It all started with mice that were captured in the wild and bred with "T mutant" lab mice which had short tails (the t in t-complex stands for tail). Somehow when you bred them together you got mice with no tail. There was some kind of "modifier" in one of the mice which was exacerbating the phenotype. People tried to map this supposed modifier, but the linkage was all messed up, and the heredity wasn't Mendelian.
Well this modifier wasn't in fact a gene it was a large region of chromosome 17. This region could either be wildtype (+) or be severely remodeled and called t-haplotype (t). The remodeling involved several rearrangements including multiple inversions which basically prevented the region from ever recombining with a normal (+) carrying chromosome 17. On top of that the region contains multiple mutant genes which are in a really strong linkage disequilibrium (IE you either get a bunch of mutants or the wt, no mixing at meiosis).
Males carrying two t-haplotypes are infertile, and so you would expect this anomaly to be quickly swept under the evolutionary rug. However there was a strange case of selfish gene interfering with the process. The het males (+/t) transmit the t to all the offsprings, and completely ignore poor Mendel and display a male transmission ratio distortion (TRD). How can this be? Obviously those genes where interfering with either meiosis or reproduction, and making sure they were passed on preferentially. Well it turns out that in that region there are a number of genes which affect sperm flagellar development.
It wasn't that the embryos getting the + allelle from the +/t father were dying, it wasn't that the +/t males were only producing t sperm, it was much more insidious. Somehow the t sperm is a better swimmer and the t-complex genes conspire to cripple the + sperm. How is that possible you may ask. Well through a quirk of sperm development: the syncithium. while the spermatocytes are maturing they share cytoplasm via bridges, so proteins and RNAs are free to flow between + spermatocytes and t spermatocytes. This way some problems of the mutants can be overcome by shared wt protein. This explain how the t survives but not how the + get crippled. Turns out that among the mutant genes being passed along together in the t-haplotype is a gain of function mutant which rescues the other mutants and is only expressed late in spermiogenesis, after the bridges between sperm have been burned. So the t spermatocyte share the poison to all connected sperm when the syncithium is present, but they save the antidote for later.
So this is how you get an entire selfish region to behave as a unit and be passed along preferentially despite being recessive lethal. Neat huh.