The paper opens by noting that although many naturalists had crossed plants and recorded the results, none had followed the offspring through successive generations in numbers large enough to reveal a general law. Mendel set out to do exactly that, choosing the garden pea (_Pisum sativum_) because its reproductive organs are enclosed in the blossom, making accidental cross-fertilisation rare and controlled artificial fertilisation straightforward.
He selected twenty-two constant varieties and identified seven pairs of sharply contrasting characters: round versus wrinkled seeds, yellow versus green seed albumen, grey-brown versus white seed-coats, inflated versus constricted pods, green versus yellow unripe pods, axial versus terminal flower position, and long versus short stems. For each pair he performed hundreds of reciprocal crossings, recording the results from individual plants and summing across entire trials.
In the first hybrid generation every plant showed only one of the two parental forms, the one Mendel calls dominant, while the other, the recessive, disappeared entirely. When he allowed these hybrids to self-fertilise, the recessive character reappeared in the next generation alongside the dominant, and the two stood in a remarkably consistent average ratio of 2.98 to 1 across all seven characters. Transitional or blended forms were not observed in any experiment.
Mendel then grew on the offspring of that second generation and found that the recessives bred true while the apparent dominants split into two classes: one-third constant, two-thirds hybrid again. This three-generation pattern, hybrid then 3:1 then 1:2:1 constants and hybrids, held for every character and for combinations of two or three characters simultaneously, where the ratios multiplied as independent factors.
He concluded that the egg and pollen cells of a hybrid are not themselves hybrid in composition. Each carries only one form of each differentiating character, and the different kinds arise in equal numbers. When such cells unite at random, the resulting proportions follow directly. Though he lacked the language of genes and chromosomes, the logic Mendel set out (discrete units, purity of gametes, random combination) is the founding statement of Mendelian genetics.