The way crossbreeding works, as far as I can tell, is like so (I'm getting most of this from SR2, but I've seen no indication that it's changed in the new edition):

Human + Human = Almost always Human, with any of the other metatypes possible, but rare and getting rarer.
Human + Elf = Elf
Human + Dwarf = Dwarf
Human + Ork = Ork
Human + Troll = Troll
Elf + Elf = Elf
Elf + Dwarf = ~50% chance of either Elf or Dwarf
Elf + Ork = Almost always Ork, though Elf is a possibility.
Elf + Troll = Almost always Troll, though Elf is a possibility.
Dwarf + Dwarf = Dwarf
Dwarf + Ork = Almost always Ork, though Dwarf is a possibility.
Dwarf + Troll = Almost always Troll, though Dwarf is a possibility.
Ork + Ork = Ork
Ork + Troll = ~50% chance of either Ork or Troll
Troll + Troll = Troll

There are occasional "throwbacks" to human, but I think they're very rare, except for giants, who fairly frequently have human daughters, and elves, who occasionally have twins, with one twin elven and the other human and usually non-viable. And I may have just come up with a theory to explain both the throwbacks and why the goblinized races appear to be dominant, but I need to think it over some more...

I've seen no evidence that it makes any difference which parent was which. I have no clear idea how metavariants work into all this.

Can we get a page number and book on that? I have never read that in a previous edition.

It's mostly from SR2 p.34 (hardcover; I assume the paperback's laid out the same way, but if not, it's the intro to the Metahumanity section), with bits and pieces thrown in from other parts of that section, or other sources. I made assumptions in a couple of places where things weren't stated outright or were ambiguously phrased, but I don't think any of my assumptions are unreasonable, and I've not seen anything canon that'd contradict them.

edit: Specifically, I'm assuming that, "In the instances of attempts at natural crossbreeding, the child is always of the same race as either the elf or dwarf parent," is referring to human + elf and human + dwarf crossbreeds as well as elf + dwarf crossbreeds. I'm also assuming that that sentence and the 50/50 chance of which parent is matched apply to orks and trolls as well (but with ork and troll substituted for elf and dwarf), which is implied but not clearly stated. I think everything else I said is straight canon without much leeway for interpretation.

Usually, yes. The cross-referencing is sufficiently obsessive and as far as I can tell accurate as to merit my sincere admiration.
He's not infallible, but when looking for a bit of SR obscura that's where I typically turn first. I have yet to be steered wrong.

~J

AnHi compiles information from canon source, sure. Here and there he sticks some conclusions and some assumptions in there, so I don't think his word is gospel, and I think I'm not the only one.

Great info!

Guys just to wrap it up with the Giants. You can NOT use the rules for the Giant metatype in this discussion, since this rule ONLY owes its existance to Norse mythology. In Norse myths the Giants often have beautiful human looking daughters. The rule is just SR's way of trying to accomodate the myths. Also Steven Kensons article on love in Earthdawn might prove to contain valuable hints to SR's future. Basically according to the article, the different races can't even procreate with each other in ED, therefore it would stand to reason that the racial differnces become more and more pronounced with the rise of the manalevel, thus making children of a different metatype than one or both parents an occurence that is fast disappearing and soon to be a thing of the past. The same thing goes with goblinization. Of course it will still be some time until the different races can no longer procreate, but it will likely, and already have become rarer and rarer.

If the way it has been implemented in canon material is nearly identical to the mythological source, and if there is no mention of the aspect occurring outside of the explicitly documented implementation (in this case Giants), then assuming it holds true outside of that (read: for other meta-humans) is just that; an assumption. So Reth has a very valid point.

I'd just like to point out that the passage on giants very directly implies it occurs among other metahumans. It says a "greater-than-average tendancy toward genetic reversions". The key phrase being "greater-than-average". If the giants are the only race which experiences this at a statisticly significant level, why was it written this way?

Good point, TG. But you could also reverse that; if it did apply to all meta-races, then why was it not more significantly stated?

Well, the statement seems to logically indicate a reversion rate of about 5% for other races (giants have approximately a 12.5% reversion rate, figuring 50% boys and 50% girls). However, it could be as low as 1-2%. If it's that low, then I can definitely understand why it's not mentioned.

Well to clarify, i didn't mean to say that genetic reversion doesn't happen amongst the other metas, just that you can't use the example of the Giants to try to calculate the reversion percentage amongst thee other metas, because of the fact that the example of the giants seems to aiming more at being in synch with the Norse myths than with the SR universe. Genetic reversion do happen ( still ) but the worldwide percentage is problaly lower than 1%, saying that the Giants express a higher than average genetic reversion percentage doesn't say anything about the exact worldwide numerical value of the meta to human genetic reversion percentage, it could be 12.4% or it could be 0.001%, there is no way to say, just that it is lower than 12.5%. In my opinion the average percentage is problaly around 1 or below, and the high reversion rate of the Giants owes its explanation to a regional and cultural mythology, not standard SR metagenetics, or else it would definitely have merited more mention in books such as SSG, since it would be a major impact on life in the 60's.

If the reversion rate is lower than 1%, the phrasing "greater-than-average" is definitely poorly chosen. "Abnormally high" or similar strong language is warranted if the occurance percentage is more than a factor of ten off of average.