Kolkata: Last year, Samserganj in Murshidabad’s border belt made national headlines after communal tension over the Waqf Bill left three people dead. This year, the constituency is in the news again, for a different but allied reason. In the special intensive revision (SIR), 91,712 voters – more than one in three names – have been deleted from the rolls of Samserganj. The highest number in West Bengal. When the Election Commission uploaded the appellate tribunal supplementary lists for 247 polling stations in the early hours of April 22, only a single name was cleared.Samserganj was never supposed to be the epicentre of a democratic controversy of this scale. It was, until recently, just another politically sensitive assembly constituency, shaped by familiar Bengal faultlines of party rivalry, poverty, out-migration and local patronage. But on the eve of polling, the constituency has become the site of something far more serious. The electoral roll revision here has been so sweeping and so uneven that it raises a disturbing possibility that the most decisive battle may already have been fought before a single vote was cast.At first glance, Samserganj is a conventional high-stakes seat. It is assembly constituency number 56, part of the Maldaha Dakshin Lok Sabha segment. The seat has changed hands with Bengal’s shifting political tides. After delimitation, the Communist Party of India (Marxist) won it in 2011, Trinamool Congress captured it in 2016 by a narrow margin, and then strengthened its hold in 2021 with a victory margin of more than 26,000. In this election, TMC, Congress and Bharatiya Janata Party are once again locked in a triangular contest.But the numbers behind the voter list suggest that the electoral field itself may have been altered before campaigning reached its final stretch.To compare the constituency with others in Bengal, click on the two dashboards here and here, to check the data for yourself.Samserganj is one of the most Muslim-majority seats in West Bengal, with the social composition of the seat deeply intertwined with its political outcomes. Out of 247 booths, 180 have Muslim populations of 80% or more. When the threshold is broadened to 50% or more Muslim voters, that number rises to 207 booths, leaving only 40 booths below that mark. In other words, the “high Muslim” category is not a marginal slice of the constituency but the constituency itself.That is what makes the booth-level deletion pattern so politically significant.A total of 91,712 voters were deleted from the draft roll in Samserganj, amounting to 36.27% of the electorate under revision. Of these, only a minority were removed through standard identification of ASDD – Absent, Shifted, Displaced, or Dead – voters. The number deleted through that route stands at 16,836. Another 2,459 names, or 0.97% of the voters could not map themselves with the 2002 electoral roll, the base document for the SIR. The overwhelming driver of the purge was the ‘Under Adjudication’ process. Out of 108,400 voters flagged as Under Adjudication, as many as 74,775 were ultimately deleted. That means that nearly 68.98% of all those pushed into adjudication were struck off the rolls.This is the heart of the story. The apparent injustice does not lie simply in the final number of deletions but in the pattern itself – in who was marked, where the cuts fell, and how sharply the burden rises with the minority concentration.The data reveals what can only be described as a flagging skew. In booths with a Muslim population of 50% or more, an average of 467 voters per booth were put in the Under Adjudication list. In booths with less than 50% Muslim population, the comparable figure was just 65. In other words, Muslim-majority booths saw more than seven times as many voters pushed into the adjudication pipeline.That initial skew is the real demographic anomaly, because once voters entered the adjudication process, the outcome itself appears to have been strikingly similar across demographic categories. Whether a booth was overwhelmingly Muslim or overwhelmingly non-Muslim, the proportion of flagged voters who were eventually deleted remained clustered in the same broad range, roughly 67 to 71%, matching the aggregate Under Adjudication deletion rate of 68.98%. On paper, the adjudication machinery appears demographically blind at the stage of final disposal.But that superficial parity only throws the deeper distortion into sharper relief. If one group of booths is feeding in seven times more names than another, then similar deletion rates at the end do not prove neutrality. They merely show that once inside the funnel, the machine behaved consistently. The discrimination lies earlier, at the point of entry.Because Muslim-majority booths were flooded with adjudication cases at such disproportionate volumes, they saw their total eligible voter base cut by roughly 39%. In booths with lower Muslim populations, by contrast, the average reduction was only around 10%.That gap is too large to be explained away as administrative noise.And then came the final blow. In the early hours of April 22, the Election Commission uploaded the appellate tribunal supplementary lists for 247 polling stations in Samserganj. Only one name was cleared, that of TMC leader and Dhuliyan Municipality Chairman Inzamul Islam. Only he will be able to vote this time. This transforms the meaning of what happened here. This is no longer merely a story of mass flagging and deletion but a story in which the final institutional avenue for correction appears to have yielded almost nothing at all.The tribunal’s sheet clearing Inzamul Islam.The political implications are impossible to miss.When the deletion pattern is mapped against historical booth-level election performance, a clear partisan effect emerges. TMC’s booth strength rises alongside voter deletion. The party shows a positive correlation of +0.30 with deletion rates. Put simply, the better TMC performs in a given booth, the higher the deletion rate tends to be. The BJP shows the opposite pattern, with a negative correlation of -0.15. The better the BJP performs in a booth, the lower the rate of deletion tends to be. Congress shows a flatter, slightly negative correlation, meaning its strongholds were not targeted with the same statistical precision as TMC’s, but it still suffers heavily because it relies on the same minority electorate that has borne the brunt of the revision.That means the deletions are not simply concentrated in Muslim-majority booths. They are concentrated in the social terrain where the BJP’s principal rivals draw their strength.The estimated electoral effect is severe. If the official total of 91,712 deleted voters is distributed party-wise according to how each booth voted in recent elections, then the losses fall overwhelmingly on TMC and Congress.Based on 2024 Lok Sabha voting patterns, the estimated loss stands at roughly 38,269 potential voters for Congress, 35,030 for TMC, and 13,402 for the BJP.Using the 2021 assembly voting pattern, the estimated loss becomes even starker. TMC is projected to have lost around 46,040 potential voters, Congress around 32,235, BJP around 4,882, and CPI(M) around 2,778.No such estimate can claim mathematical certainty. Voters do not reproduce past booth averages with perfect fidelity. But the political direction is unmistakable. The parties taking the greatest hit are the ones whose support base is concentrated in the Muslim-majority booths that were funnelled into adjudication at extraordinary volume. In Samserganj, that means TMC and Congress.Also read: Three Constituencies, Over 1.2 Lakh Deleted Voters: Realities of the Bengal SIRTaken together, the scale of that damage is enormous. The two opposition-to-BJP parties appear to have lost roughly 73,000 to 78,000 potential supporters combined, depending on whether one uses the Lok Sabha or assembly benchmark. That is the real electoral weight of the revision.The BJP, by contrast, appears comparatively insulated. Its stronger booths were generally spared the kind of mass adjudication seen elsewhere. That does not mean BJP voters were untouched. It means their voter base was left far more intact than that of its principal rivals. The effect is obvious. Even without winning over large numbers of new voters, the BJP’s proportional position in the constituency is likely to rise because opposition-leaning voter pools have been cut down far more aggressively.There is another reason these numbers cut so deep. Samserganj is not a document-secure, socially stable constituency. It is a landscape of bidi labour, informal work, fragile housing and repeated river erosion. Families here live with displacement, shifting addresses and bureaucratic precarity as part of ordinary life. In such a place, electoral revision is never a purely technical exercise. It can become a test of who is legible enough to the state to remain a voter at all.That is why the official defence of “procedure” cannot answer the democratic question. A system that looks formally administrative on paper can still be politically devastating in effect. In Samserganj, the numbers suggest exactly that. A process that selected Muslim-majority booths for vastly heavier adjudication, hollowed out their rolls, and then left almost no visible space for correction at the tribunal stage.The constituency is set to vote on April 23. When Samserganj votes, the real question may not be who wins the seat, but how more than a third of its electorate was cut down before the contest even began.