Pre-Conquest Alpaca and
Llama Breeding

By Jane C. Wheeler, Ph.D.

Based primarily on the study of fragmented bone remains recovered during archaeological excavations, the aforementioned reports provide no information on physical appearance of the domestic alpacas and llamas. Textile remnants found at sites in the Atacama desert of northern Chile record the appearance of black fiber from presumably domestic animals around 3,000 years ago, while the earliest known (but certainly not the oldest) use of camelid fiber in Peruvian weaving comes from fabrics preserved at coastal sites dated approximately 2,500-700 BP.

In northwestern Argentina, llama fiber cordage has been dated to 1450 BP. However, it is only with the discovery of 700-1300 year old naturally desiccated llama and alpaca mummies from
southern Peru that information on the physical appearance of ancient animals has become available. The discovery of 26 perfectly preserved, naturally desiccated alpaca and llama mummies at the pre-Inca (A.D. 700-1300) Chiribaya culture site of El Yaral, provides a first ever view of preconquest animals. Located in the extremely arid coastal desert of southern Peru (17°01’ south latitude, 71°00’ west longitude), the site lies 50 km inland at an elevation of 1000 meters above sea level. Covering approximately 12.3 hectares, and containing more than 330 elongated residential terraces, El Yaral rises 120 meters up the barren hillside at the extreme south-western end of the Moquegua valley. It overlooks a natural corral, approximately one square kilometer in extent, which is enclosed by sand covered hills and watered by the Osmore River (Figure 1 at left). Extensive river bottom agricultural land lies immediately to the north of the site.

The El Yaral structures were built of cane walls with wooden support posts for cane mat roofs, and subdivided by rows of vertical canes into anywhere from two to eight rooms which were used for storage, cooking, chicha (a fermented maize beverage) preparation, sleeping and ritual activities. Compacted floors of fine gravel, sand and clay covered offerings of guinea pigs, coca leaves, thread wrapped sticks, turquoise and marine shell beads, small silver plaques, feathers, fish, maize and burnt charcoal, as well as sacrificial alpacas and llamas.

These animals had been killed by a massive blow between the ears and rapidly interred. The sand in which they were buried and the extreme aridity of the climate produced exceptionally well preserved specimens. Contemporary llamas and alpacas are classified on the basis of fiber characteristics and physical appearance. Recognized llama phenotypes include the hairy, coarse, sparsely fibered ccara; the coarse, more densely fibered chaku; the long, wavy fiber suri and animals with an intermediate fiber type. Most alpacas exhibit dense, crimped huacaya fleeces, but a small number ( <10%) are distinguished by long, wavy suri fiber. Intermediate alpaca fleece types also exist but are not a recognized phenotype.

Virtually all llamas, and 75% of all alpacas, are held by traditional herders who fail to consistently breed for selected phenotypes and both inbreeding and hybridization between llamas and alpacas are acknowledged problems. Nonetheless, elaborate classification systems based on color and conformation characteristics exist among Quechua and Aymara herders, suggesting that earlier management strategies selectively bred for fiber characteristics in both alpacas and llamas. Although written records were not part of Andean civilization, detailed data on size and color of flocks was kept utilizing the quipu, a memory aid made of knotted camelid fiber strands. Under Inca rule, an annual census was taken of the state and shrine herds.

Special emphasis was placed on breeding pure brown, black and white animals for sacrifice to specific deities, as well as on quality fiber production for the state controlled textile industry and the production of sturdy pack llamas for the Inca army. Given such rigorous demands, it is likely that specific llama and alpaca breeds were maintained which subsequently disappeared during the Spanish conquest. The data from El Yaral, presented here, suggests that the origin of such breeds may predate the Inca Empire.

Pre-conquest breeds

The El Yaral llamas and alpacas were identified on the basis of phenotypic attributes (conformation, fiber distribution) and confirmed by incisor morphology. The age and sex of each specimen was determined, when possible, on the basis of dental eruption and preserved genitalia. Observations on fleece type and color were recorded, and evaluation of the fiber producing qualities of 6 llamas and 4 alpacas from the site was carried out. Samples of skin with attached fiber (1.5x1.5cm) were taken when possible, at 11 standardized sites on the left side of the animals: on the lower neck (A2), at equidistant points from the fore to hind legs at the level of the sacral tuberosity (A3-A6) and mid-way between the dorsal and ventral surfaces (B3-B6), as well as half-way down the fore (C3) and hind (C6) legs.

Fiber length was recorded, and 1mm segments of 300 to 500 fibers were cut from next to the skin of each sample and diameter measurements were recorded for 200 fibers per sample. Experimental desiccation of fresh llama skin samples was carried out under varying conditions, and statistical analysis of the results, revealed no significant alteration in fiber diameter during these processes. Fiber diameter measurements from the El Yaral alpacas and llamas (Figure 2) revealed the existence of four distinct groups of animals (Figure 3, next page).

The raw data reflect natural variation in fiber diameter across the body, with coarser/hairier samples coming from the neck (site A2) and legs (sites C3, C6). The finest fiber is located along the back (sites A3-A6), with a tendency towards a gradual increase in diameter at mid-rib height (sites B3-6). Taken together, the eight samples from sites A3-6 and B3-6, correspond to the fleece, and represent that portion of the fiber which is shorn for use in textile manufacture. Based on complete (not dehaired) samples (n=8 per animal), average fleece diameters for the two alpaca groups were found to be 17.9 (sd ± 1. 0µm) and 23.6 (± 1.9µm). Llama fleeces likewise included a fine fiber group at 22.2± 1.8µm, as well as a coarse fiber animal at 32.7 ± 4.2µm.

Analyses of variance indicate that the differences between these ancient groups are significant. To the best of our knowledge, the heritability factor of fiber fineness has not been determined for llamas and alpacas, although 0.22, 0.27 and 0.38 have been reported for fleece weight. It has been observed, however, that fiber diameter generally increases with age and number of shearing events. There is also some indication that a rich diet may increase fiber diameter, while finer fiber may be produced by animals kept on a low plane

FIGURE 2. El Yaral alpaca and llama fiber diameter in microns, by sample site, 200 fiber count.
A2-A6: adjacent to midline from neck to tail, B3-B6: at midrib from shoulder to hip,
C3: lower front; and C6: lower rear legs, - n.s. = no sample

of nutrition, and during periods of physiological stress. In the Peruvian Andes, it is generally believed that a relationship exists between elevation and alpaca fiber fineness. As altitude increases, pasture quality decreases and environmental stressors become more severe, but the
relationship of these factors to fiber diameter remains to be demonstrated.

Nevertheless, the presumed correlation between fine fiber and high altitude is often cited as evidence that alpacas can be reared successfully only in this habitat. In Peru, fiber diameter values reported for contemporary Andean llamas and alpacas vary greatly. This is due, in part, to
differences in the samples studied, measurement techniques and reporting.

It is not always clear, for example, if published figures come from complete or dehaired
fleeces, and extreme care must be taken to select comparable data sets. Even with this caveat, the range of values reported for alpaca and llama fleeces is so great that it would be misleading to give a single value for each. Alpaca fibers have been reported from 9 to 88µm, and llama fibers from 8 to 144µm, indicating a variable but significant presence of coarse guard hair in both groups. In comparing the El Yaral mummies with contemporary animals, we have chosen age matched samples insofar as possible (Figure 3).

The fiber diameters of the ancient alpaca specimens were found to be significantly finer relative to today’s animals. The Chiribaya extra fine fiber alpaca group (named after the culture to which El Yaral pertained) measured from 10-14µm less than huacaya alpacas of the same age, while the Chiribaya fine fiber alpaca group registered 4.7-8.5µm less. Comparable age and sample data on Peruvian suri fleeces is not available. It is generally considered that the wavy suri fiber is finer than the crimped huacaya fiber, but insufficient information is available to confirm this. Among the mummified alpacas, however, the finest fiber came from animal 119, a 24 month old male with crimped fiber (Figure 4), while the other fleeces were wavy (animals 317 and 228 (Figure 5) or contained both crimped

Figure 3. Mean fleece diameter of contemporary and prehispanic South American camelids.
All figures represent complete fleece counts except where noted (u = undercoat, h = hair).
El Yaral fleece data represent the mean of 8 sample locations.

Figure 4. (above) Alpaca 119 from structure 70 at El Yaral.

Figure 5. (below left) Alpaca 228 from structure 70 at El Yaral.

Figure 6. (below right) Alpaca 314 from structure 193 at El Yaral.

and wavy fibers (animal 314; Figure 6). Fiber diameter measurements of the six El Yaral llamas revealed the presence of fine and coarse fiber animals. Chiribaya fine fiber llama fleeces (n=5) averaged 22.2 ± 1.8µm. Comparative figures for 12 months old males from Puno, Peru, vary from 18.8µm undercoat/39.8µm hair in the more heavily fibered chakus to 20.1µm
undercoat/73.1µm hair in ccaras, increasing to 22.0µm undercoat/42.2µm hair in chakus and 25.2µm undercoat/77.7µm hair in ccaras at 24 months (Figure 3). Although these figures cannot be directly compared with the complete fiber counts on the El Yaral specimens, the nondehaired ancient fleeces are as fine as the dehaired modern ones.

Abundant coarse hairs were present in all the contemporary samples, while in contrast, only one of the five ancient llamas (237) had visible fine hair in the fleece. Two others had very fine hair (248, 231) and two were single coated with no perceptible hair in the fleece (243 Figure 7, 237). In contrast, the sixth El Yaral llama (247), a 3 month old male, had a coarse hairy coat with an almost equally coarse undercoat. At 32.7 ± 4.2 µm, the fiber diameter of this animal is greater than that reported in contemporary llamas (Figure 3).

Based on fiber diameter, it would appear that 3 possible fiber producing breeds of domestic camelids may have existed at El Yaral. These include the Chiribaya extra fine and fine fiber alpacas and the Chiribaya fine fiber llama. Although alpacas continue to be raised for fiber up to
the present, llama fiber is generally deemed too coarse and hairy for textile production,
and these animals are reared primarily for use as pack animals and for meat. Early Spanish writers were impressed by the large pack trains which accompanied the Inca army, and llamas were valued as beasts of burden during the colonial period. But, not all llamas are reared for cargo, and in some communities they are selectively bred for fiber production. The Chiribaya fine
fiber llamas suggest that this practice, which has virtually disappeared, may have ancient roots. Regardless, because so many environmental factors can affect it, fiber diameter data alone is insufficient for affirming the existence of breeds.

Evidence that selective breeding was practiced at El Yaral is seen in a uniformity of fleece characteristics which are lacking in contemporary Andean llamas and alpacas. In the ancient specimens, variation in fiber diameters across the fleece is minimal, and hairy patches are rare even in the coarse fiber llama, while in living animals fiber diameter tends to vary greatly and hairiness is a problem in both alpacas and llamas.

Even though the alpaca is considered to have a single coat comprised entirely of undercoat fiber, animals with up to 40% hair in the fleece are known today. Additionally, the existence of single coat llamas with uniformly fine fiber at El Yaral,provides evidence of a variety or
breed which is no longer recognized. Another indication of selective breeding for fiber production is seen in the uniform coloration of the mummies. Only two of the El Yaral specimens are multicolored, the coarse fiber llama male which was eliminated from the gene pool at 3
months, and a brown and white alpaca (317). The possibility that the coarse fiber llama represents a non fiber producing (possibly cargo bearing) breed should be

Figure 7. Llama 237 from structure 70 at El Yaral.

Figure 8. Looted llama and alpaca remains at the site
of Chiribaya Alta, Ilo.

taken into account. The Chiribaya fine fiber llamas included one pure white, two red brown, and two grey/beige animals. Under Inca rule, animals of pure color were required for sacrifice to particular deities and the shrine herds bred for this demand. White llamas were sacrificed to
the sun, red brown animals to Viracocha at the beginning of the agricultural year, and black animals were starved and sacrificed in times of crisis. Whether or not the llamas at El Yaral were sacrificed to the same deities is unknown, but with one exception, the requirement for pure
color high quality animals seems to have been met. The mummified alpacas, included the brown and white animal mentioned above, as well as a white, a brown and a vicuña-colored alpaca.
One final line of evidence suggestive that the Chiribaya fine fiber llama was selectively bred for fiber production is an indication that fiber growth may have been more rapid than in contemporary animals.

In the unshorn or partly shorn specimens (243, 273), a discrepancy was noted in the age predicted on the basis of fiber length and the age of the animal as predicted by dental eruption. Although the data on fiber growth comes from animals raised at high altitude and may not therefore be fully comparable, fiber lengths of 13 and 18 cm for llama 273 would suggest an animal more than 2 years of age, but in fact the animal is >9 (likely 12) months.

The same is true for llama 243, with fiber lengths of 14 and 16 cm, the expected age would be
more than 2 years, but the real age is <18 (likely 15) months. Clearly selection for faster fiber growth would be of economic importance if the goal of breeding is for fiber production. None the less, it should be observed that a good plane of nutrition is also required in order to sustain such
growth. Unfortunately, all four alpacas had been shorn prior to sacrifice, so comparable data is lacking.

Clearly the coastal desert was not a bad place to raise llamas and alpacas: fleece fineness and uniformity, accelerated fiber growth, uniform coloration, and the existence of a single coat llama all point to controlled breeding, reproducible results and the probable existence of breeds. The
skill of the Chiribaya culture herders is recorded in the mummified alpacas and llamas from El Yaral, and further confirmed by the faunal assemblage from Chiribaya Alta, a nearby, contemporary site (Figure 8). Examination of 140 sacrificial llamas from intact burials revealed that they had been selected in accordance with herd management priorities. Only 12.9%
of the llamas killed were of prime breeding age (>24 months), and 3 (or 2.1%) of these
had dental pathologies which would have been cause for sacrifice. Of the remaining 122 animals, 69.6% were culled before first breeding at 21 months; 15.7% were killed after the first and before the end of the second breeding season at 21-42 months, and 2.9% were old adults probably past
reproductive age. Such a pattern would be produced by the elimination of undesirable and infertile animals from the herd which was raised locally under a controlled breeding program. The end product of these efforts were fiber producing alpacas and llamas of unparalleled quality.
In contrast with the preconquest alpacas and llamas from El Yaral, today’s animals are characterized by a lack of uniformity.

Coarsening of the fiber, increased hairiness and the increased variation in fiber diameter across the fleece of huacaya and suri alpacas, as well as the apparent disappearance of fine fiber llamas, can almost certainly be explained by a breakdown in controlled breeding
accompanied by extensive hybridization produced by events of the conquest. Taken together, the four preconquest breeds bracket the range of fiber diameter measurements in today’s animals, and crossing with the hairy Chiribaya coarse fiber llama could have produced this result.

Scientific study of the role of hybridization in the evolution of today’s llamas and alpacas has only recently been undertaken utilizing DNA, and has revealed alarming levels of hybridization. On a trans-Andean level it has been found that only 20% of alpacas have not suffered hybridization
with llamas at some point in their ancestry.

As of yet unpublished data place the incidence at just less than 6% in some regions of Peru. Sadly, hybridization between alpacas and llamas - heritage of the conquest - continues apace making the genetically pure alpaca an endangered species. Although attempts to determine
the genetic purity of the mummies through DNA analysis have not yet proven possible, preliminary study in contemporary alpacas points to a possible relationship between fine fiber and purity. Much research remains to be done, but one thing is clear, we urgently need to
identify and protect the few remaining unhybridized alpacas, and these may well turn out to be the true descendents of the Chiribaya alpacas and llamas.

CQ

Author’s Note:
Much of the information contained in this article comes from: J.C. Wheeler, A.J.F. Russel and H. Redden 1995 Llamas and Alpacas: Pre-conquest Breeds and Post-conquest Hybrids. Journal of Archaeological Science, 22:833-840. 1995. The excavations at El Yaral and the other Chiribaya sites in Moquegua were directed by Jane Buikstra (Arizona State University) and
Don Rice (Southern Illinois University) with support from US National Science Foundation grants BNS85-10877 and BNS89-20769.

Jane C. Wheeler studying 1,300 year old llama mummy from the site
of El Yaral, Moquegua, Peru.

About the Author
Jane Wheeler is Vice President for Research, C O N O P A (Coordinadora de Investigación y Desarrollo de Camélidos Sudamericanos), Lima, Perú, jwheeler@conopa.org. She holds
degrees from American University, Cambridge University, and the University of Michigan, and
completed postdoctoral studies at the University of Paris. For more than 30 years she has conducted broad based research on the South American camelids, covering topics from origin, evolution and domestication of alpacas and llamas, to molecular genetics, breeding and fibre production, as well as vicuña and guanaco genetics and conservation.