Quote from john dough:
Your premise is false -- and I suggest that it is intentionally so.
In natural evolution, a group AA of species A will be geographically dissociated from some group BB, while both groups remain inter-fertile members of species A. Over time both groups will develop allele differences which will not affect intra-group fertility, but which will eventually prevent inter-group fertility. And, the result is species AA and species BB. Meanwhile, species A will no longer exist -- it will have become "extinct," as neither AA nor BB will be a genetic match to what species A once was.
Thus there is no "direct" path from two parents of species A to a single child of species B, because this would be a practical impossibility, as there would be no second member of the new species B with which any other members of species B could be reproduced.
Your required test does not occur in nature (or if it does occur, examples would be incredibly rare), because of the near zero probability of two offspring of a species A being produced simultaneously, from two different sets of species A parents, with their respective offspring capable of reproducing among each other, but not with their respective sets of parents.
However, scientific evidence exists demonstrating that successive incidences of geographical dissociation will produce a "ring species," of groups A, B, C, etc., where group members A <-> B, and B <-> C are inter-fertile, but groups A <-> C are not. Species A and C, thus originated from the same same species A, yet no "direct" species C offspring of species A ever existed.
So, Z, your premise -- is false.