Two days ago, I posted fifteen questions nobody had answered. Since then, a newspaper has put one of them to the company and been refused. Three have moved without anyone answering them. Eleven have not been touched. And eight new ones have popped into my sleepless brain. Here are all twenty-three, short.

The fifty state health departments now report 23,295 cases. The fifteen states the Washington Post reports are in the federal outbreak — Arkansas, Illinois, Indiana, Iowa, Kansas, Kentucky, Michigan, Missouri, Nebraska, New Hampshire, North Carolina, Ohio, Oklahoma, Pennsylvania and West Virginia — account for 21,393 of them. CDC’s surveillance page carries 10,468 laboratory-confirmed cases plus more than 12,255 unconfirmed. FDA’s advisory says 1,947 illnesses and 98 hospitalizations in nine states and has said so since July 24. Michigan reports two deaths. Twenty-six days remain in the Cyclospora season.

Asked, and refused

What shipped before June 29, and how many pounds were recalled? On Tuesday the Post put both halves of that question to Taylor Farms and the company declined to answer. It matters more than it did Sunday, because the company says the recalled lettuce shipped June 29 through July 16 while FDA says the illnesses began June 22. Cyclosporiasis takes about a week, and CDC puts the range at two to fourteen days. The first person in the federal government’s own outbreak ate that lettuce nine to twenty-one days before the recall window opens.

Moved, but not answered

How many people have died, and how would anybody know? Michigan reported two and CDC acknowledged them. The number moved; the method did not. There is still no published death definition for this illness, no described matching against vital records, no follow-up on hospitalized patients. And the government now says three things at once: CDC’s outbreak page says two, CDC’s investigation page on the same website says none have been reported, and FDA says zero.

Where else did the recalled lettuce go? Consumer Reports went to the companies and confirmed three of the abbreviations on the recall notice: MKTSD is Walmart’s Marketside, SY is Sysco, JB is Jack in the Box. Four others are still guesses, including mine. The House Oversight demand for the full list is due Monday. Three names out of eight, obtained by a magazine three weeks after a recall, is not disclosure.

Were the other twenty thousand ever screened against the case definition? Answered by events, not by anyone saying so. CDC’s confirmed count jumped from 6,707 to 10,468 in one update and six states were added the same week. Somebody went back through cases previously outside the count. Nobody has said how, or how many were reviewed and set aside.

Eight questions that did not exist on Sunday

Why is a newspaper announcing federal outbreak expansions? Second time in twelve days. July 24had a CDC release behind it. Tuesday had four anonymous sources and no comment from CDC. As I write this, no federal page says fifteen.

What are the six new states being added on? Missouri says its investigators interviewed patients who reported eating at Taco Bell. Nebraska’s Douglas County said the same in July, with nearly two thirds of interviewed cases naming the restaurant. New Hampshire has thirty-three cases and its only documented cluster is seventeen people at a hospital cafeteria in Keene. Arkansas, Iowa and North Carolina have published no basis at all.

Did the case definition change when they were added? CDC’s own wording is 1,947 people infected and reporting exposure to Taco Bell. That is the entry ticket, and it is why a state with 11,508 cases contributes a fraction of them. If New Hampshire is coming in on those seventeen cafeteria cases, the definition has broadened for the first time. One sentence from Atlanta would settle it.

How did Nebraska get in? Nebraska is not on the company’s twenty-seven-state distribution list. Neither is West Virginia, which has been in this outbreak from the start. The likeliest answer is not sinister: the recall notice records where Taylor Farms shipped directly, and the chain’s own distributors likely carried it the rest of the way. If so, that list is a shipping manifest and not a map of where this lettuce was eaten, and every consumer checking it has been reading the wrong document.

What is the hospitalization rate now that the case count has doubled? CDC’s last figure is 423, attached to 6,707 confirmed cases. The confirmed count is 10,468 and Tuesday’s update carried no new number. Thirteen states publish a figure of their own and those add to 433. Ohio’s ninety-six dates to July 17, when Ohio had 1,316 cases; Ohio now has 4,013. Nobody is updating the injury column while the case column triples.

Why did CDC take the state count off one page and leave it on another? On Tuesday evening the sentence on the surveillance page describing this outbreak stopped saying nine states. The investigation pagestill says nine. If the number came out because it is about to become fifteen, that should have come with a sentence saying so. Silent edits to a live outbreak page are how a public record becomes unreliable.

Did anyone in this government call the British? UKHSA reported sixty-seven cases in returning travelers against an annual average of ninety-three for whole years. Forty-eight of the fifty-one with travel information had been to Mexico, at different hotels in the same two regions — a supply chain, not a kitchen. Ontario has warned its residents about travel to Michigan. Neither CDC nor FDA has mentioned either country.

What happened to North Carolina’s parsley? A week ago, the state said its increase did not appear associated with the Taco Bell outbreak and that patients were reporting parsley and cilantro more often than expected. Today it is in the lettuce outbreak. Both cannot be the whole story for 718 cases. FDA’s own closed-investigation table carries reference 1325 from August 2025: parsley, sixty-nine cases, closed. Washington named parsley as a vehicle for this parasite twelve months ago.

Eleven that have not been touched

Why did the outbreak’s start date move, and how many left the count? Onsets of May 13 to July 13 became June 22 to July 20 while the count rose by 303 and four states were added. A tightened case definition is a fair explanation and probably the right one. It is also unstated. Everyone who got sick before June 22 left this outbreak and nobody has said how many that was.

Do the two peaks in CDC’s own epidemic curve carry the same genotype? A second hump normally suggests person-to-person spread, and this parasite cannot do that. A second wave therefore requires a second exposure event. CDC has run targeted amplicon deep sequencing on this organism since 2018 for exactly this purpose. Nothing has been published.

What is in the other five investigations? FDA’s outbreak table carries six Cyclospora files. Five read Not Yet Identified: ninety-three cases, ten, eighteen, two and eight. That is 131 people where the government opened an investigation, ran a traceback and named nothing. Add the outbreak and every illness attached to a federal file this season comes to 2,078, against 23,295 the states report.

Did the 2013 sampling program ever run, and what did it find? The same plant reopened in 2013 on a stated condition: comprehensive sampling of product and water, plus continued monitoring of sanitary conditions. Thirteen years later the company’s defense is that no product sample has ever tested positive. Whether it ran, and what it showed, has a paper answer somewhere.

Has anybody sampled the water? FDA’s 2013 environmental assessment recommended this firm determine whether Cyclospora is a reasonably likely hazard in that growing region and, if so, re-evaluate the wash step. That is a standing obligation. What has changed since Sunday is that I can now say where the records are. Rainfall and river levels for 2026 sit in CONAGUA’s Sistema de Información Hidrológica. The long record sits in the national meteorological service’s climatological archive, more than 5,400 stations. Guanajuato’s state water commission administers 419 climatological stations and 395 monitoring wells across sixteen aquifers. Drainage and wastewater-treatment records for the town beside that plant belong to the municipal utility of San José Iturbide. The question is no longer whether anyone sampled. It is why nobody has gone and looked, when every one of those filing cabinets has an address.

What is Mexico doing? In 2013 the investigation was run jointly with COFEPRIS, SENASICA and Guanajuato state health authorities. In 2026 there is no published Mexican case count, no Mexican agency statement, and no sign anyone on that side of the border is investigating. Every document in this outbreak is American. The contamination is not.

Were the harvest workers sick? The 2013 assessment tabulated prescriptions written for harvest crews and found the antibiotic combination used to treat cyclosporiasis being dispensed during the outbreak window. That proves nothing standing alone and I have said so. It is also the most direct question anyone could ask about a fecal-oral parasite, and in 2026 nobody has asked it.

What happens to patients who cannot take the only recommended drug? Two of the four clients whose cases we have filed are sulfa-reactive. Both were prescribed nitazoxanide; one could not obtain it, because it was on backorder, and her husband drove roughly seven hours to fill it. She was hospitalized. At 23,295 counted illnesses, several hundred Americans are in that position now, and no federal agency has said a word about second-line supply.

What does this illness actually leave behind? The produce industry’s consumer FAQ says no significant long-term health consequences have been reported. I would like to know what study that rests on. Nobody is following this cohort, and the government values a case of this disease at four hundred sixty-five dollars — a number that survives only if nobody checks month three.

When did Taco Bell first hear from sick customers? Store signs went up in Michigan in early July saying the restaurants could not serve lettuce, cilantro, onions, pico de gallo or guacamole. On July 14 the company said no link had been confirmed to any ingredient or supplier. Both were true at once. Product does not come off a prep line for nothing, and the date the company first had enough to act exists in a ticketing system.

Why is any of this being answered in litigation instead of by a public agency? The 2024 romaine outbreak that killed one person and hospitalized thirty-six is known to involve this company only because we sued and posted the pages. This week the country learned its largest Cyclospora outbreak had grown by two thirds from a newspaper quoting people not permitted to say their names, and Food Safety News is publishing the fifteen-state count because the states publish it and the federal government does not.

Twenty-six days

Nine members of Congress have written about this outbreak. Not one has been publicly answered. The Oversight deadline falls Monday. CDC’s confirmed count more than doubled this week and now sits 572 cases short of what the fifty states have been reporting all along, which means the states were right in July and confirmation had not caught up. The illness was always there.

The season closes August 31. In 2013 the same parasite from the same valley produced an investigation that ended without an answer and stayed that way. Twenty-six days from now the counts will flatten, the reporters will move on, and nineteen of these twenty-three questions will still be sitting there unless somebody makes them expensive to ignore. I have written down what I think ought to be done. Answering these comes first.

At 4:13 this afternoon the Washington Post named the other five states. Arkansas, Iowa, Nebraska, New Hampshire and North Carolina, joining Missouri. That is the expansion of the largest cyclosporiasis outbreak this country has ever recorded, from nine states to fifteen, and it is attributed to a fourth person familiar with the investigation who is not permitted to say their name. The Post adds that the investigation continues to evolve and the number could change again. CDC did not respond to its request for comment.

I have just been back through every federal page. CDC’s outbreak page says nine. Its investigation update says nine and still puts the count at 1,947. Its map of where people got sick says nine. Its surveillance page says nine. FDA’s advisory says nine, is titled Investigation of 9-State Outbreak, and has not been touched since July 24.

This is how you found out today that your state is part of the largest outbreak of this parasite in American history. You read it in a newspaper. Behind a paywall. Sourced to somebody who is not allowed to be named.

None of that is a criticism of the Post. Lena Sun and Rachel Roubein have been ahead of this outbreak for three weeks and they have been right every single time. On July 24 they had the last expansion too, and that one at least had an agency update posted behind it. This one does not. What has changed in twelve days is not the quality of the journalism. It is that a newspaper has quietly become the announcement channel for a federal foodborne-illness investigation, and the federal government has become the place you go afterward to see whether it has caught up.

Now the arithmetic, because it is worse than it looks. Those fifteen states report 21,393 cases among themselves — Michigan alone reports 11,508. The federal figure for the outbreak is 1,947. Put the other way: thirty-five states are not in this outbreak, and they report 1,902 cases between them.

Two things in the new list are worth stopping on.

The first is Nebraska. Nebraska is not on Taylor Farms’ distribution list. The company’s own recall noticenames twenty-seven states that received the recalled shredded iceberg and Nebraska is not among them. Neither is West Virginia, which has been in this outbreak since the beginning. Two of the fifteen outbreak states are therefore places where, according to the company that made the product, the product never went.

Douglas County, which is Omaha, reported on July 20 that eating at Taco Bell was the most common exposureamong the cases it had interviewed, with nearly two thirds naming the restaurant in the two weeks before they got sick. Nebraskans were eating this lettuce at Taco Bell in a state where, according to the company, the lettuce was never sent. The explanation is not sinister, and it is not complicated. The recall notice records where Taylor Farms shipped directly, and Taco Bell’s own distributors likely carried it the rest of the way. Which means the twenty-seven-state list is a shipping manifest and not a map of where this lettuce was eaten, and everyone who has been checking it to see whether their state was affected has been reading the wrong document.

North Carolina is the reversal in the list. A week ago, NCDHHS told the public that its increase did not appear to be associated with the Taco Bell outbreak, and that its patients were reporting parsley and cilantro more often than expected. Today the state is in the lettuce outbreak. It may still turn out that North Carolina was right about there being more than one thing happening — 718 cases is a great deal to explain with one restaurant chain, and the state said plainly that many of its patients ate neither herb.

And New Hampshire raises the question I flagged this morning, before the names came out. New Hampshire has thirty-three cases and seventeen of them are tied to recalled Taylor Farms lettuce at Cheshire Medical Center in Keene — a hospital, not a Taco Bell.

The fifty states are reporting 23,295 cases today. Michigan added 274 today alone and has recorded two deaths. CDC’s outbreak page acknowledged those two deaths yesterday. Its investigation update, on the same site, still says no deaths have been reported, and FDA’s advisory still says zero. All of them still say nine states and have since July 24.

The Post says the number of states could still change. I believe that. And when it does, we will all read about it in the Post.

The Washington Post reported at 2:21 this afternoon, Lena H. Sun and Rachel Roubein reporting, that CDC is expanding this outbreak from nine states to fifteen, adding six where illnesses have been traced to shredded iceberg lettuce from central Mexico. Three individuals familiar with the investigation are the source for the expansion itself, and they spoke anonymously because the investigation is ongoing. CDC did not immediately respond to the Post’s request for comment.

Missouri is not one of the anonymous parts. Missouri is on the record. The state health department’s spokesperson, Lisa Cox, told the Post that Missouri is one of the states being added, and that patients interviewed at length by state and local investigators said they consumed Taco Bell before getting sick. Information on the other five states was not immediately available.

That distinction matters more than it looks. A state health department putting a named spokesperson behind the statement that it is joining a federal outbreak — while no federal page yet says so — is a state getting out in front of the agency that is supposed to be counting it. Yesterday a CDC spokesperson told a Kansas City television station that the remaining thirty-two states, Missouri among them, were unrelated to this outbreak. Yesterday Missouri said it expected to be added. Today Missouri says it is being added, by name, on the record. The correction took about twenty-four hours, and it still has not reached a federal website.

I checked every federal page before writing this. None of them says fifteen. CDC’s outbreak page says nine and is titled nine. Its investigation update says nine. Its surveillance page says nine. FDA’s advisory says nine, is titled Investigation of 9-State Outbreak, and is still stamped July 24 — eleven days — with Total Illnesses 1,947, Hospitalizations 98, Deaths 0. On July 24 the same reporter had the last expansion and attributed it to an agency update already posted, with a CDC newsroom release behind it.

Now the paragraph in the Post story that I think is the most important thing published about this outbreak in two weeks, and it has nothing to do with the state count.

Taylor Farms told the Post that the recalled lettuce was distributed from June 29 through July 16, through food-service channels and retail stores in twenty-seven states. FDA’s own outbreak advisory says the illnesses in this cluster began on dates ranging from June 22 to July 20.

Read those two sentences against each other. The first person in the federal government’s own outbreak got sick seven days before the company says the recalled product started shipping. Cyclosporiasis takes about a week to make you sick, and CDC puts the range at two to fourteen days. Work it backwards: a June 22 onset means that person ate the lettuce somewhere around June 15, and no later than June 20 even on the fastest incubation the agency recognizes. That is nine to twenty-one days before the recall window opens.

There are only a few explanations. The recall window is drawn too narrowly. Or the earliest cases in the cluster ate product that shipped before June 29 and is not covered by the recall. Or the onset dates are wrong. I do not know which, and neither, apparently, does anyone who has published on it. What I do know is that the Post asked Taylor Farms how many pounds of product the recall covered and where shredded lettuce was distributed before June 29, and the company declined to provide it. That second question is the one that matters, and the refusal to answer it is the most consequential thing Taylor Farms has done since the recall.

Nine days is not a technicality in an outbreak this size. Michigan alone reports 11,508 cases. If contaminated product moved before June 29, then the recall — the only recall of this entire season — never reached it, and it was eaten, and nobody has ever gone looking for it. A recall is a boundary drawn on a calendar. Taylor Farms drew this one and will not say what is on the other side of the line.

Two more things in the story worth pulling out. The Post notes that federal investigators have publicly linked illnesses to only a portion of the states where the company said the recalled lettuce went. That has been true since July 20 and it is good to see it in print. And federal officials are investigating outbreaks and illnesses in more than two dozen additional states that have not been linked to the fifteen-state outbreak, with sources that have not been identified or tied to Taylor Farms.

Which raises the question I have been asked more than any other since the Post story went up: is this outbreak Taco Bell, or does it reach Taylor Farms’ other customers? As the federal pages are written tonight, it is Taco Bell and nothing else. FDA’s advisory says 1,947 people infected with Cyclospora and reporting exposure to Taco Bell. Its list of affected restaurants names exactly one chain. No other company appears anywhere in the case definition.

The recalled product went a great deal further than that. FDA’s own recall notice says the shredded iceberg went out between June 29 and July 16 to twenty-seven states. Consumer Reports reported that alongside Taco Bell, the recalled lettuce went to Walmart stores, Jack in the Box restaurants and major foodservice distributors including Sysco, in that same window. Jack in the Box says it learned of the problem on July 16, a day before FDA’s announcement, pulled product from restaurants served by its Texas distribution center, and has had no known illnesses. And the proportion matters: thirty-three of the thirty-five recalled items went to foodservice and only two went to Walmart retail. The overwhelming majority of this lettuce went to restaurants and distributors, and the federal count covers one restaurant chain.

Missouri is a good illustration, because it had two routes and not one. Lisa Cox says Missouri is joining on the strength of patient interviews pointing to Taco Bell. But FDA’s own advisory lists the states where recalled Marketside bags were sold at Walmart — Alabama, Arkansas, Florida, Georgia, Indiana, Kansas, Kentucky, Louisiana, Missouri, Mississippi, Oklahoma, Tennessee, Texas, Virginia and West Virginia — and Missouri is on it. The recalled bags were on Missouri shelves at the same time the recalled lettuce was in Missouri Taco Bells.

And we already have the proof that non-Taco-Bell exposure to this product makes people sick, because New Hampshire published it. Seventeen people, all at Cheshire Medical Center in Keene, all tied to recalled Taylor Farms lettuce. A hospital cafeteria, not a Taco Bell, in a state that is not among the nine. That is not a theory about what the case definition might be missing. It is a documented cluster sitting outside it.

A note on the twenty-seven. The chart at the bottom of this post counts twenty-eight states as having received the recalled lettuce, and the difference is West Virginia. Taylor Farms’ own distribution list runs to twenty-seven states and West Virginia is not on it. West Virginia is one of the nine outbreak states. It has been in the federal outbreak since the beginning, and it is not on the company’s list of places the recalled product was sent. I have flagged that since July 20, and it has never been explained.

Michigan updated this morning. MDHHS now reports 11,508 cases, up 274 from yesterday, with 193 hospitalizations as of July 30 and two deaths. The fifty-state total is now 22,957. Michigan was the only state that moved: North Carolina’s dashboard, which refreshes Tuesdays, was still stamped July 28 when I fetched it directly, and CDC’s surveillance page had not refreshed either.

When the expanded page does post, here is the test. The 1,947 is not a count of illness. It counts people who were infected and reported eating at Taco Bell, in nine named states. Widen the list to fifteen and either the number grows substantially, in which case the parasite was always in those states, and the case definition was hiding it, or six states get added and the number barely moves, in which case the definition is doing the work rather than the disease. Missouri alone reports 1,095 cases and its investigators have already tied patients to Taco Bell by interview. Watch what happens to the federal figure.

On the federal injury numbers. CDC’s surveillance page, updated July 28 with data through July 27, reports 6,707 laboratory-confirmed domestically acquired cases across forty-five states, more than 11,500 awaiting confirmation, and 423 hospitalizations. FDA’s advisory says ninety-eight. Thirteen states publish a hospitalization figure of their own and those thirteen come to 426.

These are not identical measures, and I am not adding 426 to 423. CDC counts hospitalizations among laboratory-confirmed domestically acquired cases; the states count among everything they have, probable cases included, across thirteen different dates, and Michigan’s 193 are certainly inside CDC’s 423. What the arithmetic establishes is direction. Thirteen states already publish more hospitalizations than the national surveillance total and more than four times what the federal outbreak advisory describes. And 426 is a floor, because the state hospitalization figures are older than the state case counts: Ohio’s ninety-six date to July 17 when Ohio had 1,316 cases, and Ohio now has 4,013 with no update since. On deaths, Michigan says two, CDC acknowledged two on August 3, and FDA still says zero.

Where each state stands now, with the source for every number. Michigan is the only row that changed today.

Thirteen states in that table publish no count at all. Eleven more trace to a national roundup last refreshed July 17, and several of those are bands rather than counts — Georgia and Tennessee published as eleven to thirty, Louisiana and Utah as one to ten. I use the bottom of every band, which is one of several reasons the total is a floor.

Twenty-seven days remain in the season CDC defines as May 1 through August 31. Tonight the outbreak is fifteen states according to a newspaper, nine states according to the government, and 22,957 cases according to the fifty health departments doing the counting. And the company at the center of it will not say how much lettuce it recalled, or where the lettuce went in the weeks before the recall window it drew for itself.

Look at the picture above before you read another word. Ninety-two boxes. Fifteen points of service across seven states, feeding fourteen distributors, four processors, sixteen farms and forty-three individual ranches. Arrows in every direction, legs that cross, and a handful of places where several lines converge on one box. That is what a completed federal traceback investigation looks like when somebody draws it.

It comes from a paper called An Overview of Traceback Investigations and Three Case Studies of Recent Outbreaks of Escherichia coli O157:H7 Infections Linked to Romaine Lettuce, written by eight people at FDA’s Center for Food Safety and Applied Nutrition and three at CDC. The paper says outright that FDA is home to traceback investigation experts who use a standardized process to initiate, execute and interpret these investigations. It then walks through three romaine outbreaks — Yuma in 2018, Santa Maria in 2018, Salinas in 2019 — and publishes the diagram for each one. Those three outbreaks together caused 474 illnesses, 215 hospitalizations and five deaths, and traced back to 45 farms and 84 ranches.

Now look at the diagram again and notice what is not in it. There is not a single company name. Every box is a letter. POS A. Distributor H. Farm C. Ranch AQ. FDA published the entire supply chain, across a hundred pages of records and dozens of businesses, without naming one of them. Whatever the reason may be that we have no such diagram for the Cyclospora outbreak, confidentiality is not it. The agency hid that problem years ago and showed its work anyway.

Here is where we actually are in the Cyclospora outbreak. Fifty state health departments now report 22,683 cases of cyclosporiasis this season. Two people are dead in Michigan. CDC’s outbreak page describes 1,947 of those people in nine states, and its surveillance page carries 6,707 lab-confirmed cases plus more than 11,500 additional cases awaiting analysis. FDA’s advisory lists product distributed to 28 states. What nobody has published is a diagram. Not a redacted one, not an anonymized one, not a partial one for a single leg. Nothing.

What we have instead is a recall covering only seventeen days of shipped product (we still have nothing about product shipped earlier when illnesses happened in April, May and June), a list of the states that product reached, and an assertion that the epidemiological data is overwhelming as to the Taco Bell and Taylor Farms cluster. I am not disputing that assertion. I am saying it is an assertion. The traceback is the evidence, and the diagram is how the evidence gets shown to people who did not sit in the meetings.

The paper is also candid about why these investigations fail, and every reason it names is present in this outbreak. It lists the time it took to initiate a traceback, limited product-identifying information moving through the supply chain, a lack of compatibility between record-keeping systems, and the commingling of product from multiple suppliers. Those four things, the authors write, delayed identification of the farm source and left the root cause of contamination unidentified. FDA said that about itself, in a peer-reviewed journal, five years ago. 

By the way, the rule that would fix the record-keeping half of it has been pushed to 2028.

And this is not a technique that only works on bacteria. It has been done for this exact parasite. In 2013, investigators in Nebraska and Iowa ran a regional traceback on 227 confirmed cases and a multistate traceforward that tied romaine shipped between June 2 and June 6, from a single Mexico-origin field to 189 illnesses across seven states. They published flowcharts of the distribution channels. FDA’s own environmental assessment that year narrowed the source to three ranches of interest among two growers in Guanajuato. Same parasite, same country of origin, the same company (Taylor Farms) at the center of it, with worse tools and a fraction of the illnesses. They did the work and they showed it.

There is now a specific reason this matters beyond principle. A CDC spokesperson has told reporters that cases in the remaining thirty-two states are unrelated to this outbreak linked to Taco Bell and Taylor Farms and are being investigated as separate clusters driven by different foods. That is not an epidemiological aside. That is a traceback claim, and it is the strongest one anybody at the federal level has made all summer. If the legs from New Hampshire, Alabama, Missouri, North Carolina and Iowa genuinely do not converge on the same processor, then FDA has drawn diagrams showing exactly that, and those diagrams are the single most reassuring document the government could release. If it has not drawn them, then nobody knows whether they converge or not, and the claim rests on something other than a traceback.

The 2015 import alert on Puebla cilantro is the model for how this reads when an agency is being straight with the public. FDA said in that document that no single supplier, distributor, farm or packing house could explain all of the illnesses, and that contamination was attributable to multiple farms and packing houses. That is a hard thing for an agency to say. It is also exactly the kind of finding a diagram makes legible, because you can see the legs that fail to meet.

What I am asking FDA to publish is not complicated, and it does not require naming anybody, although there are many times they should and do not. How many points of service were traced, and from which states. How many distributors, processors, farms and ranches the traceback identified. The level of convergence — whether any single processor, farm or ranch appears in every leg, or whether the lines stop at a processor, or whether they do not meet at all. Whether any one entity can account for all the illnesses, or whether this is the Puebla answer again. Which of the six open Cyclospora investigations FDA is now running each diagram belongs to. And the diagram itself, lettered exactly like the one above.

FDA drew that picture for an outbreak that sickened a few hundred people, and it took a peer-reviewed article to get it into public view. Twenty-two thousand six hundred and eighty-three people have been counted this season by their own state health departments. Two families are burying somebody. Twenty-eight days remain in the season, and when it closes the counts flatten and the attention leaves. A traceback diagram is a piece of paper. We are owed the piece of paper.