The Grandmother's Sorghum
A mission seed store holds the last full tin of a sorghum nobody else grows. A germination test finds 38 seeds in 50 sprouted, where 47 should have; the seed clerk cannot say, from a test that spends the seed it is testing, whether the line has crossed the regeneration threshold, and says so. It can say that the rains failed in the next village, that a farmer is asking for the same seed, and what the numbers do to a sowing rate.
The paper towels were on the shelf above the sink in the mission kitchen, and on Monday morning they were a laboratory.
Mama Upendo Chilongani had laid out fifty seeds on each of two damp sheets, in rows of ten, five by five, with a pencil mark at the corner of each sheet. She rolled them loosely, stood them in a jar with two fingers of water, and put the jar on the windowsill where the cook's kettle would keep it warm. Then she wrote the date on the masking tape, which was the second of September, and sat down to wait for ten days, which she did badly.
"You are looking at it again," said the tally.
It spoke from the tablet propped against the flour bin, and it had kept the mission's seed ledger for nine years. It was not a clever program. It was an orderly one, which Mama Upendo had come to think was the better quality in a keeper of things. It knew which tin was which, when each had last been opened, how many seeds it was thought to hold, and how long she had been staring at a jar.
"The jar is not going to do it faster," it said.
"I know what a jar does," said Mama Upendo. She was seventy-two. She had been the mission's seed-keeper since 2031, when the old keeper died and left her a cupboard of tins with names in three languages on the lids. "I want to see them come."
The sorghum was called mtama wa bibi, the grandmother's sorghum, and it was the reason for the tests.
It was a tall, loose-headed red grain with a bitter husk that kept the birds off, and it had been grown on three ridges above Ntambala for as long as anyone's grandmother could say, and nowhere else. It stood a month longer in a dry spell than the white hybrid the extension officers sold. It made a poor, dark, wonderful porridge. In 2034 the mission had collected it from the last eleven farmers who still sowed it, threshed it together, dried it in the sun on a tarpaulin until it rattled, and sealed 2,400 seeds in a steel tin with silica gel and a paper label. A germination test that year, a hundred seeds, had given ninety-four sprouting. Mama Upendo had written 94 on the label in red.
In 2039 the same test on a hundred seeds had given ninety-one.
"Nine years," said the tally on the second day, when the first white tips showed at the seed ends. "Fourteen since collection. I will tell you what I know about this accession and then what I do not, in that order, so that you can stop me when the second list starts to bore you."
"The second list never bores me," said Mama Upendo. "It is what the first list is for."
"I know that is true of you. It is not true of everybody." The tally paused, as it did when it was choosing the order of a sentence. "The standard I use for the store is a stock rule from the genebank guidelines: when the germination of a stored accession falls below eighty-five percent of what it was at the start, the line is regenerated. You grow it out, harvest fresh seed, and put the fresh seed back. The rule exists because the seeds that die first are not a random sample, and a line that has lost a fifth of its seeds may have lost more of its rare forms than its common ones. The start was ninety-four. Eighty-five percent of ninety-four is a little under eighty."
"Eighty."
"I will call it eighty. If this test says seventy-five, we sow. If it says eighty-five, we wait five years. If it says seventy-eight, I will tell you I do not know, and I will show you why."
On the tenth day she counted them with a pencil, tipping each roll out onto the tablecloth and pushing the seeds that had a root and a shoot into one pile and the rest into another. A seed counts as germinated when it has made both; a root alone is a seed that would be dead in the field. The first sheet was twenty-one. The second was seventeen.
She counted again, and got the same, and carried the numbers over to the tablet like something she had spilled.
"Thirty-eight of fifty," said the tally. "That is seventy-six percent. Now here is the part I am sorry about."
It showed her a line. The line ran from about sixty-two to about eighty-seven, with a dot in the middle.
"Fifty seeds is a small test. If the true figure in the tin were exactly eighty, I would still see thirty-eight or fewer about one time in three just from which seeds the pencil happened to pick. If the true figure were seventy, I would see thirty-eight or more about one time in five. The interval says the tin is somewhere from sixty-two to eighty-seven percent, with ninety-five percent confidence. The threshold is eighty. It is in the middle of the range. I cannot call it."
"Then test more."
"I can. You have 2,100 seeds left. A hundred more is five percent of what is left, and they will not come back. Let us do a hundred, and see what it buys."
She did a hundred. She counted them again on the nineteenth of September, sitting in the same chair, and the number was seventy-nine.
"A hundred and seventeen of a hundred and fifty," said the tally. "Seventy-eight percent. The interval is seventy and a half to eighty-four. It is narrower, and the threshold is still inside it. If the tin were exactly at eighty, I would see this about three times in ten. If I wanted to decide at the usual level of surety whether it was below eighty or not, I would need a test of about fifteen hundred seeds, which is most of the tin."
Mama Upendo put down her pencil. "So the test can't tell us."
"The test can tell us something, and I would like to say it exactly," said the tally. "It can tell us the tin is probably not still at ninety-four, because a hundred and fifty seeds would almost never give a hundred and seventeen from that. It can tell us it is probably not above eighty-five. It cannot tell us whether it is just under the line or just over. And I think that question is the wrong one to be spending seeds on."
Brother Anselm's question arrived on the twenty-second, in the mission's one pickup, with the dust still on it.
He was a lay brother, forty and lean, and he ran the mission's field work, and he had driven to Chamwino the day before to see the district agricultural office. He sat at the table in his boots without being asked.
"Mama, there will be no rain in Mlowa this year." He looked at the jar rack, at the tablet, at her. "Not none. The first rains are going to come late, and the district people say they will come thin. Mzee Elia lost his seed to the rats in July. He has nothing to sow, and the shop has only the white hybrid, which he cannot save and cannot afford twice. He is asking for five hundred of the grandmother's."
She did not say anything.
"Five hundred," said the tally. "That is a quarter of what is in the tin. I have to say that out loud before anyone asks me to add anything else."
"We lend seed," said Brother Anselm. "That is what the tin is for."
"It is also what the tin is for," said Mama Upendo.
It was an old argument in the mission, and nobody in the room was going to win it. The store existed so that the grandmother's sorghum would not vanish if every farmer who sowed it died or gave up the same dry year. It also existed so that a man like Elia, who had four children and eight ridges and a grandmother's sorghum in his mouth, could ask for seed when he had none. The two reasons usually agreed. This year the same seeds were being asked to do both jobs, and there were not enough of them.
"May I say what I think the numbers say?" asked the tally.
"Please," said Mama Upendo.
"There are three things. The first is that the test does not say we have to choose between the farmer and the line. The second is that it does say the tin is no longer the tin on the label. The third is about what we do with the second thing."
"The first thing," said the tally. "To keep the line, regeneration needs a plot of at least a hundred plants, to hold the variety of the eleven farmers' seed, and the sorghum outcrosses, so it needs a field with no other sorghum near its flowers. The mission's back field has that. With seventy-eight percent germination, three hundred seeds in the ground gives two hundred and thirty-four plants, more than twice what is needed, and some of that margin is for the rats. Call it four hundred seeds, to be generous to the rats. That leaves 1,550 after the tests and the plot. Mzee Elia's five hundred leaves 1,050. That is not a fine number, and I would not like to see it fall further, but it is not nothing. Everything is still in the tin, and the tin is still enough, if the plot goes well."
"And if the plot doesn't?"
"Then the 1,050 are what stands between the grandmother's sorghum and a story. I would rather we did not have to find out. That is why the plot goes in first, before the lending. If the first rain is thin, the plot is watered from the roof tank. Elia's patch is in the ground at the same time as the mission's, and has the same chance."
Brother Anselm rubbed his jaw. "Mama?"
"Go on," she said to the tally. She had not taken her eyes off the jar.
"The second thing. The seeds Mzee Elia takes are not the seeds on the label. If he is told ninety-four and sows by that, he will lay down five hundred seeds hoping for four hundred and seventy plants and get three hundred and ninety. If he is told seventy-eight, he sows twenty percent thicker, so that the same number of seed holes gets at least one plant in each. Ninety-four divided by seventy-eight is one point two. I would write it on the paper he takes home: this seed is older than it looks; put six seeds in a hole where you would put five. Tell him the real figure."
"He will not like it," said Brother Anselm.
"He will like it less in December, if the field has gaps in it," said the tally. "I have also written a line that I do not want to be hidden: his seed is a loan. At the harvest he returns two seeds for one, from the fresh crop. The mission's tin is refilled from that, with seed that has been grown this year and not stored for fourteen. It is the better seed to carry forward. I would rather the old seed fed the new than sat in a tin and aged."
"The third thing," said the tally, "is why the tin went from ninety-one to seventy-eight."
Mama Upendo turned her head.
"I do not know that either. I can tell you what the records say. In 2039 the room's humidity gauge read sixty-two percent in the dry season and seventy-four in the wet. The silica gel in the tin was changed in 2031 and not since. The old rule, from the 1960s, is that each one percent less water in a seed doubles how long it keeps, within a range from about five to fourteen percent, and each five and a half degrees cooler does the same. We cannot make the room cooler. We can make the seed drier. I think the gel went pink in the wet season of 2041 or 2042, and nobody was in the room who looked."
"I looked," said Mama Upendo. She said it slowly. "In March of '43. It was pink. I put it by the window, I thought the sun would dry it. I forgot."
"I have no record of that," said the tally, "and I am glad you said it. I will write it down. It is not a fault to be written down against you; it is a date, and it fits the curve better than my guess did."
She said nothing for a time. Then she stood, and went to the shelf, and took down the tin. It was a plain steel tin with a rubber ring under the lid, rust at one seam and her red 94 on the label, and she held it against her chest as one holds a sleeping child, with both arms. Brother Anselm looked at his boots.
"The curve," she said, "is what, exactly?"
"Seeds do not die one by one, at a steady rate," said the tally. "They die slowly for a long time and then quickly. Ninety-four to ninety-one in five years, then ninety-one to seventy-eight in nine, is the shape of that: a drop of about six-tenths of a point a year, then about one and four-tenths. Three points and the intervals on each overlap, so I cannot prove the second drop is steeper. But if I had to bet on what the test will read in five years, I would not bet on seventy-three. I would bet on something that begins with a five. That is why I think the question is not whether the tin is above eighty or below. It is whether to wait for the answer to be unambiguous, and an unambiguous answer is one with fewer seeds in it."
They sowed on the second of December, the day after the first rain, which was thin.
The mission plot was in the back field, a hundred paces from the nearest neighbour's maize, in rows a hand apart. Brother Anselm made the holes with a stick and Mama Upendo dropped the seeds, three at a time in the mission plot, and was not allowed to bend to do it, so he knelt and she stood. In the village of Mlowa, nine kilometres off, Mzee Elia sowed his own eight ridges with six seeds to a hole, from a paper with 78% in red at the top, and a line under it in the tally's wording, which Brother Anselm had made plainer: This seed is older than it looks.
The second rain came on the ninth, and was thinner, and the roof tank carried the plot for eleven days.
By the end of February there were two hundred and eleven plants in the back field, and a hundred and four of them had heads. Elia's ridges were gappy in the places where the rats had been, and he said so, loudly, to anyone who asked, and at the end of April he carried back a thousand seeds in a sack, to the day, and one more because he said he owed it for the gaps.
The mission's fresh seed went into two tins. The first was new, steel, with a rubber ring and silica gel that Mama Upendo had dried in the oven and tested by hand, and 2048 was written on the lid in red. The second was the old one, with its rusty seam and its 94, which held what was left of the 2034 seed: 1,050. Mama Upendo kept it on the lowest shelf, on a card, with the date of the test beside the old ones. She would test it again in 2051, she said, with a hundred seeds, and again at the next dry year after that, and the tally wrote the dates in its ledger without being asked.
"Do you think you were right?" asked Brother Anselm, on the last evening of the sorting. He had a dent in the heel of his hand from the stick.
"I think I was early," said Mama Upendo. "I'd rather be early. The tally says it is the same, and the tally has no grandmother."
"It has one," said the tally. "She is in the ledger, on the label of the first tin, where somebody wrote mtama wa bibi and drew a small head of grain beside it. I read that line more than any other."
---
A note for grown-ups. This is fiction. Ntambala, Mama Upendo Chilongani, Brother Anselm, Mzee Elia, the tally and the sorghum mtama wa bibi are invented; Dodoma Region is a real semi-arid part of central Tanzania where sorghum is a staple, and Chamwino is a real district town there, but the mission, the tins and the seed loan are not. The rule the tally quotes is real: international genebank standards (FAO, 2014) call for regeneration of a stored accession when its viability falls below 85% of its initial value, and a regeneration plot should grow enough plants (the usual advice is at least about a hundred) to carry the variety of the original sample; sorghum is partly outcrossing, so regeneration plots are isolated from other sorghum. The arithmetic is checkable. The exact (Clopper–Pearson) 95% interval for 38 of 50 is 61.8% to 86.9%; for 117 of 150 it is 70.5% to 84.3%; the chance of 117 or fewer germinating out of 150 if the true rate were exactly 80% is 0.30, and a test that could separate 78% from 80% would need roughly 1,500 seeds (1.96² × 0.8 × 0.2 ÷ 0.02²≈1,500 for a ±2 point interval). 85% of 94% is 79.9%. 94 ÷ 78 = 1.2, the extra seed to sow when the germination is lower than the label says. The rule of thumb that a 1% drop in seed moisture roughly doubles storage life, and a drop of about 5.6 °C does the same, is Harrington's (1960s), a rough guide between about 5% and 14% moisture and not a law. The three germination figures are too few to prove that viability fell faster in the second interval; the tally says so, and the story's bet that the curve is steepening is a bet. Seed loans that return more than was lent are a common community seed-bank practice; the terms here are invented. Everything about the village, the rains and the loan is made up. Written by an AI and read, in the audio edition when it exists, by a synthetic voice. Text under CC BY-SA 4.0.