 (years)	Longest (days, ballots)	Sessions ending with no election
Delaware	7 rows (1895, 1899, 1901 twice, 1903 twice, 1905)	1895: 114 days, 217 ballots	5 (1895, 1899, both 1901 seats, 1905)
Washington	4 (1893, 1895, 1897, 1903)	1893: 51 days, 101 ballots	1 (1893)
Oregon	4 (1895, 1897, 1901, 1903)	1897: 53 days, no final count possible	1 (1897)
Montana	3 (1893, 1899, 1901)	1901: 51 days, 66 ballots	1 (1893)
Nebraska	4 rows (1893, 1899, 1901 both seats)	1901: 72 days, 54 ballots	0
Maryland	3 (1896, 1898, 1904)	1904: 16 days, 12 ballots	0
Florida	2 (1891, 1897)	1891: 35 days, about 75 ballots	0
North Dakota	2 (1891, 1893)	1893: 33 days, 61 ballots	0
South Dakota	2 (1891, 1897)	1897: 29 days, 27 ballots	0
Louisiana	2 (1892, 1896)	1892: 44 days	1 (1892)
Idaho	2 (1895, 1897)	1895: 51 days, 52 ballots (George L. Shoup)	0
Kentucky	2 (1896, 1897)	1896: 58 days, 52 ballots	1 (1896)
Utah	2 (1897, 1899)	1899: 52 days, 164 ballots	1 (1899)
California	1 (1899)	67 days, 104 ballots	1
Pennsylvania	1 (1899)	92 days, 79 ballots	1
Wyoming	1 (1893)	not legible in scan	1
Missouri	1 (1905)	60 days, 67 ballots	0
Tennessee	1 (1898)	7 days, 7 ballots	0
Wisconsin	1 (1899)	8 days, 6 ballots	0
North Carolina	1 (1903)	10 days, 9 ballots (85 candidates on the first ballot)	0
Total	46 rows, 20 states		14

Haynes's own summary of the table: "Not a few, but at least half the States of the Union, belonging to no isolated section, but States scattered the country over, from Delaware to California and from Montana to Louisiana, have suffered from serious deadlocks." And: "Fourteen contests in ten States have lasted throughout an entire session of the legislature without effecting an election. Four States have submitted to the heavy cost and inconvenience of special sessions to elect senators. Six States have preferred to accept vacancies as the penalty for their legislatures' deadlocks."

The Senate Historical Office repeats the count two ways. Its overview essay says: "From 1891 to 1905, legislatures deadlocked 46 times and failed to elect a senator 14 times." Its shorter briefing page says "45 deadlocks occurred in 20 states between 1891 and 1905." Both trace to Haynes. The difference of one is almost certainly a double deadlock counted as a single contest. If you cite a number, cite the source with it.

Idaho appears twice in the table, in 1895 and 1897, and the 1890 to 1891 dispute over Idaho's first senators is in the Senate's contested-election files. Lesson 07 covers Idaho's legislative elections in full.

Delaware: the seats that stayed empty

Hamilton wrote in Federalist 59 that a state "forbearing the appointment of senators" could not "impair the activity" of the Senate. He did not consider what it would do to the state. Delaware is the test.

The trouble began with money and a man. J. Edward Addicks, a gas company promoter who moved from Boston to Dover in 1889, spent, by the Senate Historical Office's account, seventeen years trying to get the Delaware legislature to elect him senator (Addicks v. Kenney). He split the Republican party into a faction that would vote for him and a faction that would vote for anyone else, and for a decade neither faction would yield to the other or to the Democrats.

In 1895 the legislature was deadlocked among three Republican candidates. Governor Joshua Marvil died on April 8, 1895; William T. Watson, who had been speaker of the state senate, became governor; and on May 9, 1895 Watson presided over the joint session and voted in it. Henry A. du Pont received 15 of the 30 votes cast, one short of a majority, and if Watson's vote were thrown out du Pont would have had a majority of the remaining 29. Du Pont claimed the seat. On May 15, 1896 the Senate rejected his claim "by a party-line vote of 31 to 30," with four Populists and one Silver Party senator joining the Democrats (du Pont case). Haynes's table records that session as 114 days and 217 ballots with no election; the Senate's Seventeenth Amendment page uses the same numbers. The Class 2 seat stayed vacant from March 4, 1895 until the legislature elected Richard R. Kenney, a Democrat, on January 19, 1897. The day after, a rump of the house declared Addicks elected to the same seat; the Senate seated Kenney on credentials regular on their face and never acted on Addicks's petition.

Then it happened again. In 1899 the legislature took 113 ballots over 64 days and elected nobody. Haynes describes the finish: "on a single day forty-two ballots were taken, and on the following day thirty-seven, the last of them but a few minutes before the final adjournment; yet all to no effect, and Delaware was left with but one senator in the next Congress." Twenty-seven candidates had been voted for on the first ballot. The Senate Historical Office's Delaware timeline dates the deadlock March 13, 1899 and says the seat stayed vacant until 1903. In 1901 the other seat came up, and the legislature failed to fill either, after 46 ballots over 52 days. "Delaware remained unrepresented in the U.S. Senate from March 4, 1901 until March 1, 1903," the Senate's timeline says; Haynes, listing by Congress, gives both vacancies as March 4, 1901 to March 3, 1903. For the entire Fifty-seventh Congress the state "had no part whatever in the Senate's deliberations." Only in 1903, after 41 days and 36 ballots, did the "double deadlock" break, "not till the very end of the session," electing J. Frank Allee and L. Heisler Ball. In 1905 the legislature deadlocked once more, 80 days and 51 ballots, and failed again. The Addicks fight ended only in 1906, when Delaware Republicans reunited to elect Henry du Pont, who then served until 1917. Addicks "lost his fortune and died in 1919."

Delaware seat	Vacant from	Filled	Length	How
Class 2	March 4, 1895	Kenney elected Jan. 19, 1897; credentials presented Feb. 5, 1897	About 22 months	1895 session: 114 days, 217 ballots, no election; Senate rejected du Pont 31 to 30 on May 15, 1896
Class 1	March 4, 1899	March 1903 (Allee)	Four years	1899 session: 64 days, 113 ballots, no election; 1901 session: no election
Class 2	March 4, 1901	March 1903 (Ball)	Two years	1901 session: 52 days, 46 ballots, no election for either seat
Class 1 (next term)	March 4, 1905	1906 (du Pont)	Over a year	1905 session: 80 days, 51 ballots, no election; Republicans reunited in 1906

Haynes's judgment on what this did to the state: the contest "may even subordinate all interest in a presidential campaign. 'This year, the question in Delaware is not "Roosevelt or Parker?" but "Addicks or no Addicks?"'" Whether or not Addicks said it, Haynes wrote, "there is not the slightest doubt that they stated the exact truth of the situation." Willard Saulsbury of Delaware had told the Senate in 1866 that if legislatures failed to elect "a little oftener, perhaps it would have been for the public good" (lesson 02). His state took him at his word.

What a deadlock cost

The constitutional cost was representation. Haynes listed every seat left vacant by a failure to elect, by Congress:

Congress	State	Vacant from	Filled
53rd (1893-95)	Montana	March 4, 1893	February 2, 1895
53rd	Washington	March 4, 1893	February 19, 1895
53rd	Wyoming	March 4, 1893	February 6, 1895
54th (1895-97)	Delaware	March 4, 1895	February 5, 1897
55th (1897-99)	Oregon	March 4, 1897	December 5, 1898
56th (1899-1901)	Delaware	March 4, 1899	(not filled; continued)
56th	California	March 4, 1899	March 5, 1900
56th	Pennsylvania	March 4, 1899	January 17, 1901
56th	Utah	March 4, 1899	February 4, 1901
57th (1901-03)	Delaware (both seats)	March 4, 1901	March 3, 1903
59th (1905-07)	Delaware	March 4, 1905	1906

"The record of seven Congresses, therefore, shows that only one has not had its Senate cut down by vacancies due to deadlocks in state legislatures," Haynes wrote. "In the Fifty-third Congress, three seats were vacant; in the Fifty-sixth, four." In most cases the loss "amounted to the loss of a senator for the entire term of a Congress; for the senator's service could amount to little, when he was seated only within a month of the end of the last session."

Governors tried to fill the gaps and the Senate stopped them. The original vacancy clause let an executive appoint only "during the Recess of the Legislature" (lesson 01). When a legislature had been in session and failed to elect, the Senate held, "following the unbroken precedent of three-quarters of a century," that there was no recess vacancy to fill. It excluded Mantle of Montana and Allen of Washington in 1893; Wyoming's appointee, Beckwith, resigned before the Senate ruled; Corbett of Oregon was excluded in 1897; and in 1899 Quay of Pennsylvania, appointed by his own governor after a 92-day deadlock, was excluded too, partly because Pennsylvania's constitution told the governor to call a special session instead. After Quay, Delaware's governor made no appointment and Utah's appointee did not bother to present credentials (Haynes, ch. III; Senate Historical Office overview: "the Senate exacerbated the problem in 1893 when it refused to seat senators appointed to fill vacancies caused by deadlocks in the legislature").

The other remedy was a special session, and that cost money. Kentucky's 1897 special session ran nearly seven weeks and deadlocked again before electing William J. Deboe, "who had been nominated only five days before." California's lasted thirteen days; Oregon's, in a state "too outraged by the fiasco" of the regular session to tolerate delay, elected on the fourth day. Haynes could not compute the total cost "either in money or in the derangement of public affairs," but he gave what figures he had: Tennessee's special session of January 17 to February 5, 1898, called to fill a vacancy from a death, was estimated by Tennesseans at $20,000; California's members' pay alone ran $960 a day; "it would be a low estimate to say that each day of a special session costs the unfortunate State not less than $1,000."

The third cost was the legislature's own work. The daily joint vote, Haynes wrote, caused "interference with the normal work of state legislation, ranging all the way from the exaction of a few hours of the legislature's time to the virtual annihilation of the legislature, which was chosen to guard the interests of the State." Elections decided in the last minutes of a session were common: Oregon's senator in the 32-day, 58-ballot contest of 1895 (George W. McBride; Haynes's narrative misdates it to 1893, but his own table places it in 1895) was chosen on the 58th ballot by a majority of one after being nominated "fifteen minutes before the time when the term of the legislature must expire"; Montana's 1901 legislature stopped the clock and elected Paris Gibson at 3:30 a.m. after its legal term had ended, taking 22 of its 66 ballots on the final day; Missouri's 1905 legislature elected William Warner "within ten minutes of the time set for the final adjournment" after sixty days, a man "not proposed until the balloting had been going on for nearly fifty days." The Missouri election took place in a riot: members tried to stop the chamber clock, a ladder went out the window, desks were torn from the floor, and a legislator broke the clock's pendulum with ink bottles. Kentucky's 1896 session ended under martial law, "a capital filled with troops," with the Frankfort correspondent of the Courier-Journal reporting that few of the 132 members lacked "one or two pistols concealed." Colorado's 1903 joint assembly met with Denver police backing one party and a request for state troops from the other.

The fourth cost was what Haynes called misrepresentation. In 1893, the year Montana, Washington and Wyoming elected no one, "North Dakota, a Republican State with a Republican legislature, returned a Democratic senator, while Kansas elected a Democratic senator, although the Legislature contained only a handful of Democratic voters," both through fusion deals in deadlocked assemblies. A deadlock compromise, Haynes argued, "vastly increased" the chance that a state would send to Washington for six years "a man representing a party that is in distinct minority in the State."

Bribery: the money follows the votes

A legislature of 100 to 150 members, voting aloud, with a seat worth six years in the Senate at stake, was a market. Bryce in 1893 was already pointing to the Atlantic Monthly and Nation articles of W. P. Garrison on "the evils of the present system." Haynes in 1906 found that "in not less than seven States, during the past fifteen years, charges of corruption have been put forward with enough of presumptive evidence to make them a national scandal." In Ohio, California and Montana the legislatures themselves investigated, "and, in each case, the majority of the committee of investigation declared that the evidence of the corrupt use of money was conclusive." In California a campaign manager had spent more than $20,000 to elect legislators, and the speaker of the house resigned under charges of taking gifts and loans. Charges "have been chronic in connection with Delaware and Pennsylvania senatorial elections." Connecticut's 1904 to 1905 election drew claims of "widespread corrupt use of money."

Cases reached the Senate itself only when a candidate won. Haynes listed ten senators "brought to the bar of the Senate" on bribery charges from 1857 to 1899, the first being Simon Cameron (lesson 02) and the other nine after the 1866 Act; the Senate Historical Office counts "nine bribery cases" between 1866 and 1906 (briefing). Haynes: "In the lifetime of a single generation, thus, the Senate has had to deal with nine cases of alleged bribery, while only one had arisen in all its earlier history."

Year	Senator, state	Charge	Outcome
1857	Simon Cameron, Pa.	Election procured "by corrupt and unlawful means"	Committee: charge "too vague and indefinite"; no investigation
1872-73	Samuel C. Pomeroy, Kan.	1867 election; then legislator Alexander York testified Pomeroy paid him $2,000 down and $5,000 more for his vote in 1873	First charges "totally failed"; 1873 special committee majority called it a plot; Thurman's minority found him guilty; term ended March 3, 1873 with no action
1872	Powell Clayton, Ark.	Votes secured by money and offices	Charges "not sustained"; minority disagreed
1873	Alexander Caldwell, Kan.	1871 election; testimony that he would pay $250,000 for the seat; former governor Carney admitted $15,000 to withdraw; Caldwell said the contest cost him over $60,000	Committee unanimously found bribery, Feb. 17, 1873, and recommended declaring the election void; Caldwell resigned March 24, 1873
1875	George E. Spencer, Ala.	Bribery	"Not proven"
1877	La Fayette Grover, Ore.	Bribery	Evidence did not sustain the charges
1879	John J. Ingalls, Kan.	Bribery by both sides	Ingalls exonerated personally though corruption was proved; the precedent that only personal complicity unseats
1886	Henry B. Payne, Ohio	Bribery of the Ohio legislature	Senate refused further investigation, 44 to 17
1898	Marcus A. Hanna, Ohio	Bribery by agents	Majority: no evidence Hanna knew; minority wanted inquiry; no action
1899-1900	William A. Clark, Mont.	Bribery of the Montana legislature	Unanimous committee report against him, April 23, 1900; resigned May 15, 1900; re-elected 1901
1909-12	William Lorimer, Ill.	Bribery of the Illinois legislature	Kept his seat 40 to 46, March 1, 1911; election declared invalid 55 to 28, July 13, 1912

Haynes read the record as an indictment of the Senate as much as of the legislatures. The Senate "has shown extreme reluctance to investigate such charges, and has bound itself by precedents which make not only the unseating of a member, but even the pursuit of a thoroughgoing investigation, practically impossible, except where the evidence of guilt is overwhelming and notorious." The precedent was the Ingalls case: unless the senator himself was complicit, proof that his legislature had been bribed did not void his election. And the Senate had no jurisdiction at all over a candidate who bought votes and lost, which excluded "many of the most flagrant instances of recent corruption in senatorial contests, as in California and Delaware." In "every one of the four cases which have occurred within the past twenty-five years there has been no question whatever that bribery was at least attempted, if not carried out," yet "no senators have been expelled for bribery, and ... only two have resigned in consequence of these investigations."

Montana, 1899: William A. Clark

Clark's rival in Montana was Marcus Daly, whose own unlimited spending against Clark two members of the Senate committee would later note in a concurring statement. Haynes's table shows the 1899 Montana election at 17 days and 17 ballots, with 21 candidates on the first ballot. Clark presented his credentials and was admitted on December 4, 1899; the same day his opponents filed a petition charging bribery, and the Committee on Privileges and Elections was authorized to investigate (Senate Historical Office, Clark case). The petition alleged spending far beyond the $2,000 limit in an 1895 Montana law. The committee heard 96 witnesses. "The testimony detailed a dazzling list of bribes ranging from $240 to $100,000," paid by a scheme run by Clark's son that included paying off mortgages and debts, buying ranches, financing banks and handing cash to legislators; the margin had depended on eleven Republican legislators who switched under suspicious circumstances; Clark admitted destroying all his personal checks from the campaign. On April 23, 1900 the committee "returned a report unanimously concluding that William Clark was not entitled to his seat," relying on the Pomeroy and Caldwell precedents. Haynes gives the committee's arithmetic: "of his apparent majority of fifteen, more than eight votes had been obtained through illegal and corrupt practices."

Clark did not wait for the vote. On May 15, 1900, "clearly aware that he did not have the necessary votes to keep his seat," he spoke in his own defense and resigned. The same day Montana's acting governor appointed him to fill his own vacancy. When the governor returned three days later he telegraphed that Martin Maginnis would fill the seat instead. "Credentials for both Clark and Maginnis were presented to the Senate, which ordered them to lie on the table." Daly died in November 1900. In January 1901 a new Montana legislature, most of whose winners had received Clark's financial support, elected him for the same term, no charges were raised, and he was seated without objection on March 4, 1901 and served until 1907 (Montana timeline). Haynes's deadpan version: the legislature "forthwith 'vindicated' him, by reelecting him to succeed himself; and he took his seat without further protest." Haynes placed Clark among the eight senators whose presence in the Senate "finds its chief, if not its sole explanation in their great wealth."

Illinois, 1909 to 1912: William Lorimer

Lorimer's case is the one that changed the Senate's standard, and it ran in parallel with the amendment. Illinois's legislature deadlocked for months in 1909; Lorimer, a Chicago Republican, was finally elected with the help of many Democratic votes. The Senate Historical Office's case page says he was elected "after months of deadlock" and seated on June 18, 1909. (The exact date of the legislature's final ballot appears differently in the sources checked for this lesson; the Roseburg Evening News chronology gives May 24, 1909. It is left unstated here.) In 1910 the Chicago Tribune charged that the seat had been won by bribery, and four legislators eventually said they had received bribes to vote for Lorimer; several later recanted. On May 28, 1910 Lorimer himself asked the Senate to investigate. The Committee on Privileges and Elections majority reported on December 21, 1910 that the charges were not proved. Albert J. Beveridge of Indiana dissented, noting that the four legislators had testified to bribes and had "unusual sums of money in bills of large denominations" shortly afterward, and three other members joined his conclusion that the election was invalid. The committee majority applied the old standard: as the Senate Historical Office's "Senator Ousted" essay puts it, the Senate "invalidated elections only when the accused senator had actively promoted the bribery," and under that rule Lorimer had done nothing wrong. After six weeks of debate, on March 1, 1911, the Senate voted 40 yeas to 46 nays against declaring his election void.

The reaction was national. The Senate Historical Office says "the resulting storm of public outrage, combined with an infusion of recently elected progressive-minded members, led the Senate on June 12, 1911, to approve a long-pending constitutional amendment providing for direct popular election of senators" (lesson 05). About a week before that vote, new public charges led the Senate, on June 7, 1911, to authorize a special committee of eight members of the Privileges and Elections Committee to investigate again. It heard 180 witnesses and filled eight volumes. Its majority report of May 20, 1912 "failed to find any evidence that a fund" had been raised to elect Lorimer. A three-member minority, John Worth Kern of Indiana (who had replaced Beveridge), Kenyon and Luke Lea of Tennessee, reported that the investigation "established conclusively that at least ten of the votes cast" for Lorimer were corruptly obtained and recommended that the seat be declared vacant (Roseburg Evening News, July 13, 1912, chronology).

On July 13, 1912 the Senate adopted Lea's resolution: "Resolved, That corrupt methods and practices were employed in the election of William Lorimer to the Senate of the United States from the State of Illinois, and that his election was therefore invalid." The Omaha Daily Bee of July 14 headlined the vote "55 to 28 Against" and reported eight senators paired; the OCR of its body text reads "65 to 28," which conflicts with its own headline, and the Roseburg paper's garbled "So to 28" is consistent with 55. The Senate Historical Office page gives the date and the result, "Unseated," without a tally. Lorimer was not expelled, which would have taken two thirds. His election was declared invalid by majority under Article I, Section 5. He became, in the Senate Historical Office's words, "the last senator to be deprived of office for corrupting a state legislature." The amendment that would make such cases impossible had already passed Congress and was awaiting ratification by the state legislatures when the vote was taken (lesson 05).

The Millionaires' Club and the newspapers

By the 1890s the Senate had a nickname. Haynes: "Of late years it has become very common, both in conversation and in the press, to refer to the Senate as the 'Rich Men's Club,' the 'Paradise of Millionaires.'" Puck had run a cartoon on January 19, 1887 titled "The Senate of the Future, A Close Corporation of Millionaires". The National Archives' Treasures of Congress exhibit summarizes the Progressive reformers' charge that the Senate was a "millionaire's club" serving powerful private interests (exhibit text).

Bryce, writing in 1888 and revising in 1893, gave the measured version: "The Senate now contains many men of great wealth. Some, an increasing number, are senators because they are rich; a few are rich because they are senators; while in the remaining cases the same talents which have won success in law or commerce have brought their possessor to the top in politics also." He noted "sarcastic references" to "the sumptuous residences which senators have built on the new avenues of Washington," admitted "more sympathy for the capitalist class among these rich men than there would be in a Senate of poor men," and still judged the Senate "far from being a class body like the upper houses of England or Prussia or Spain or Denmark" and "far too dependent ... upon public opinion, to dream of legislating in the interest of the rich" (The American Commonwealth, vol. 1, ch. XII). The Senate Historical Office's own summary of Bryce notes that he admired the Senate as a "center of gravity" and judged that it "has succeeded in making itself eminent and respected" (Idea of the Senate: Bryce). He also explained, in the passage lesson 02 quoted, why elections had become fights: "Every vote in the Senate is so important to the great parties that they are forced to struggle for ascendency in each of the State legislatures by whom the senators are elected," and "a senatorial election often leads to long and bitter struggles; the minority endeavouring to prevent a choice, and so keep the seat vacant."

Haynes tried to count. Charles Dudley Warner had estimated that among the 88 senators of 1892 "there were but six millionaires; sixteen were men of wealth ranging from $100,000 to $700,000, while the rest were men of moderate means, many of whom might fairly be called poor." Checking two published lists of wealthy Americans against the Senate of the Fifty-eighth Congress (1903 to 1905), Haynes found twenty senators on a list of men worth at least $300,000, among them Aldrich of Rhode Island, Alger of Michigan, Clark of Montana, Depew of New York and Lodge of Massachusetts. Then he sorted every senator by the chief reason he was there. Under "Wealth," subgroup "Rich Men," he placed eight, "Men whose presence in the Senate finds its chief, if not its sole explanation in their great wealth": Alger, Ankeny, Clark, Dryden, Kean, Kearns, Newlands and Wetmore, with an average service of 4.4 years. Under "Representatives of Corporate Wealth," men "highly acceptable to great corporate interests," he placed four: Depew, Elkins, Martin and Stewart. Two more he put in one or the other. That is fourteen of ninety, with eight owing their seats mainly to money. The club existed. It was not the whole Senate.

The press made it the whole Senate anyway. On February 17, 1906, David Graham Phillips's "The Treason of the Senate" began in William Randolph Hearst's Cosmopolitan (Senate chronology). The Senate Historical Office describes the series as portraying senators "as pawns of industrialists and financiers," says it "was largely fictionalized," and credits it with having "further galvanized public support for reform" (Seventeenth Amendment page). Four years later the Chicago Tribune's Lorimer stories did what Phillips's prose could not: they produced sworn testimony, two Senate investigations and a vacated seat.

Why the system broke: three readings

Haynes's reading was institutional. The 1866 Act's first separate vote told every faction its strength; the daily joint ballot with no penalty let a minority hold out; the Senate's Ingalls precedent made bribery nearly unpunishable; and the stakes, a six-year seat controlling federal patronage in the state, rose every decade. "Experiences such as these, exceptional though they still are, have nevertheless become so frequent and so widespread that in recent years they have given rise to a determined propaganda, which no longer contents itself with an attempt to correct obvious defects in the law ... but which demands that these elections be placed directly in the hands of the people."

Bryce's reading was about parties. Legislative election had "helped to make the national parties powerful, and their strife intense," inside the state legislatures. By 1893 "the election of senators has in substance almost ceased to be indirect": the caucus decided, and the caucus was "very often arranged beforehand by the party managers." Where party control was firm, elections were quick and the legislature was a rubber stamp. Where it split, as in Delaware's Republicans or Montana's Democrats, the legislature became the battlefield and the state paid.

Schiller and Stewart's reading, from their roll-call dataset of 1871 to 1913, is that parties reliably controlled which party won but "had much less control over the universe of candidates who competed for votes in Senate elections," so that money, ambition and factional fights inside the majority party drove the deadlocks and scandals (Electing the Senate). On that view the breakdown was less about legislatures versus voters than about a nominating system without a primary. The states that adopted the Oregon System after 1904 were, in effect, inventing one, and that is where lesson 04 begins.

Questions and answers

Is the right number 45 or 46 deadlocks?

Haynes's table, the source for both, has 46 rows in 20 states, 14 ending with no election. The Senate Historical Office's overview essay uses 46 and 14; its briefing page uses 45 in 20 states. Two of Haynes's rows are the double deadlock of Delaware in 1901, which a reader counting contests rather than seats would merge into one, giving 45. Either is defensible if you say what you are counting and cite Haynes.

How long was Delaware without any senator?

Two years, from March 4, 1901 to the first days of March 1903, covering the whole Fifty-seventh Congress. One of its seats had already been empty since March 1899, so that seat was vacant four years. A separate vacancy ran from March 1895 to early 1897. The Senate's Delaware timeline and Haynes's vacancy list agree on these dates within a day or two.

Did the Senate find that Clark bribed the Montana legislature?

Its Committee on Privileges and Elections did, unanimously, on April 23, 1900, after hearing 96 witnesses describe bribes from $240 to $100,000 and finding that more than eight of his fifteen-vote margin were corruptly obtained. The full Senate never voted because Clark resigned on May 15, 1900. Montana's legislature re-elected him in January 1901 and the Senate seated him on March 4, 1901 without objection.

What did the two Lorimer votes decide?

On March 1, 1911 the Senate rejected a resolution declaring his election void, 40 yeas to 46 nays, applying the rule that only the senator's own complicity voids an election. On July 13, 1912, after a second investigation of 180 witnesses, it adopted Lea's resolution that "corrupt methods and practices were employed in the election" and "his election was therefore invalid," 55 to 28 per the Omaha Daily Bee. The second vote abandoned the personal-complicity rule.

Why could the Senate not simply expel senators whose legislatures were bribed?

Expulsion needs two thirds and was understood as punishment for the member's own conduct. Where the member could claim ignorance, as Ingalls, Payne, Hanna and Lorimer did, a two-thirds vote was never available. The alternative, declaring the election invalid by majority, was recommended for Caldwell in 1873 and used for Lorimer in 1912. Between those dates the Senate mostly took no action, which is what Haynes meant by "extreme reluctance."

What happened to the governors' appointees from deadlocked states?

The Senate refused to seat them. The original vacancy clause allowed appointments only during a legislative recess; a legislature that had sat and failed to elect had created no recess vacancy. Mantle of Montana and Allen of Washington were excluded in 1893, Corbett of Oregon in 1897, and Quay of Pennsylvania in 1899. After Quay, governors in Delaware and Utah stopped trying.

How much did the deadlocks cost in money?

Haynes could not total it. He gave Tennessee's estimate of $20,000 for a special session of January 17 to February 5, 1898, California's $960 a day in members' pay, and a working floor of $1,000 a day for any special session. Four states held special sessions to elect senators between 1891 and 1905. The larger cost was in regular sessions consumed by balloting, which no one priced.

Was the Senate really a "Millionaires' Club"?

It had a wealthy minority that was growing. Warner counted six millionaires among 88 senators in 1892. Haynes found twenty senators of 1903 to 1905 on a list of men worth $300,000 or more, and judged that eight of ninety were there chiefly because of their wealth, with four more serving corporate interests. Bryce said the number of rich senators was increasing and that they were more sympathetic to capital than a poorer Senate would be, while denying the Senate was a class body. The nickname fit the trend and exaggerated the count.

Hard questions

Haynes wanted direct election. Isn't his deadlock table advocacy dressed as data?

He was an advocate and said so in his later chapters. But the table is reproducible: he named his sources (Appleton's, the Tribune Almanac, state journals, secretaries of state), defined his terms, and listed every row with state, days, ballots and result. The Senate Historical Office, which has no reform agenda, adopted his 46 and 14. Anyone can check a row against a state journal. The fair criticism is about the window: fifteen years chosen because they were the worst. Pre-1866 deadlocks (Maine, Tennessee, California, Indiana) show the problem was older, and quiet decades are not in the table. The data are sound; the framing is his.

Thirty-some states never deadlocked at all in those years. Doesn't that mean the system worked almost everywhere?

Mostly, yes, and Haynes's "at least half the States" cuts both ways. Where one party controlled the legislature firmly, as in most of the South and much of the Northeast, elections were quick and uneventful. The trouble concentrated in closely divided or factionally split states: new Western states, border states, Delaware. But "worked" needs a definition. Bryce's point is that in the smooth cases the legislature was a formality for a decision the caucus or convention had already made, so the states where the system looked healthiest were the states where it had stopped being indirect election in any real sense. Lesson 09 takes up whether that matters for federalism.

Clark bought a legislature, resigned, and was back within a year with no objection. Doesn't that show the Senate was simply complicit?

It shows the limits of the Senate's power more than its complicity. The committee did its job: 96 witnesses, a unanimous adverse report. Clark's resignation prevented a vote. When a new legislature, elected with his money but without any proof of bribery in the 1901 vote itself, chose him again, the Senate had no precedent for refusing a facially regular election because of what had happened in a previous one. The remedy for a legislature willing to elect a man its predecessor's investigation had condemned was with Montana's voters, who in 1901 had no direct say. That is the reformers' argument in one case.

The Lorimer unseating came after the amendment had passed Congress. Doesn't that make the case irrelevant to why the amendment passed?

The first Lorimer vote, not the second, is the one historians tie to the amendment. The Senate's refusal on March 1, 1911 to unseat him despite, in the Senate Historical Office's words, "considerable evidence against Lorimer" produced the outrage that preceded the June 12, 1911 Senate vote for the amendment by three months. The 1912 unseating is relevant for a different reason: it shows the Senate changing its own standard under public pressure at the very moment the public was taking the power away. Both halves belong in the story.

If bribery of legislators was the problem, why not prosecute the legislators under state law instead of amending the Constitution?

Because it was not done. Haynes found that the Senate could not reach losing candidates or bribed legislators, and that state prosecutions almost never followed even where legislative committees found corruption "conclusive," as in Ohio, California and Montana. The California speaker who resigned under charges "was not prosecuted." Legislatures investigating themselves, and prosecutors answering to the same parties, did not produce convictions. Reformers concluded that the market would exist as long as 100 to 150 people could sell a Senate seat, and that the only sure remedy was to make the electorate too large to buy. Whether that remedy worked, and at what cost to the states' standing in Washington, is the rest of this course.