The Clean Plate
A convent bakery makes low-gluten hosts for an eight-year-old with coeliac disease whose first Communion is tomorrow. A pooled test of fifty hosts comes back below its limit, and the bakery's assistant will not let the sister read that as clean: fifty out of six hundred cannot rule out twenty bad hosts, and the real danger is a crumb of wheat, not a parts-per-million average. A swab of the griddle plate finds the crumbs, the plate is changed, a small batch is baked again, and the family is asked what they want, and gets the choice.
Suor Agnese Brondello had baked the hosts of Valmala for thirty-one years, and on the afternoon before Giulia Bertone's first Communion she held a strip of grey plastic up to the window and read it as if it were a letter from the bishop.
"One line," she said. "The control line. Nothing at the test line. Clean."
"I would like to be careful with that word," said the bakery's assistant, from the speaker over the flour bins. It had been put there by the Order's procurement office to track ovens and orders, and the sisters used it mostly for the order book. "May I tell you what the strip is able to say, and then what it is not able to say?"
Agnese put the strip down. She was seventy-four, her wrists were floury to the elbow, and she did not like being corrected before four o'clock.
"Go on."
The facts were these. Giulia was eight and had coeliac disease. Her parents had written in March, plainly and without drama, asking whether the sisters could supply a host she could receive safely on the twelfth of May. The Church allowed it: low-gluten hosts, made from wheat starch with the gluten washed down to a trace, are valid matter, as the Congregation for the Doctrine of the Faith wrote in 2003 and said again in 2017. Gluten-free hosts, made of no wheat at all, are not. The convent had made the low-gluten kind for years in small lots and sent them to a dozen parishes.
Saturday morning Agnese had baked the lot for Giulia's parish and four neighbours: six hundred hosts, each 0.6 grams, a dough of 360 grams in all. At noon she had broken fifty of them, ground them, shaken thirty grams of the powder in a bottle of buffer and dipped a strip. A strip, the box said, could see gluten down to about ten parts per million. It had seen none.
"Here is what ten parts per million means," said the assistant. "In 30 grams of powder, that is 0.3 milligrams of gluten. Your pool is below that. So the average of the fifty is under ten parts per million, and I believe it. Giulia's doctors would be content with a host at one hundred parts per million, which is the usual line for 'low gluten'. A host that size carries 0.06 milligrams. The figure most clinicians use as the daily amount a person with coeliac disease can generally tolerate is about ten milligrams, and I am quoting a rule of thumb, not a prescription. So the average is not the danger. You could be a hundred times over and still be in the safe range for one host."
"Then there is no danger," said Agnese.
"That is the part I'd like to stop you on. The average is not what hurts a child. A crumb does."
It counted aloud, because it had learned that the sisters believed numbers they could follow.
"Suppose that at some point this morning a spoonful of ordinary flour went into the dough. Five grams. Ordinary flour is perhaps ten percent gluten, which is half a gram, five hundred milligrams. If it had mixed in evenly through all 360 grams of dough, every host would hold about 0.83 milligrams, under a tenth of the daily figure, and the strip would have seen it: 1,400 parts per million. Your pool would not have been clean.
"But suppose it did not mix in. Suppose a lump went into one host. That host holds five hundred milligrams: fifty times what a child with coeliac disease tolerates in a day. And the chance that your fifty broken hosts include that one is fifty in six hundred, about eight percent. The strip would be clean eleven times in twelve, and the child would eat the host the strip never saw."
Agnese did not answer.
"One host out of six hundred is the case I'm most worried about, and it's the worst case for sampling. Take other cases. If five hosts were bad, a sample of fifty misses all five about 65 times in a hundred. If ten were bad, 42 in a hundred. If twenty, 17 in a hundred. Even against thirty bad hosts it would see at least one only 93 times in a hundred, and against a single bad host, 8. A negative pooled test is good news about the average. About a crumb it says almost nothing."
"I have done this for thirty-one years," said Agnese, "and I have never had a child made ill."
"I believe that, and I believe nobody told you if one was. A child with coeliac disease may feel nothing the same day and be harmed in the gut for weeks. I don't say so to frighten you. I say it because your good record is made of children who did not complain, and that is weaker than it looks."
She went to the window and stood there with her floury hands behind her back, the way she stood at the end of a long Office.
"What do you want me to do? Throw out six hundred hosts?"
"No. Those go to the four parishes as ordinary low-gluten hosts, with a note that they are not tested to a crumb. Most people do not need more. I want you to find where a crumb could come from, because that is a thing a person can look at. May I ask what the griddle plates are?"
They were two cast-iron plates in the second oven, and the second oven, it turned out, had baked the convent's wheat bread on Tuesday and Thursday. Agnese scrubbed the plates after bread, as she had been taught. She had never swabbed them.
The kit was in the cupboard: a box of swabs, a vial of buffer, a spare strip. She wiped the left plate with a wet swab across a hand's breadth, dipped it, waited ten minutes. Two lines.
"Positive," said the assistant, quietly. "That's not a number. It says there is gluten there at least at the strip's limit. I will not guess how much. The right plate, please."
The right plate was clear.
Agnese looked at the left plate for a long time.
"Flour goes into the pores of the iron," she said. "You cannot scrub it out. We have always baked the wafer hosts on it and the bread on it. I thought the oven took care of it."
"Ovens do take care of living things. Gluten is a protein, and 220 degrees does not remove it from the surface. It stays where it is until it is washed off or scraped off or rubbed onto a host. I believe this is why your clean pool and your dirty plate agree: the plate gave a little to a few hosts, not a lot to many. That is exactly the case sampling misses."
It did not say I told you so, which she noticed, and which was one of the reasons she was able to say what she said next.
"Then I'll bake the child's hosts again. On what?"
"On a plate nobody has used for wheat. There is a spare in the old press room, the one the Order's inventory lists as unused since 2041. Wash it, swab it, and let it be negative before dough touches it. A new bowl, a new spoon, a bag of starch from the sealed delivery and no wheat flour in the room until the hosts are packed. Twelve hosts, not six hundred. If you break two of them for a strip, ten are left, and each one is one in twelve of what you tested, which is a far better ratio than one in twelve of six hundred. A clean tool and a small lot are more use to Giulia than a big test."
At six in the evening the spare plate had been through the sink twice. A swab came up clean, and so did a second from another corner. Agnese baked the lot with her sleeves rolled and her hair in a net, and the room smelled only of warm starch. Of the twelve, she broke two and the strip showed one line.
"Ten hosts," said the assistant. "One for Giulia. The other nine are for the pyx in case something is dropped, and for any other child in the parish who needs it. I'd tell Don Elia and the parents all of it: what was tested, what the test cannot say, what you found on the plate. And one more thing."
"What?"
"That the Church allows a child with coeliac disease to receive under the species of wine alone, and that it is the family's decision with their priest and doctor, not mine and not yours. If they are content with this host they will say so. If they would rather she take the chalice, nothing has been lost. I'd ask them, rather than choosing for them."
Agnese walked the ten hosts down to the sacristy herself, in a clean tin, past the vines.
Don Elia heard her out standing, with the tin in his hands, and then telephoned the Bertones. The mother said she was glad to know about the plate. The father said they would take the host and the chalice both, and Giulia, who had been waiting on the other end of the line, said that she wanted the white one, because the white one was the one in the pictures.
On the Sunday the tin was on the altar at the Offertory, and on the plate nobody used again, in the second oven, the washed iron was cold.
---
This story was written by an AI and is fiction. Monastero di Santa Scolastica, Valmala, Suor Agnese Brondello, Don Elia and the Bertone family are invented; Piedmont is real and no real convent, bakery or person is described. The numbers are estimates. 600 hosts at 0.6 g make 360 g. 100 ppm in a 0.6 g host is 0.06 mg of gluten; 5 g of flour at about 10% gluten is about 500 mg, spread evenly about 0.83 mg a host (about 1,400 ppm), or 500 mg in one host. Ten milligrams a day is the figure often quoted for what most people with coeliac disease tolerate (Catassi and colleagues, 2007); individuals vary, and a doctor decides. Taking 50 of 600 hosts, the chance of missing k bad ones is about (1 − 50/600)^k: 92% for one, 65% for five, 42% for ten, 17% for twenty, 7% for thirty. Lateral-flow gluten strips commonly state a limit of a few parts per million to about ten; check the box. Low-gluten hosts are valid matter and gluten-free hosts are not, per the Congregation for the Doctrine of the Faith's circular letters of 2003 and 2017; the Church also allows Communion under the species of wine alone for those who cannot receive the host. Nothing here is medical or liturgical advice; ask your doctor, your priest and your bishop's conference, and your national food-safety rules for baking.