Except they didn't; they alleged that the numbers had been fudged on rather scant evidence.

As I understand it, Fisher had two complaints. First, that the trifactorial data was just wildly too accurate (using a chi-squared approach). This seems to be resolved in the scientific community as an off-by-one error: if instead of 10 seeds, Gregor had used 11 on occasion, that's enough to resolve the issue.

The other issue is the chi-squared analysis of Mendel's experiments in aggregate, which when pooled also too neatly fit the pattern he wanted to illustrate. And as best I know, nobody has adaquately disputed that analysis.

The only actual analysis I saw came up with a p-value of about 0.07, which is just a complete nothingburger. There's no explanation for that because there's nothing to explain.