The earliest domestic pigs in Neolithic China were closely managed

The domestication of large mammals is a key topic for archaeologists. “Domestication” may be defined either (a) behaviorally, animals being habituated to and managed by humans, or (b) biologically, as a morphological divergence from the wild progenitors, usually a size decrease. Whichever definition is used, the problems of interpreting the archaeological evidence are manifold. They are greatest regarding wild boar because their behaviors and diets are highly diverse, and the potential range or relationships between domestic pigs and humans is broader than for other species except dogs (1). The importance of using multiple methods of analysis is universally acknowledged. The paper in PNAS by Wang et al. (2) combines new and established methods to overcome many of the problems and demonstrates close management of pigs in the Lower Yangtze Valley at an unexpectedly early date.

Archaeologists usually view domestication as a gradual process, with behavioral domestication preceding biological domestication by centuries. Domestication is sometimes argued to start with commensalism, where animals scavenge food waste from around human settlements, so people and animals gradually become habituated to each other—this is particularly likely for pigs and dogs (3). From this develops a long series of putative stages through increasing control, leading to extensive breeding, then intensive breeding, and finally pet keeping (4). An example for pigs is shown in Fig. 1, Top, from the Early Neolithic site of Çayönü in Turkey. Length of lower third molar decreases gradually through time. This is argued to show the domestication process extending over two millennia at least (1, 5).

Fig. 1.

(Top) Chart showing size decrease of pig lower third molars from Çayönü, Turkey, in the Pre-Pottery Neolithic (PPN) and Pottery Neolithic. Values from ref. 5. (Bottom) Bishnoi woman in Rajasthan (India) breastfeeding an orphaned gazelle fawn (Nature Picture Library). Image Credit: Reprinted with permission from Axel Gomille/Nature Picture Library (“Çayönü M3,” 01604001 RM). Copyright 2018.

Wang et al. (2) combine new and existing methods to show that the process was much faster in the Yangtze Valley. At Jingtoushan around 8000 y ago and Kuahuqiao shortly after, analysis of dental calculus shows that while some pigs ate a wild-type diet, others ate much human domestic and fecal waste. The latter animals were evidently living actually on the human settlement and must have been habituated to people—they were behaviorally domestic. Metrical analysis of the teeth reveals that they were also biologically domestic: Pearson’s coefficient of variation (V) shows that the pig size range was so large that two pig populations of differing sizes were present, even though there was no absolute size difference between them. The domestic pigs were evidently reproductively isolated from the surrounding wild boar and had already become smaller. Biological and behavioral domestication apparently proceeded in parallel among these, the earliest domestic pigs yet known in China.

This forces us to reconsider pig domestication not just in China but across Asia. An initial acquisition of animals from the wild would not be a problem. Early descriptions of many hunter-gatherer societies reveal that people often brought home infant wild animals, usually after killing their mothers. Many species were involved. In Asia, this included pigs and also bamboo rats, squirrels, monkeys, cats, bears, and others; in Australia, dingoes, wallabies, possums, bandicoots, and cassowaries; in South America, sloths, capybaras, coatis, ocelots, and others; and in North America, various deer species, bison, raccoons, wolves, turkeys, and bears (6–8).

In PNAS, Wang et al. (2) combines new and established methods to overcome many of the problems, and demonstrates close management of pigs in the Lower Yangtze Valley at an unexpectedly early date.

This practice is rarely discussed by archaeologists, perhaps because of the perceived difficulty that hunter-gatherers had no domestic dairy animals to produce the milk for rearing such infants. This is not a problem, however: Women breastfeed these small creatures. This practice seems outlandish to modern western sensibilities but is reported from a very wide range of societies. It is most common for puppies and piglets, but also includes beavers, gazelle fawns (Fig. 1, Bottom), possums, and many others. Underwood, an early 20th-century American woman in Maine, suckled a bear cub named Bruno alongside her own daughter, rearing both to adulthood (8). In Malaya about the same time, ethnographer Paul Schebesta “often saw a little monkey, jealous of the child at the breast, push it away to get the milk, or a mother with a child at one breast and a monkey at the other” (9). Such animals grow up among and are habituated to people, who become fond of them and keep them into adulthood as pets. Many of these animals stay with the people and show no inclination to run away and rejoin their wild conspecifics (8).

Individual infant rearing is a far cry from establishing a domestic population, and archaeologists usually prefer the commensal route for dogs, pigs, and others (3, 4, 10). Some wolf biologists have however recently argued for pup adoption as the origin of domestic dogs (11). It is best if wolf cubs are taken from the wild before their eyes open aged 12 d and breastfed, and are weaned by 11 wk (12). Adult wolves are afraid of humans and stay away from settlements, while the more amenable reared pups stay near the humans. This would achieve reproductive isolation—something the commensal hypothesis cannot explain (11). It is not known how quickly biological changes might occur in the earliest domestic dogs, but among captive silver foxes, selective breeding from the most amenable animals produced foxes happy to be handled by humans in as little as six generations and some biological changes in 15 to 20 generations (13).

Something analogous could lead to early pig populations both behaviorally and biologically domestic, of the kind identified by Wang et al. (2). Wolf domestication involves relocating their ranges—including birthing dens, rendezvous sites, feeding areas, etc.—onto and around the human settlement (11). Wild boar have similar home ranges, with birthing dens, feeding and defecation points, scratching sites, etc. (14), so domestication would involve this being similarly “mapped on” to the human settlement. Like wolves, wild boar are afraid of humans, while the released offspring of pen-reared wild boar stay close to home and tend not to spread far (15, 16). Reproductive isolation of the domestic pig population could thus easily be established, something not possible for a commensal population.

Wang et al. (2) identify a further crucial factor, namely settlement sedentism. Piglets (and wolf cubs) are helpless for a few weeks after birth and remain in and around the birthing den during this time (11, 14). Campsite moves in the middle of this would destroy the human–animal relationship. Dogs can and do move with nomadic people, but there are few examples of pigs doing so. Year-round settlement occupation would be optimal. This applies not just to Jingtoushan and Kuahuqiao but also to sites in western Asia such as Çayönü (Fig. 1, Top) and Abu Hureyra in Syria, a sedentary site (17) with the earliest evidence for sheep tending (18). Selection for animals amenable to close human contact (i.e., behavioral domestication) would be intense in such situations, and the little evidence available (see above) suggests that morphological change (i.e., biological domestication) could also occur rapidly. Behavioral change may be deep rooted and long-lasting. In Norway, descendants of domestic reindeer were released into the wild in 1964 to supplement declining wild herds, and it was assumed they would rapidly revert to wild behavior. But even now, vigilance and flight response among the descendants of the released domesticates shows that multiple generations later, these animals remain more tolerant of human presence even though they are extensively hunted (19).

Wang et al. (2) have thus substantially raised the bar for future work on pig domestication. Their paper shows that the earliest domestic pigs so far known in China were not just behaviorally but also biologically domestic, so the domestic population was already reproductively isolated from its wild conspecifics. This shows the value of Pearson’s coefficient of variation (V) for detecting the presence of two reproductively isolated populations with overlapping sizes (20). This method is all too rarely applied to western Asian samples. When V is calculated for the pigs from Çayönü in Turkey (see Fig. 1, Top for the values), it shows that there was increasing divergence of two reproductively isolated populations. Therefore, there was no long, slow domestication process stretching over millennia. The domestication event must have occurred rapidly, perhaps in Pre-Pottery Neolithic two times. What was long and slow was the subsequent size reduction of the domestic pig population. Calculus analysis of the kind pioneered by Wang et al. would also be most revealing if applied to Çayönü and other sites in western Asia. The paper by Wang et al. (2) thus has major reverberations for pig domestication all across the Old World.