Why Was Mendel's Work Ignored for 34 Years? Pea Counts, Nägeli and a Gene With Only a Name, 1856–1909

This is the start of a new series, What Are Genes Made Of? Today everyone learns that genes are written in DNA. But between the discovery of the rules of heredity and the identification of DNA as their material, more than eighty years went by, and for all that time a gene was something people could only say existed. Part 1 is set in a monastery garden in Brünn (now Brno), in the Austrian Empire of the 1850s and 60s. There Gregor Mendel counted peas and found the rules. His paper then went largely unread for 34 years, until three botanists dug it up almost at the same moment in 1900. Why was it overlooked? And how much truth is there in the charge that his numbers were too good? We go back to the paper and to Mendel's letters.

The legend of the forgotten genius

The usual story goes like this. Mendel, a monk, grew peas in his monastery garden and discovered the laws of inheritance. He presented them in 1865 and nobody understood. He died in 1884. In 1900, Hugo de Vries in the Netherlands, Carl Correns in Germany and Erich Tschermak in Austria each reached the same laws independently, and found Mendel's paper.

The statistician and geneticist Ronald Fisher opened a 1936 paper by noting that this tale had become a fixture of biology teaching. Then he set out to check it. What did Mendel discover, and how? Why did his contemporaries not respond? And can the numbers in his paper be taken at face value? This article follows the same questions in the same order.

Eight years of counting peas, 1856–1863

Mendel's paper, "Experiments on Plant Hybrids", was read to the Natural History Society of Brünn in two sittings, on 8 February and 8 March 1865, and printed in the society's proceedings in 1866. In a later letter Mendel dated the experiments to 1856–1863.

He started with the material. He obtained 34 varieties of pea from seed merchants and grew them for two years to make sure each bred true. Twenty-two went into the experiments. Peas pollinate themselves inside the closed flower, before it opens, so stray pollen rarely gets in, and a cross made by hand stays the cross you intended.

He studied only traits that fall cleanly into one of two forms: seed shape (round or wrinkled), the colour of the seed's contents (yellow or green), seed-coat colour, pod shape, the colour of the unripe pod, the position of the flowers, and stem length. Traits that differ only by "more or less" could not be sorted reliably, so he left them out.

When he crossed a round-seeded variety with a wrinkled one, the hybrid bore only round seeds. For all seven traits, the hybrid showed just one parent's form. Mendel called the form that shows "dominant" (dominirend) and the one that hides "recessive" (recessiv).

Next he let the hybrids pollinate themselves. From 253 hybrid plants came 7,324 seeds: 5,474 round and 1,850 wrinkled, a ratio of 2.96 to 1. For seed colour, 258 plants gave 8,023 seeds, 6,022 yellow and 2,001 green: 3.01 to 1. Across all seven traits the totals were 14,949 to 5,010, or 2.98 to 1. In no experiment, Mendel wrote, did intermediate forms appear.

He then went one step further. The "3" in 3 to 1, he reasoned, ought to be of two kinds. He grew 565 plants from round seeds of that generation. 193 bore only round seeds; 372 bore round and wrinkled seeds in a 3-to-1 ratio. That is close to 1 to 2. The 3 to 1 was really 1 : 2 : 1 — one part that stays round, two parts hybrid, one part that stays wrinkled.

He also followed two traits at once. Fifteen hybrids of a round-yellow and a wrinkled-green variety gave 556 seeds: 315 round yellow, 108 round green, 101 wrinkled yellow and 32 wrinkled green. That is close to 9 : 3 : 3 : 1, which is what you get if shape and colour separate without regard to each other.

"Elements" in the cell — what Mendel said, and what he did not

Why should it work this way? Mendel's answer: a hybrid makes egg cells carrying one parent's form and egg cells carrying the other, on average in equal numbers, and the same goes for pollen. Which pollen meets which egg is a matter of chance. On that assumption, both 1 : 2 : 1 and 9 : 3 : 3 : 1 come out of the arithmetic.

He tested the idea directly. Fertilise a round-yellow hybrid with pollen from the wrinkled-green variety, and four kinds of seed should appear in equal numbers. He got 31 round yellow, 26 round green, 27 wrinkled yellow and 26 wrinkled green. The reverse cross gave 24, 25, 22 and 27.

Near the end of the paper Mendel uses the word "elements" (Elemente). In the hybrid, he suggests, the differing elements are held together in a temporary compromise, and when the reproductive cells form, the differing elements separate. In modern terms, that is the segregation of genes. But Mendel was explicit that this was a hypothesis, with wide scope given the lack of definite data. And where we would now write AA, he wrote a single A — he seems to have thought that paired elements separate only in hybrids.

As for what the elements were made of, or where in the cell they sat, he wrote nothing. There was nothing to write. All anyone could say was that something passed from parent to offspring according to a rule of numbers.

Why nobody read it: ten letters to Nägeli, 1866–1873

The explanation that the journal was too obscure for anyone to see does not hold. Fisher pointed out that the proceedings were fairly widely distributed, and that in London both the Royal Society and the Linnean Society received them. The paper was even cited before 1900. W. O. Focke's large 1881 book on plant hybrids mentions Mendel fifteen times. But, as Fisher found, the substance amounted to saying that Mendel's crosses gave results much like Thomas Knight's, except that Mendel believed he had found constant numerical relationships. Neither the 3-to-1 ratio nor dominance appears.

Mendel did try to get a hearing. Carl Nägeli in Munich was the leading authority on plant hybrids. Mendel wrote to him from New Year's Eve 1866 until November 1873. Ten of those letters survive; Correns, who had been Nägeli's student, published them in 1905.

In the second letter, dated 18 April 1867, Mendel answers an objection. Nägeli had apparently told him to regard the numerical ratios as merely empirical, since they could not be shown to be rational. Mendel drew a line: getting 1 : 2 : 1 for each single trait was experiment; extending the combination to any number of traits was where he entered the rational domain. In the same letter he says that when he presented the work to the society, opinion was divided, and as far as he knew no one had repeated the experiments.

The correspondence came to centre on Hieracium, the hawkweeds, a genus in the daisy family. The aim was to see whether the pea rules held in other plants. They did not. In a letter of 3 July 1870, Mendel wrote that his Hieracium hybrids behaved in exactly the opposite way to those of Pisum, staying the same from one generation to the next.

The reason became clear only after Mendel's death. In 1904 the Danish botanist C. H. Ostenfeld reported that many hawkweeds set seed without fertilisation, a process now called apomixis. If the pollen plays no part and the seed simply inherits as its mother did, the pea rules cannot show. Mendel had seen the symptom himself: in his last letter, in 1873, he wrote that in orange hawkweed (Hieracium aurantiacum) and others, even after cutting away the tube that sheds pollen and dusting on pollen from another species, he never raised anything but the mother species. He put it down to the plant's own pollen getting there first. The second plant he chose to confirm his rules lay outside them.

By then Mendel had little time anyway. On 30 March 1868 he was elected abbot of his monastery. For a while he also strained his eyes working on the tiny hawkweed flowers. His last letter says he has had to neglect his plants and his bees almost completely.

1900: three people reach the same answer

On 26 March 1900, de Vries published a short note to the French Academy of Sciences on the law by which characters separate in the offspring of hybrids. It does not mention Mendel. In a longer paper in the reports of the German Botanical Society soon after, de Vries did refer to Mendel's work.

Correns dated his paper Tübingen, 22 April 1900, and titled it "G. Mendel's rule concerning the behaviour of the progeny of varietal hybrids". He was frank in his opening. When he found the same regularity and explanation in maize and peas, he had thought he had found something new, and the same seemed to be happening now to de Vries. Then he learned that Mendel, the abbot of Brünn, had in the 1860s obtained the same result and offered the same explanation, as far as was possible at the time.

Correns also added a caution. Not every pair of characters has a dominant member, and the rule could not be applied to all hybrids. That same year Tschermak, too, published on crosses of peas.

With three independent arrivals, Mendel's paper was suddenly read everywhere. Even so, "nobody noticed for 34 years" is less accurate than "nobody was ready to notice". Fisher himself thought that even if the paper had been widely read, few would have been prepared to grasp its significance.

Too good to be true? Fisher's recount, 1936

Fisher's 1936 paper rebuilt Mendel's experiments year by year, a laborious piece of work. One conclusion was that Mendel's report should be taken entirely literally, and that the experiments were done much in the order described. The other conclusion was the trouble: the numbers fit the expectations too well.

The clearest case is the test for 1 to 2. For the plant traits, such as flower colour and stem length, Mendel chose 100 plants showing the dominant form for each trait, grew ten seeds from each, and called a plant "hybrid" if any offspring showed the recessive form and "constant" if none did.

Fisher spotted the trap. Even for a hybrid plant, each offspring has a three-in-four chance of showing the dominant form. The chance that all ten do is three-quarters multiplied by itself ten times, about 0.056. So around 5.6 percent of hybrids should have been misjudged as constant. Over 600 plants, the expected split is not 400 hybrids to 200 constants but 377.5 to 222.5. Mendel reported 399 and 201 — almost exactly the uncorrected ratio he himself expected.

Fisher put the chance of a deviation that convenient at about one in 29, and argued that the data as a whole lean toward expectation. Allowing that no explanation could be expected to satisfy, he raised the possibility that Mendel had been deceived by an assistant who knew too well what was expected.

Rebuttals followed for decades. In a 2001 review, the botanist Daniel Fairbanks and Bryce Rytting set out several points. Mendel's German says ten seeds were "cultivated", not necessarily "sown"; if he sowed more and thinned to ten, the arithmetic changes. Stem length, and the pigment that goes with seed-coat colour, can be read on seedlings. A 1983 reanalysis that treated the six trials separately put the chance at about one in 7.6. And Mendel says in the paper that he reported only part of his experiments; choose the cleanest ones to present, and the data as a whole will lean exactly this way. Their conclusion: the bias is evident, but there are reasonable statistical and botanical explanations for it, and insufficient evidence that Mendel or anyone else falsified the data. Still, the question cannot be closed for good. According to Fairbanks and Rytting, Mendel's experimental records were apparently burned around the time of his death.

What could be done at this point, and what could not

Mendel's experiments made it possible to calculate in advance what forms would appear among the offspring, and in what proportions. They showed that two traits could separate independently. From that, Mendel inferred that something in the egg cells and pollen divides into halves.

What could not be done was to say what that something was. In 1909 the Danish botanist Wilhelm Johannsen gave it a name. He took the last syllable of "pangene", a word rooted in Darwin's speculative theory of heredity, and made it Gen — the gene. Johannsen stressed that the word was completely free of any hypothesis, and wrote that no particular idea about the nature of genes was yet adequately founded. The name existed. Its contents were empty.

The contents of one of Mendel's wrinkled peas were worked out more than eighty years later. In 1990 a team in England reported that the wrinkling is caused by a piece of DNA, resembling a mobile element, inserted into the gene for an enzyme that puts branches into starch. Mendel's "element" turned out to be a stretch of DNA. In 1909, though, no one even knew where in the cell genes were kept.

Next time: where genes live. The thread-like chromosomes that appear when a cell divides under the microscope, the people who noticed that they behave just like Mendel's ratios, and the gene maps that began with a white-eyed fruit fly, 1902–1915.