Read 2 of 13

Ward 18 and the counting problem

Published 7 October 2026

Bandu Sonule's last day is the best-documented single case of the outbreak, because The Hindu's Serish Nanisetti went back to Manoli village and reconstructed it hour by hour from the family. In that account, Sonule came to the cotton farm of Shankar Ramakrishna Chowdhuri on September 19 looking for work, filled a container with the pesticide he was handed and started a petrol-powered sprayer. The crop was taller than the five-and-a-half-foot labourer. His daily wage was Rs 300. He covered his nose and mouth with a handkerchief and sprayed from about 9 a.m., stopping only for tobacco. The date was Tuesday, September 19, 2017.

He came home as it was getting dark, bathed, made a ritual offering to his late father, and then began to complain of dizziness and nausea. His wife Geeta, quoted in the same report, said he had been fasting for the new-moon observance and she first put the nausea down to that. He then vomited through the night and said he could not see. In the morning she took him to the rural health centre at Ghatanji, 5 km away, where he was given eye drops and told he would be fine. By the time they returned home he was convulsing, and, she said, they could not take him back to a hospital that night because there was only Rs 15 in the house.

The next morning a brother lent them money. An autorickshaw from Manoli to Yavatmal costs Rs 200, the report says, and the family instead went back to Ghatanji, waited two hours for an admission card to the Vasantrao Naik Government Medical College, took a bus, and reached the hospital on the evening of Thursday, September 21, 2017. The district headquarters is 42 km from Ghatanji, and the bus ride takes more than an hour. Sonule died on Saturday, September 23, 2017, according to the Hindu report of October 7. Where sources differ, it is on small things: The Hindu's 2017 report gives the widow's age as 39 and the later feature gives the dead man's full name as Chandrakant Bandu Sonule, aged 40, while the village and the date of death are consistent across both.

The Hindu's later feature also reconstructs the first admissions. It says a farmer named Dashrath Rama Chavan was brought to the medical college on August 6, and that Devidas Ramaji Madavi was brought in by relatives the next day. August 6, 2017, was a Sunday. The Indian Express, put Madavi's death on August 19 and his first spray about twelve days before. This is the earliest death cited in the contemporary press. By the end of November, the Hindu feature says, 21 people (13 farmers and eight farm workers) had died of pesticide inhalation in Yavatmal district alone, while activists put the dead at 50 in the other cotton-growing regions of Maharashtra, such as Akola, Amravati, Dhamangaon, Nagpur and Wardha.

Those two numbers, 21 and 50, have different owners and different scopes, and they are not meant to be added. The first is a count of deaths attributed to inhalation in one district, as reported by the paper. The second is an activists' estimate for other districts. The BBC's reporter Jaideep Hardikar wrote something close to the second when he described at least 50 farmers dead in Maharashtra since July, according to officials and media reports, of whom 19 were in Yavatmal.

The BBC's report has some of the earliest clinical detail. It says doctors first noticed suspected poisoning in the last week of July and admitted 41 patients with similar symptoms: vomiting, dizziness, respiratory problems, visual impairment and disorientation. It says at least ten of the farmers then being treated were on life support and 25 others had suffered visual impairment, as hospital authorities put it. It quotes Dr Ashok Rathod, the dean of the local government hospital, calling the phenomenon "very unusual" because the hospital usually treated farmers who had deliberately swallowed pesticide. That is a distinction that matters for what follows: poisoning by swallowing is a familiar emergency for Indian district hospitals, and this was a different kind of case.

The same BBC report says that in September district officials asked agriculture scientists to carry out a field study, and that the resulting report blamed farmers for not taking precautions while spraying, with no protection kit, eyeglasses or gloves. The reporter set out the open questions: whether farmers had used a spurious pesticide, whether they were using a new formulation they did not know about, and whether any advisories had been issued.

The Hindu's report of October 2017 adds the farmers' side of the height argument, and a retailer's rebuttal. It reports that over 1,000 farmers in Yavatmal had been exposed since July and that 23 had died, a figure that differs from other counters of the same weeks. Ramdas Vadai, a farmer in Manoli, told the paper that the cotton plants had grown unusually tall that year, up to six feet, and attracted pests; some farmers said the lack of rain meant the plants had received more fertilizer. Maruti Barbhate, the patient who had been tied to a bed in ward number 12, had by then recovered.

The retailer's view in the same article came from Bhavesh Gandecha, owner of Jalaram Krishi Kendra in Ghatanji, who said there had been no variation in the ingredients of the pesticides sold that year and that shops had put up large boards telling farmers about precautions. "Why would the Krishi Kendra owners want to kill farmers when their livelihood depends on them?" he said. He was apprehensive of government action, the paper noted. The report also records that the district administration had booked five Krishi Kendra owners for "selling unauthorised pesticides".

The counting problem is best seen by lining the figures up. The table below lists each count as its owner stated it, with date and scope, and does not total anything.

Where sources differ, then, they differ because they count different things on different days. Some of these counts are of deaths the hospital attributed to inhalation, some are of deaths the press attributed to pesticide poisoning, and some are of people exposed and not necessarily admitted. The 21 in the late-November feature is lower than the 23 in the October piece, and the reports do not say why. The record does not reconcile them.

A hospital record exists that is more systematic than any of the press counts. A retrospective study by community medicine doctors at the same medical college, published in 2020, reviewed the case papers of patients admitted with inhalational organophosphorus poisoning between January and December 2017. It found 535 admitted patients with clinical features of poisoning. Of these 517 (96.6 percent) were men and 18 were women, and 391 were aged 21 to 40, with a mean age of 33.3 years. The most common presenting symptoms were nausea and vomiting (325 patients), giddiness (221) and loose motions (151). Cholinesterase was tested in 407 patients, and of those 13 died against 394 who lived, and the study notes that death was far more common where cholinesterase was lowest. The study counts a full calendar year, so its 535 is not comparable to the press counts for the Yavatmal season, and the 13 deaths are among the 407 tested only.

The Pesticide Action Network India sent its own team on Friday, October 13, 2017, and Saturday, October 14, 2017. Its report, Pesticide Poisonings in Yavatmal District in Maharashtra: Untold Realities, says the two-day visit covered local organisations, one village, the family of a farm worker who died in September, farmers who had been poisoned, cotton fields and the Yavatmal medical college, where the team spoke to patients, physicians and nurses. Its participants included the PAN India director D. Narasimha Reddy and a lawyer from the Human Rights Law Network.

The PAN report lists the products named by victims and doctors. Its table gives Polo as a thiourea, Profex Super as an organophosphate plus pyrethroid, Starthene (spelled Starcene in the table) and Monocil as organophosphates, and Police as a neonicotinoid plus pyrazole, with two brands, Gayathri and Tonic, whose technical ingredient the team could not trace. The report adds that farmers and labourers do not memorise the names of the chemicals, so the names they gave to the team might be trade names that came to mind rather than the products they sprayed. It also reports that a total of 16 brands are used on cotton and that farmers say they often mix them. The team said it did not get consumption data for Yavatmal.

The report is sceptical of the hospital's treatment pathway. It says all the pesticide poisoning cases are treated as medico-legal cases and that the hospital categorises the accidental ones as "inhalational poisoning". The team said this could be the first step in administering wrong therapy, because staff told them they treat all of these as organophosphate poisonings, and when asked how they knew, said they asked the patient or attendants for the name of the chemical sprayed. That is a finding about the hospital's method, offered by an advocacy group, and the medical college did not respond in the report itself.

The Hindu's feature gives a view of the ward from the inside. The hospital had set aside Ward 18 for the poisoned patients, and a nurse who asked not to be named said it had become "like a VIP ward" for the number of visitors it drew. At the time of the report only seven patients remained, but for three months the ward had been full. Some patients screamed in pain and complained of burning eyes, the nurse said, and relatives were asked to hold others down while some had to be tied to beds. Doctors at the hospital declined to speak, citing the inquiry.

The patients' own accounts explain the economics. Vinod Shirji Dhanke, a farm hand from Dhamangaon, told the paper that with a litre of petrol he could spray 30 tanks, that on one day he sprayed 50 times and earned Rs 1,500, and that a few weeks later he began to lose his sight. He said he was selling his Rs 5,500 pump for Rs 2,000 and would never spray again. The Hindu's report says farm workers were paid on the number of times they emptied pesticide containers, so they "stretch themselves to the limit".

A surviving farmer, Avdhut Laxman Wadgure, 48, from Manoli, told the paper he sprayed Profex Super on his pigeon pea crop on November 14, from 9 a.m. to 3:30 p.m., without a break because he was fasting, and three days later, when he took his mother to the hospital, he felt feverish and could not see properly. The date shows the poisonings had not stopped when the government's inquiry began. Wadgure's case sits in the hospital's late-season admissions, after the main spraying period that the Indian Express put between July 1 and about September 3.

Which chemicals were involved was itself unsettled. The Hindu's feature says agriculture department officials identified organophosphates, pyrethroids, neonicotinoids and pyrazoles, and that district officials found a cocktail named Polo, Steamreach and Tonic at Chowdhuri's farm. Polo is made by Syngenta and has diafenthiuron as its active ingredient, the report says, while the bottles labelled Tonic and Steamreach had no identifiable chemical names. All of them had a blue band, indicating toxicity. This is the first mention in the contemporary press of the product that would later anchor the Swiss legal case against Syngenta, and in 2017 it appears only as one of several names on one farm.

At this point the story's dispute is already visible. The state's early reports pointed to farmers' handling and the height of the crop. Hospital staff reported an unusual pattern of illness. Advocacy groups said that combination products, weak labels and absent training were the cause. None of these accounts was complete, and the government's own special team (the SIT), constituted in the first week of October, was supposed to reconcile them within three weeks.

Not every case came from Yavatmal's main road. PARI's reporter went to Chikhali (Kanhoba) village in Ner tehsil in January 2018, and wrote up what villagers there told him. Sanjay Bhorkhade, 35, an Adivasi farm labourer of the Andh community, said he had sprayed for five to six hours in October 2017, on and off for a week, a cocktail of pesticides on a 10-acre cotton farm where he worked on an annual contract known locally as saldari, which paid him Rs 70,000 a year. He said he "almost lost my sight" and still felt eye irritation and fatigue months later. His wife Tulsa told PARI she had been hearing about pesticide sickness and deaths from women in the village and was terrified he would lose his eyes.

His employer, the village sarpanch Uddhavrao Bhalerao, told PARI that around 5 that evening Sanjay came to his house with swollen eyes and he took him by motorbike to Arni town, about 10 km away, and from there to the Yavatmal district hospital, another 40 km. Bhalerao said the farm had used the same chemicals as in earlier years, the only difference being the extreme humidity and the robust growth of the cotton plants. That statement, from an employer, is one of the few in this record from a landowner whose worker was poisoned, and it supports the height and humidity explanation that the state later adopted.

PARI reported that five workers in the village became sick in September and October, and that four survived. The fifth, Dnyaneshwar Tale, 45, a marginal farmer, was taken to the sub-district hospital in Arni on October 1, 2017 and to the Yavatmal hospital the next day. According to PARI, he stayed in hospital for over two months, improving for a day and then slipping back, his brother Gajanan said, and died with an official cause of "septicaemia". The report said that the medical records showed infections took over after the insecticides had damaged his immune system and vital organs. That reading is PARI's account of the records as it understood them, and the hospital's own account is not in the piece.

The death certificate matters for counting. A death on the record as septicaemia after two months in hospital may or may not appear in a count of poisoning deaths that runs from July to October or November. PARI said the state collated its figure of around 50 dead and 1,000 sick from public and private hospitals, and described the dead as dying between September and November 2017. Where sources differ, The Hindu's feature gave 21 for Yavatmal by the end of November and PARI's figure covers Vidarbha, so the two cannot be compared one to one. The table above stays unsummed for the same reason.

The first paper to put numbers on the cluster, before the task force and the opposition were quoted, was the Times of India's Nagpur edition. Its report said that in the last three weeks at least nine farmers or farm labourers had died after accidentally inhaling toxic fumes while spraying pesticides in Yavatmal, where the Bt cotton crop was infested with pink bollworm and needed regular spraying. Hospital officials confirmed at least 160 patients treated for toxic inhaling that month and over 300 in the last three months, with around 60 under treatment at the time, according to the dean, Dr Ashok Rathod, who called the number unusual and said the wards were overflowing.

The same report quoted the head of forensics at the hospital, Dr Manish Shrigiriwar, confirming at least six deaths from pesticide poisoning in the previous two weeks. It said the district administration had no definite information: the resident district collector, Narendra Fulzele, said that after a farmers' delegation registered a complaint on Monday the administration began an inquiry, and on Tuesday wrote to the civil surgeon and the district health officer asking for data within two days. The paper added that the situation got official attention only after the farm distress task force took it up, and that Kishor Tiwari had warned almost a month earlier that Bt cotton was no longer immune to pest attack and could lead to heavy pesticide use. The report does not carry a date in the text retrieved here.

Dr Rathod told the paper that if proper protective gear was not used and precautions were not observed, sprayers could be in for trouble. Set beside his later comment to the BBC that the phenomenon was very unusual, the two statements show a hospital that saw the pattern but framed it first as a matter of precaution. The hospital's early figures, 160 in a month and 300 in three, are far lower than the 472 over three months he gave The Hindu a week or two later, and the 438 the fact-finding team recorded on October 12. The numbers rose because the season and the count were both still running, and no early report claimed to be final.

Get the next file

One email when a new file is published. No paid service needed.

Email Subscribe

You are subscribed.