A MEASURE OF LOSS: PREHISPANIC AND BREEDS

A MEASURE OF LOSS: PREHISPANIC LLAMA AND ALPACA BREEDS

RAZAS PREHISPANICAS DE Y ; LA MEDIDA DE LO QUE SE HA PERDIDO

Wheeler*, J.C., A.J.F. Russel* and H.F. Stanley**

* The Macaulay Land Use Research Institute. Hartwood Research Sta. Sholls. Lanarkshire ML7 AJY, U.K. (Present address Faculty of Veterinary Medicine. San Marcos University. Apartado 5137, Lima 3, Perú) ** Conservation Biology Group, Institute of Zoology, Regents Park, London NW1 4RY, U.K.

Additional Keywords Palabras clave adicionales Domestic South American Camelidae. Fiber Camélidos domésticos sudamericanos. Produc- production. Perú. Natural mummification. DNA. ción de fibras. Perú. Momificación natural. DNA.

SUMMARY

The discovery of 900-1000 year old naturally huacaya fleeces are 32.3 ± 2.5 µ (Carpio, 1991), mummified llamas and alpacas at the archaeologi- but the range of variation throughout the cal site of El Yaral (, Perú) has permitted is undoubtedly greater. Degeneration in both lla- detailed study of native andean breeds prior to ma and alpaca fiber quality can be explained as a european contact. The analysis of skin and fiber result of hybridization due to a breakdown in samples provides evidence of the previous controlled breeding, a not improbable scenario existence of two llama and two alpaca breeds. amid the chaos and destruction of the conquest. The llamas included a fine fiber breed (fleece Sixteenth and seventeenth century spanish sour- mean 22.4 ± 2.3 µ) which is apparently extinct, ces document the virtual disappearance of llama and a coarse fiber breed (fleece mean 34.8 ± 7.3 and alpaca herds, as well as up to 80 p.c. of the µ) which is similar to contemporary animals but human population, within little more than a cen- with a more uniform fleece. Among the alpacas tury following the conquest of Cuzco in 1532. extra fine fiber (fleece mean 18.0 ± 1.1µ) and a Analysis of DNA from the prehispanic specimens fine fiber (fleece mean 23.6 ± 1.6 µ) breeds were will provide baseline data for evaluating the ex- distinguished. The prehispanic specimens exhibit tent of hybridization in present llamas and alpacas. a uniformity of fiber color, distribution and fine- ness characteristic of controlled systematic bree- ding which is absent in contemporary animals. RESUMEN Published statistics for contemporary peruvian llama fleeces vary from 27.0 ± 15.6 µ for woolly El descubrimiento de llamas y alpacas natural- ccaras to 35.3. ± 17.8 µ for nonwooly chakus mente momificadas hace 900 a 1000 años en el (Vidal, 1967); while long wool suri alpacas average sitio arqueológico de El Yaral, Moquegua, Perú, 26.8 ± 6.0 µ (Von Bergen, 1963) abd short wool ha permitido por primera vez la descripción de

Archivos de Zootecnia,Arch. Zootec. vol. 41, 41núm. (extra): 154 (extra), 467-475. p. 4671992. WHEELER et al. sus razas autóctonas prehispánicas. Mediante an environmental crisis of unpreceden- análisis de muestras de piel y fibra, se han obteni- ted magnitude. With the introduction do evidencias de la preexistencia de dos razas de of Old World cultigens, animals, llamas y dos de alpacas. Existía una raza de llamas, farming practices, and legislation the aparentemente ya extinta, de fibra fina sin pelos landscape was irreversibly altered (promedio del vellón 22,4 ± 2,3 µ; y otra de fibra and native genetic resources impove- gruesa (promedio del vellón 34,8 ± 7,3 µ) seme- rished (Crosby, 1972). In the Andes, jante a los animales de hoy pero más uniforme. previously unknown diseases reduced Había alpacas con fibra extrafina (promedio del the human population by an estimated vellón 18,0 ± 1,1 µ) y otras de fibra fina (prome- 80 percent leading to economic and dio del vellón 2,6. ± 1,6 µ). Las llamas y alpacas social disintegration of native societies antiguas fueron seleccionadas intensivamente, (Wachtel, 1977). Native llama and pues se caracterizan por uniformidad de finura, alpaca herds also virtually disappeared conformación del vellón y color, mientras que las within little more than a century fo- actuales se caracterizan por su gran variación en llowing the conquest of Cuzco in 1532 el grosor, pilosidad y color de su fibra, producto (Flores Ochoa, 1977 and 1982). Euro- del cruzamiento no controlado. En las llamas pean livestock (sheep, goats, cattle, peruanas actuales el diámetro promedio del and pigs) rapidly displaced llamas vellón varía desde 27,0 ± 15,6 µ para las ccaras and alpacas from their coastal and lanudas hasta 35,3 ± 17,8 µ para los chakus pela- highland valley range, and pushed dos (Vidal 1967); mientras que en las alpacas las them into marginal high elevation cifras van de 26,8 ± 6,0 µ para la suri (Von Bergen lands where the newly introduced 1963) y 32,3 ± 2,5 µ para la huacaya (Carpio animals could not survive. Early tax 1991). Sin embargo, es probable que la variación records, laboriously kept by the spa- a través de los Andes sea aún mayor que lo re- niards, illustrate the extent of this portado. El deterioro de la calidad de fibra de las disaster. In 1567, Garci Diez de San actuales llamas y alpacas andinas se puede expli- Miguel, inspector for the spanish car como producto de la hibridación entre am- crown reported a single, privately bas, hecho probablemente ocurrido durante el owned herd of 50000 llamas and alpa- caos y la destrucción que acompañó la conquista cas in the province of Chucuito on the española. Documentos de los siglos XVI y XVII western shore of Lake Titicaca. Five registran la virtual desaparición de la llamas y years later, in 1572, Pedro Gutiérrez alpacas, junto con 80 p.c. de la población humana, Flores counted only 160000 head in en menos de un siglo después de la conquista de the entire province, reflecting the de- Cuzco en 1532. El análisis del ADN de animales vastation of native livestock in what prehispánicos permitirá medir los cambios gené- was undeniably the most important ticos producidos por la conquista y la incidencia center of llama and alpaca production de hibridación en las llamas y alpacas de hoy. in the Andes. The consequences of this catastrophe on contemporary ca- melid production are rarely considered INTRODUCTION and poorly understood. The lack of written records from The European settlement and colo- preconquest times, and the loss of nization of the New World unleashed orally transmitted specialist knowledge

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 468 A MEASURE OF LOSS: PREHISPANIC LLAMA AND ALPACA BREEDS about llama and alpaca rearing makes to be uniform across the body. Today, it specially difficult to evaluate the all llamas and 80 p.c. of alpacas in the extent to which ideas of european Andes are under the control of origin about stockrearing have in- traditional pastoralists (Novoa, 1981), fluenced management practices. Now, who do not selectively breed for llamas are primarily bred for use as specific phenotypes. The european pack animals and three different concept of breeds, with herd books phenotypes are known, although it is and written records is not part of likely that more exist. Nearly all An- native andean stockrearing practice. dean llamas are of the q’ara or non- wolly type, characterized by lack of fiber on the face and relatively sparse PRECONQUEST LLAMA AND fiber growth. Less common is the ALPACA PRODUCTION ch’aku or woolly llama, which as its name implies, has a heavier fleece Sixteenth century documents indi- with fiber growing on its forehead cate that llama and alpaca production and in its ears. The fiber characteristics and breeding was rigidly controlled of the third variety are intermediate by the Inca government (Murra, 1965). between the q’ara and ch’aku llamas. State herds supplied pack llamas for Color tends to be patchy and varies the royal armies and alpaca wool for from white, to brown, black or grey. textile production. Shrine herds, divi- The fiber is generally coarse and ded according to color, provided sa- hairy, and the consensus of opinion is crificial animals for state religious that the llama is not now and never observances. A hereditary cast of has been bred for fiber production. In State herders, the yana, was apparently contrast, alpacas are kept specifically emerging prior to the conquest. fiber production, and two phenotypes Detailed herd records were kept using can be distinguished based on fiber the quipu, a mnemonic device. and characteristics. Now, approximately textile production and redistribution 90 p.c. of alpacas exhibit the shorter, was largely State controlled. Offerings crimped fibers of the huacaya variety, or sacrificial burning of cloth accom- while 10 p.c. have the long wavy panied all social, religious, military fibers of the Suri type (Wheeler, and political events. Communally and 1991). In appearance, the huacaya is individually owned herds also existed. reminiscent of Corriedale, and the The fundamental importance of suri of Lincoln breeds of sheep. Ani- llama and alpaca production to the mals with intermediate fiber types Inca economy is seen in their repeated exist, but are rare. The genetic factors attempts to conquer the circum Titica- which control these characteristics ca kingdoms to the south. The wealth are not known, and the 9/1 huacaya/ of herds in this region descendend suri ratio is a product of random from the earlier tiwanaku empire, breeding. Fiber color varies from and was unequalled elsewhere in the white to black and brown, including Andes. In contrast with the Inca, ho- all the intermediate shades, but tends wever, details concerning Tiwanaku

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 469 WHEELER et al. and earlier stockrearing practices are origin, and a direct result of their limited because there are no written displacement by spanish introduced records. Such information must be livestock. To the best of our reconstructed from analysis of the knowledge, llama herding began on material remains recovered in the coast approximately 1400 years archaeological sites. ago (Pozorski 1979, Shimada and The recent discovery of well Shimada, 1985) but the first clear preserved 900 to 1000 year old natu- evidence of alpaca breeding in this rally mummified llamas and alpacas zone comes from the site of El Yaral at El Yaral, in the Moquegua Valley with and antiquity of 900 to 1000 of southern Perú, has provided a years. Alpacas may also have been unique opportunity to examine the present on the north coast of Perú at physical appearance of these animals this same time (Shimada and Shimada, prior to European contact. Excavations 1985). Both llamas and alpacas were at this site began in late 1986 under originally brought under human con- the direction of Don Rice (Southern trol as domestic animals in the high Illinois University) and Geoff Conrad andean puna at least 6.000 years ago (University of Indiana) with support (Wheeler, 1984 and 1986; Wing, 1986), from the United States National and the commonly held belief that Science Foundation (grant BNS85- they cannot survive at low elevation 10877), and more recently research is a heritage of the spanish conquest. on biocultural remains from El Yaral El Yaral is located on the west has been undertaken by Jane Buikstra bank of the Osmore River, (University of Chicago) (BNS89- approximatley 20 km south of the city 20769). See García Márquez (1988), of Moquegua, Perú. The area is Rice, Stanish and Scarr (1989) and extemely dry, with condensation from Watanabe, Moseley and Cabieses winter fog supplying most of the (1990), for details about the annual 15 to 20 mm precipitation total excavations at El Yaral and other sites in during the months from May through the Moquegua valley. August. Winter temperatures average Located just 50 km from the 8oC, but during the dry, sunny summer Pacific coast, at an elevation of only months from January through March 1000 meters above sea level, in one of may reach 28oC. Agriculture cannot the worlds driest deserts, El Yaral be practiced here without irrigation may seem an unlikely place to find and fields are largely confined to the llama and alpaca remains. None the river bed. Just to the south of El Yaral less, evidence from animal bones the Osmore River disappears into recovered during archaeological subterranean channels, emerging 20 excavations at sites in Perú, and km downstream towards the coast. Ecuador clearly indicates that the The Chiribaya settlement of El present restricted distribution of Yaral began aprroximately 1000 years domestic camelids to areas of extreme ago and lasted less than 400 years. elevation in the Andes is of recent Three radiocarbon dates have been

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 470 A MEASURE OF LOSS: PREHISPANIC LLAMA AND ALPACA BREEDS obtained from the excavations: 970 + subfloor alpaca and llama burials. 50 B.P. (Beta 22845), 960 + 80 B.P. These animals had been ritually (Beta 19323) and 790 + 50 B.P. (Beta sacrificed by a blow between the ears 22846). Ceramic remains indicate that which fractured the skull, then imme- Chiribaya developed near the present diately buried in shallow pits and co- port of Ilo, out of the earlier (A.D. vered with sand. Due to the extreme 600) tiwanaku occupations in the aridity of the environment, this process Osmore drainage (Bermann et al. resulted in the desiccation and preser- 1989, Stanish 1989). These settlers vation of these animals as natural from the Lake Titicaca basin brought . Other offerings, indluding highland traditions and livestock to cuyes (guinea pigs), coca leaves, sticke the coast, forming the basis of what wrapped in thread and camelid fiber, evolved into the Chiribaya culture spondylous (a marine shell imported following the breakup of the Tiwanaku from Ecuador) and turquoise beads, empire. Since there is no evidence small pieces of beaten silver, feathere, that Chiribaya was in contact with fish bones, marine shells and corn other contemporary highland groups, were found with the alpacas and the alpacas and llamas of El Yaral can llamas. be considered descendants of the original Tiwanaku stock. This is particularly important because it means PRECONQUEST LLAMA AND that not only do these animals predate ALPACA BREEDS the , but also that they came from what both the inca and The exceptional state of spanish conquerors considered to be preservation of the El Yaral alpaca the heartland of alpaca and llama and llama mummies has made possible production. the first systematic analysis of the The site of El Yaral covers appro- fiber characteristics of preconquest ximately 12.3 hectares. It consists of animals. In all, we have been able to a series of steep, elongated terraces, study skin and fiber samples taken at extending approximately 300 meters 12 different locations across the body along the river bank, which were (6 along the midline from neck to tail, built to provide a base for architectural 4 at mid-rid from fore to hind limb, contruction. The remains of cane and 2 midway down the fore and hind wall, earthen floor buildings of varying limbs) from 4 alpaca and 6 llama size were found strung out along the mummies. Special attention has been terraces. Although the number of placed on the analysis of fiber diameter rooms varied from structure to struc- in order to obtain information about ture, living, cooking, storage and chi- fleece composition, but microstructure cha (corn beer) preparation areas studies utilizing the scanning electron could be identified within the houses. microscope and skin histology have Excavations in two of the larger, also been undertaken. A total of 200 more complex buildings revealied fibers (the textile industry standard)

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 471 WHEELER et al. were examined from earch sample fleeces lack uniformity, ranging from utilizing a projection microscope, 32.5 + 17.9 µ (o) to 35.5 + 17.8 µ (o) IWTO standards and the MLURI for coarse nonwoolly ccaras, 30.5 + GWBASIC FIBRE2 program. This 18.5 µ (o) to 30.5 + 17.9 µ (o) for inter- data has permitted detailed evaluation mediates, and 27.0 + 15.6 µ (o) to 29.1 of fiber tineness and uniformity in + 12.7 µ (o) for woolly chakus (Vidal the mummies, and provided the first 1967). The variability of present day direct information on the fleece cha- llama fiber is related to an increase in racteristics of preconquest llamas and hairs and general coarsening of the alpacas. fleece, which probably began at the Quite unexpectedly, five of the six time of the spanish conquest. Increased El Yaral llama mummies studied were hairiness is produced by lack of found to belong to an apparently controlled breeding, and crossing extinct fine fiber breed. The quality between the two prespanish llama of fiber from these animals is breeds from El Yaral could account impressive. Based on mean fiber for the entire range of fleece variation diameter at 8 sample sites per animal, observed in todays animals. their fleeces average 22.4 µ with a As with the llamas, two probable between sample standard deviation of breeds of preconquest alpacas were 2.3 µ, compared to fleece measu- found among the El Yaral mummies. rements of 27.0 + 15.6 µ for female Based on mean fiber diameter at 8 and 29.1 + 12.7 µ for male woolly cha- sample sites per animal, fine fiber ku llamas in the Andes today (Vidal and extra fine fiber alpacas were dis- 1967). The small range of variation in tinguished. The former have fleeces fiber diameter of the ancient llamas averaging 23.6 µ with a between sam- reflects selection for a uniform, single- ple standard deviation of 1.6 µ, while coat fleece through reduction in the the latter fleeces average 18.0 µ with size difference between primary hairs a between sample standard deviation and secondary undercoat fibers. The of 1.1 µ. Both groups exhibit uniform fleece of the sixth llama was single-coats. In comparison, contem- coarse, averaging 34.8 µ with a standard porary andean Huacaya and Suri deviation of 7.3 µ between the 8 fleeces, at 31.2 + 3.8 µ (Carpio, 1991) sample sites. Although only one and 26.8 + 6.0 µ (Von Bergen, 1963) specimen of this class was included in respectively, are coarser, may have a our sample, this animal clearly tendency to hairiness, and are of represents a second breed of llama. uneven quality. Some coats containing The variation in diameter indicates a up to 40% hair have been reported for tendency to a double-coat with coarser both living varieties, and considera- hairs and a finer undercoat, but this ble variation is reported in published is much less pronounced than in statistics on fiber diameter. The origin present day llamas. of this degene-ration almost certainly In comparison to the preconquest lies in the spanish conquest, but a llamas, contemporary Peruvian llama breakdown in controlled breeding

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 472 A MEASURE OF LOSS: PREHISPANIC LLAMA AND ALPACA BREEDS between the fine and extra fine El as the animals themselves have been Yaral breeds would alone account for modified by hydridization. We have the variation observed today. known for a long time that the pri- The most probable cause of mary reason why llamas and alpacas coarsening and hairiness in both are said to have a low reproductive huacayas and suris would be through success rate is because native andean hybridization with the coarse fiber herders tend to raise them as if they llama breed, a not improbable scenario were sheep not because they are in- amid the chaos and destruction of the herently infertile or maladapted to conquest. Clearly, however, such a the high andean environment as has process would not have affected only been claimed. We are now beginning the alpaca gene pool. The El Yaral to learn from the El Yaral mummies mummies indicate the possibility that what the real potential of these species extensive crossbreeding between as fiber producers is. Unfortunately, alpacas and llamas may have occurred too little is known about alpaca and during the sixteenth century and has llama variation in the Andes to even played a much more important role in determing if the preconquest variants the formation of todays livestock still exist in the general population than has been realized. and can be rescued by careful breeding. The absolute urgency of identifying DISCUSSION and preserving relict populations before any further genetic loss or The truly unexpected results modification occurs cannot be empha- obtained from fiber analysis of the El sized strongly enough. Genetic mar- Yaral mummies indicate both how ckers must be identified which can be much has been lost, and how much used to distinguish breeds and hybrids, remains to be learned about alpacas and research underway on both and llamas. Many of the confused mitochondiral and nuclear DNA ideas and contradictory statements sequences from all four south american about these animals coming from camelids is producing promising both native andean herders and north results (Stanley and Wheeler 1992). american breeders, to say nothing of The inclusion of ancient DNA from camelid specialists, are more a heritage the El Yaral llamas and alpacas will be of the spanish conquest than we have of crucial importance in documenting realized. The crux of the problem lies the historical development of todays in the lack of written records in pre- llamas and alpacas. Because they hispanic andean society, and the loss include apparently extinct breeds, of orally transmitted specialist infor- the mummies may provide baseline mation about breeds and breeding data against which to measure the within the first century after european probable effects of hybridization and contact. What has survived is probably reduced genetic diversity which are as distorted by poorly understood thought to have accompanied the european-origin ideas and traditons Spanish conquest.

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 473 WHEELER et al.

REFERENCES

Bermann, M., P. Goldstein, Ch. Stanish and Novoa, C. 1981. La conservation de especies L. Watanabe. 1989. The Collapse of the nativas en América Latina. In: Animal Genetic Tiwanaku State: A View from the Osmore Resources Conservation and Management, Drainage. In: D.S. Rice, C. Stanish and P.R. pp. 349-63. FAO Animal Production and Scarr, Eds. Ecology, Settlement and History Health Paper No. 24, FAO, Rome. in the Osmore Drainage, Perú. Oxford, BAR International Series 545, pp. 269-285. Pozorski, S. 1979. Late Prehistoric Llama Remains from the Moche Valley, . Annals Carpio, M. 1991. La fibra de camélidos. In: C. of Carnegie Museum 48: 139-170. Novoa and A. Florez, Eds. Producción de rumiantes menores: Alpacas, Lima, Resu- Rice, D.S., C. Stanish and P.R. Scarr, Eds. men, pages 295-359. Ecology, Settlement and History in the Osmore Drainage, Peru. Oxford, BAR Crosby, A.W. 1972. The Columbiam Exchange International Series 545. Biological and Cultural Consequences of 1492. Westport, Greenwood Press. Shimada, M. y I. Shimada. 1985. Prehistoric Llama Breeding and Herding on the North Flores Ochoa, J.A. 1977. Pastores de Alpacas Coast of Peru. American Antiquity, 50: 3-26. de los Andes. In: J.A. Flores Ochoa, Ed. Pastores de Puna, pp. 15-52. Lima, Instituto Stanish, C. 1989. An Archaeological Evaluation de Estudios Peruanos. of an Ethnohistorical Model in Maquegua. In: D.S. Rice, C. Stanish and P.R. Scarr, Eds. Flores Ochoa, J.A. 1982. Causas que origina- Ecology, Settlement and History in the ron la actual distribución espacial de las Osmore Drainage, Peru. Oxford, BAR alpacas y llamas. In: L. Millones and H. Tomoe- International Series 545, pp. 303-320. da, Ed. El hombre y su ambiente en los Andes Centrales. Osaka, Nat. Museum of Ethnology, Stanley, H.F. and J.C. Wheeler. 1992. Molecular Senri Ethnnological Studies 10: 63-92. Evolution and Genetic Diversity of the Camelidae. International Conference on García Márquez, Manuel Enrique. 1988. Molecular Evolution, University Park, Excavaciones en dos viviendas Chiribaya en Pennsylvania. el Yaral, Valle de Moquequa. Tesis para optar el grado de Bachiller en Ciencias Histórico Vidal, O. 1967. La crianza de la llama y algunas Arqueológicos, Universidad Católica Santa características de su fibra. Tesis, Universidad María, Arequipa, Perú. Nacional Agraria La Molina, Lima, Perú.

Murra, J.V, 1965. Herds and Herders in the Inca Von Bergen, W., Ed. 1963. Wool Handbook. State. In: A. Leeds y A.P. Vayda, Eds. Man, New York, Interscience. Culture and Animals, pp. 185-215. Washing- ton, D.C. Publication 78, American Wachtel, N. 1977. The Vision of the Vanquished. Association for the Advancement of Science. New York, Barnes and Noble.

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 474 A MEASURE OF LOSS: PREHISPANIC LLAMA AND ALPACA BREEDS

Watanabe, L.K., J.E. Moseley and F. Cabieses, Prehistoriques des Andes 1. Paris, Editions Eds. 1990. Trabajos Arqueológicos en Recherces sur les Civilisations, ADPF, pp. Moquequa, Perú. 3 vols. Lima, Imprenta Es- 21-59. cuela Nueva. Wheeler, J.C, 1991. Origen, Evolución y Status Wheeler, J.C. 1984. On the Origin and Early Actual. In: S. Fernández-Baca, Ed. Avances y Development of Camelid Pastoralism in the Perspectivas del Conocimiento de los Andes. In: J. Clutton-Brock y C. Grigson, Camelidos Sub Americanos. Santiago, FAO Eds. Animals and Archaeology, Vol. 3, Early Oficina Regional de la FAO para América Herders and Their Flocks. Oxford, BAR Latina y el Caribe, pp. 11-48. International Series 202, pp. 395-410. Wing, E.S. 1986. Domestication of Andean Wheeler, J.C. 1986. De la Chasse a l’Elevage. En: Mammals. In: F. Vuilleumier and M. Monaste- D. Lavallee, M. Julien, J.C. Wheeler and C. rio, Eds. High Altitude Tropical Biogeography. Karlin, Telarmachay Chasseurs et Pasteurs Oxford, Oxford Univ. Press, pp. 246-264.

Archivos de Zootecnia, vol. 41, núm. 154 (extra), p. 475