I recently found an interesting DNA study done in Japan that could help provide more insight into the very far back origins of several cold-hardy Japanese citrus varieties.
Some things to bear in mind, this genetic marker analysis is not exactly indicative of precise ancestry percentages. It is only looking at certain markers, so this is only going to give us a very rough idea of the probable ancestry.
Apparently citron (C. medica) is nearly identical to Ichang papeda (C. ichangensis) in the DNA marker analysis, such that the study did not bother to differentiate them.
This means that wherever you see "C. medica" in the ancestry of these cold-hardy Japanese citrus varieties it is actually C. ichangensis.
The following percentages are not exact, they are rough estimates I copied from a visual graph:
C. ichangensis: 91.5% C. medica, 8% C. maxima, 0.5% reticula
Ichang lemon: 60% C. medica, 40% C. maxima
Yuzu: 99% C. medica, 1% C. reticula
Hyuganatsu: 26% C. reticula, 40% C. medica, 34% C. maxima
Kunenbo: 30% C. reticula, 65% C. medica, 5% C. maxima
Kunenbo (II) : 35% C. maxima, 65% C. reticula
Sudachi: 33% C. reticula, 67% C. medica
Kabuchi 33% C. reticula, 51% C. medica, 19% C. maxima
Kabosu: 34% C. reticula, 58% C. medica, 8% C. maxima
Kinkoji: 36% C. reticula, 0.5% C. medica, 63.5% C. maxima
Shiikuwassha 44% C. reticula, 56% C. medica
Keraji: 50% C. reticula, 16% C. maxima, 34% C. medica
Natsudaidai 52% C. reticula, 0.5% C. medica, 47.5% C. maxima
Satsuma: 25% C. maxima, 75% C. reticula
Hirado Buntan: 100% C. maxima
Hybrid Origins of Citrus Varieties Inferred from DNA Marker Analysis of Nuclear and Organelle Genomes,
Shimizu T, Kitajima A.
https://www.ncbi.nlm.nih.gov/pubmed/27902727Some thoughts.
Yuzu is believed to share about half-and-half ancestry from C. reticula and C. ichangensis (or possibly even an earlier papeda species forerunner of C. ichangensis), so the fact that C. reticula barely showed up in the genetic analysis is a clear example of how imprecise the ancestry results of such an analysis are.
Yuzu probably wasn't just a simple cross of C. ichangensis with C. reticula; there probably had to be a few generations sexual propagation for the C. reticula gene markers to get bred out.
Apparently there are two very different forms of kunenbo, one with C. ichangensis ancestry, the other without. It looks like Kinkoji doesn't have any recent C. ichangensis ancestors, so it probably didn't descend from the kunenbo type in the study that showed C. ichangensis ancestry. The same is probably true of Satsuma as well.
Natsudaidai apparently doesn't have any close connection to Yuzu.
(Nansho-daidai I believe is Tiwanica lemon)
The overall ancestry composition in the graph is consistent with the theory that Hyuganatsu resulted from buntan getting pollinated by yuzu. However, if you look at the flow chart, the study inferred that Tachibana-B was one of the parents of Hyuganatsu. The graph shows Tachibana-B to be about 31% C. reticula, 69% "C. medica" (remember represents C. ichangensis here), so it may be that buntan (C. maxima) was pollinated by Tachibana-B, rather than yuzu. Although with that high a percentage of C. ichangensis I suspect Tachibana-B originated from a yuzu cross.
And the study does confirm the leading theory that Ichang lemon is a hybrid between C. ichangensis and C. maxima (though still doesn't prove whether it was a simple cross).
Keraji displays a surprisingly high percentage of C. ichangensis. It was my understanding that keraji originated, over a progression, from a triple backcross of Kunenbo with Shiikuwasha (C. depressa). It's possible that the C. ichangensis genes were positively selected for over time, since those genes conferred cold hardiness.
This isn't from this study but is just some things I've been able to put together from other studies, that may help you make some more sense of those varieties in that list:
Shiikuwasha x kunenbo = kabuchi; kabuchi x kunenbo = kikaimikan; kikaimikan x kunenbo = keraji
kishu x kunenbo = Satsuma; buntan x kunenbo = kinkoji ( kinkoji = Bloomsweet)
kabosu and sudachi are almost certainly hybrids from yuzu
Another DNA marker analysis done in China did not seem to show a connection between Changsha mandarin and C. ichangensis, which is interesting because the fruits/seeds of Changsha mandarin appear very morphologically similar to clementine-yuzu hybrid. The analysis did suggest Changsha mandarin might have just a little C. maxima ancestry though (maybe 15%)
Genomics of the origin and evolution of Citrus, Guohong Albert Wu, Javier Terol
Of course it's also possible the gene markers could have been completely bred out over numerous suceeding generations, since Changsha mandarin originated from growing in the wild.
A note about availability of these varieties in the U.S.
Most of these varieties on this list are fairly available to those in the U.S. The main exceptions are Kunenbo and Hyuganatsu.
Kabosu can be harder to find. Shikuwasa and Keraji seem to be fairly prevelent in Georgia and North Carolina, but I don't believe they can be found in the rest of the country. (Shikuwasa is sometimes written shikwasa, different spellings) C. ichangensis used to be more popular, but currently I don't think it is available from any mail order nurseries. It can still be found in Europe. Bloomsweet was introduced into the U.S. from Texas, so can be found in that part of the country.
Hyuganatsu isn't too difficult to get in Japan, it is sometimes sold as a seasonal specialty fruit. (I don't know whether it actually displays any exceptional cold hardiness traits) Kunenbo used to be the popular fruit in Japan hundreds of years ago before it was replaced by what is today called Satsuma. It's not widely available any longer but can still be found in some botanical and historical collections.
The UCR collection in California supposedly has one but apparently from the descriptions the type they have is not the delicious tasting one that exists in Japan, and in any case it's not available to the public.
Hirado Buntan isn't really that cold hardy but is more cold tolerant than other pomelos.