It’s the first week of November, and we’re at Peak Pumpkin. Jack o’lanterns are passé, but Thanksgiving (in the U.S.) and traditional winter-solstice-adjacent holidays will keep pumpkin pie and its infamous espresso-based brethren in style for almost two more months.
The cucurbit family, which encompasses cucumbers, melons, and squashes, is kind of a taxonomic mess with relation to common English names. For instance, watermelon is in genus Citrullus, while most other melons are in genus Cucumis with cucumbers. (This makes sense if you’ve ever scooped out a muskmelon’s cucumber-like central mass of seeds and pulp, or if you consider that both honeydew and cucumber are pretty terrific in a gin cocktail.) The pumpkin you usually get in a can for pie filling is a variety of Cucurbita moschata, which has also been bred into hard winter squashes and something called the “Long Island cheese pumpkin.”; the pumpkin the size of a Chesterfield that took a blue ribbon at your state fair is almost certainly the aptly named C. maxima, a species that includes Hubbard squash and several other domesticates. Meanwhile the species from which domestic jack o’lantern pumpkins are derived, C. pepo, is also the source of gourds, zucchini, and summer squashes — “pepo” is the botanic name for the general structure of cucurbit fruits, a specialized form of berry.
All of this reflects the messiness of our use of the word “pumpkin” as well as our use of actual pumpkins. Millennia of domestication and selective breeding have left the wild origins of these species obscure, even though they’re globally important crop plants. The Cucurbit Genomics Database compiles results from projects sequencing genomes from cucumber, melon, watermelon, and the three species that include things we call pumpkins. Many of these are still in early draft stages — for Cucurbita pepo there’s just an expressed sequence tag library. But the genus-wide data is enough to start improving our understanding of the complicated history of all these different, useful species, and a paper that came out earlier this year in Molecular Phylogenetics and Evolution makes some good headway towards that.
Heather Kates, working with Pam and Doug Soltis at the University of Florida, used comparisons across the available cucurbit genomes to identify short regions suitable for phylogenetic reconstruction — introns in conserved, single-copy genes. The authors sequenced 44 of these regions to reconstruct the relationships among diverse samples of each major domestic cucurbit lineage and a number of their wild relatives. The resulting evolutionary tree shows domestic Cucurbita species intermingled with wild species in a monophyletic clade.
The fact that the domestic Cucurbita species don’t form a monophyletic clade separate from their wild relatives is consistent with prior evidence, including archaeological remains, that cucurbits were domesticated multiple times in different parts of the Americas, the family’s native range. The species in that clade are all annuals growing in mesophytic (relatively wet) habitats; because still earlier-branching Cucurbita species are perennials growing in xerophytic (relatively dry) conditions, the common ancestor of the domestic Cucurbita was most likely a perennial in a dry habitat. The move into wetter conditions, and to an annual life history, may have made facilitated domestication when humans arrived in the Americas and started picking pumpkins.
The rapidity of diversification in that new habitat is a big part of the reason that it took data from several dozen independent loci to resolve these relationships as well as Kates and the Soltises were able to — but more fine-scaled geographic sampling will be necessary to pin down the geographic origins of the domestic pumpkins. Until then, we’ll have to carry right on enjoying pumpkin pie without knowing which wild population of Cucurbita was sincere enough to give rise to C. moschata.
Kates HR, PS Soltis, and DE Soltis, 2017. Evolutionary and domestication history of Cucurbita (pumpkin and squash) species inferred from 44 nuclear loci. Molecular Phylogenetics and Evolution, 111, 98-109. doi: https://doi.org/10.1016/j.ympev.2017.03.002