Per Roose and Traugh (1988), writing at UCR and thus presumably studying the CRC's yuzu accession, historical estimates are that more than 10% of yuzu seedlings (when crossed with P. trifoliata) were believed to be zygotic. There are lots of papers on the methodology used to distinguish zygotic seedlings in polyembryonic citrus using gross morphology, namely seedling height. Their study looked at established plantings to see if rootstocks which were propagated in the field, and thus, presumed by morphological selection to be nucellar, and found that they were about 25% zygotic. So if previous estimates were at most 90% nucellar, and those estimates were only 75% correct, you'd expect that to mean around 67.5% nucellar. I don't know if that's counting all embryos from each seed, because I don't know which author suggested "more than 10%" zygotic, and haven't been able to turn it up despite reading through a few dozen papers on the subject.
Fatta Del Bosco, Matranga, and Geraci (1994) suggest that using isozyme analysis, about 19.4% of yuzu seedlings appear to be zygotic. I don't know what their yuzu source was.