Finding. Broom hares in Iberia and Italian hares carry someone else's mitochondria. The donor is the mountain hare, an Arctic species that used to live across southern Europe. Costa, Melo-Ferreira and colleagues date that replacement in the two species' common ancestor with whole mitogenomes, including Iron Age samples. Then they ask the nuclear question that mitochondrial capture papers often skip: did a genome come with the organelle?
Coalescent modeling of complete genomes says yes: about 24 percent mountain-hare ancestry in that ancestor, about 900 thousand years ago, roughly 5N generations. Phylogenetic segmentation still sees mountain-hare-like blocks over about 17.8 percent of the autosomes, which is close to the 18 percent you expect if drift, not a purge, handled a pulse of that size. Genome polarisation agrees in outline (25-28 percent of the mountain-hare genome groups with broom and Italian hares rather than the sister Iberian hare, 15-18 percent at diagnostic sites). The leftover blocks are gene-dense with sharp ancestry edges. The deserts are repeat-rich. That is a recombination landscape. Inside the gene-dense, highly introgressed set sit mitochondrial-function genes that look like the mito-nuclear list already discussed in Iberian hares.
Why this paper matters Mitochondrial capture is easy to see and easy to over-interpret. A foreign mitogenome can sweep because of selection, because of sex-biased hybridization, or because a small population took a ride. Nuclear data are how you tell those stories apart. Here the nuclear fraction is large and close to a drift expectation, which argues against a tidy 'only the mitochondrion was wanted' sweep. At the same time, the retained nuclear loci are not a random smear: they concentrate in gene-rich, high-recombination sequence and include mitochondrial partners. Selection can still have played at the edges.
The methods sentence is the other product. They do this without a non-introgressed living reference for the captured lineage, by stacking demography, segmentation, and polarisation. That is reusable in other hybrid mammals where the 'pure' population is gone.
How to read the score Low seventies. Real mitochondrial capture, dated, with a quantified nuclear payload and a mito-nuclear gene hint. It is evolutionary genomics, not a respiratory-chain mechanism paper. Confidence is medium for the percentages (models move) and lower for functional coadaptation (no enzyme assays). Heuristic copy would have said 'hares have introgressed mtDNA.' The paper says the organelle came with a genome-sized dowry, and some of the leftover nuclear DNA looks like it belongs with that organelle.
Caveats No oxygen-consumption curve, no cybrid, no claim that today's climate still favors the mountain-hare mitogenome. Gene-list overlap is not function. Incomplete lineage sorting can mimic introgression; they use polarisation and diagnostic sites to push back, not to erase the risk. Do not turn this into a human mito-nuclear disease analogy.
What to do with it If you curate mitochondrial capture, hybridization, or mito-nuclear coadaptation, take the 24 percent / 17.8 percent / Iron Age dating package and the candidate mitochondrial-function genes. If you build introgression methods, steal the no-clean-reference stack. The experiment this brief cannot replace is a functional test of those nuclear partners against the captured mitogenome.