Premise of the study: Hybridization between diploids and tetraploids can lead to new allopolyploid species, often via a triploid intermediate. Viable triploids are often produced asymmetrically, with greater success observed for maternal-excess crosses where the mother has a higher ploidy than the father. Here we investigate the evolutionary origins of Mimulus peregrinus, an allopolyploid recently derived from the triploid M. x robertsii, to determine whether reproductive asymmetry has shaped the formation of this new species. Methods: We used reciprocal crosses between the diploid (M. guttatus) and tetraploid (M. luteus) progenitors to determine the viability of triploid hybrids resulting from paternal- versus maternal-excess crosses. To investigate whether experimental results predict patterns seen in the field, we performed parentage analyses comparing natural populations of M. peregrinus to its diploid, tetraploid, and triploid progenitors. Organellar sequences obtained from pre-existing genomic data, supplemented with additional genotyping was used to establish the maternal ancestry of multiple M. peregrinus and M. x robertsii populations. Key results: We find strong evidence for asymmetric origins of M. peregrinus, but opposite to the common pattern, with paternal-excess crosses significantly more successful than maternal-excess crosses. These results successfully predicted hybrid formation in nature: 111 of 114 M. x robertsii individuals, and 27 of 27 M. peregrinus, had an M. guttatus maternal haplotype. Conclusion: This study, which includes assembly of the first Mimulus chloroplast genome, demonstrates the utility of parentage analysis through genome skimming. We highlight the benefits of complementing genomic analyses with experimental approaches to understand asymmetry in allopolyploid speciation.