Means of Ascent

by Robert A. Caro

  The “bloc vote” that Lyndon Johnson received with the help of the men in the Brown Building was substantial even for Texas. In the counties where George Parr “just counted ’em”—Duval, Starr, La Salle, Jim Hogg, Zapata and Brooks—the total reported on Election Day was 10,323 for Johnson, 1,329 for Stevenson, a plurality for Johnson of almost 9,000 votes. (To that plurality could be added the 700 votes from Box 13 in Jim Wells County.) Parr-controlled “Mextown” and “Niggertown” in Corpus Christi added another 3,000 votes to Johnson’s advantage. And these were not the only “ethnic votes” that Johnson received from the area bordering the Rio Grande. He came out of the Valley—an area that included not only Webb County, where Judge Raymond “just wrote down what he wanted,” and the six counties in Parr’s domain but also Cameron, Hidalgo, Dimmit, Maverick and Nueces—with a margin of more than 27,000. The “remarkable turnaround” in San Antonio was achieved through the 10,000 votes he received in “ethnic” areas. Although key officials in both camps say that 10,000 was the number of votes controlled—on Johnson’s behalf—in that city, the number of votes that were, in a term euphemistically used by these officials, “changed over” between the two primaries is only about 8,000. Even if the lower—8,000—figure is used, however, that figure, added to the Valley’s 27,000 plurality, means that Johnson’s plurality from the “bloc” totaled 35,000 votes.

  If his total was impressive, his percentages were more so. In Duval County, Lyndon Johnson received 99 percent of the vote; in the six counties of George Parr’s domain combined, he received 93 percent. He was given 82 percent of the vote in Judge Raymond’s Webb County. In the key Mexican-American and black sections of Corpus Christi, his percentage was 80 percent. In each of these areas, in other words, at least four out of every five voters voted for him. So he received a 35,000-vote plurality from areas in which the voting pattern was dramatically out of keeping with the normal patterns in a democracy. How much of that 35,000-vote edge can be said to have been “bought,” either by payments to jefes and other political bosses who wrote down voting totals with little or no reference to the actual votes that had been cast, or by payments to bosses who rounded up men and women and made sure they voted as they were told to, or by payments to election judges who counted for Johnson votes that had actually been cast for Stevenson, or by payments to individual voters who voted in accordance with the payment they received, or by payments to deputy sheriffs who transported to the polls and directed the voting of men and women who did not even know whom they were voting for—how many of those votes were “bought” and how many would have been cast for Lyndon Johnson even had no money changed hands cannot, after forty years, be determined. But from the descriptions given by men familiar with the voting in the Valley and on the West Side that Election Day in 1948, it is apparent that the overwhelming majority of those votes—not merely thousands of votes but tens of thousands—fall into that category.

  The dramatic disparity between the returns from the West Side and the Valley and the returns from the rest of Texas indicates the disproportionate significance that this area played in the election.

  Coke Stevenson carried three counties in the state by a margin of at least 80 percent, two of them so tiny as to be of almost no significance in the election. In one, Kenedy County, the vote was 52–8, and in the other, Roberts County, it was 181–12. The third county was Gillespie, where the Johnson Ranch is located; Gillespie gave Stevenson 80 percent of its vote, 1,014–250. These three counties produced for Stevenson a plurality not of 35,000 votes but of 977; they were a minor factor in the election. Johnson’s bloc voting was a major factor—the decisive factor. Without that plurality from the West Side and the Valley, he would have trailed Stevenson by a substantial margin.

  Despite Johnson’s harvest of these votes, however, he was still behind. His months of ceaseless campaigning, and his injection into the Texas political scene of unprecedented amounts of money, had narrowed the huge advantage which Coke Stevenson had enjoyed at the beginning of the long campaign, and had brought him within striking distance of “Mr. Texas.” His purchased votes had brought him almost even. But he was still behind by 854 votes, and the men in the Brown Building feared that Stevenson was likely to increase his lead during the blizzard of corrections—honest mistakes and normal Texas “re-counting”—that were to come. And if their own candidate didn’t make up that narrow margin, his loss would be as final as if he had lost by 100,000. Lyndon Johnson’s political career was all but over—for want of fewer than a thousand votes.

  THE ELECTION BUREAU closed at 1:30 a.m. Lady Bird begged Lyndon to get some sleep; he went into the bedroom and changed into pajamas, but a few minutes later he burst out, his hair as wild as his eyes. He spent the night in the little paneled den of the Dillman Street house, still in his pajamas, pacing back and forth or sprawled across the day-bed, and a telephone was always in his hand. In the Brown Building, Alvin Wirtz and Everett Looney were telephoning, too. Ed Clark was calling from his home. Lyndon Johnson had tried to buy a state, and, although he had paid the highest price in Texas history, he had failed.

  So now he was trying to steal it.

  The telephone calls were to local Johnson managers in thousands of precincts all across Texas. Some of the calls were to ask the managers to be vigilant against any Stevenson attempt to steal votes in their counties. But other calls had a different purpose: the local managers were being asked to “re-check,” and, in the re-checking, to “find” a few more votes for Johnson. So far as can be learned—or at least proven—forty years later, Johnson personally made no telephone request that votes be added to his totals. Clark explains that the men in the Brown Building did not want the candidate to make such requests himself; when questioned as to whether that was because Johnson “might be asked if he had called,” Clark nodded. Rather, the purpose of Johnson’s calls was to reassure the local leaders that they would be with the winner when all the votes had been counted, and to give them a sense of personal contact with the candidate. “We weren’t asking him to say anything illegal, but it was important that he give them recognition, that he tell them everything was going to be okay,” Clark says. “Because we still needed them—to validate their votes, to stand behind them.” Of course, it was not necessary for Johnson to make requests himself; others were making them on his behalf.

  Some of the men who received these calls replied that technological considerations prevented them from helping. In San Antonio, Sheriff Kilday’s chief deputy, George Huntress, received a telephone call from “the Johnson camp … inquiring into the possibility of additional votes.” Huntress reminded “the person inquiring” that San Antonio now used voting machines, and that the results had been recorded. For other recipients of these calls, the considerations were ethical. The Johnson man who took the call in San Marcos replied that “of course he couldn’t ‘find’ any more votes; the election was over.” But other calls went to a more receptive man, a man who could “find” not just a few more votes, but a few hundred more. A politician in San Diego, county seat of Duval, who took one of those calls remembers: “It was Lyndon Johnson’s office. It was very important that they get hold of George.” Early Sunday morning a rumor began to circulate in Austin, and later that morning the rumor was confirmed. The Valley had already done a lot for Lyndon Johnson: a 27,000-vote plurality. But the Valley was going to do more. In particular, George Parr was going to do more. The Duke of Duval had given Lyndon Johnson a 100–1 plurality, more than 4,000 votes, in his county alone. Now, on Sunday, with Johnson behind despite the Duke’s efforts, Duval County announced that there were more votes from the county to come. It had just been discovered, Parr’s election officials told the Texas Election Bureau, that the returns from one of Duval County’s precincts had not yet been counted; it was hoped to have them later in the day.

  OUT ON THE BANKS of the South Llano, over a telephone that had been installed—on a tree—for Election Night, Stevenson had been hearing rumors of th
e Johnson camp’s maneuvers. Confirmation came from one of his county managers in the Panhandle, who told the former Governor that he had just received a telephone call from someone in Johnson’s headquarters who was under the impression that he was Johnson’s county manager and who had asked him if he couldn’t change at least three votes. “That was how we first learned what Lyndon was doing,” Ernest Boyett recalls. “When he was behind, he got on the telephone to his managers around the state, and asked them if they couldn’t change votes.”

  Stevenson remained calm; he might not be participating in, or even aware of, what his own allies might be doing in Austin, but he was certainly aware that such maneuvering was within the normal parameters of Texas politics. But then, early Sunday morning, he heard about the Duval announcement—which went beyond the parameters.

  Stevenson and his advisers understood at once the significance not only of the fact that Parr was announcing that new votes would be coming in from the Valley, but of the fact that now, more than twelve hours after the polls had closed, these votes had not yet been counted, so that no one could tell how many each candidate had received. In later years, Johnson aides would argue that such a late finding of votes in the Valley was customary in Texas politics, but it wasn’t. As Allen Duckworth wrote in the Dallas News that day, “The thing that brought protest from the Stevenson camp was the lateness of Duval’s final report this time. The county usually reports, practically complete, on Election Night.” George Norris Green, whose book The Establishment in Texas Politics analyzes Texas elections in the 1940s, notes that Stevenson had won the Valley vote in his previous races but comments that “all of those votes had at least been turned in on time.” Since few constraints limited Parr in the number of votes he reported (“he just counted ’em”), the effect of his turning in votes late would be that he could report almost any number of votes needed—which would mean that to a considerable extent the Duke of Duval could decide the result of any close statewide election all by himself; that the only limit would be the number of poll taxes he had paid. The 27,000-vote margin already given to Lyndon Johnson by the Duke of Duval and his allies in the Valley hadn’t been enough to elect Johnson. So now, this Sunday, what, really, was George Parr doing but just waiting to see how many more votes would be needed? Jumping into his car, Stevenson accelerated so fast that he left a plume of dust behind him as he sped to Austin. But there was nothing he could do once he got there. All day Sunday, the Election Bureau was inundated in the usual blizzard of vote changes from all over the state, a few large, most small; but all day, looming over the counting of these ballots was that missing box down in the Valley.

  For Coke Stevenson, Duckworth wrote, “The news came at dusk Sunday. It was bad, as expected.” And it came at a strategic moment. All day, minor changes had in general balanced themselves out, and Stevenson had stayed ahead, by approximately the same 800-vote margin. Then, about seven o’clock, twenty-four hours after the polls had closed, upward revisions for Johnson came in from several Houston precincts—revisions in themselves so suspicious that there were immediate calls for an investigation. Suddenly, Stevenson was leading by a bare handful of votes. And then, at this crucial point, Duval announced that it now had its returns ready. The vote it had reported on Saturday night had been 4,195 for Johnson, 38 for Stevenson. Now Duval election officials said there had been 427 previously unreported votes in that “uncounted” precinct. Stevenson, they said, had received two of them; Johnson had received 425. Four hundred and twenty-five new votes (Duval’s count was now Johnson 4,620, Stevenson 40) and for the first time Stevenson was no longer ahead in the statewide totals. “In the closest major race in the state’s long political history,” the Fort Worth Star-Telegram reported, “Lyndon Johnson rode into the lead of the U.S. Senate race Sunday night on a sudden tide of votes from Duval County.” (No newspaper commented on a remarkable aspect of the Duval vote. Since another two votes would later be found there for Johnson, making Duval’s final vote 4,622 to 40, 4,662 persons thus voted in a county in which only 4,679 poll tax receipts had been issued—the 99.6 percent turnout was an astonishing display of civic responsibility.)

  On Monday, however, the Election Bureau received in the mail written returns from counties that had previously reported over the telephone, and there were the usual corrections made necessary by transpositions, discoveries of double-counting and simple mistakes in arithmetic—as well as, perhaps, other kinds of “mistakes.” But Monday brought only one new return from the Valley—80 new votes for Lyndon Johnson in Starr County (final returns from Starr were Johnson 3,038, Stevenson 170; Parr’s two key counties alone therefore gave Johnson an edge of more than seven thousand votes: 7,660 to 210)—and Stevenson received more benefits from the corrections than Johnson. Hour after hour all that day and into the evening, Johnson worked the telephone in his den, alternately sprawled, half reclining, on the day-bed and pacing back and forth across the little room, telephone in hand, lighting one cigarette from the end of another, the skin on his gaunt face drawing tighter and tighter, the circles under his eyes growing darker and darker—a candidate by El Greco. But by the end of the day, Stevenson had 494,396 votes, Johnson had 494,277—Stevenson led by 119 votes in what Duckworth, in a view echoed in newspapers throughout the state, wrote was “the most exciting Senate race Texas has ever known.”

  Tuesday brought more transpositions—including a sizable one from Eastland County in West Texas. When someone at the Bureau re-checked Eastland’s 2,645–2,317 figures, which had had the higher figure in Johnson’s column and the lower in Stevenson’s, it was found that they should, in fact, have been reversed, and that change pushed Stevenson further ahead. And under Texas election law, seven p.m. Tuesday (seventy-two hours after the polls closed) was the deadline for precinct election judges to turn over to their county chairmen all ballot boxes, containing not only the actual ballots that voters had cast but the tally and poll sheets kept by election judges and clerks. At that hour, the Election Bureau announced it had “complete” (complete in the sense that all boxes had reported) returns in all of the state’s 254 counties except for one box in sparsely populated Borden County, where about forty votes were supposed to be still unreported (and, of course, except for the two “Stevenson” counties Hansford and Kinney, which hadn’t held primaries because Stevenson’s men had not considered them necessary). At seven, the Election Bureau announced its final returns for the day: Stevenson, with 494,555 votes, was 349 votes ahead of Johnson, who had 494,206. Election Bureau Manager Bob Johnson told a reporter that “there was little doubt the bureau’s unofficial count would give Stevenson a majority.” And, as the Houston Chronicle pointed out, “The [Election] Bureau, although it cannot actually elect an officeholder on the basis of its unofficial returns, has been consistently accurate during the thirty-two years of its existence. It has never missed naming the eventual winner.” Indeed, the returns from each county had been checked and re-checked so many times by both sides that no further major changes were expected. Lyndon Johnson’s second Senate race apparently had ended like the first; “Should Johnson lose by a few hundred votes,” as Duckworth wrote, “it would be the second time he has been nosed out on his quest for a seat in the upper house.” The blaring headlines—“STEVENSON LEADS BY 349 WITH ABOUT 40 VOTES OUT”—seemed to trumpet the death of his dreams.

  LITTLE CHANGE in the situation occurred on Wednesday, September 1, the fourth day after the election. A day of checking and re-checking returns produced only seventeen revisions—most of them downward because of discoveries of double-counting—and these were so evenly divided between the two candidates that they virtually canceled themselves out, producing a net change of only thirteen votes. These were in Stevenson’s favor, and his lead increased to 362 votes, “STEVENSON’S MARGIN FIRM,” a typical headline said; the vote in the race for United States Senator from Texas seemed to have all but hardened into its final form.

  Little change occurred on Thursday, the fifth day a
fter the election. Eleven revisions reduced Stevenson’s lead by a total of 11 votes to 351. Many newspapers were treating the race as over; “STEVENSON HOLDS FINAL VOTE LEAD,” said the Brownsville Herald. His supporters exulted; in an editorial entitled “Good Senator by Bad Margin,” the Dallas News said that “Texas has made a good senatorial choice.… It is a remarkable victory when it is clearly in the face of what in mere politeness may be termed ‘bloc voting’ in four [sic] Southwest Texas counties. Duval County had come quite close to being represented in Congress by a Senator of its own.” Reporters found Stevenson at the Travis County Courthouse, performing legal work for a rancher who was his neighbor in Kimble County. Pressed for a statement, he said that “All the counties have been re-checked carefully,” and he was sure he had won. And Stevenson’s feelings were echoed by even the most cautious Texas politicians; the election seemed to be over.