In a glass terrarium in Bratislava, a female harvestman with a male on top of her rotates downward, puts her mouthparts on his penis, and by doing so brings the encounter to an early end. Twenty three of the thirty five females in the experiment did it at least once, and the ones who did it in the first pairing tended to do it again in the second, which makes it a strategy rather than an accident.

Scientific American covered the finding on Monday. The study is sound and the reporting is fair, but the emphasis lands on the novelty. What reads as a curiosity is the visible edge of a sexual arms race that somebody has now measured, on both sides, in units of force. What follows is the study, one correction to the coverage, and the considerably stranger literature sitting a single citation away.

Both papers come from Pavol Prokop's group at Comenius University Bratislava. The first, in Scientific Reports in 2024, worked with Phalangium opilio, a harvestman common across North America and Europe and probably resident on your porch. The team collected roughly 230 juveniles, raised them to adulthood in isolation so every female started as a virgin, and then ran twenty minute pairings in glass terrariums.

One more result the coverage left out, because it complicates the story. Females that performed fellatio produced fewer eggs. Prokop's group reads the behavior as a cost of sexual conflict rather than a benefit the female extracts, and they explicitly reject the nuptial feeding explanation for this species. So the tidy version, in which the female trades a favor for a meal and walks away ahead, is not what the data show. She is paying to get out of something worse.

Here is the thread that runs under all of it, and it belongs mostly to Mercedes Burns at UMBC, who is an author on essentially every paper worth reading in this field.

Harvestmen in the eastern North American Leiobunum group come in two anatomical arrangements. Sacculate males have those gift sacs on the penis and open with a face to face embrace, handing the secretion directly to the female's mouth. Nonsacculate males have lost the sacs (though they kept the glands that fill them) and have instead evolved femoral hooks on the pedipalps that clamp the female's front legs and hold her still.

Burns, Hedin and Shultz mapped this across 29 species in 2013. The ancestral condition is gift giving males and unarmored females. Then, on at least four separate branches, the same thing happens in the same order. Males lose the gift sacs and gain the clamps. Females gain a sclerotized pregenital barrier, a plate of hardened cuticle that mechanically obstructs forced entry. The losses and the gains co-occur. Persuasion is replaced by leverage, and leverage is answered with armor.

Two years later the same group went and measured the forces. They dissected muscle cross sections, pennation angles, lever mechanics and cuticle mass across 28 species, and computed how hard a male can protract his penis and how hard a female can hold her operculum shut. The two scale together, R² = 0.482, p < 0.0001. This is a sexual arms race with units.

The 2015 paper also delivers the finding that ruins the clean version. The species do not fall into two camps. They distribute along a continuum, and the analysis failed to partition them by strategy at all. So the romantics and the brutes are the two ends of a gradient, and most harvestmen live in the middle, negotiating.

The myth is durable enough to have its own dedicated debunking pages at two universities. The claim is that daddy longlegs carry the most powerful venom in the world but have fangs too short to break human skin. Rod Crawford at the Burke Museum calls it "a full-fledged Urban Legend, with no basis in fact whatever," and the sentence does a lot of work, because the myth is wrong twice in three different animals.

Three unrelated creatures answer to the name. Harvestmen (Opiliones) have no venom glands, no fangs, no silk and one apparent body segment. They tear up solid food and eat it, which no spider does. Cellar spiders (Pholcidae) are actual spiders and do have venom. Crane flies (Tipulidae) are insects, and many adults have no functional mouthparts at all.

The pholcid is the only candidate that could in principle be dangerous, so somebody checked. Zobel-Thropp and colleagues sequenced and assayed the venom of Physocyclus mexicanus in 2019 and found it made mostly of metalloproteinases and peptide neurotoxins, with an effective paralytic dose in crickets of 3.9 micrograms per gram. That is unremarkable among spiders. Rick Vetter at UC Riverside is on record that he is aware of no publication showing any toxic effect of pholcid venom in humans.

My favorite detail in the whole file is that the 2019 venom proteomics paper, in a peer reviewed journal, cites MythBusters. The show tested it in 2004, verdict Busted, and the spider did bite Adam Savage. He reported a mild burning sensation that went away.