Contributions of Heterosis and Epistasis to Hybrid Fitness in The American Naturalist¹:

...Hybrid survival surpassed that of inbred lines and was equal to or greater than outbred lines' survival, and more F1 than parental plants reproduced. Reductions in hybrid fitness due to Dobzhansky-Muller incompatibilities) (epistasis among divergent genetic elements) were expressed as differences in vegetative growth, survival, and reproduction between plants from reciprocal crosses for both F1 and backcross hybrid generations. Comparing performance of hybrids against parental genotypes from intra- and interpopulation crosses allowed a more robust prediction of F1 hybrids' success and more accurate interpretations of the genetic architecture of F1 hybrid vigor.

I tend to think that heterosis or hybrid vigor is often too easy a fallback theoretical position and am not averse to waving the wand of either genetic incompabilities² or positive fitness syngeries even though the latter tend to throw an element of uncertainty into most generalizations. Speaking of which, plants are easy to breed and track, but they shouldn't be prime candidates for extrapolation to the rest of the biological world (remember that many plants "self" and become hyper-inbred, the viability of which might give some people bizarre ideas). Even so, you have to start somewhere, though I note that the abstract doesn't mention subsequent F1 X F1 matings which would flesh out the extent of the Walhund Effect where less fit homozygote genotypes reemerge from heterozygote matings. Anyhow, I'm rather vexed at the fact that I don't have academic access to The American Naturalist. [Thanks to those who sent me the paper! God bless you!]

1 - This is the journal where William Hamilton published a verbal precis of the two altruism papers that were later to be so influential. I just wanted to note that because the title makes it sound like The Smithsonian magazine or something of that sort (ie; general interest).

2 - A simple way to conceive of how Dobzhansky-Muller incompatibilities emerge is to imagine two separate breeding populations derived from common ancestral population. Over time, imagine that they fix for alternative alleles on the same loci, AA and aa (doesn't matter how or why). In this case the "genetic background" against which other loci might interact with on the aforementioned locus is disjoint between the two populations. So if a second locus fixed for two alternative alleles which are fitness optimized for the population-specific polymoprhism on the first locus, when the populations are brought together and begin to interbreed you might have a novel genetic combination which might be for ill or good.