Tamar Haspel makes the case against lettuce in the Washington Post this morning. It makes people sick, it takes land and water, we throw away roughly as much as we eat, and it delivers almost nothing nutritionally. Her illness numbers are real, and I went and found the paper behind them. Her conclusion is where I part company, and I say that as the lawyer who has spent nearly thirty years suing companies that sold contaminated lettuce, and who does not eat bagged salad himself.

It is a serious column, and it asks a question the produce business would rather not hear asked out loud, which is whether a food this fragile earns its place at all. I am not going to pretend that is unfair. More than thirty-two thousand Americans have been counted sick this summer with a parasite that arrived on shredded iceberg.

Her illness figures come from a real paper, and they are the top of a range.

The 2024 analysis is Yang and Scharff, Foodborne Illnesses from Leafy Greens in the United States: Attribution, Burden, and Cost, in the Journal of Food Protection. It attributes to leafy greens as many as 2,307,558 illnesses a year in this country at a cost of as much as $5.278 billion, and it finds that lettuce accounts for up to 75.7 percent of those illnesses and 70 percent of the cost. Run the multiplication and you land near 1.75 million illnesses and $3.7 billion, which is where the column’s figures come from. Two honest notes. Those are upper bounds. The paper’s own ninety percent interval runs from 1,077,815 to 4,075,642 illnesses and from $3.23 billion to $8.221 billion, so the lettuce share at the bottom of the range is closer to 820,000 illnesses and $2.3 billion. And the paper is two years old. It does not include this summer.

This summer is worse than the paper that already put lettuce at the top.

As of its most recent update, CDC counts 10,930 people sick in the multistate Cyclospora outbreak traced to iceberg lettuce grown in central Mexico, at least 454 of them hospitalized, two dead, across seventeen states, with illness onsets running from June 14 through August 11. On its separate surveillance page, with data through August 17, the agency counts 15,716 laboratory-confirmed cases acquired in this country since May 1, at least 11,841 more awaiting confirmation, and cases in forty-seven states plus the District of Columbia and Puerto Rico, against 1,180 for the same stretch of last year. National hospitalizations on that page now stand at 828, up eighty-eight in a single week. My own fifty-state tally, built from what each health department publishes, was 32,135 on Sunday night. Nobody needs to convince me that lettuce can hurt people. I have spent the summer counting the people it hurt.

She is right that there is no kill step, and that is the honest structural fact underneath all of it.

Everything that makes lettuce worth eating is what makes it dangerous. It reaches the table raw, so nothing along the way kills what is riding on it. Washing is not a kill step. Ortega and Sanchez wrote in Clinical Microbiology Reviews that washing and sanitizing may reduce the risk but cannot be expected to eliminate it, and have been demonstrated not to remove oocysts completely. FDA says in its own voice that chlorine and other common antimicrobial treatments are not effective against this parasite. The National Restaurant Association tells its own members that washing produce under running water cannot guarantee removal of it. I wrote before IAFP this year that there is no kill step for Cyclospora, and that is still the shortest true sentence anyone can write about this food. It is also why I skip the bag.

The pathogen is not a property of the plant. It is a property of the water.

Cyclospora cayetanensis lives only in people. It reaches a lettuce field exactly one way, which is human waste getting into the irrigation water or onto the ground. That is not something lettuce does. It is something we do upstream of lettuce, and the record says so every time. FDA’s advisory changed this month from an onsite inspection at the processing plant to onsite inspections at iceberg lettuce growers in Mexico, which is the agency saying in public that the problem sits upstream of the plant. The 2018 romaine outbreak was traced to an irrigation canal in the Yuma growing region. In 2020 FDA found Cyclospora in a canal serving an Indiana farm that supplied a salad processor, though the agency said its genetic typing could not establish that the canal parasite caused the illnesses. The same Guanajuato facility sat at the center of a Cyclospora outbreak in 2013, and FDA’s environmental assessment then told the firm to determine whether the parasite was a reasonably likely hazard in that growing region and to re-evaluate the wash step. Stop eating lettuce tomorrow and every one of those water sources is still there, still running onto something somebody eats.

Quitting the category is what a traceability failure already does, and we watched it happen in July.

The column proposes as policy what this market already did to itself by accident. After the recall, Chopt traffic fell about 12 percent and weekly spending at Sweetgreen dropped roughly ten percentage points against last year, and neither chain has been linked to anything. Paul Sellew, who grows lettuce in the eastern United States and has no connection to Mexico or to Taylor Farms, watched his sales fall by double digits and said his customers were motivated by fear. NielsenIQ measured fresh lettuce unit sales down 9 percent in a single week. Not one of those declines made a single person safer. That is the mechanism worth understanding. When nobody can say which lot, people stop buying the category, and the loss lands on growers who did nothing wrong while the contaminated water stays exactly where it was. It is also the best economic argument there is for the traceability rule Congress passed and then pushed to 2028.

The cost-benefit only works if you price one side of it.

The column prices carefully what lettuce costs us and never prices what fixing it would cost. Those numbers exist, and they are small. USDA’s Economic Research Service puts the annual national cost of cyclosporiasis at $5.3 million, a figure resting on a case estimate from 2011 that nobody has updated. Read what that model actually assumes: about twenty Cyclospora hospitalizations a year in the entire country, eleven of them domestically acquired and foodborne. CDC counted 828 this season. The Microbiological Data Program, the produce-testing program killed in 2012, ran about $4.5 million a year. FDA costed the traceability delay itself in the Federal Register at 90 FR 38084 on August 7, 2025, and the arithmetic is the agency’s own: forgone public health benefit of $91 million to $112 million a year against $54 million to $73 million in industry savings, which means the delay destroys roughly $37 million to $39 million a year more than it saves. Set all of that next to Spalding and Sexton in the American Journal of Agricultural Economics, who put the total societal loss from the single November 2018 romaine incident at $276 million to $343 million, and whose own conclusion was that the answer is mandatory standards and better traceability. One of my degrees is in economics, and it does not take one to read that ledger. Outbreaks are expensive. Preventing them is cheap.

Adding E. coli and Listeria to the Ledger Makes the Case for Fixing It, Not for Quitting It.

The obvious objection to everything above is that I have priced one parasite. Fold in the pathogens that made lettuce famous and the numbers do change, though not in the direction the column needs. Start with the correction. The $3.7 billion is already the multi-pathogen figure. Yang and Scharff built it out of norovirus, Shiga toxin-producing E. coli both O157:H7 and non-O157, Campylobacter, nontyphoidal Salmonella and the rest. E. coli and Listeria do not stack on top of that number. They are most of it. Cyclospora is the pathogen barely in it. What changes when you go pathogen by pathogen is severity, and the spread is enormous.

PathogenERS cost per caseWhere it shows up in lettuce
Cyclospora cayetanensis$464.6310,930 people in this summer’s federal outbreak alone
STEC E. coli O157:H7$7,23812,496 illnesses a year attributed to romaine
Nontyphoidal Salmonella$16,668Among the largest contributors to the leafy green total
Listeria monocytogenes$2,504,296About 1,500 US infections and 250 deaths a year, with packaged salad a recurring vehicle

Per-case costs from USDA’s Economic Research Service, Cost Estimates of Foodborne Illnesses, in 2023 dollars, read August 2026. Lettuce attribution from Yang and Scharff and from CDC. Check the arithmetic: 10,930 times $464.63 is $5,078,406, which is 2.03 times what ERS assigns to a single case of listeriosis.

Read the two ends of that column together. The entire federal Cyclospora outbreak this summer, all 10,930 people and 454 of them hospitalized, comes to roughly $5.1 million at the government’s own per-case figure. Two cases of listeriosis match it. That is why a cost-benefit run on this parasite by itself always loses. It is the highest-volume and lowest-priced pathogen on the page, and no program aimed at it will ever clear a threshold built that way.

The intervention set does not change with the pathogen, and that is the whole answer. Treating agricultural water, traceability, restored surveillance and a water rule that measures something other than generic E. coliwork on all of them at once. FDA’s arithmetic on the traceability delay was never a Cyclospora calculation in the first place; it covers every food on the traceability list and every pathogen that rides in on them. Add a pathogen to the benefit side and the cost side does not move. Each one folded in improves the ratio for fixing this, and not one of them improves the ratio for quitting.

E. coli is the strongest card the other side holds, and it deserves saying plainly. The Interagency Food Safety Analytics Collaboration attributes more than 85 percent of E. coli O157 illnesses to vegetable row crops and beef, with row crops significantly higher than every other food category. FDA maintains a standing Leafy Greens STEC Action Plan whose stated purpose is breaking a cycle of reoccurring outbreaks. The same O157 strain has come back since 2016. A commodity that produces the same genotype for a decade is not having bad luck.

And the answer to it is the same water. USDA researchers found Listeria monocytogenes in 43 percent of environmental samples taken in a watershed connected to a major leafy green growing region. Yuma in 2018 was a canal. Guanajuato was sewage in 2013 and again this summer. Three pathogens, one failure mode, which is an argument that gets stronger with E. coli and Listeria in it rather than weaker.

One asymmetry belongs here honestly. Listeria in packaged salad is a plant problem rather than a field problem. The 2015 outbreak traced to a processor whose own records showed Listeria in the facility across nineteen months. That is the one place the record runs the column’s way, and it is exactly why I do not eat the bag. The answer to harborage in a plant is environmental monitoring and enforcement, not abstinence.

I already skip the bagged salad, and I still will not tell anyone to give up vegetables.

My list of foods I do not eat has been on this blog for decades, and prewashed, precut produce is number four on it. I told the Washington Post the same thing in December. That is a personal calculation made with information most people do not have, and it is not public health advice. Most Americans already fall short of the vegetable recommendation, and the last thing anyone in this field should want is a country that is afraid of produce. The column concedes the point itself, that the salad on the table is partly what keeps the second helping of pasta off it. People who give up salad do not switch to lentils.

The column closes on a choice between a solar-powered home and explosive diarrhea. That is a false choice, and both halves are available. Treat the agricultural water. Name Cyclospora in the water rule instead of scoring that water with a bacterial indicator that cannot see it. Restore the surveillance that tells us how many people are sick. Turn on the traceability rule that is already law, so the next recall can name a lot instead of a vegetable. Do that and my docket empties, which has been the point since 1993. Nobody has to give up a salad to get there.

Phyllis Entis published the inspection history of Everything Sprouts, LLC at eFoodAlert on August 23, 2026 — the Minneapolis company whose alfalfa sprouts have now sickened 55 people in 15 states with Salmonella Agona or one of three strains of Shiga toxin-producing E. coli, four of them hospitalized, with onsets running May 31 through August 8, 2026. Her piece sent me to FDA’s own citation data. What is in that data is worse than what is in the warning letter, and the difference is a single line: on November 14, 2024, FDA cited this facility for failing to document the treatment of its seeds. The warning letter that followed never mentions it.

Phyllis Entis found the history, and FDA’s own citation data confirms it.

Credit where it belongs. Entis has been at this for fifty years, and she did the thing almost nobody does with a recall notice, which is to go look at what the government already knew about the company before anyone got sick. She reported the ownership chain — a facility operating in Minneapolis since 2009, bought by Everything Sprouts after it registered in Minnesota in 2022 — and she reported the classification arc, which is the part that matters. FDA inspected the building repeatedly under the prior owner and closed each of those inspections as Voluntary Action Indicated. It inspected again in November 2023 and closed that one the same way. It came back a year later, found that essentially nothing had been fixed, and closed that inspection as Official Action Indicated, which is the agency’s highest level of concern.

None of that appears in the warning letter. I went looking for the underlying record and found the next best thing, which is FDA’s Inspection Observations citation data — the file behind the agency’s inspections dashboard, 279,860 citation lines covering 76,079 inspections at 47,890 firms between October 1, 2008 and August 13, 2026. Every citation FDA’s electronic inspection tools wrote onto a Form FDA 483 in that period is in it, listed by firm, by date, by regulation, and by a one-line description of what the investigator found. Everything you are about to read comes out of that file.

This one building has been cited thirty-six times since 2010, and the same problems keep coming back.

The facility carries FDA establishment identifier 2130526. Eight inspections there have produced citations, thirty-six in all. One caution before the table: FDA’s file stamps the current registrant’s name onto the entire history, so every row back to 2010 reads Everything Sprouts even though the company did not exist until 2022. These are facts about a building and the people running it at the time, not about one owner. FDA says as much in the warning letter itself, noting that its 2018 and 2023 observations were made while the sprouting operation was under different ownership in the same facility.

Inspection endedCitationsRegulationWhat the citations covered
September 2, 2010621 CFR 110Contamination precautions, pest exclusion, plumbing as a source of contamination, non-food-contact equipment, holding temperature
May 9, 2011421 CFR 110Contamination precautions, equipment that could not be properly cleaned, handling of containers and utensils
August 15, 2014821 CFR 110Plant design, fixtures in disrepair, pests, plumbing, rubbish, missing thermometer, no microbial testing
December 4, 2017221 CFR 112Contamination through fixtures and ducts; plumbing not adequately installed or maintained
January 29, 2018221 CFR 112Food contact surfaces not maintained; contamination through floors, walls, fixtures, drip and condensate
September 22, 2021221 CFR 112Equipment not designed to be cleaned; plumbing not of adequate size and design
November 27, 2023521 CFR 112Equipment storage and sanitizing, contamination through ceilings and condensate, environmental monitoring corrective actions, water quality test method
November 14, 2024721 CFR 112Equipment cleanability, food and non-food contact surfaces, contamination through ceilings and condensate, environmental monitoring sites and corrective actions, seed treatment records

Table 1. Every FDA inspection of FEI 2130526, 2751 Minnehaha Avenue, Minneapolis, that produced citations. Source: FDA Inspection Citations data, downloaded August 24, 2026. Check: 6 + 4 + 8 + 2 + 2 + 2 + 5 + 7 = 36, the total number of citation lines the file carries for this facility.

Read down the last column and the building tells on itself. Contamination through fixtures, ceilings, drip and condensate is cited in 2017, in 2018, in 2023 and in 2024. Plumbing is cited in 2010, 2014, 2017 and 2021. Food contact surfaces are cited in 2018, 2023 and 2024. A regulation changed in the middle of this — Part 110’s general sanitation rules gave way to Part 112, the Produce Safety Rule — and the citations changed number without changing subject.

One thing the table cannot show, and I want to be straight about it: this is a record of citations, not of inspections. An inspection that found nothing generates no line. What the file does establish is that after November 14, 2024, FDA cited this facility exactly zero more times through August 13, 2026, which is where the data ends. That is the documentary version of what Entis reported — the agency never came back.

The November 2024 inspection produced seven observations. The warning letter carried two.

FDA inspected from October 30 through November 14, 2024. The warning letter, MARCS-CMS 699480, issued to owner Melissa L. Mann on March 28, 2025, sets out two significant violations and then adds one closing comment. The citation file says the investigator wrote up seven.

CitationWhat it saysWhere it ended up
112.126(b)Did not prevent contamination through walls, ceilings, fixtures, ducts, pipes, drip and condensateViolation 1 of the warning letter
112.123(d)(1)Did not inspect, maintain and sanitize food contact surfaces as often as necessaryMoved to the closing comment, which carries no fifteen-day response obligation
112.123(a)Equipment and tools not of adequate design or workmanship to be cleaned and maintainedNot in the warning letter
112.123(d)(2)Non-food-contact surfaces not maintained or cleaned as often as necessaryNot in the warning letter
112.145(c)(3)Written environmental monitoring plan lists too few sampling sites to tell whether measures workNot in the warning letter
112.145(e)Environmental monitoring plan has no adequate written corrective action planNot in the warning letter
112.150(b)(1)Did not establish documentation of treatment of your seeds or beansNot in the warning letter

Table 2. The seven citations recorded for inspection 1254250, ended November 14, 2024, matched against the March 28, 2025 warning letter. Sources: FDA Inspection Citations data, downloaded August 24, 2026; FDA warning letter 699480. Check: one carried forward as a violation, one demoted to a comment, five absent — 1 + 1 + 5 = 7.

Two honest notes on that table. The warning letter’s second violation cites 21 CFR 112.143(b), the sprout-specific requirement to clean and sanitize food contact surfaces before they touch sprouts or seed, and that section does not appear among the seven. The facts under it — the black organic buildup on the irrigation hose over exposed soy sprouts, the ripped plastic sheeting patched with tape, the sprouts stuck to the tape — read like the same observations the 483 recorded under 112.123. FDA’s own caveat on the dashboard says the citation file comes from its electronic inspection tools and will not always match a 483 that was edited afterward. The likeliest reading is that the agency recast those facts under the sprout subpart when it wrote the letter. That is a lawyer’s judgment, not a finding, and the way to settle it is to read the 483.

The second note is that a warning letter is not required to carry every observation. It is a discretionary enforcement document, and FDA says plainly that it is not an all-inclusive statement of violations. Nobody did anything improper here. That is exactly the problem. The published document is a selection, and the public has no way to see what was left on the cutting room floor.

One of the five left out was the seed.

21 CFR 112.150(b)(1) requires a sprout operation to keep documentation of the treatment of its seeds or beans. The one-line description FDA’s investigator selected reads: you did not establish documentation of treatment of your seeds or beans. That is not a paperwork foot-fault in a sprouting facility. Seed treatment is the control. Alfalfa seed is grown in open fields, harvested with no kill step, and then held for days in warm water at the exact temperature a pathogen would choose. Everything downstream of the seed is containment. The seed treatment is the intervention.

I wrote last week that the warning letter says nothing about seed and nothing about alfalfa, and that the missing subject was the tell. That was true of the letter and I am correcting it now anyway, because it understated what happened. FDA did not overlook the seed. Its investigator cited the seed records, in writing, sixteen months before this outbreak, and the citation did not survive the trip from the 483 to the letter the public gets to read.

Three of those citations are rare enough to count on your fingers.

Rarity is the part you cannot see without the whole file, and it is what turns a list of code sections into a finding. Across all 279,860 citations in the dataset, 21 CFR 112.150(b)(1) — the seed treatment records — has been cited thirteen times, to eleven firms, in the entire period from March 2018 to today. The November 14, 2024 citation at this facility is the most recent of the thirteen.

21 CFR 112.145(e), no adequate written corrective action plan in the environmental monitoring program, has been cited nine times to eight firms. Everything Sprouts holds two of the nine, in November 2023 and again in November 2024, and it is the only firm in the file cited twice. That is the citation Entis flagged as unchanged year over year, and the data confirms it precisely: the same subsection, the same facility, twelve months apart.

And 21 CFR 112.126(b), contamination through walls, ceilings, fixtures, drip and condensate — the one that became Violation 1 — has been cited fifty-six times to forty-eight firms since the Produce Safety Rule took effect. Two facilities in the United States have been cited under it four separate times. This is one of them. The other is also a sprouter.

FDA told this company twice that it would check the corrections at the next inspection. There was no next inspection.

The warning letter closes each of its two violations the same way: FDA is unable to evaluate the adequacy of the corrective actions, and will evaluate them at the next inspection. That sentence is written twice. The next investigator to walk into 2751 Minnehaha Avenue arrived on August 19, 2026, seventeen months later, after two state health departments had already traced illnesses to the product. Everything Sprouts issued a recall on August 22 — five-ounce containers of alfalfa sprouts and two sprout mixes, lots 222, 223, 225, 226 and 230, distributed May 27 through August 21, 2026. FDA’s posted version of that notice finally carries the lot numbers, the bar codes and the Calco brand that the company’s own Friday night notice left out. It also says the product went to wholesale distributors and grocery stores in two states, Minnesota and Wisconsin, while CDC counts sick people in fifteen.

Nine days before this warning letter went out, FDA sent another one to Jack & The Green Sprouts of River Falls, Wisconsin — the sprouter behind the last alfalfa outbreak to hit Minnesota and Wisconsin together, in 2016. Two sprout operations on the same commodity in the same two states, warned in the same nine-day window, and one of them is now the subject of a fifteen-state outbreak. Somebody at FDA was paying attention in March 2025. What did not happen next is the follow-up.

FDA’s own regulation says a 483 is a public record. The agency simply does not publish them.

Here is what is odd about all of this. FDA posts every warning letter it issues, in a searchable index, within weeks. It posts the citation data I have been quoting all the way down to the subsection. It posts the classification — No Action, Voluntary Action, Official Action. What it does not post is the one document that sits between the inspection and the letter: the Form FDA 483 the investigator handed across the table on the last day.

Not because the law forbids it. 21 CFR 20.101 is FDA’s own regulation on administrative enforcement records, and it names the document. All FDA records relating to administrative enforcement action that have been disclosed to any member of the public, it says — including Forms FD-483 furnished to companies after factory inspection — are available for public disclosure at the time that disclosure is first made. A 483 is handed to the firm at the close of every inspection. That is the disclosure. The regulation goes on to say that to the extent such records fall within the investigatory records exemption, the Commissioner has already determined they are subject to discretionary release. The determination has been made. It was made decades ago.

The timing objection is real and it is also finite. FDA’s dashboard warns that some inspection records may be withheld until a planned compliance action is taken, so that disclosure does not interfere with enforcement. Fine. The compliance action here was taken on March 28, 2025. Seventeen months have passed. The warning letter has been public for fifteen of them. Whatever interest that exemption was protecting expired long before anyone in Minnesota ate a sprout.

What we have instead is a system where a 483 becomes public if somebody pays for it. The document goes into the reading room only after it has been requested, processed and cleared, which means the public record of American food inspection is assembled out of whichever documents a competitor, an investor or a subscription database thought were worth a FOIA fee. Nobody paid the fee on a small Minneapolis sprouter. In 2010 FDA’s own Transparency Task Force put out twenty-one proposals; two of them covered publishing the fact of an inspection and its classification, and summaries of the most commonly cited observations. Both got built, and I have been using both all morning. Publishing the documents themselves was never on the list.

The fix costs nothing, requires no rulemaking, and would have put this in front of buyers in April 2025.

Post the 483 with the warning letter. Same page, same day, same index — the letter already quotes the observations at length, so the confidentiality argument is spent by the time the letter goes up. Then set a rule for food facilities as a class: every 483 published within a fixed period after the inspection is classified, redacted for anything genuinely confidential, whether or not anyone asks. The agency already runs the pipeline. It already writes the citations into a database. It is holding back only the paper.

Think about who would have read this one. A grocery buyer deciding whether to keep a sprout supplier. A restaurant chain’s food safety director. A seed broker. An insurance underwriter. A reporter in Minneapolis. Any one of them could have opened a document in April 2025 that said this facility had been cited seven times, that the ceiling was leaking onto exposed sprouts, that the environmental monitoring plan was a repeat failure, and that nobody could produce records showing what had been done to the seed. Instead they got a two-violation letter about a roof and a hose, four months late, and no way to know what else was on the page.

Fifty-five people are sick in fifteen states. FDA says it is still working out whether other products or firms are involved, and whether a common seed supplier reaches beyond this one company. That question has been sitting in a filing cabinet since November 14, 2024, in a document that FDA’s own rules say has been public the entire time. It was never secret. It was just never published. Thank you, Phyllis, for going and looking.

Want to Put Us Out of Business? Come Help Us Try.

Marler Clark has been suing the companies that poison people since 1993 — since a nine-year-old girl named Brianne Kiner nearly died from a hamburger. Sixteen years later it was Stephanie Smith, a children’s dance instructor paralyzed by a hamburger, whose case put the grinding of American beef on the front page of The New York Times and won Michael Moss a Pulitzer Prize. Thirty-plus years later we are still the only law firm in the country that does nothing but foodborne illness — tens of thousands of clients and over $850 million recovered — and there is more work than there are lawyers to do it. The names on the other side are ones you already know: Jack in the Box, Cargill, Dole, Nestlé, Walmart, Wendy’s, McDonald’s, Nebraska Beef, Yum Brands, and this year Taylor Farms, Taco Bell and Boar’s Head.

As I write this, on August 19, 2026: a Cyclospora outbreak the fifty state health departments now count at 31,521 people and rising. The federal government called it 1,947 in nine states until it changed what qualifies as a case in August; its own number is now 9,481 in seventeen states, 398 hospitalized, two dead. Michigan by itself reports 13,909 cases — more than the federal government counts for the entire seventeen-state outbreak. We have five cases filed against Taco Bell, four different franchise operators and Taylor Farms, and more than two hundred other people in twenty-six states have retained us to find out how they were poisoned. Salmonella Javiana in jalapeños grown in Sinaloa — 431 sick in 32 states, Chipotle and QDOBA — which Minnesota epidemiologists and a company traceback found before it ever reached FDA’s outbreak table; a month later the grower still has not been named and the recall of what it sold has spread to more than 160 grocery products. Infant botulism litigation against ByHeart and Nara Organics. Listeria out of Boar’s Head. E. coli O157:H7 from beef kofta and from romaine. Salmonella from nineteen million recalled eggs and from a Lexington pie shop. E. coli O145 in frozen blueberries. Vibrio from raw seafood in Queens. Read a month of marlerblog.com. Clients in need find us; that has never been the hard part.

Finding the defendant is. Nobody hands you one in a foodborne case. You will read FDA inspection reports and environmental assessments, USDA establishment records, whole genome sequencing and PulseNet clusters, epidemic curves built by state health departments, and — this summer — the sanitation and rainfall records of a Mexican municipality where the lettuce was grown and the traceback on a lot of peppers out of Sinaloa. That work gets done before the complaint is filed, not in discovery. You will also have the Outbreak Database, which this office built and still maintains — decades of outbreaks, searchable, and no other firm in the country has it. Spanish is a real plus.

And half of this job is not litigation at all. E. coli O157:H7 became an adulterant in American beef in 1994; six more strains followed because we petitioned USDA to make them so. We helped pass FSMA in 2011, petitioned — so far without success — to have outbreak strains of Salmonella declared adulterants in 2019, and pushed hepatitis A vaccination for food workers in 2020. This year alone: an infant formula bill in the House, a food safety bill in the Senate, the Stop Foodborne Illness petition pending at FDA asking it to tell the public where recalled food went, comment letters on the traceability rule Congress delayed to 2028, and a standing demand that CDC count Cyclospora again. I publish Food Safety News, I have written over 6,000 posts on Marler Blog since 2004, I speak all over the world, and I teach food safety at the Harvard T.H. Chan School of Public Health. You will do this too — including writing here under your own name, where reporters will read it. Within a few years it will be you the conference invites, the health department calls and the reporter quotes; our lawyers speak to industry groups, public health audiences and university classes, and that platform is handed over, not loaned. Work out of this office has been cited in more than a thousand media reports. We also correct in public: when a poultry veterinarian wrote in last month to tell me a post of mine was wrong, we ran the correction louder than the original. Being right matters less here than getting it right. Cases fix one company; the rest fixes the industry.

If you want to know what that adds up to, read the testimonials page on billmarler.com and pay attention to who is on it: a former FDA food safety czar, a former USDA food safety official, the retired head of the Produce Marketing Association. Michael Pollan wrote that the federal watchdog on food safety is asleep at the switch and we are what is left. One of the lawyers who defended the other side in the Jack in the Box cases became my law partner. You will spend your career arguing with the food industry, and some of the people you argue with will end up thanking you for it.

We are looking for an associate with at least five years of litigation experience — personal injury, product liability, med-mal, or complex tort — who is ready to run cases, not just paper. We are a handful of lawyers, not a floor of them: you will carry your own docket from the start.

Here is what you will not do here. There is no billable hour requirement and no origination credit to chase — the clients find us, and the firm carries the cost of every case out of its own pocket until it resolves. There is no committee, no partner standing between you and the person who decides, and no waiting your turn: when you need an answer you walk down the hall and get one. And there is no volume. We turn down cases every week that we could not do right, because a file nobody has time for is worse than no file at all.

Our office is on Bainbridge Island, and the honest truth is that we want you here — living on the island, not commuting to it. We are a few minutes’ walk up from the ferry terminal, thirty-five minutes across the water from downtown Seattle, in a town where you can walk to work, walk to lunch, and walk home. Most people who make the move do not move back. But the right lawyer matters more to us than the right zip code: if you are elsewhere in Puget Sound, or willing to relocate, and can be here most days, we should talk, and we will help the right person get here. 

What you’ll actually do: file in federal and state courts across the country, write the settlement demand letters that resolve most of our cases before anyone sees a courtroom — the medical chronology, the liability analysis, the damages model, all of it yours — take and defend depositions of plant managers and epidemiologists, work up hemolytic uremic syndrome, Listeria meningitis and Salmonella reactive arthritis cases with the best experts in the field, sit with families in hospital rooms, and try cases. The docket is national; you will be on airplanes. It has been international as well — outbreaks in Europe, Asia and Africa. You will be trusted early and often.

Many of our clients are children, and on those cases the last thing you do is the hardest. A settlement for a child has to be approved by a judge, cleared of Medicaid, Medicare, ERISA and sometimes federal military liens, and structured — now and then into a special needs trust — so the money is still there decades later, when the kidney an outbreak took finally has to be replaced. If you have never done a minor’s compromise, you will.

What you need: licensed and in good standing somewhere, admissible pro hac vice everywhere else — Washington admission is not required; real deposition experience; the writing chops to draft a complaint that reporters quote; and the stomach for what an outbreak does to a child. A science background is a bonus, not a requirement — we’ll teach you the microbiology.

What we offer: Competitive salary depending on experience, plus bonus; medical, dental and vision; 401(k) with match; paid vacation and holidays; bar dues and CLE; and cases that end up in The New York Times, in front of Congress, and in Poisoned, the Netflix documentary about this work — watch it before you apply. Also: a commute that involves a ferry or no ferry at all, and a firm that goes home for a safer dinner because of you.

One last thing, and it is the whole thing. Nobody here treats this as a practice area. Everyone here can name the child whose case they will never forget. That is why they stay. A client wrote to us that I came to the hospital while their son was there, and that somebody from this office was in touch every step after; that somebody is going to be you as often as it is me. I am not looking for a lawyer who wants a job. I am looking for one who cannot look away from it and wants to spend the next thirty years making a difference.

The job is open now. Resume, writing sample and a short note to bmarler@marlerclark.com — in the note, pick any case or outbreak off marlerblog.com and tell me what you would do with it first. The rest of it is public too: marlerclark.com and billmarler.com. No recruiters. Equal

Equal opportunity employer.

Between July 14 and August 13, members of the House and Senate sent thirteen separate letters about the 2026 Cyclospora outbreak. Five came out of the House and eight out of the Senate. They went to the Secretary of Health and Human Services, to the acting heads of FDA and CDC, and twice to the chief executive of Taylor Farms. Seven of them set a dated deadline that has now passed. One deadline is still live, and it falls this Thursday. As of this evening I cannot find a published answer to a single one of the thirteen.

What has changed since I last wrote about this is not that anybody answered. It is that the answers started showing up somewhere else.

Start with the credit, because it is owed

A letter from a member of Congress is not a subpoena and it is not a rule. I have spent thirty-three years watching letters like these go into a drawer. But a letter puts a specific question in writing with a name underneath it, and it creates a date on which somebody either answered or did not. Every one of these members chose to spend staff time and political attention on a parasite most of the country had never heard of in May. The questions in them are the most specific anyone has asked about this outbreak, and several of them surfaced facts that would otherwise still be buried — the parasitology bench at CDC cut from eleven people to three, the food safety laboratory network down from forty-one active labs to fifteen, the USDA research center that does much of the country’s Cyclospora genotyping being closed and relocated.

Here is the whole list. The column on the right is the one to read.

DateWho wrote, and to whomWhat was asked forDeadline and status
Jul 14Sen. Amy Klobuchar (MN) to CDC and FDARestore FoodNet, FERN, inspection staffing and public health grantsNone set. No answer found
Jul 15Rep. Greg Stanton (AZ) to Sec. KennedySix questions: who ordered the surveillance change, what funding was cut, how many staff lostAug 3. Passed, no answer found
Jul 16Sen. Jon Ossoff (GA) to Sec. KennedyThree questions on restoring Cyclospora tracking and the state of the investigation72 hours, Jul 19. Passed, no answer found
Jul 17Reps. Dingell, McDonald Rivet, Scholten and Tlaib (MI) to CDC and FDATen questions from inside the outbreak, including whether the agencies agree with Michigan on the sourceJul 24. Passed, no answer found
Jul 20Rep. Josh Gottheimer (NJ) to CDCRestore FoodNet to mandatory active surveillance, plus two billsNone set. No answer found
Jul 20Sens. Blumenthal (CT) and Booker (NJ) to CDCA briefing on how the outbreak was tracked and whether staffing cuts caused delayUndated. No answer found
Jul 24Sen. Ossoff, second letter, to Sec. KennedyThe same three questions, plus the fact that the first letter went unansweredNone set. No answer found
Jul 27Rep. Robert Garcia (CA), Oversight Ranking Member, to Bruce TaylorFive categories of documents and five written questions, including facility records back to 2013Aug 10. Passed, no production found
Jul 28Sen. Kirsten Gillibrand (NY) to CDC and FDASix questions on the collapse of the federal laboratory networkJul 31. Passed, no answer found
Jul 30Sen. Blumenthal, PSI Ranking Member, to FDAAll records, including every communication with the White House and with the companyAug 13. Passed, no answer found
Jul 30Sen. Blumenthal, PSI Ranking Member, to Bruce TaylorAll records, including every communication with the White House and with FDAAug 13. Passed, no production found
Jul 30Reps. Veasey, Dingell and Sewell with about fifty colleagues, to CDC and FDATransparency and immediate action to identify the sourceNone set. No answer found
Aug 13Sen. Elizabeth Warren (MA) to Sec. KennedyNine questions on the retracted test, the White House call, what the outbreak has cost, and which states still reportAug 27. Pending

I believe that is every letter and I am not certain it is. I built the list from press releases, from posted PDFs, and from reporting where offices posted nothing. A letter sent quietly without a release is one I would not have seen — I have had to add one four separate times already. Twelve of the thirteen links in that table go to the letter itself or to the sending office’s own release. One does not: for the Blumenthal and Booker letter to CDC I have never been able to find a posted copy or a press release, and the link goes to the New Haven television station that reported it. If either office posts the letter, I will swap the link. And the status column says no answer found, which means exactly that and nothing more. An agency can answer a member privately and often does, and neither side has to publish it. If you have an answer to any of these, send it to me and I will correct this post and say plainly that I had it wrong. That is the list I would like to see grow.

The answers exist. They are being given to reporters instead.

On the day Senator Warren’s letter went out, HHS press secretary Emily Hilliard told The Hill that the agencies’ response has been guided by scientific and public health evidence, that the FDA investigators responsible for outbreak investigations were not affected by staffing changes or reductions in force, and that political contributions or relationships have no bearing on FDA or CDC outbreak investigations. That is a direct answer to the central allegation in at least four of the thirteen letters. It was given to a newspaper.

Taylor Farms did the same thing. The company told Scripps News that it acts in the interest of science and consumer health rather than politics, and that it categorically rejects any suggestion that it sought or received favorable regulatory treatment through political contributions or improper influence. That is a direct answer to Representative Garcia and to Senator Blumenthal. It was given to a television network.

And then there is the FAQ page, which is the part that should decide this

Go to the company’s own Cyclospora information hub and scroll to the questions. One of them asks whether Taylor Farms used political connections or contributions to influence the FDA investigation, and answers no. Another lays out a day-by-day account of July 15 through July 19 — the date FDA and CDC first shared traceback data, the call on which the company says it raised scientific questions about the government’s assumptions and methodology, and then July 16, when those questions were discussed further with government counsel and officials including career staff, lawyers, and officials from FDA, HHS and the White House. The company states that it decided to recall that same day.

Read that against the letters. Representative Garcia asked for every communication with the White House since January 2025. Senator Blumenthal asked FDA and the company, separately, for all communications with the Executive Office of the President and the Domestic Policy Council, in a records definition that named Signal and WhatsApp. Senator Warren asked for a transcript of the reported call. None of them has received anything. And a version of the answer — the company’s version, unsworn, undated as to who said what, and written by the party with the most at stake — has been sitting on a corporate FAQ page, revised at least five times, overwritten in place, with no version history and no correction notes.

I am not saying the company did anything improper by putting it there. I would rather have it than not have it. I am saying that this is what the record looks like when nobody with subpoena power has asked for it: the regulated party gets to write the first draft of history on its own website, and Congress reads it there like everybody else.

A statement to a reporter is not an answer to a letter, and the difference is the whole point

A press statement is not made under oath. It is not a document production. It can be revised or deleted without anyone being told, which has already happened repeatedly on this outbreak, on both the government’s pages and the company’s. It does not tell Senator Warren how many states stopped reporting cyclosporiasis after the surveillance change in July 2025, state by state, which is the single most useful question in all thirteen letters. It does not tell Representative Garcia what the facility records at Doctor Mora show going back to 2013. It does not produce the call.

Those things exist. They are in federal file cabinets and on company servers right now. Nothing in a letter can make anybody walk to one.

What would actually work, and every bit of it belongs to the majority

The Senate Permanent Subcommittee on Investigations can issue a subpoena. Senator Blumenthal wrote as its ranking member and copied Chairman Ron Johnson on both July 30 letters; a ranking member cannot subpoena alone and a chairman can. The House Oversight majority could adopt Representative Garcia’s document request tomorrow and it would carry the committee’s weight instead of the minority’s. The Senate HELP Committee has jurisdiction over both agencies, and so does the Appropriations subcommittee that funds FDA, where Senator Ossoff already sits and where the Senate wrote language, in its own report, directing that enforcement of the traceability rule is not delayed past 2028 — language that did not survive.

And there is precedent that is almost eerie. The Permanent Subcommittee on Investigations held a hearing on Cyclospora in imported raspberries in July 1998. Same subcommittee, same parasite, imported produce, twenty-eight years ago. Rosa DeLauro wrote FDA and CDC during the 2013 outbreak traced to this same Guanajuato facility, and again in 2018 about Fresh Express. The questions do not change. What changes is whether anyone is ever made to answer them out loud.

A hearing does three things a letter cannot. It puts a human being in a chair on a date certain. It puts them under oath, so the answer costs something if it is wrong. And it creates a public transcript that cannot be revised in place at two in the morning the way a webpage can. That last one matters more in this outbreak than in any I have worked on, because nearly every primary document in it — FDA’s advisory, CDC’s outbreak pages, the company’s statement — has been rewritten on the same URL with no record of what it used to say.

Thursday is the last date on the board

Senator Warren’s nine questions are due August 27. After that there is no outstanding deadline in front of any agency or any company on this outbreak, and the Cyclospora season closes on August 31. FDA counts 10,930 illnesses, 454 hospitalizations and two deaths in the outbreak it attributes to Taylor Farms. CDC counts 15,716 laboratory-confirmed illnesses nationally since May 1 and at least 828 hospitalizations, against 1,180 cases in the same stretch of 2025. Adding up what the fifty states publish about themselves, my own count is 32,135 people this morning, and it is a floor rather than an estimate.

Thirteen letters produced a great deal of good work and no answers. I wrote in an earlier version of this list that the questions were good questions and the problem was that nobody had reached for the one thing that would make anybody answer them. A month later that is still true, and the calendar has nearly run out. Hold a hearing. Put the acting commissioner, the acting director and Bruce Taylor at the same table, under oath, and ask them the questions that are already written down. Thirteen members have done the drafting. Somebody with a gavel has to do the rest — the same somebody who could also strike the provision that pushed the traceability rule to 2028 and made this outbreak so much harder to trace in the first place.

My Sunday night office.

North Carolina has posted 940 cases, thirty-four more than my chart carried, which takes the fifty-state floor from 32,135 to 32,169. The North Carolina section below is new.

Adding up what each of the fifty state health departments reports on its own website, the floor under the 2026 Cyclospora outbreak this evening, Sunday August 23, is 32,169 people. Four numbers moved since Thursday and everything else held. Missouri went to 2,152, North Carolina to 940, West Virginia to 313, and Washington turned out to have been carried far too low on my own chart. The federal outbreak figure has not changed since Thursday: FDA still reports 10,930 illnesses, 454 hospitalizations, two deaths and seventeen states, with the last illness onset on August 11.

Missouri crossed two thousand today, and its weekly increase has finally slowed

The Missouri Department of Health and Senior Services publishes its communicable disease report on Sundays, and today’s puts the state at 2,152 cases. My chart carried 1,950 from last Sunday, so that is 202 more people and the first time Missouri has cleared two thousand.

The shape of that number matters more than the number. Missouri ran 216 in late July, then 1,095, then 1,577, then 1,950, and now 2,152. Those are increases of roughly four hundred percent, forty-four percent, twenty-four percent and ten percent.

North Carolina is at 940, and one county holds more than half of it

North Carolina has posted 940 cases for the year, according to state health data last updated August 18 and reported Sunday by Michael White at Spectrum News 1. My chart carried 906. That is thirty-four more people and it is what moves the floor to 32,169.

The weekly curve underneath that total is the cleanest evidence any state has published that this outbreak is ending. North Carolina reported 15 cases the week ending June 5, then 26, then 97. By the week ending July 3 it was 244, and 254 the week after. In the first two weeks of August the state identified twenty new cases in total. That is a fall from roughly two hundred and fifty a week to about ten. Michigan had its smallest week of the outbreak on Thursday and Missouri’s increase has dropped to ten percent, and both point the same way, but North Carolina’s curve is the steepest decline on the board.

The other North Carolina number is one nobody has written about. Wake County alone reports 519 cases as of August 21 — more than half of the entire state total, in one county. Put that against Michigan, where Wayne County’s 2,029 is about fourteen percent of the state’s 14,277. North Carolina’s outbreak is far more concentrated than Michigan’s, and a cluster that tight in a single county is the kind of thing that usually has a findable source. Wake County says its own counts are preliminary, and the state says recent weeks understate the total because of reporting delays.

And here is the part that does not sit comfortably. North Carolina is one of the seventeen states FDA and CDC count in the multistate outbreak, and has been since August 5. But state officials say North Carolina’s illnesses are not part of the cluster tied to lettuce served at Taco Bell. Both of those things are on the record right now. A state with 940 cases, more than half of them in one county, is on the federal list without agreeing that it belongs there, and no one has publicly reconciled it. North Carolina’s own baseline makes the size of the question plain: 306 cases in 2025, 332 in 2024, and 151 in 2023.

West Virginia added nine on Friday and its hospitalizations did not move

West Virginia’s Office of Epidemiology updated Friday, August 21: 313 total cases, twenty of them hospitalized. That is nine more people than the 304 the state reported the previous Friday, and the hospitalization count held exactly where it was. West Virginia publishes once a week now, on Fridays, and it is the only state on the chart that has ever revised its hospitalization figure downward, which it did two weeks ago.

My Washington row was wrong by thirty-five people, and the fix is the biggest single change this weekend

I carried Washington at 36 for weeks. That was never a case count — it was the number of people the state had interviewed back in July, and I misread it. The Washington State Department of Health reports 71 cases since May 1 through August 21, two hospitalized, none dead. Of the sixty-five people interviewed, forty-four had traveled to a country where the parasite is endemic and eight never left the state at all. Three Washington residents are counted in the federal outbreak, all after eating at Taco Bell in Michigan and Indiana.

That correction alone adds thirty-five people to the floor, which is more than Missouri and West Virginia added between them last week. I would rather find my own errors in public than have somebody else find them, and Washington publishes more detail about where its residents were exposed than almost any state in the country, so the mistake was mine and not theirs. And, Washington was a state that MAY have received Taylor Farm’s product.

CDC’s national hospitalization figure is now 828, and that is the number nobody is quoting

CDC’s national surveillance page, updated Tuesday with data through August 17, reports 15,716 laboratory-confirmed domestically acquired cases since May 1, at least 11,841 more awaiting confirmation or investigation, two deaths, and cases in forty-seven states plus the District of Columbia and Puerto Rico. The comparison CDC prints in its own key points is the one to remember: 1,180 cases in the same stretch of 2025.

The figure that changed is hospitalizations. CDC now says 828 nationally, up from 740 in the August 10 data — eighty-eight more people in a hospital in one week. A separate 1,681 travel-associated cases with sixty-one hospitalizations sit on top of that and are not part of this outbreak. Add the two series and CDC has confirmed 17,397 people sick and 889 hospitalized this season. The outbreak attributed to Taylor Farms accounts for 454 of those hospitalizations, a little over half.

Also worth noting from the same update: the domestic onset range now runs through August 13, two days later than FDA’s outbreak curve. People are still getting sick.

What to expect this week

Arkansas and Virginia normally post Monday. CDC updates national surveillance Tuesday, and North Carolina and Oklahoma follow the same day. Kansas and New York publish Wednesday. Michigan and Florida report Thursday, which is also Senator Warren’s deadline. West Virginia comes Friday and Missouri next Sunday. FDA and CDC have updated the outbreak figure on Thursdays for the last two weeks, so expect the federal number to move on the twenty-seventh.

And the season ends August 31. CDC treats May 1 through August 31 as the Cyclospora season, so next Sunday’s Missouri report is the last one inside it. That does not mean the counting stops — the six-week lag between when somebody gets sick and when they land in a federal total means illnesses from early August will still be arriving in October. The season closing and the outbreak closing are two different events, and only one of them has a date.

Three questions people are actually asking

Is it safe to eat lettuce now?

The recalled product is gone. Best-by dates ran out at the beginning of August and FDA says it is confident all recalled iceberg lettuce is off the market. Nothing currently on a shelf is part of that recall.

I have had diarrhea since July. Could this be it?

Possibly, and it is worth asking. Cyclospora does not show up on a routine stool panel — a clinician has to order the specific test for it. Untreated, the illness can relapse for weeks. If your symptoms date to this summer, ask for the Cyclospora-specific test by name.

Why is the state number so much bigger than the CDC number?

Because they count different things. CDC and FDA report only laboratory-confirmed illnesses they have tied to the recalled lettuce. States report every case of cyclosporiasis found during the investigation period, confirmed and probable. Missouri says 2,152 and CDC credits Missouri with 927. Neither is wrong.

The floor is 32,169 this evening and it is a floor, not an estimate. It counts only what fifty state health departments have published about themselves. CDC’s own literature puts the underdiagnosis multiplier for this parasite in the dozens, so the real figure is far higher than anything on my chart.

As I said the Guardian last week:

But for some advocates like Marler, whose law firm handles cases of victims who may be severely injured or even killed by food contamination, the delays are nothing short of inconceivable.

The traceability rule “allows product to get off the market faster so less people get sick – all of those things are good for industry, whose product is being besmirched by one entity”, said Marler.

“These grown white men who run these organizations are acting like 2-year-olds… They just don’t want to be told what to do.”

I have been asked a version of the same question all week, most recently by someone who reads this blog (1 of a few) and could not make it add up. I was being a bit too flip to the Guardian.

The fixes are cheap. They plainly help the people paying for this. Why would an industry that just watched a third of the California lettuce crop go back into the ground fight the traceability rule, kill the sampling program and stay quiet while the inspection money got cut? It does add up. 

It only stops adding up if you assume the industry is one actor with one interest. Break it into the people who actually made each of those decisions and every one of them was rational for whoever made it. Here is the arithmetic, as fairly as I can state it, and then the part that changed in July. 

Having one of my three degrees being in economics helps with this.

Compliance has a number next to it. The outbreak that never happens has nothing next to it.

The clearest illustration is a federal document, not an accusation. When the FDA proposed pushing the Food Traceability Rule from January 2026 to July 2028, it ran its own numbers. The agency put the forgone public health benefits of the delay at roughly 91 to 112 million dollars a year and the cost savings to industry at roughly 54 to 73 million. By FDA’s own primary estimate the delay destroys about 37 to 39 million dollars a year more than it saves.

It is worth saying plainly what that first number is made of, because the post-it-note version of this debate never does. Forgone public health benefit is not an abstraction and it is not a rounding item. It is illnesses that happen instead of not happening, hospital admissions, and deaths, converted into dollars by a federal agency using a published method. When somebody weighs ninety-one million against fifty-four million and concludes the delay pencils out, what is on the losing side of that ledger is people. Everyone running the calculation knows that. The convention of stating it in dollars is what makes it possible to run at all.

Traceability does not change how much loss there is. It changes who pays it.

This is the part I think most people miss, and it is the reason the incentive survives even when the total cost is obviously worse. Suppose you can find the lot in a day. The loss is privatized. It lands on the firm that shipped the bad product, which takes the recall, the judgment and the lost account. Now suppose you cannot. The loss is socialized across the whole category. Dick Peixoto in Watsonville pays it. Ryan Kelly in Salinas pays it, and so does whoever was going to fill his cancelled October order. Larry Cox pays it in 300,000 pounds of romaine. A cilantro grower in Chualar pays it on a two percent margin.

Any single firm running the odds that it will be the firm at fault is better off with the loss spread. Opacity functions as a subsidy, and the people paying the subsidy are growers who never touched the product. I am not describing a conspiracy. I am describing what the incentive says when you actually run it, which is why it persists across administrations and across commodities and why arguing about anyone’s motives is a waste of everyone’s time.

And there is a second column of payers that the growers list leaves out. As of August 20 the federal count of this outbreak is 10,930 people in seventeen states, at least 454 of them hospitalized, and two dead in Michigan. State health departments have counted more than thirty thousand. Cyclosporiasis is not a bad night. Untreated it runs for weeks and often a month or more, it remits and comes back, and the specific antibiotic that ends it only gets prescribed if somebody thinks to test for a parasite. The first person in this outbreak got sick on June 14. The food was not named until July 17. Thirty-three days, and the product kept moving the entire time, because nobody could say which lot it was.

A trade association aggregates the median dues payer, not the industry.

The members with the most complex supply chains carry the highest absolute compliance cost and generally the most weight in the room. The organic grower in Santa Cruz County who just laid off harvest crews carries neither. The position that gets filed therefore reflects the members who would pay for the rule rather than the members who pay for its absence, and those are two different sets of people. That is not corruption; it is arithmetic. It is also exactly why I keep saying the associations are the ones who could move tomorrow without anyone’s permission. Their own membership has changed sides underneath them this summer.

That is not a guess about where the membership sits. It is on the record. The International Fresh Produce Association, the largest produce trade association in the country, said in its own statement on July 15 that it has long championed end-to-end traceability and that it and its members are on the record against continued delays with the Traceability Rule, and its revised version asks Congress and the Administration to implement end-to-end traceability and secure resources for state produce safety programs. The National Restaurant Association, whose sector usually gets named as the source of the pressure, calls the rule on its own member guidance page a measure expected to result in fewer foodborne illnesses and deaths, and tells restaurants to prepare now. The editorial board of the Delmarva Farmer, a paper written for the people who grow this food, called for exactly this infrastructure in the middle of the season — electronic purchase records, standardized lot coding, interoperable databases — without ever naming the rule it was describing. Western Growers now has three officers pointing the same way. And Jeff Church at Church Brothers Farms has asked retailers to stop letting silence fill the gap and start telling shoppers about the food safety and traceability programs their suppliers already run. Post outbreak, Taylor Farms says it fully supports and endorses the Food Traceability Rule and that it has already implemented traceback protocols across its operations ahead of the compliance deadline, calling any suggestion that it opposes traceability legislation absolutely false. On the same page the company describes every lot as carrying continuous chain-of-custody documentation from field to shipment, which it says enables real-time traceback rather than after-the-fact reconstruction.

The Leafy Greens Marketing Agreement is the most interesting case, because its answer is in what it built rather than in what it has said this summer. It exists because of 2006. After the spinach outbreak the leafy greens industry wrote traceback and recall protocols, farm water metrics and mandatory government audits into a voluntary agreement that now covers roughly ninety percent of the nation’s lettuce, and it has revised those metrics repeatedly since, including more than fifty changes to farm water use and field sanitation approved in a single vote. This summer it hosted a public webinar on this parasite and posted the slides, and its chief executive said the point was to give the industry practical, evidence-based information it could use that day. The same webinar conceded that important questions remain about how Cyclospora gets into the production environment in the first place. An organization that already writes water standards, already revises them on new science, and says out loud that it does not yet know how this parasite arrives is one metric revision away from what I am asking for. That one does not need Congress and it does not need FDA. It needs a vote.

The Microbiological Data Program was killed for working.

The MDP cost about five million dollars a year. It pulled 120,887 produce samples across 42 states between 2002 and 2011 and triggered 23 produce recalls in 2010 and 2011 alone. It was zeroed out in the fiscal 2013 budget. Read that from a single firm’s chair and the logic is immediate. A federal program that samples your product and generates recalls is pure downside, because the illnesses it prevents are invisible and the recall it causes is not. Nobody ever got a headline for the outbreak that the trip wire stopped. The program’s effectiveness was its liability, and it is the cleanest example in this whole file of a rational actor destroying something valuable to itself.

Their objections are not frivolous, and treating them as though they are loses the argument.

I want to be fair here, because the strongest version of my case has to answer the strongest version of theirs. FDA scored the delay as worth 16 to 22 million dollars a year to covered small entities, and compliance genuinely lands hardest on operations that do not have a compliance department. 

And the government has not held up its own end. Section 204 also directed HHS to build a system inside FDA capable of receiving traceability data and tracing food faster. GAO reported in January that no such system has been established, that FDA has completed 41 of 46 key requirements under the law, and that the agency now expects the tracing system to be finished by July 2028. Telling growers to hit a date the agency has missed for fifteen years is a real objection and it deserves a real answer, which is to fund and finish both halves rather than to delay the half that already exists. I have laid out how the fifteen and a half years actually went by and who filed for the delay, by name, from their own disclosures.

This exact thing happened in 2008, and the industry admitted it on the record.

In 2003 and 2004 the food industry pressed the Bush White House to limit produce recordkeeping, and a plan to require electronic tracking records was killed. Companies argued the proposals were too burdensome. Then came the 2008 Salmonella Saintpaul outbreak. More than 1,400 people were sickened, investigators chased tomatoes for weeks before the answer turned out to be peppers, and the paper recordkeeping system slowed them down. The Associated Press put the business losses at 250 million dollars and headlined the story as an industry bitten by its own lobbying success. Robert Brackett, then at the Grocery Manufacturers Association and previously a senior FDA food safety official, told the AP that broader and more far-reaching rules could have helped, and that they would not have hurt.

That concession is eighteen years old. The tomato growers who lost that season had no more to do with those peppers than Watsonville had to do with Guanajuato. The pattern is not new, and it is not a mystery. It is a rational choice, made repeatedly, that produces the same result every time.

The calculation stopped working in July, and the buyers noticed first.

All of the above held right up until the cost of being untraceable exceeded the cost of being traceable, and this summer it did. Sysco, the largest food distributor in the country, stopped buying iceberg from Mexico entirely and moved sourcing to United States growers, and its chief executive said in the same breath that Taylor Farms is a high quality, high integrity operation. Both things at once. When a buyer cannot show a customer which case came from where, walking away from the whole category is the only risk management left, even for a buyer who trusts the supplier. That is the private market imposing the discipline the rule was written to impose, and it is far more expensive than the rule would have been.

Set the numbers side by side. A peer reviewed analysis in the American Journal of Agricultural Economics put the total societal loss from the November 2018 romaine outbreak at 276 to 343 million dollars and concluded in its own words that the episode demonstrated the economic benefit of mandatory food safety standards and improved traceability. California lettuce alone was a 3.7-billion-dollar crop in 2024, and about a third of the harvest ready crop went into the ground this summer. Against that, the traceability rule is already written and merely delayed, and the sampling program cost five million a year.

The people best positioned to see that the price signal flipped are already saying so. Dave Puglia at Western Growers has said the outbreak showed a need to open the doors wider and give consumers greater visibility into the food safety practices the industry has built. He is right, and it is a notable thing for the head of that organization to say out loud in August of 2026. The answer to the question I keep being asked is not that anyone was foolish. It is that a calculation which was defensible for twenty years quit being defensible somewhere around the middle of July, and the institutions that made it have not caught up yet. They can. Nothing is stopping them but the admission.

Here is what catching up would look like, and none of it needs a new law. The trade associations that asked for the delay can withdraw the ask in a letter, this week, to the same appropriations subcommittee that received the original one. It costs nothing but the sentence. The Leafy Greens Marketing Agreement can put a Cyclospora specific water standard into its own metrics without waiting for FDA to write one, the way it wrote traceback and recall protocols into a voluntary agreement in 2007 after the spinach outbreak. That is what this industry does when it decides its collective reputation is worth protecting, and it has done it before within living memory.

The big buyers can finish what Sysco started. A traceability lot code written into a purchase agreement is not a regulation, does not need Congress and does not need a comment period, and it is far cheaper for everyone in the chain than a distributor walking away from an entire growing region. Every large retailer and foodservice buyer in the country already has the leverage to require it. And Congress can strike the rider, fund FDA to finish its own tracing system on the same clock it is asking growers to hit, and put back the five million dollar sampling program. Funding both halves on the same schedule removes the one objection to this rule that I think is genuinely fair, which is that nobody should be told to meet a deadline the agency itself has missed for fifteen years.

Every one of those moves does the same thing, which is why they belong together. They make the lot findable. And the ability to find the lot is the only thing that puts a loss back on the company that caused it. Do that and the next one is smaller, and it costs a single firm a bad year instead of costing Watsonville, Salinas, Chualar and Yuma a season they had no part in. Nobody has to concede they were wrong for twenty years to get there. They chose the economic case, which was their right, and for twenty years it was a defensible choice. They just priced it wrong. They only have to notice that the arithmetic changed in July, and that the cheap choice and the right choice are finally the same choice.

Might get a kick out of this – nearly 20 years ago – https://www.youtube.com/watch?v=Aq3X5PsUheg

It’s 3:00AM – you are hopefully asleep amd I am obsessing about traceability.

As of its August 20, 2026 update, FDA counts 10,930 people sick in the multistate Cyclospora outbreak traced to iceberg lettuce grown in central Mexico and recalled by Taylor Farms de Mexico, at least 454 of them hospitalized, two dead, across 17 states, with illness onsets running from June 14 through August 11. Lower on the same page, under a heading FDA itself stamps updated 8/13/2026, the agency says distribution of the recalled lettuce was confirmed in thirty-one states and that, based on additional information, the product may have been further distributed to California, the District of Columbia, Delaware, North Dakota, Puerto Rico, Rhode Island, South Dakota, Vermont and Washington. That sentence has been sitting on a federal outbreak page for ten days and I missed it.

The nine appeared on August 13, and the coverage that found them came a week later

I want to be precise about the date, because it is the whole point. FDA did not add those nine jurisdictions on Thursday. The Taylor Fresh Foods recall notice and the confirmed thirty-one-state list were both revised on August 13, when the agency also raised the case count to 9,481 and reported inspectors going into the plant in Guanajuato. The nine may-have entries went up in that same revision. Last Thursday’s update changed the counts, dropped the line about adding new states, and moved the inspection language from the processor to the lettuce growers themselves. It did not touch the distribution section at all.

What is new is that somebody finally read it. Patch ran separate state editions over Thursday and Friday telling readers in California and the District of Columbia that the recalled lettuce may have reached them, and framing it as something federal officials confirmed on Thursday. The date is off by a week. The story underneath it is not. Their point was that residents of those places went weeks without being told, and that is correct, and it is a fair thing to have noticed. It took a local news chain publishing automated state-by-state editions to put a federal outbreak footnote in front of anyone.

Washington’s own health department still tells residents the lettuce did not come here

I live in Washington, which is on the may-have list. The Washington State Department of Health outbreak pagecarries a background section marked updated August 7 that ends with a sentence set in bold: there has been no reported distribution of the recalled iceberg lettuce to Washington state. That sentence was accurate when it was written. FDA’s page has said something different since August 13, and as of this morning the Washington page still reads the way it did on the seventh.

This is not a page nobody maintains. The case block right below that paragraph is current through August 21 and reports 71 cyclosporiasis cases in Washington since May 1, two hospitalized, none dead, with 44 of the interviewed cases tied to travel to a country where the parasite is endemic and 8 people who did not leave the state at all. Three Washington residents are counted in the federal outbreak after eating at Taco Bell in Michigan and Indiana. Washington publishes more detail on where its residents were exposed than almost any state in the country, which is exactly why the stale paragraph stands out. The counts get refreshed weekly. The sentence that tells a person whether the recalled food reached them is sixteen days old and now sits crosswise with the federal page it links to. That page, incidentally, still points at FDA’s nine-state and fifteen-state advisory URLs, both of which have been renamed twice since.

I owe a correction of my own here. My fifty-state chart carried Washington at 36 for weeks, which was never a case total at all — it was the number of people the state had interviewed back in July. The real figure is 71.

I went looking for a statement from any of the nine and did not find one

California, the District of Columbia, Delaware, North Dakota, Puerto Rico, Rhode Island, South Dakota, Vermont, Washington. I searched for a health department notice, a press release, a page update, anything from any of them acknowledging that they are on FDA’s list. I did not find one. That is not proof none exists, and if a reader sends me one I will post it and say so. But the absence is consistent with what Patch found and with what Washington’s page still says.

A may-have is not a scare, and it is not nothing either

The honest counterweight belongs right here. The best-by dates on this product ran out at the beginning of August. Nobody is going to find recalled lettuce in a refrigerator now, and a notice issued today would not pull a single bag off a shelf. A may-have is also not a confirmation; it means FDA received information suggesting the product moved further down the chain, not that it landed on a plate in Providence.

A notice would still do two things worth doing. It would tell somebody in Sacramento or Burlington who has been miserable since July that there is a reason to ask for a Cyclospora-specific stool test, which a routine ova and parasite panel does not reliably catch — that is CDC’s own advice to clinicians. And it would tell that state’s epidemiologists to go look. Both matter more this week than they did last week, because CDC said on August 20 that genotyping of samples from sick people shows the parasites in this outbreak are genetically related. One day later San Francisco used that finding to report that a resident’s infection was genetically associated with the multistate outbreak, in a city and a state with no confirmed distribution and no reason to have been looking. A case in a non-outbreak state can now be tied to this lettuce in a laboratory after the fact, which was not true a month ago.

There is a structural reason the home states stay quiet. CDC counts a sick person in the state where the exposure happened, not where they live, so a Californian who ate at a Michigan Taco Bell is a Michigan number. The state where that person actually lives, sees a doctor and shops for groceries has no entry for them and no particular prompt to say anything. The seventeen-state map reads tighter than the outbreak is, and it always has.

Nobody decided not to tell Delaware

I do not think anyone at FDA or at any of these nine health departments made a decision to keep this quiet. There was no meeting. The information was published, in public, in a federal document, by people doing the job in front of them, and then it sat there because putting it in front of the residents of nine jurisdictions was not on anyone’s desk. That is what a thin system looks like from the outside — not a scandal, just a gap where a person used to be. It is the same gap I have been writing about all summer, whether the subject is a parasite that got dropped from FoodNet, a plant that went seven years between inspections, or a traceability rule pushed out to 2028.

Traceability that ends in a footnote nobody is assigned to read is not traceability. It is a filing. The whole argument for knowing where food went is that somebody then tells the people it went to. My running fifty-state count has passed thirty-one thousand people this summer, and the reason I keep publishing it is that the numbers are the only part of this anyone can check. The distribution list ought to be the same. Nine states and territories are on it. Nine of them should say so.

As of August 20, the Centers for Disease Control and Prevention and the Food and Drug Administration count 10,930 illnesses in seventeen states, at least 454 hospitalizations and two deaths, with onsets running from June 14 through August 11, all of it traced to shredded iceberg lettuce processed by Taylor Farms de Mexico in Guanajuato and recalled on July 17. Not one of those illnesses has been tied to lettuce grown in California, in Arizona, or in a greenhouse in Brooklyn. Retail lettuce prices fell 16.4 percent in July anyway, the steepest one-month drop the Bureau of Labor Statistics has ever recorded for the category.

The federal record names one processor in central Mexico and nothing else.

The CDC outbreak page added 1,449 people on August 20 without adding a single state, which is the first update all summer that did not widen the map. It also carried something new. Public health investigators are running Cyclospora genotyping on samples from sick people, and CDC says the results show the parasites that caused these illnesses are genetically related. Read that precisely, because it matters in both directions. It ties the patients to each other.

FDA moved in the same update, and the move is the most interesting sentence either agency has written this month. A week earlier the agency described an onsite inspection at the Taylor Farms plant. On August 20 it described onsite inspections and sampling at iceberg lettuce growers in Mexico, plural, growers, alongside federal, state and international partners. The investigation went upstream from the building to the fields, which is where a waterborne parasite gets onto a leaf in the first place, and it is the one place no wash step can undo it.

Nothing in that record touches Salinas, Watsonville or Yuma.

California lettuce was never implicated. The California Department of Food and Agriculture and the Leafy Greens Marketing Agreement said so plainly on August 12, and they were right. It did not help. California growers destroyed about a third of the lettuce crop that was ready to harvest. Dick Peixoto at Lakeside Organic Gardens in Watsonville laid off harvest crews and plowed his lettuce under and told Ag Alert his farm has about as much relationship to Taylor Farms bagged salads as it does to Oscar Mayer wieners. Ryan Kelly at Boutonnet Farms in Salinas watched the collapse erase his margin for the year and had his October orders cancelled outright. Larry Cox at Coastline Family Farms chopped up 300,000 pounds of romaine hearts and ran them back into the soil.

It is not only lettuce, and it is not only California. Sabor Farms in Chualar plowed under cilantro after demand sank by roughly 30 percent, on margins of about 2 percent, which is another way of saying the year is gone. Strawberry sales fell 14.7 percent, blackberries 16.4 percent and raspberries 28.1 percent between July 14 and August 1. California lettuce alone was a 3.7-billion-dollar crop in 2024. Every one of those growers is being priced for a parasite that arrived on somebody else’s product from another country.

The receipts show shoppers walked away from the category, not the lot.

The clearest evidence yet came out of verified grocery receipts. Fetch, a rewards platform reading receipts that represent roughly 212 billion dollars in annual United States retail sales, found shopping trips containing fresh lettuce down 10 percent over two weeks, salad mixes down 23 percent and packaged salad kits down 27 percent. Counted by individual shopper, it is worse. A quarter fewer people bought fresh lettuce at all, and buyer counts for mixes and kits fell by roughly 40 percent.

Then it spread. Of six additional produce categories Fetch examined, five were down or flat. Cucumbers fell 13.4 percent, cauliflower 5.6 percent, root vegetables 4 percent. Carrots were the only category to grow, by about a quarter of one percent. Frozen produce rose 17 percent in the same window with no change in promotions, so that was shoppers deciding, not retailers discounting. Pam Cerrone at Market 32 and Price Chopper confirmed the same pattern at the shelf: bagged salad down, fresh produce down overall, frozen up. NielsenIQ has iceberg down 25.4 percent and prepacked salads down 34.2 percent from July 18 to August 8, with all fresh produce down 9.9 percent.

Which is what the price collapse actually measures. Lettuce fell 16.4 percent from June, and it is still up 7.5 percent against a year ago, so the cheap lettuce is not a bargain arriving, it is inventory nobody would take at any price. Retailers cut hard on product that was never recalled and it did not bring the shoppers back.

Growers cannot plan, and that becomes next winter’s problem.

The forward-looking damage is worse than the summer damage. Nicole Minnich-Zapata at Misionero told The Packer that demand lifts one week and pulls back sharply the next without settling into any rhythm, and that this reads as a consumer confidence problem rather than a category problem. She is explicit that the acres being left in the field trace to soft orders and not to quality. Jeff Church at Church Brothers Farms says foodservice held up better than retail and expects another six weeks of soft demand, or longer. And in Yuma, where the winter crop gets decided right now, nobody is allocating acreage with confidence, because no grower can find out where the buyers intend to source. Seed gets bought and crews get hired on a guess.

I have made this argument before and I will keep making it, because it is the whole point of reaping what you sow. When you can name the grower in a day, one farm takes the loss. When you cannot, the category takes it. That is not a slogan. It is the mechanism that turned one processor’s iceberg into a national boycott of cucumbers.

Fund the genotyping, because CDC just proved on August 20 that it works.

Here is the ask nobody has made yet, and it is the one industry should want most. On August 20 CDC produced the first laboratory evidence of any kind in this outbreak, and it did it with genotyping rather than whole genome sequencing, which does not work on this parasite. One day later San Francisco used the same tool to tie a single resident of a state that is not in the outbreak back to the outbreak. That is a capability that did not exist in a usable form a month ago.

Scale it and it cuts both ways, which is exactly why growers should be lobbying for the money. Genotyping is the first tool in the history of this organism that can clear a supplier instead of merely accusing one. A California grower who is currently reduced to issuing press releases about not being implicated could instead be excluded by a laboratory. New York has 939 cases since May and no named source at all. Nobody has asked Albany whether its isolates have been typed, and nobody has appropriated a dollar to make sure they can be.

Count the parasite again and put the counting where the parasite lives.

On July 1, 2025, CDC cut its Foodborne Diseases Active Surveillance Network from eight pathogens to two. Cyclospora was one of the six dropped, for budget reasons, twelve months and eighteen days before the largest Cyclospora outbreak in American history started filling Michigan emergency rooms. Restore it. Restore the state and local inspection funding two administrations cut, because state and local agencies do roughly half of all food processing inspections, ninety percent of produce safety inspections, every retail inspection, and all of the interviewing that attribution is actually built out of. Michigan interviewed more than 2,900 people to get where it got.

Then put the parasite where it belongs on the government’s own lists. Name Cyclospora in the agricultural water rule, where the only numeric criterion is still generic E. coli, an indicator organism that tells you nothing about whether this parasite is in the water. Add it to the National Wastewater Surveillance System, which runs about 1,500 sites across half the country and does not test for it. And rebuild something like the Microbiological Data Program, which cost about five million dollars a year, pulled 120,887 produce samples across 42 states between 2002 and 2011, and triggered 23 produce recalls in 2010 and 2011 alone before it was zeroed out. Set five million a year against a 3.7-billion-dollar California lettuce crop that just lost a third of its harvest.

Get the traceability rule in force, and make the states say where the cases came from.

FDA priced the delay itself. In its August 2025 proposal to push the Food Traceability Rule from January 2026 to July 2028, the agency put the forgone public health benefits at roughly 91 to 112 million dollars a year against industry cost savings of roughly 54 to 73 million. By its own primary estimate the delay destroys about 37 to 39 million dollars a year more than it saves, and Congress then wrote the non-enforcement date into an appropriations rider on top of it. Strike the rider and let the rule take effect. Every trade association that asked for the delay can withdraw the ask tomorrow, at no cost but the admission, and I have already gone through what each of them said and whether they still say it.

One more, and it is cheap. Make acquisition reporting routine. Florida already publishes, as an ordinary field in its weekly disease report, how many cases were acquired in state, out of state and abroad. Washington publishes the most detailed version in the country. Forty-five states publish nothing of the kind, which is why a category gets condemned nationally on the strength of seventeen states doing the asking. If every state reported where its cases came from, the map would stop looking like a map of which health departments had the staff to ask the question.

None of this is expensive. The outbreak is expensive.

I said in the last one of these that I would rather work with these organizations than shout at them, and I still would, and I said as much to Western Growers back in July. The surveillance line, the inspection money, the genotyping, the wastewater panel, the MDP rebuild and a rule that is already law and merely delayed add up to a rounding error against a single lost lettuce season. The growers in Watsonville and Salinas who plowed under a third of a crop they had nothing to do with paid more than that this summer, and they will pay again in Yuma this winter, and they have no way to stop it on their own.

The fastest way to make sure the next one lands on one grower instead of every grower is to be able to find the lot. I would like nothing better than to be put out of the fresh fruit and vegetable business. Nobody has taken me up on it yet.

I am working on a post as to why this keeps happening and what we all can do to prevent this.

As of Saturday, August 22, 2026, four foodborne outbreaks are running at once and every one of them is still open. The Cyclospora outbreak linked to recalled Taylor Farms de Mexico iceberg lettuce stands at 10,930 illnesses in seventeen states, with a fifty-state count of 31,972. The Salmonella Javiana outbreak traced to Sinaloa jalapeños distributed by Coast Citrus Distributors stands at 431 in thirty-two states. The alfalfa sprout outbreak from Everything Sprouts of Minneapolis stands at 55 in fifteen states, and the company finally issued a recall Friday night. San Diego County’s Ramona outbreak, traced to raw eggs from the Happy Hens ranch, stands at twenty-three, and there is still no recall there at all. Here is where each one sits and what the coming week should tell us.

The Cyclospora count has not moved since Thursday, and two states have gone quiet.

Both federal pages froze on August 20 and neither has moved since. CDC’s outbreak page and FDA’s advisoryboth carry 10,930 illnesses, at least 454 hospitalizations, two deaths and seventeen states, with illness onsets running June 14 through August 11 and FDA stating outright that no additional states have been added. On the surveillance side, CDC counts 15,716 laboratory-confirmed cases acquired in the United States since May 1, with 828 hospitalizations, data current through August 17. The subtraction I ran Friday still holds: 828 hospitalizations counted this season, 454 of them attributed to the recalled lettuce, leaving 374 Americans hospitalized with this parasite and no food named for their illness.

What to watch. The reporting calendar does most of the work — New York reports Wednesday, Michigan and Florida on Thursday, West Virginia on Friday, and both federal pages have been updating Thursdays. Beyond that, two states have stopped talking. Ohio has published no statewide figure since August 13 despite reporting weekly all summer, and Indiana skipped its Friday update after telling readers on its own page that it posts Mondays, Wednesdays and Fridays by one o’clock. If either one is still dark by midweek, that is a reporting story rather than a counting one. The larger thing to watch is genotyping. After San Francisco tied a single resident to this outbreak by laboratory relatedness rather than by interview, the question worth putting to every state with a large unattributed series is whether its isolates have been typed and what they showed. New York, with 939 cases and no source ever named, is the biggest of those. The season CDC defines closes August 31, nine days out. The counting does not.

FDA caught up to CDC on the jalapeño outbreak, and the Coast Citrus recall turns out to be 53,760 cases.

The gap I wrote about Wednesday is closed. FDA’s advisory now carries 431 illnesses in thirty-two states with 57 hospitalizations and no deaths, illness onsets running June 19 through August 2, and California added to the state list. It names Coast Citrus Distributors in its opening sentence. And the Coast Citrus recall that had no public notice anywhere for two weeks does exist in the enforcement record: Phyllis Entis at eFoodAlert reportsthat the firm recalled 53,760 cases of jalapeño peppers on July 22, covering all product with pack dates of July 1 through July 31, classified Class I, distributed primarily to wholesale distributors in the central and East Coast states. That is the document the downstream companies have been recalling against since August 8, and consumers have not been able to read it.

What to watch. First, whether FDA ever posts a Coast Citrus company announcement on its recall page. An entry in a weekly enforcement report is not a public notice; it is a database record, and no shopper holding a tub of pico de gallo is going to find one. Second, whether the grower is named. FDA stopped hedging about the distributor this week, and the Sinaloa farm at the end of every one of these threads is still anonymous six weeks in. Third, and this is the one I would chase, Colorado. It is the second-largest state in the outbreak at 118 illnesses, and it appears on none of the downstream recall distribution lists. That is either a hole in the traceforward or a second vehicle nobody has looked for.

Everything Sprouts issued a recall on Friday, and FDA’s blank sample template is still attached to the bottom of it.

The counts hold at 55 sick in fifteen states with four hospitalized and no deaths, illness onsets May 31 through August 8, per CDC and FDA. The news is that the company acted. Everything Sprouts, LLC issued a voluntary recall notice on August 21 covering alfalfa sprouts in five-ounce plastic containers distributed from May 27 forward to third-party distributors and retail grocery stores in Minnesota and Wisconsin. Read the notice itself. It says the recalled products “include the following lot numbers” and then lists no lot numbers at all. There are no best-by dates, no UPC, no product photograph and no store list. The Calco brand, which both federal agencies name, is not mentioned anywhere. Only two of the four organisms are named. The company calls itself “Everyday Sprouts” twice, including in the sentence stating that no illnesses have been reported to it. Retailers and distributors are told to stop selling and dispose; consumers are given no instruction to return anything or get their money back. And below a row of hash marks at the foot of the document sits FDA’s blank sample press release, XYZ Inc. of Anywhere, Mississippi, recalling five-ounce packages of “Snackies” with lot number 666666 and an expiration date of 12/12/99. They published the form with the instructions still in it.

A correction of my own belongs here. On Thursday I wrote that FDA’s table is built one row per pathogen, so a genuinely multi-pathogen outbreak could never appear on it as a single entry. FDA’s active investigations tablehas since merged references 1394 and 1404 into one incident, stating that CDC identified four sick people infected simultaneously with three E. coli strains and Salmonella Agona and combined the associated clusters. The architecture bent. I was wrong about that, and it is a better outcome than the one I predicted. What to watch this week: whether FDA posts the recall on its own recall page, whether a corrected or expanded notice issues carrying actual lot codes and a distribution list, whether Calco gets added, and what came out of the inspection and sampling FDA began at the plant on August 19. The distribution list is the one that matters most, because it is still the only document that would say where this product went.

San Diego County split the Ramona outbreak in two, and the egg ranch has now made its actual argument.

The County raised the count to twenty-three confirmed and suspected cases on Friday afternoon and, as I wrote Friday night, broke the outbreak into its halves for the first time: eighteen illnesses tied to the house-made mayonnaise the County calls Ramona Red Sauce, sold at Ramona Family Naturals Market, and five in people who got sick from Happy Hens eggs directly. Seven have been hospitalized. The most recent case was detected August 16. Nothing has moved since. The County’s last full public update remains August 17, neither CDPH nor CDFA has issued anything of its own, and there is still no recall and no published customer list four weeks in.

The farm’s position has hardened into something more useful than a denial. Chloe Nevarez told NBC 7 that nine of the ill people share the same genotyping but that the health department never obtained a test sample from her farm. That concedes the human clustering and attacks the missing product isolate, which is the argument I expected and the wrong burden. Civil causation has never required a match between a patient isolate and a food; sequencing that ties the sick to each other, exposure histories, and a traceback converging on one supplier is the ordinary proof package. It is also worth saying plainly that a negative environmental swab pulled in August tells you very little about eggs laid in June, because this organism is shed intermittently and can reach the inside of an egg through the hen. What to watch: whether the farm’s own swab results, which its owner said were expected back Friday, ever become public; whether the five non-market cases grow, because that number is the entire reason a recall and a customer list still matter; and whether any record surfaces from the FDA inspection that CDFA confirmed took place at the ranch. An inspection generates an establishment inspection report. It may generate a Form 483. Those are the documents that would turn four weeks of argument into a record.

Four outbreaks, four different foods, four different agencies in the lead. The thing they have in common is not the pathogen. In every one of them, the piece of paper a person would need in order to protect themselves — a distribution list, a lot code, a customer list, a recall notice with a product photograph on it — arrived late, arrived incomplete, or has not arrived at all. The illnesses get counted either way.

Mia de Graaf published a piece at Business Insider this morning asking why the summer of 2026 has felt like one long food scare — Cyclospora in iceberg lettuce, Salmonella in jalapeños and shell eggs, E. coli in frozen blueberries, metal fragments in bread and glass in fruit bars — and whether any of it means the food supply is actually getting worse. She called me. She also called three people who spend their working lives counting this: Craig Hedberg and Melanie Firestone, both foodborne‑disease researchers at the University of Minnesota, and Jennifer McEntire, who consults for the industry side. The answer she came back with is more careful than most of what has run this month, and it is the right answer. Go read it.

She opens on the fact that my business is booming, which is the worst review a food system can get.

The first line of the piece is that my business is booming and that this is not good news for the rest of us. That is the correct way to read it, and I said as much to her. I have been doing this for thirty‑three years. She describes me as a towering figure in food safety, which is generous, and which I would point out is another way of saying that a single plaintiffs’ lawyer on an island in Puget Sound has had a thirty‑three‑year run of work. Nobody should want that job to be steady. This year it is busy enough that we are hiring for the first time in years — another lawyer and a couple of paralegals — because, as I put it to her less elegantly, we have too much to do. Every one of those files is a person who ate something.

The other line of mine she kept is the one I would want kept. Looking at this summer as a consumer, you are entitled to ask what is actually safe to eat. That is not hysteria. It is the reasonable reaction of somebody who read four advisories in six weeks and got no clear account of any of them. My whole answer to her question about whether this summer is as bad as it looks was five words: it is real, and it looks bad. Both halves matter. The illnesses are real and they are being undercounted; the appearance is also worse than the underlying trend, for the reasons she goes on to lay out.

The recall count is not a count of problems.

McEntire makes a point in the piece that belongs at the top of every recall story: when one contaminated ingredient goes into many finished foods, each of those foods gets its own recall notice, and the tally climbs while the number of underlying failures stays at one. Her example is the California Dairies recall in April of bulk powdered milk and buttermilk for possible Salmonella, which rippled downstream into more than a dozen products, from potato chips to ranch seasoning. US PIRG found the same shape in its own review: ten recalls in 2025 set off thirty‑one more because the contaminated ingredient had already been shipped, and all but two of those involved foods on the FDA traceability list. That is exactly the problem the traceability rule was written to solve, and enforcement of it now begins in July 2028.

We are better at finding outbreaks than we used to be — except for the one that has defined this summer.

Hedberg and Firestone both make the surveillance point, and it is an honest one. Whole genome sequencing has let public health laboratories connect Salmonella and E. coli cases that ten years ago would have looked like unrelated stomach bugs in five different states. Some of what looks like more illness is really more detection. De Graaf then does what most coverage skips: she notes that none of it applies to Cyclospora. The parasite will not grow in culture, there is no genome match to be had between a patient and a product, and the investigation therefore rests on what sick people can remember eating six weeks earlier. Hedberg’s summary of this year is the sentence to keep — “an order of magnitude bigger than anything we’ve seen in previous years.” CDC did say this week that genotyping shows the parasites from sick people are genetically related, which is the first laboratory evidence of the season. Read it precisely, though: it ties the patients to each other, not to any particular lettuce.

One farm’s problem becomes a national outbreak because that is how the system is built.

The 2006 spinach outbreak is the example I gave her, and twenty years later I still cannot improve on it. Investigators traced it to a single twenty‑acre field that had been visited by wild pigs. Not every plant was contaminated, but harvesting and processing spread what was, and the bags went to the whole country. Hedberg’s version of the same idea is the more useful one for a consumer: once fresh produce is contaminated, there is very little anybody can do to decontaminate it. There is no kill step between the field and the plate. A system optimized for year‑round lettuce is also a system that turns a local failure into a national one.

The cuts to food‑safety surveillance did not start last year.

She reports the 2025 reductions across the federal health agencies and the grant cuts to state and local health departments, and she prints FDA’s response that frontline investigators were exempt from past workforce reductions. Then she goes back further, which is the part I want people to notice. The USDA Microbiological Data Program — which sampled fruits and vegetables for pathogens and reported what it found, before anyone got sick — ended in 2012 when its funding was eliminated. That was a different administration and a different party. Proactive sampling has been a bipartisan casualty for fifteen years, and the case for bringing it back does not belong to either side.

Hamburger is the proof that this is fixable.

Thirty years ago, more than nine out of ten of my E. coli cases came from a hamburger. Today it is close to none. She has me giving the reason in six words — government intervention, and industry stepping up. That did not happen because consumers learned to use a thermometer. It happened because O157:H7 was declared an adulterant in ground beef, because the industry-built testing and process control around that rule, and because both sides of the table stayed on it for years. The pathogens sickening people this summer are the ones nobody has done that work on yet. Which is another way of saying the work is available.

The one number I would add.

The piece uses the federal figures for the lettuce outbreak — 10,930 illnesses, 454 hospitalizations and two deaths as of the August 20 update — and those are the correct federal figures. They are also a subset. They count the people investigators have been able to connect to the recalled lettuce. CDC’s national surveillancecounts 15,716 laboratory‑confirmed cases acquired in this country since May 1, with 11,841 more awaiting investigation, against 1,180 for the same stretch of last year. My own tally of what the fifty state health departments have published is past thirty‑one thousand. All three are floors, and none of them is the number of people who got sick.

I would trade the whole docket for a quiet summer, and I have said so in print often enough that nobody has to take my word for it. Until then, reporters who take the time to call four people and write the careful version instead of the loud one are how the public actually learns what is going on with its food. I have said the same thing about the epidemiologist who made CDC’s Cyclospora data readable when nobody was paying her to. Thank you, Mia de Graaf. The piece is worth your ten minutes.