A house sample becomes a federal warning
Published 13 October 2026
Del Toral's interim report, dated June 24, 2015, assembled the known water-quality problems and the missing protection against lead. It listed coliform and trihalomethane violations, then described the absence of corrosion-control treatment after the change from Detroit water. The memo said high lead results were to be expected in a system without mitigating treatment, particularly where homes had lead or partially lead service lines. It also raised concern about ferric chloride used to help remove organic matter, noting the potential effect of a higher chloride-to-sulfate ratio on corrosion. The memo did not claim that one house measured the exposure of every resident. It argued that the system's treatment conditions and the material in its pipes made the household findings a public-health concern rather than a result to dismiss as isolated. F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
The memo compared the new operation with what Flint had stopped receiving. Detroit's finished water contained orthophosphate to control lead and copper. When Flint changed source, that treatment was not continued. Del Toral wrote that all large systems serving more than 50,000 people were required to install and maintain corrosion control. His central point was therefore not simply that the river was dirty or that a test result looked unusual. A protective treatment used in the earlier supply had disappeared from the new one, while the water still passed through lead-bearing infrastructure. The memo describes a discontinuity in protection. It also puts the obligation on the system, rather than requiring each resident to discover independently which sections of pipe might release lead into the water reaching their home. F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
Sampling instructions were part of the warning. Del Toral said residents were told to flush before collecting compliance samples, a practice that could minimise lead capture and underestimate the amount present. He acknowledged that the rule did not specifically prohibit the practice, but argued that it defeated the intention to collect under worst-case conditions. The distinction was between a procedure that could be defended against a narrow reading of the text and a procedure that served the health-protection purpose of the sampling. His memo warned that low reported compliance results could give residents a false sense of security and discourage precautions. This was not an assertion that every result had been fabricated. It was a technical objection to how the water was selected for measurement and how that selection affected what the public was told. F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
The city had found 104 and 397 micrograms per litre of lead at Walters's home, according to the memo. The environmental department initially said the source was her plumbing, while she had said the interior pipes were plastic. EPA visits on April 27 and May 6 examined that explanation. Del Toral reported that, apart from minor metallic connectors, the interior network used plastic certified for drinking-water applications. Further sampling suggested the faucets met the newer lead-free requirements and were not the cause of the high results. He concluded that the interior network did not explain the measured lead. The work did not make a citywide claim from the construction of one home. It did remove an explanation officials had offered for that particular home, forcing attention back towards the service line outside it. F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
The service line supplied the next part of the explanation. Sections removed in May included galvanised iron, and inspection confirmed lead between the water main and the external shutoff. The memo records that the home was disconnected in early April and briefly turned on for further collection later that month. Before thirty sequential samples were taken at different flows, the kitchen tap had been flushed for twenty-five minutes the previous evening. Those samples ranged from 200 to 13,200 micrograms per litre. They were results from a defined home and collection sequence, not a distribution of values across Flint. Keeping that scope prevents the worst measurement from becoming a claim that every resident's water reached the same level. The result was alarming in its own right and relevant to a system containing other lead service lines. F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
Del Toral also distinguished the question his memo raised from earlier consultant work. He said a November 2014 operational review by Lockwood, Andrews and Newnam and a March 2015 Veolia water-quality report addressed trihalomethanes and operating issues, not the newly discovered high lead results. The reports were prepared before those results became known and did not account for their possible implications. He recommended involving EPA specialists with experience in corrosion and distribution systems. The point was not that any report about Flint water automatically proved knowledge of every later problem. It was that reports focused on one set of issues could not be used as an answer to another. The memo sought expertise directed at the missing control and lead-bearing network, rather than allowing the existence of previous technical reviews to end the discussion. F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
The proposed next steps included checking whether Flint was violating the corrosion-control requirements and whether the state was applying them properly. Del Toral also asked for review of customer-requested results that might belong in the official percentile calculation. His memo cited EPA guidance saying those samples should be included when collected during the monitoring period from qualifying sites and under qualifying procedures. This introduced another route by which a reassuring city statistic could diverge from what residents had measured. The compliance number depended both on the collection method and on which eligible samples were counted. The recommendation was to test those decisions against the rule, not to replace the official calculation with the highest result anyone had found. It asked the agency to review whether resident-requested samples were being included in the compliance calculation. F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
Questions about the memo continued inside EPA the following day. In released correspondence, Rita Bair asked Del Toral how many homes supported his statement about high lead and what data showed that the problem was widespread. His reply said his judgment drew on decades working with lead, the absence of corrosion control for more than a year and the presence of lead service lines. He argued that those conditions, together with the limited results already collected, were enough to recognise a problem. The exchange is a primary record of technical disagreement about the evidential threshold for action. His judgment was not a statistically complete household survey. It was an argument that the agency did not need to wait for such a survey before accepting the implications of known chemistry and an unprotected system. F086, courtlistener.com, EPA FOIA email chain, Interim Report on High Lead Levels in Flint, June 25, 2015
Del Toral's reply also made the collection problem vivid. He explained that flushing before a small first-draw sample could leave the lead-rich water in a service line outside the reach of the kitchen sample. He described particulate lead as sporadic, making it possible for particles to appear even when procedures reduced the chance of capturing dissolved lead. He asked to sample one hundred homes without pre-flushing if colleagues still doubted the wider risk. The released email contains angry accusations about the handling of evidence. Those accusations remain his words, not a criminal finding by this account. What the exchange establishes more directly is the depth of his objection and the practical alternative he proposed: collect water in a way designed to reveal the problem rather than rely on a procedure he considered likely to conceal it. F086, courtlistener.com, EPA FOIA email chain, Interim Report on High Lead Levels in Flint, June 25, 2015
On June 25, 2015, MDEQ reminded Flint Utilities that more compliance samples were due by June 30. The task force records a June 25 email reminding Flint staff that sixty-one more samples were needed by June 30. With thirty-nine results in hand, the city's ninetieth percentile was above the action level. The email's wording anticipated that the remaining results would be below it. That anticipation matters in the task force's account of sampling practice. It is not, on its own, proof that the later samples were invented or that a particular resident changed a result. It does show officials discussing how the outstanding results would affect the reported percentile while a federal specialist was warning that the collection procedure could miss lead. The compliance sampling reminder and the researcher's concern appear in separate records of the same period. F090, michigan.gov, Flint Water Advisory Task Force final report, March 2016 F086, courtlistener.com, EPA FOIA email chain, Interim Report on High Lead Levels in Flint, June 25, 2015
The memo reached the public dispute in July. The task force records the American Civil Liberties Union contacting the environmental department about Del Toral's concerns, then EPA regional administrator Susan Hedman writing to the mayor that the federal agency would work with the state on lead. Later that month, the governor's chief of staff wrote that community members were raising concerns but felt unheard. Health-department correspondence repeated the state environmental department's position that compliance had not changed and described Del Toral as acting outside his authority. On July 24, Wurfel wrote that residents did not need to worry about lead in the supply and that recent sampling did not indicate a health threat. These were contemporaneous assurances within the official correspondence, and they contrast with the treatment and sampling concerns already documented in the federal memo. F090, michigan.gov, Flint Water Advisory Task Force final report, March 2016 F084, flintwaterstudy.org, EPA Region 5 memo, High Lead Levels in Flint, Michigan - Interim Report, June 24, 2015
A separate warning was present in the state's own childhood blood-lead data. On July 28, epidemiologist Cristin Larder found that summer 2014 test results lay outside the expected control limit compared with earlier years. The task force says she considered the result worthy of further work. Other staff analysed a longer period and concluded that water was not a major factor, then communicated that there was no problem with Flint children's lead levels. This was not yet the public presentation later made by the Hurley doctor. The record shows MDHHS staff reaching different conclusions from the same state blood lead data. Signals existed inside the health department, and the departmental response depended on which analysis it accepted. F090, michigan.gov, Flint Water Advisory Task Force final report, March 2016
Virginia Tech's August account of Walters's home made the extraordinary results public with more detail about method. It reported that the city found 104 parts per billion in February and 397 in March, even under methods that tended to lower measured lead. The university team used bottles allowing more normal flows and collected thirty bottles at low, medium and high rates. It reported the same 200-to-13,200 range and said additional quality checks confirmed the results. The researchers compared measurements above 5,000 parts per billion with a hazardous-waste criterion. That comparison belongs to their account of this household's samples, not a declaration that the entire city supply had been legally classified as hazardous waste. The figures should be read as evidence of a severe local exposure and a reason to test the claim that the home was the only problem. F030, flintwaterstudy.org, Hazardous Waste-levels of Lead found in a Flint household's water, 2015-08-24
The August 31 update extended the household evidence beyond Walters's address. Virginia Tech reported another seventy-two samples, bringing the collection analysed to 120 homes. Twenty-three of those first-draw samples exceeded fifteen parts per billion, and six homes still had values above that level after two minutes of continuous flushing. The team's interim ninetieth percentile was thirty parts per billion. These were the research team's results for its participating homes, not the state's official monitoring round. The update also said the survey had not specifically targeted the worst-case lead plumbing required for regulatory sampling. The distinction sharpened the researchers' concern: even a collection not designed around the highest-risk homes was returning results less reassuring than the official programme. The team continued to seek returned kits and give residents their individual results. F031, flintwaterstudy.org, Analysis of water samples from an additional 72 Flint homes are concerning, 2015-08-31
The state moved towards requiring corrosion control at the end of August. The task force dates an EPA-state conference call to August 31 and records a second-round ninetieth percentile of eleven parts per billion. It says treatment was now considered necessary and implementation steps were set out. Eleven remained below the fifteen-parts-per-billion action level, which again shows why the duty to treat could not be reduced to whether the headline percentile crossed that line. The same day's chronology records Wurfel raising concern about the university testing and press coverage with departmental and governor's-office staff. It finds no apparent communication with the health department about that issue. Technical agreement that control was needed thus coexisted with concern about how the external results were being presented, rather than an immediate unified public-health response. F090, michigan.gov, Flint Water Advisory Task Force final report, March 2016
The university team's September 2 experiment addressed the mechanism rather than the geographic spread. Copper pipe pieces joined with lead solder were placed in Detroit water, current Flint water and Flint water with orthophosphate. Replicates were sampled after one and two weeks. The team reported much more lead release in the Flint-water conditions and only a slight early improvement with phosphate. It described the experiment as continuing, with the early results limited to the first two weeks. Those were controlled tests of solder and water, not household measurements or a final prediction of how quickly the whole city network would recover after treatment changed. Their value in the sequence was explanatory: the different water could interact differently with lead-bearing material, and adding a chemical did not necessarily restore a protective condition instantly. F026, flintwaterstudy.org, Flint River water is very corrosive to lead, and causing lead contamination in homes, 2015-09-02
The team's September 3 account of its August visit shows the sampling effort developing on the ground. It had mailed three hundred lead kits before travelling from Virginia to Flint. A group of students and scientists equipped a van as a mobile laboratory, brought supplies for biological sampling and visited Walters's home. Their account says tests there found no E. coli or coliform, while they also measured chlorine residuals. The absence of those particular bacteria in those samples did not invalidate the earlier lead results: they were different measurements of different hazards. The visit extended into testing across the water system rather than being only a collection of first-draw bottles. In the chronology of the crisis, residents' homes and local access were becoming part of the route by which an outside team could examine a supply that public agencies continued to describe through their own compliance programme. F024, flintwaterstudy.org, Our Flint Sampling Trip (Aug 17-19 2015), 2015-09-03