Indeed--the cytoplasm of the sperm is negligible, so (essentially) all the mtDNA comes from the ova. The contradiction would be why, in this limited area, would there be so much diversity of mtDNA? One way to possibly explain it would be for the dominant males to raid other islands and bring captive women back--women who would have different genetic lines.
Interbreeding wouldn't result in specific mtDNAs on "specific islands and even specific populations". This has to be the result of population isolation and I'm guessing long isolation at various locals. Long isolation accounts for the high number of mtDNA mutations not found in other populations.
FWIW Melanesians have the greatest genetic distance from Africans according to Cavalli-Sforza.