SIMFEROPOL, Crimea — On March 16, as Crimeans voted in a referendum on joining Russia, a convoy of Russian minibuses and cars drew up to the center of Lytvynenkove, a village about 15 miles northeast of the peninsular capital. Members of the local self-defense committee of Crimean Tatars, the Muslim minority group who were exiled under Stalin but returned here when Communist rule collapsed, watched with trepidation as about 50 men, some in track suits and others in military uniform, got out of the vehicles. But the passengers hadn’t come to bully the local Tatar population, which had announced a boycott. Instead, they headed into the local polling station. The two white vans and the several cars were registered in Krasnodar, Russia. The men’s accents were Russian, and so from their appearance were they — those in uniform were Don Cossacks, a famed fighting force that served the czars and now, experts say, has become a sort of Pretorian guard for Russian President Vladimir Putin. “Political tourists” traveling by the van-load from one polling station to the next have been a feature of Ukrainian elections going back more than a decade — locals call it “carousel voting” — but this was the first time that anyone had heard of foreigners getting into the act, a Tatar organizer said. Tatar observers alerted their counterparts in nearby villages, and the convoy next turned up in Zuya, just three miles south. Later that afternoon, the scene repeated itself later in Petrove, a few miles further down the road, said Zair Smedlyayev, head of central election commission of the Tatar’s Kurultai, an unofficial council, who fielded many of the calls from Tatar observers in the provinces that day. The “Cossack Carousel” vote is but one of many tales that add to deep doubt about the validity of the snap secession referendum, which took place only nine days after it was announced. The United States and its European allies have said from the beginning that the referendum was illegal and that its results would not be accepted. But until now no one has examined the conduct of the referendum or the accuracy of Putin’s claim that more than 82 percent of the electorate took part and that 96 percent of them favored joining Russia, a result some U.S. politicians seem to accept. On Sunday, U.S. Senate Intelligence Committee chair Dianne Feinstein, a California Democrat, indicated she viewed the referendum as legitimate. “The Crimea is dominantly Russian, a referendum was passed,” she said during an appearance on CNN’s “State of the Union.” “That, I think, has been done,” adding later, “I get the Crimea thing.” But there are many reasons to doubt that the referendum was conducted fairly and that the result was what Putin announced. A reporter for the local Tatar ATR television station said he was able to register and vote in four separate polling stations. In Sevastopol, the home of the Russian Black Sea fleet, initial results spoke of a 123 percent turnout. The ATR television station showed film of Russian “bikers” visiting one polling location and a man departing carrying a submachine gun. Putin’s claim of 96 percent favoring secession out of an 82 percent turnout also appears to be mathematically impossible. Both the local Ukrainian community and the Tatars, who make up just under 30 percent of the 2.2 million residents of Crimea, organized a boycott. Even with defections, and there were some, both the alleged turnout and margin of victory would be impossible without many pro-Russian voters casting ballots multiple times. Putin “was using percentages out of North Korea,” said Vladimir Kazarin, a professor of journalism at Simferopol’s Russian language Vernadsky University and a former high official in the Crimean and Sevastopol government. “It’s not truthful.” While it is impossible to know from publicly available data, Kazarin estimated that at most 60 percent of Crimeans took part in the referendum, but said it could be as low as 50 percent. Mustafa Dzhemilev, 70, a survivor of the Tatar deportation of 1944 and a member of the Ukrainian parliament, said last week that data sent by local officials to the Russian FSB intelligence agency showed that only 34.2 percent of the Crimean population took part in the referendum. He didn’t say how he got the number, but Smedlyayev of the Tatars’ election commission endorsed the claim. “I always trust the information that Dzhemilev gets from his secret channels,” he said. As for local pro-Russian officials’ claims that 40 percent of the Tatars voted, Tatar leaders here reacted with disbelief. Nariman Dzhelyalov, deputy leader of the Tatar Mejlis, of the 185,000 Tatars eligible to vote, at most 1,000 took part. He said that estimate is based on reports from 300 observers deployed to watch the turnout in every town or village where Tatars make up half or more of the population. Even two weeks after the referendum, detailed polling-place results have not been released. Residents here also tell of efforts to control any anti-Russia campaign before the voting, which was originally set for May 25, then quickly advanced to March 16 after masked gunmen without insignia, presumed to be Russian special forces, occupied the Crimean legislature. All television broadcasts by Ukrainian channels were blocked, and only Russian or pro-Russian channels were allowed to broadcast, along with Tatars’ ATR. Posters with the slogan “Together with Russia,” were slapped on every available billboard. Russia also pressed a campaign promising Crimeans, 600,000 of whom are retirees, that their pensions would more than double if the peninsula joined Russia. “It was a very aggressive campaign, and it appealed to many older voters who don’t think critically,” said Svetlana Kocherga, a Ukrainian professor of philology at the Crimean State University at Yalta, on the Black Sea coast. Although Ukrainian law requires that all advertising and rallies cease the day before any vote, it continued at fever pitch through election day, she said. Smedyaeyiv also charged that organizers of the referendum offered bribes of $100 each to members of local voting commissions for favorable results. “One day before the referendum, officials gathered the heads of the voting commissions and leaders of different organizations and demanded that the the results show an 80 per cent turnout and 80 per cent approval rating,” he said. Such allegations are difficult to confirm. There was no systematic monitoring of voting practices on March 16. Armed pro-Russian “self-defense” forces prevented the deployment to Crimea before the vote of observers from the Organization for Security and Cooperation in Europe. “This is one of the biggest voting stations in Yalta, where more than 2,000 people ordinarily vote,” she said. But there was a long list of people who had signed the voting register. Also missing from the material that voting board usually collects were control tabs that are detached from each ballot before it is cast. “I asked them where is the control tag,” Potapova said. She said the workers responded that they were doing as instructed.