🎓 On Geldings and the 'Natural' Social Order of Horses

The tension between natural horse behavior and the deeply embedded human modifications to horse society in pastoral regions.

In 2020, a team of archaeologists led by William Taylor published a paper on early pastoral economies in Eastern Eurasia. I dug into it back in 2021 when I was catching up on the academic debates about early horse domestication. This month, I reviewed my notes in preparation for finally getting around to writing a proper review of The Horse, The Wheel, and Language by David Anthony and re-read this seemingly unremarkable sentence about contemporary Mongolian horse herds: “Free-range horses largely organize themselves in line with their natural social structure, with a lead stallion and a harem of mares, geldings, and juveniles.” It is a fairly banal description of how Mongolian horse herds work.

It also, if you think about it for more than a second, cannot possibly be how natural horse societies work. Geldings are castrated males. Castration is surgery performed by humans. There is nothing “natural” about a gelding.

What Truly Wild Horses Actually Do

To understand why geldings don’t belong in a “natural” herd, it helps to look at what horses do when humans aren’t involved at all.

Przewalski’s horses are the only surviving truly wild horse species. They were never domesticated, and are genetically distinct from domestic horses. Not even my beloved Polish Konik horses hold this distinction. Przewalski’s horse social structures have two components: harems and bachelor bands. A harem consists of one dominant stallion, several mares, and their offspring up to about two or three years old. When young males reach reproductive age, the dominant stallion chases them out. The evicted males then join bachelor bands, which are groups of young stallions who spar, play-fight, and form surprisingly stable social relationships while they wait for their shot at collecting a harem of their own.

This is the actual natural social structure of horses. Note the distinctive lack of geldings.

In truly wild populations, the bachelor band solves the “surplus male” problem. Most stallions never breed successfully — research on feral horse herds in Utah with the ungainly title ‘Effect of adult male sterilization on the behavior and social associations of a feral polygynous ungulate: the horse’ found that just 7% of males sired 44% of foals at one site — but bachelor bands give non-breeding males a social life. They hang around, they establish dominance hierarchies through ritualized sparring, and occasionally one of them gets strong enough to challenge a harem stallion and take over.

Giraffes do this, elephants do this, hell even spear-nosed bats do this. You know what they don’t do? Get their testicles cut off without human help.

That Utah feral horse study offers a ‘natural’ experiment in what happens when you geld some males in a population and leave others intact. Gelded males who already held harems gradually lost them. Gelded bachelors stayed bachelors, while intact bachelors of the same age mostly went on to acquire harems.

In a truly wild horse population, bachelor bands are a training ground. Young males learn to fight, establish dominance, and eventually challenge for breeding rights. Remove that system and replace it with geldings, and you’ve fundamentally altered the social dynamics of the species — in a way that seems (is?) fine because conflict between male horses is no longer as necessary to the process of working out which horses are reproductively fittest.

What Mongolian Herders Do Instead

Mongolian pastoral horse management replaces the bachelor band system with gelding. Instead of letting young males form their own social groups and compete for mates, herders castrate most colts at around two years old, keeping only a few select stallions for breeding.

The practice has deep ritual dimensions. According to ethnographic accounts from Mongolian herders, castration is a communal springtime activity. The colts are caught, their legs tied, and the surgery performed with a knife cleaned in boiling water. The wound is rinsed with mare’s milk — which sounds strange unless you know that mare’s milk has antimicrobial properties… much like human breast milk, which gets pushed pretty often in mom’s groups as useful for healing wounds and rashes. Dabbing a bit of breast milk as a topical ointment does actually accelerate healing as far as I know, even compared to ‘normal’ topical medicines. Anyway, the way the castration works with horses is that one removed testicle gets pierced and tied to the new gelding’s tail; the idea is that by the time it dries, the wound will have healed (sort of similar to how umbilical clamping works with human babies, I guess?).

The other testicle is cooked in ashes and eaten by the men present, the ritual purpose being to absorb the stallion’s vitality.

Then, getting back to the practical side of things, geldings are calmer and more tractable than stallions. This is why they’re the default riding horse. This distinction is so fundamental that it is baked into the Mongolian language itself. The word mor’ — which refers to a gelded horse — is also the general term for a riding horse. You can ride a stallion (azrag) or a mare (güü), but the most useful horse, the horse you picture when someone says “saddle up,” is a gelding.

Meanwhile, each stallion watches over 15 to 50 mares. The stallions herd, protect, and breed. The geldings slot in underneath them socially — occupying in a totally different niche than how bachelor stallions would fill in a wild population. They lack the hormonal drive to challenge the harem stallion or disrupt the herd’s structure and so tend to act like juveniles forever.

From the herder’s perspective, this is far more convenient than maintaining volatile groups of intact young males. I guess since humans deal with most of the other steppe predators that testosterone probably helped primordial stallions fight off, it all works out.

Origins of the Gelding Rituals

The ancient Greek geographer Strabo, writing in the first century BCE, noted that “it is a peculiarity of the whole Scythian and Sarmatian race that they castrate their horses to make them easy to manage; for although the horses are small, they are exceedingly quick and hard to manage..” Aristotle mentioned geldings too, and the Scythians (roughly 900–200 BCE) are generally credited as some of the earliest documented practitioners of horse castration.

But detecting gelding in the archaeological record is hard. Prior methods relied on the slenderness of limb bones — the assumption being that castrated males develop more gracile skeletons than intact stallions. A 2023 study in iScience developed a DNA methylation–based method for detecting castration in ancient horse remains, which seeeeeeeeems a genuine breakthrough? The researchers tested a bunch of ancient horse specimens and if I’m reading this right, at least some of the horses from the famous Pazyryk burials in Kazakhstan (circa 300 BCE) weren’t actually castrated. These Scythian-era ritual sacrifice horses, which previous researchers had classified as geldings based on their slender leg bones, appear to have been intact males after all.

If the Pazyryk horses weren’t gelded, then maybe either Strabo was generalizing about a practice that wasn’t as universal as he claimed, or castration was reserved for working animals while intact males were preferred for ritual sacrifice? The archaeological picture of when and where gelding became standard practice on the steppe is pretty murky, alas, and honestly a lot of these papers are way over my head.

Why “Natural” Stops Being a Useful Category

That’s okay though, because although horses and steppe nomads are cool and all, the main reason this dug into my brain is how the word “natural” breaks down, in this context and in others. As Alice Roberts argues in Tamed, the distinction between natural and artificial selection is a tricky one:

“Describing artificial and natural selection separately is a false distinction. It doesn’t really matter that it’s humans — rather than the physical environment or other species — that are mediating the assortment of individuals into those more or less likely to successfully reproduce. You wouldn’t make this distinction for any other species.”

Roberts uses the example of honeybees selecting for flower traits. Nobody calls that “artificial” selection. But when humans do functionally the same thing — choosing which stallions breed, castrating the rest — it does start to feel “unnatural.” Usually I find myself arguing that domestication is just a particular type of symbiotic relationship... but the natural vs. unnatural distinction is too useful to disregard even if I do genuinely think that humans are “just animals” in some important senses.

This whole mess is complicated by the way recent scholarship frames horse domestication not as an event where wild became domestic, but as an ongoing process of mutual adaptation. Humans modified horse populations through selective breeding and castration. Horses modified human societies by enabling pastoral mobility, mounted warfare, and long-distance trade. The result is a co-evolved system where asking “what’s natural?” is a bit like asking whether a coral reef’s structure is “natural” given that it’s built by organisms that also modify it. Yes? I guess? But “humans are unnatural!” can back you into some weird mental corners if you take it too far.

Mongolian horses live free on the steppe. They break through ice to find water in winter. They choose their own pasture. Their herders take a remarkably hands-off approach; there aren’t any stables, grain, or grooming. Yet the stallion leading a herd of mares and geldings is operating within a social structure that humans have been engineering for thousands of years. The geldings exist because a human decided their genes weren’t worth passing on. It’s decidedly unnatural selection!

I Hate Unclear Hedges

Look at Taylor’s original sentence again: horses “largely organize themselves in line with their natural social structure.” That word “largely” is doing a ton of work there. It’s a claim that the fit between free-range Mongolian horse herds and truly wild horse social behavior is close but not perfect. But reality has a much bigger gap than that, imo. “Largely” is not doing enough to make the reality clear.

Yes, the stallion-harem unit is genuine horse behavior. And geldings replacing bachelor bands is a human invention so old and so deeply integrated that it has become functionally invisible (because wild horse populations largely don’t exist anymore). But I think to call the total loss of bachelor bands “largely” natural is a mistake for the same reason that calling mature male elephants “largely” useless is wrong.

In the 1970s, conservationists relocated young elephants from Kruger National Park to Pilanesberg. They left the mature bulls behind. The adolescent males grew up, entered musth, and with no older bulls around to knock them out of it, went on a rampage that killed over 50 rhinos. When six large bulls were finally brought in from Kruger, the young males dropped out of musth almost immediately.

Maybe it’s the mother-of-a-young-boy in me, but I would bet that bachelor bands aren’t a footnote to herd structure. They’re where young males learn what they can and can’t get away with. They’re where they practice how to defend their herds, from other horses but also predators. Removing that system has implications, whether we replace it with geldings or just don’t think it matters if we leave some difficult old bulls behind.

It’s just that in the case of horses, those ‘consequences’ favor us and the horses for the same reasons that retarding the emotional growth of canines is how we get dogs, which have in a meaningful sense out-competed wolves as thoroughly as “civilized” men with “unnatural” things like email jobs and atomic families have outcompeted the raiding warbands of the Eurasian steppe.

I have complicated feelings about this. You?

Further Reading

• When Did Horses Transform Mongolians’ Way of Life? from Sapiens has some nice infographics and spans early domestication to Genghis Khan.

• Carolyn Willekes’ The Horse in the Ancient World (which I found via this review) might be my next horse read if I can get ahold of a copy from the library — it’s over $100 on Amazon!