The Puquios of Nasca

by Katharina J. Schreiber, Josué Lancho Rojas
The Puquios of Nasca
Katharina J. Schreiber, Josué Lancho Rojas
Latin American Antiquity
Start Page: 
End Page: 
Select license: 
Select License


Katharina J. Schreiber and JosuC Lancho Rojas

The puquios of'Nu.7~~gulleries that provide waterjbr irrigation and domestic uses in

Los puquios de Nuscu son un .si.stenlu de ga1eriu.s .subterruneus de jiltrucion, ubicrrdos en 10s vulles de Nuscu, Turugu, y Las Truncus, en lu costu sur del Peru. Los puyuios proveen uguu puru irrigucidn de los cunzpo.7 de cultivo y para uso dom6.stico en lus porciones de 10.7 vu1le.s que carecm de uguu .superficial. A la fkcha 36 puquios de 10s 50 yue pudieron huher eri.stido en tiempos pusudo.~siguen funcionundo. El presente truhujo describe Ius curucteristicus ji,rmale.s ,v lus t6cnicus constructivus, asi conlo cudu uno de 10s puquios que e.xi.sten actuulnlente. Los resultudos de diversas prospecciones uryueeol(jgicu.sredizadus en 10.7 tres vulles indicun yue 10.7 puquios no e.xi.stiun en el periodo Nuscu Ewzpruno (Periodo Intermedio Temprano 2-4). pero si /uncionuron antes de la conqui.stu de estu regidn por el imperio incaico, hacia fi'nes del sigh XCi Sugerimos yue lu construccicin iniciul y uso de estas ohrus corresponderiu ul periodo Nuscu Turdio (Periodo lntermedio E~mpruno6-?), pudiendo huherse iniciudo en Iu 6pocu Nuscu 5.

n 1853 the young English traveler Clements region by the amount of rainfall and/or the vol-Markham reached the Nasca Valley on his ume of water flowing in the rivers would certainly journey south from Lima. He marvelled at the agree that the Nasca Valley is exceptionally dry, verdant landscape and described it as "the most even by Peruvian coastal standards. Yet the Nasca fertile and beautiful spot on the coast of Peru" Valley supports a substantial modern population (Markham 1991:SO). He found it particularly and supported a perhaps even larger late prehisnotable because the Nasca Valley seemed to be toric population. It was the locus of major Inka one of the driest places he had seen. Not only is and Wari occupations, and was also the core the Nasca region lacking in rainfall, as is typical region in which developed the well-known Nasca of the Peruvian coast, but "all that nature has Culture. How can this be possible? In Markham's given it is a small watercourse, almost always words, dry" (Markham 1991:50). The fertility is due to the skill and industry of

It seems ironic that the portion of the valley the ancient inhabitants. Under their care an arid

wilderness was converted into a smiling par-

with the broadest expanse of arable land, and in

adise, and so it has continued. This was effected

which the modern town was founded by the

by cutting deep trenches along the whole length

Spanish, is also the portion of the valley in which

of the valley and so far up the mountains that the the river is most deficient of water. Anyone accus-present inhabitants do not know the positions of tomed to judging the availability of water in a their origin. High up the valley the main trenches

Katharina J. Schreiber Department of Anthropology, University of California, Santa Barbara, CA 93106 JosuC Lancho Rojas .Calle Callao 771, Nasca, Peni

Latin American Antiquity, 6(3), 1995, pp. 229-254. Copyright ((2by the Society for American Archaeology

Figure 1. Map of the Rio Grande de Nasca drainage of the south coast of Peru.The shaded area indicates the focus of

the present study.

Markham did not understand precisely how the puquios worked, but luckily their functions are

still observable puquios are nearly horizontal trenches or galleries that tap underground water sources and act as conduits to transport the water to the gradually sloping surface. In this paper we discuss the puquios found on the south coast of Peru, in the three southern valleys of the Rio Grande de Nasca drainage (Figure 1): the Nasca Valley proper, the Taruga Valley, and the Las Trancas Valley. The Nasca

puquios have been described by other writers, most notably Gonzalez (1934), Mejia (1939), and P.osse1 (1942, 1977), and most treatments of Nasca at least refer to the puquios. Unfortunately, none of the published works is complete, and some contain serious errors. Our first goal is therefore to provide a more up-to-date and com- plete discussion of the puquios and to correct some of those earlier errors. In addition, the date of construction of the puquios has recently become a subject of some interest and disagree- ment (Barnes and Fleming 1991; Bray 1992; Clarkson and Dorn 199 1, 1995; Dorn et al. 1992; Schreiber and Lancho 1988), and our second goal in this article is to present new archaeological data that shed light on this issue, and to put the

puquios in their proper archaeological context.

Physiographic Setting

The Rio Grande de Nasca drainage comprises some nine separate rivers that flow together and through a single pass in the coastal range of mountains. The drainage can be conveniently divided into a northern and a southern group of tributaries. The northern group includes the Santa Cruz, Grande, Palpa, Viscas, and Ingenio rivers; of these, the Santa Cruz flows only intermittently, whereas the Grande has the greatest volume of water. The southern group includes the Aja, Tierras Blancas, Taruga, and Las Trancas rivers; the Aja and Tierras Blancas join (via three sepa- rate channels) to form the Nasca Valley. The southern tributaries are substantially drier than the northern group, and all the rivers are deficient in water compared to other coastal valleys. The Aja River, which has the greatest volume of the four, has an average annual flow of only 30.27 million m3 of water, compared to 198.05 million m3 of water that flow down the Rio Grande (ONERN 1971). These, in turn, pale in compari- son to the Chicama Valley (839.43 million m3) of the north coast (ONERN 1973).

Fed by seasonal precipitation in the Andes Mountains at elevations above about 2,000 m asl, the catchments of the southern tributaries are sub- stantially smaller than those of the northern tribu- taries and very small in comparison to other coastal valleys; this accounts for the low initial river volume in the southern valleys. The southern tributaries flow down the western flanks of the Andes until they reach the deep alluvial valley bottom in the lower foothills. The valley alluvium has a moderate to high infiltration capacity that results in a substantial transmission loss in river volume, especially at elevations below 1,200 m asl. For this reason the rivers are what are termed "influent streams," which means that they flow partially on the surface, and for some stretches drop completely below the surface.

The initial point at which the rivers drop below the surface varies from valley to valley, depending on water volume, and varies within each valley seasonally and annually. For example, between 1989 and 199 1, after one year of good rain and two years of drought, the AjaINasca River reached only elevations of 800 m asl, 900 m asl, and 1,050 m asl, respectively, in the month of September.

Local informants and our own observations indicate that the rivers are through-flowing only two years out of seven, on average, and hence in most years the middle portions of all of the valleys are devoid of surface water for the entire year.

At an elevation of 400 m asl, the Nasca River re-emerges and continues to flow on the surface. It appears that this location does not vary, either seasonally or with prolonged droughts. From this point on the stream is perennial for a distance of about 9 km; thence the stream variously appears and disappears on the surface until its confluence with the Rio Grande. Maps and aerial photographs indicate that the Taruga and Las Trancas rivers also re-emerge at about 400 m asl.

For present purposes each valley can be divided into four sections, based on availability of surface water: 1) the upper valley, in which water is avail- able year-round; 2) the zone of infiltration, in which water is available most years but may grad- ually dry up in years of prolonged drought; 3) the dry middle valley that is devoid of surface water except during times of flood; and 4) the lower val- ley, where the river re-emerges (Figure 2).

The distribution of surface water has two pro- found effects on human occupation. First, the rivers cannot be relied upon to provide irrigation water every year in the dry middle portion of the valley. Irrigation canals can be extended from higher up in the valley, but the volume of water is

I /

c 3000 1

Stream flow characteristics:


perennial contour interval
500 meters



I I/


@uquios) appear, some four feet in height,

roofed over and floored with stones and also

with stone sides. Descending from the moun

tains, these covered channels separate into

smaller conduits which ramify over the valley,

supplying every hacienda with water all the year

round and feeding the little streams wh~ch irri-

gate the fields and gardens. The larger puquios

are many feet below the surface, and at intervals

of about two hundred yards there are

(oios) or manholes by which workmen can

descend and clear away obstructions [Markham 1991:50].

ure u system ofsuhterruneunJiltrution the middle portions ofthe Nusca, Turugu, und Lus Truncus valleys ofthe Rio Grunde de Nuscu drainage ofthe south coust of Peru. At present 36 puyuios function in these three vulley.~;in the pus? their number may have exceeded 50. We discuss the /ijrnlul churucteristics and the construction ofthe puqrrios, and describe each ofthe extant puquios. The results c?farchaeologicul settlement surve,v.s conducted in the three vu1ley.s indicute that the puquios did not yet e.xist in Earl)>Nuscu (Eurly lntermediute Period 2 4) times, hut were ulmost certainl,~in use by the time ofthe Inka ofthe region in the lute f~f: teenth centurv. We suggest that the initiul construction and use ofthe puyuios muy have occurred us eur1,v us Nuscu 5 times, and prohah12v not Iuter thun Late Nusca (Eurly lntermediute Period 6-7) times.

------- I



0 5 10 15


I 1000


\ I I

i. DRY I




.::,:-* .

    ' I  
7, --RIOTARUGA -.-..,. -




KJ;. 9

Figure 2. The southern tributaries of the Rio Grande de Nasca drainage, indicating the availability of surface water: the upper valley with perennial flow, the zone of infiltration, the dry middle valley, and the lower valley where peren-

nial flow resumes.

so low that they can only extend irrigation water a kilometer or two. The problem is further com- pounded by the fact that the lower reach of the rivers is constantly changing. As a result, large, long-distance canals do not occur in these valleys, and there is no evidence for their existence in the past. Agriculture in the middle valleys is therefore limited to years of flood roughly two years out of seven. On the other hand, land in the upper and lower valleys is arable year-round thanks to perennial stream flow. Unfortunately, although the arable portion of the middle valley is generally 2-3 km wide, the lower valley narrows to less than 1 km and the upper valley is even narrower; hence most of the arable land in Nasca lies in the portion of the valley that normally lacks water.

Second, there is no water available for domes- tic purposes in the dry middle valley, and it is therefore not suitable for human habitation except during periods of flood. In the dry season, and during periods of drought, the aquifer is too deep (about 10 m deep in the central portion of the val- ley, from Nasca to Majoro) to be tapped conve- niently by wells. However, at the upper and lower ends of the middle valley the aquifer is shallower, and domestic water can be drawn from shallow wells dug in the dry river bed. This is common practice in Nasca as well as in other valleys with influent streams; the wells we have observed are usually about 1 m deep. In contrast, the upper and lower valleys do have available year-round sup- plies of water for domestic purposes.

The inhabitants of the region overcame the lack of reliable water in the middle valley by developing the system of puquios to allow access to the subterranean aquifer-the underground flow of the rivers. A puquio is essentially a hori- zontal well, an open trench andlor a subterranean gallery that connects a point on the surface with subsurface water. The underground water filters into the puquio, flows through it, and empties into either a small reservoir (kochu) or directly into irrigation canals. Puquios provide not only a reli- able source of irrigation water, but also a year- round supply of domestic water. At present puquios water land between 675 and 450 m as1 in the Nasca Valley, a horizontal distance of some 16 km; as we shall see, the prehistoric distribution of puquios was certainly greater.

History of Research

The first scientific mention of the puquios of which we are aware was made by Kroeber (1993) in 1926, who wrote in his field notes that he had been told of puquios in Taruga and Las Trancas. In 1927 Mejia Xesspe (1939) investigated several puquios; he provides a description of six puquios in the Las Trancas Valley and one in the Taruga Valley, and lists but does not describe eight in the Nasca Valley.

In 1934 the Consejo Superior de Aguas com- missioned a study of the puquios by the en&'


M. Francisco Gonzalez Garcia; the report, or at least part of it, was published that year by the Directorate of Waters and Irrigation (Gonzalez 1934; reprinted 1978, text and diagrams unchanged but with critical omissions on the location map). Gonzalez's study is the most com- plete and accurate treatment of the puquios in the Nasca Valley, although he did miss at least two, and other writers have made direct use of his data, not always with attribution. His published report does not include descriptions of puquios in the Taruga or Las Trancas valleys, although we have reason to believe that he included the valleys in his study. He noted at the beginning of his report, "Disgracefully, not all of the puquios that they built have been preserved. There are many that have been destroyed and only vestiges are found of others" (Gonzalez 1934:207).

Regal (1 943) provides a table summarizing the descriptions and measurements of the puquios in the Nasca Valley. It is likely that most of his data were taken from Gonzalez's article; nearly all measurements are exactly the same as those pub- lished by Gonzalez, he uses precisely the same names, and he is missing the same two puquios. However, the fact that he provides some measure- ments that Gonzalez did not indicates that he actually observed some puquios himself.

In 1942 Rossel Castro published an article on the puquios, which is reprinted in his 1977 book about the archaeology of the south coast of Peru, with only minor differences. His data are some- times unreliable, but it is clear that he did exam- ine many of the puquios, and is perhaps the only researcher who crawled through several of them; his observations are invaluable.

In 1960, a University of Tokyo expedition con- ducted a study of irrigation systems in coastal Peru; its report includes a discussion of the puquios in the Nasca Valley, and a reproduction of Gonzalez's diagrams and the map that omits only minor details (Kobori 1960:83).

In 1986- 1987 the Peruvian development agency CORDEICA (1987) sponsored a study of the puquios, which accomplished the cleaning of nine of them, and the detailed mapping of five.

Our ongoing collaboration in the study of the puquios began in 1985 (Schreiber and Lancho 1988). In 1986 Lancho conducted a survey of the puquios for the Consejo Nacional de Ciencia y Tecnologia (Lancho 1986). That same year Schreiber undertook a project designed to docu- ment and map all the existing puquios, and began to look at the archaeological setting in which they occur (Schreiber 1987). The authors have made additional observations of the puquios between 1987 and the present. Since 1986 Schreiber has undertaken several seasons of systematic settle- ment survey in the Nasca, Taruga, and Las Trancas valleys that have provided data on sites directly associated with the puquios and evidence of shifts in settlement patterns correlated with the construc- tion and use of the puquios. The data and interpre- tations presented here should be regarded as our most current assessments, and supercede all docu- ments we have written or published in the past.

We have made extensive use of aerial photog- raphy in our studies. The earliest photos covering the entire region, available from the Peruvian Air Force National Air Photo Service, were taken in 1944. These photos enable us to corroborate (or correct) the articles written by Mejia, Gonzalez, and Rossel. In addition, the 1944 photos provide invaluable data regarding puquios that have been destroyed or altered since that date, and they also provide clues to the existence of puquios that had already fallen out of use by 1944, but whose traces were still visible. The second set of Air Force aer- ial photos we used were taken in 1970; these pro- vide a good contrast with the photos of 1944, and enable us to fix the time of certain alterations in the quarter century that separates those two dates.

open trench


.: . .



a . .

tunnelled gallery

Figure 3. Cross sections of a trench-type puquio (top) and a gallery-type puquio, with both filled and tunneled sections


In our descriptions of the puquios we frequently draw contrasts between conditions in 1944 and 1970, based on these two sets of photos.

Puquio Function and Construction

To the best of our knowledge, 36 puquios still hnc- tion: 29 in the Nasca Valley, 2 in the Taruga Valley, and 5 in the Las Trancas Valley. All the existing puquios are named, usually for the land they water. About one-third have Spanish names; the rest carry indigenous names. Some puquios have been sub- stantially altered during the last century, and numerous puquios have been abandoned or destroyed. At this point we cannot know the total number of puquios that once existed, but we esti- mate there were at least 41, probably as many as 44, and possibly more than 50 puquios in the past.

Puquios are one component of a dual-source irrigation system in Nasca. River water, as in other valleys, is conveyed in acequias (irrigation canals) directly to fields that are to be irrigated. Puquios tap ground water, and deliver the water either directly into irrigation canals or into small reser- voirs. When the river flows, everyone has access to irrigation water; in years of drought, only those people who control puquios have irrigation water. Puquios are of two general types: open trenches, and combinations of trenches and subterranean galleries. This distinction corresponds to the one between puquios and acueductos made by Gonzalez, although he was not always consistent in his use of the terms. Ten puquios, all located in the Nasca Valley, seem to be open trenches for their entire length (Figure 3a); this is typical of the shorter, shallower puquios. The base of the trench is usually only a meter or so wide, but trenches widen to 10 m or more at the surface. The sides of the trench are usually lined with river cobbles that form retaining walls that are stepped back, much like terraces. Some trench-type puquios have mul- tiple branches.

Most puquios begin as tunnellike galleries, and then become open trenches at their lower end (Figure 3b); these gallery-type puquios often have multiple branches. The sides of the gallery are faced with river cobbles fitted together without mortar, and at the uppermost end the water filters between the stones into the gallery. The roof of the gallery is made either of dressed stone slabs or wood; because the wood does not survive for more than a few decades, it is often removed and replaced during the annual cleaning. We have never observed any sort of prepared surface along the base of a trench or gallery, although Rossel (1 977: 172) states that the base of a gallery might be lined with wood or stone to prevent erosion. According to Gonzrilez (l934:208), some galleries were excavated in hard conglomerate and had no need of facing stones along their sides; we, on the contrary, have never observed a gallery without stone sides, and none has been reported by infor- mants. The dimensions of the galleries vary according to the manner in which they were built.

Spaced along the galleries at varying intervals are ojos, holes dug from the ground surface down to the gallery (Figure 3b), which provide a means of access during the annual cleaning, and let in air and light. The forms and dimensions of the ojos also vary according to the manner of construction of the individual galleries. Our mapping of the puquios depends on the ojos to indicate the sub- terranean path of the gallery and any major branches. We have been told that some galleries have short side branches without ojos, but we can- not observe these branches from the surface. A comparison of our observations with those made by Gonzalez (1934), and with aerial photos taken in 1944, indicates that there are many more ojos today than there were in the past. New ojos are typically built when a cave-in has blocked the flow of water through a gallery, because it is much safer to dig a new ojo than to clear the obstruction from inside the gallery.

Galleries apparently were constructed either by tunneling or by filling an open trench (Figure 3b). The deepest, uppermost portions of galleries were usually excavated as tunnels, generally less than 1 m in height and width. Ojos into tunneled galleries are generally very large and conical in profile, and spaced a few tens of meters apart. Some exceed 15 m in diameter at the surface, but narrow to less than 1 m at the junction with the gallery. Such ojos are usually closed at the bottom to prevent soil or other obstructions from falling into the galleries. We rarely found the ojos open, and many had a meter or more of accumulated soil and vegetation that made exact measurement of their depth impossible.

Other galleries were created by first digging an open trench, then constructing the gallery walls and roof at the bottom of the trench, and finally refilling the trench. Galleries built in this manner are sometimes of much greater height-up to 2 m in the case of El Pino, according to informants- than the tunneled galleries. The ojos are much smaller, usually less then a meter in width, with vertical sides. They are square in form, made of cribbed logs with cobbles in the interstices, and were built before the trench was filled in. In one case, Totoral, we were able to observe a section of the puquio where the fill had been washed away by a recent flood and the cribbed ojo structure was standing free. Sometimes the surface above a filled-trench gallery is slumped down a meter or so as a result of the gradual compaction of the fill.

We cannot know at this point whether a trench was filled at the time of the original construction of the puquio, or later. Local informants state that they were filled within the past century. Open- trench puquios are inconvenient barriers to the dis- tribution of irrigation water on the surface, and they take up much potentially arable land so there is good reason to fill them in. Unfortunately, although use of wood for the ceilings of the tun- nels and the walls of the ojos might have been an expedient measure at the time of filling, within a few decades or so the wood deteriorates and must be replaced. Trees are a very limited resource at present, and the filled-trench sections of the puquios are therefore in the poorest state of preser- vation. It is impractical to replace the great quan- tities of wood they contain, and so as sections fall in they are left open. The most recent innovation in trench-filling can be seen in the cases of the Kopara and Chauchilla puquios of the Las Trancas Valley, and Santo Cristo in the Nasca Valley. In these cases cement tubes were laid as conduits for the water in the bottoms of the trenches, vertical cement tubes were spaced about every 50 meters, and the trenches were filled in, sometime after

1944. These puquios are visible on the surface only as a line of wells, each about 1 m in diameter.

The depth of a puquio is of course determined by the depth of the aquifer. Puquios are shallower at the upper and lower extent of their range, and deeper in the middle section. The length of a puquio is determined by the depth of the aquifer and the slope of the land: the deeper the water, the longer the puquio; the steeper the slope of the sur- face, the shorter the puquio. Puquios with rela- tively high output are located either immediately next to the riverbed, sometimes with galleries that extend under it, or else at the valley margin along- side impervious rock outcrops that extend into the valley. Puquios located in the valley bottom, away from either the riverbeds or the valley sides, have lower output.

Most puquios have small reservoirs, kochas, at their lower end from which waters are directed into irrigation canals, acequias. Kochas seem to be in a constant state of renovation; outtakes are regularly replaced with cement slabs and wooden sluice gates, and some kochas have been enlarged and completely lined with cement.

There is a modern version of the puquio, called the pozo-kocha, that is still constructed in this region. In areas where the aquifer is not too deep, typically near the upper and lower ranges of the distribution of the puquios, a deep straight trench is excavated down to the aquifer with heavy machinery. Water filters into the pozo-kocha, and is then extracted with motor-driven pumps. Some puquios have been transformed into pozo-kochas, in which the lower end of the puquio is destroyed and just the uppermost end, where the puquio intersects the aquifer, is left open-an approach that serves to increase valuable arable land. Such alteration is most common in the Las Trancas Valley, where at least three puquios have been converted. At least one puquio in the Nasca Valley has also been converted into a pozo-kocha.

Description of Extant Puquios

In this section we provide a short discussion of each extant puquio in the Nasca, Taruga, and Las Trancas valleys. In each valley we proceed from east to west, upstream to downstream; in the mid- section of the Nasca Valley we first describe the puquios located north of the Tierras Blancas River, and then those to the south. Where previous reports have used different names, they are indi- cated in parentheses. Our measurements of the puquios are provided in Table 1.

Our studies of the puquios were conducted pri- marily from the ground surface. We have made plan-view maps of all of the extant puquios except Chauchilla and La Joya (of Nasca). Where possible, usually in the case of trench-type puquios, our measurements began exactly at the zone of filtration. In the case of galleries, we began our measurements at the uppermost ojo, which is seldom the true beginning of the puquio. Informants tell us that most galleries begin some distance before the first ojo, and our measure- ments of galleries should therefore be taken as minimums. It was only possible to measure the depth of the galleries below the surface when ojos were left open, which is rarely the case. With very few exceptions, which were made possible by the absence of water, we did not enter the subter- ranean galleries ourselves.

Nasca Valley

At present there are 29 puquios functioning in the Nasca Valley. They are distributed from east to west over a distance of 16 km, from Orcona to Soisonguito (Figures 4 and 5). Gonzalez (1934) describes 27 of the 29, plus Soisongo, which has since been converted in apozo-kocha;he does not mention Santo Cristo or La Joya. Rossel (1942, 1977) describes the same ones as Gonzalez, and suggests the existence of Santo Cristo. Mejia (1 939) lists eight puquios in the Nasca Valley, but does not describe any of them.

Existing puquios. The uppermost extant puquio, Orcona, begins as a gallery under the bed of the Aja River; a second branch begins at the bedrock outcrop called Cerro Orcona. After they join, the gallery passes under the riverbed and continues to the SSW until it opens into a trench. Because it lies at the upper end of the dry valley, Orcona is rather shallow; the aquifer lies at only 5 m below the surface. The course of the trench was altered in the late 1960s to follow along new field lines.

The next three puquios are all relatively small; each begins as a short gallery and becomes an open trench. In the case of Vijuna (Matara), infor- mants report short side galleries, but there are few traces of them on the surface. The Vijuna kocha, which was functioning in 1944, has fallen out of use, but large mounds of earth indicate its loca- tion. The gallery portion of Cortez (El Cerco, Uchuya) was only recently discovered during cleaning efforts; it once had a short trench branch, still visible but not in use at the time of our study. Gonzalez's length measurements are significantly greater than ours, and it may be that the newly discovered gallery is actually much longer than presently known. Both Gonzilez and Rossel reported that the puquio Tejeje was in poor condi- tion, and had multiple branches. At present it is in good condition and has only one branch visible from the surface. The acequia fed by the Tejeje puquio crosses the Quebrada de Belen and waters the lands of Beltn.

The puquio Wachuka (Figure 6) has two major branches, both of which begin as galleries. The

Table I. Puquio Measurements.

Puquio Kocha Trunk Branches End Point Trench Gallery Trench Gallery Filled Tunnel Indet. Filled Tunnel Indet. Nasca Valley Orcona yes 793.9 .....70.2 ....131.8 ............................. .12.0 zf

\ ......70.0 OJO Vijuna no 229.2 ......................60.2 OJO Cortez yes 244.4 ......................57.4 OJ0 Tejeje yes 368.8 ..................... 151.5 OJ0 Wachuka yes 294.4 ............................,265.3 ...,172.9 berm \ ..... .47.8 ...,298.8 ............32.6 ojo \ ....,150" zf Visambra yes 209.0 .................... ,870" OJO Aja Alto yes 245.5 zf Aja N yes 487.7 ............................. .82.65 zf \ .....169.6 zf Aja S yes 509.0 ........... .266.0 0j0. Cuncumayo yes 536.5 zt Curve yes 541.0 zf Anklia yes 555.0 .....35.85 .....4.3 ojoAchako N yes 353.8 ............................. 222.5 zf \ ....-294.0 zf Achako S yes 1025.0 zf Cantalloq yes 101.7 ........... ,104.1 ............................. .70.8 OJO

\ ....,265.2 0-10 Santo Cristo yes 301.3 ....2 19.7 OJO Gobemadora yes 180.4 ...,123.5 ...,372.2 OJO Kayanal no 80" ......................37wb

OJO Pangaravi yes 262.2 ............................ ,291.8 ...,207.2

OJ0 \ ....,256.0 zf San Antonio yes 488.1 .....22.4 berm Wayrona yes 493.7 .....43.7 ...,304.0 ojo Majorito no 823.3 zf Majoro yes 907.2 ..................... .95.9 ............................. 31.6 oJo \ ....,246.1 OJO San Marcelo no 336.8 .....96.3 OJO Llicuas 2 yes 272.0 ......................27.0 ............................. .42.5 OJO \ ....-172.9 OJO Llicuas I yes 348.9 ....127.5 ..................................... ,127.2 \ ....,131.3 La Joya yes 260" zf Ocongalla yes 592.0 zf Agua Santa yes 552.0 zf Conventillo yes 559.0 zf Soisonguito yes 597.0 zf Taruga Valley Santa Maria yes 2 18.7 ...,168.8 ............................. .23.2

0Jo \ ......39.0 ...-216.6


San Carlos yes 65.6 ...,140.7 ...,183.6


Las Trancas Valley Totoral yes 96.3 ...,467.6 ............ .90.0 OJO Pampon no 279.9 ...,630.5 ............................ -536.I ............ .41.4 OJO

\ ....,340" berm El Pino yes 435.6 ...-122.7 .......................... ,238.2 ............123.4 zf \ .....-54.4 berm Kopara ves 82gb nln

-J, '

chauchilla yes 1280" 01 o Notes: Measurements are in meters. For example, the puquio Orcona begins as two tunneled galleries, 12 and 70 m long,

~ ~

respectively. These flow together to form a single tunneled gallery 13 1.8m long, which becomes a filled-trench gallery-for

70.2 m, and then an open trench for 793.9 m; it terminates in a kocha.
Gallery types are filled =filled-trench gallery; tunnel =tunneled gallery; indet. =indeterminate gallery construction.
End points, the uppermost point measured on each puquio, are zf =zone of filtration, in cases of open trench puquio; ojo =
first ojo, in case of gallery (puquio could be significantly longer); berm =end of berm from original trench construction, in
case of filled-trench gallery.


"Measurement was taken from aerial photographs or CORDEICA maps '~uquio is considerably longer than what we were able to measure.

Andean foothills

Andean foothills

t I I 4 I 1  
desert pampa 0         5 km  
              K.15 95

Figure 4. Map of the Nasca Valley showing limits of arable land, dry riverbeds, and locations of puquios from Orcona to La Joya: A Orcona; B Vijuna; C Cortez; D Tejeje; E Wachuka; F Visambra; G Aja Alto; H Aja; I Cuncumayo; J Curve; K Anklia; L Achaco; M Cantalloq; N Santo Cristo; 0La Gobernadora; P Kayanal; Q Pangaravi; R San Antonio; SWayrona; T Majorito; U Majoro; V San Marcelo; W Llicuas #2; X Llicuas #I; Y La Joya.

south branch is formed by the confluence of two galleries, one of which begins under the bed of the Tierras Blancas River. Cleaning of the puquio by CORDElCA in 1987 allowed the mapping of the portion under the riverbed. The large berms along most of the south branch and all of the north branch are remnants of the original trench construction; in these sections the gallery is of filled-trench type.

The puquio Visambra provides domestic water to the land on which the town of Nasca was established in the 1590s; it has the highest output of any puquio measured (Gonzalez 1934:220). Visambra figures prominently in local beliefs about water. It is said that a person who drinks the water of Visambra will become enamored of Nasca, and should they ever leave Nasca they will always return.

Visambra begins as a gallery on the south side of the Tierras Blancas River and runs west parallel to the riverbed. Originally the gallery was located along the southern margin of the riverbed itself, but modern levees have reclaimed this land which is now built up with houses. The gallery then bends to the northwest and crosses to the north side of the riverbed and continues to the west under reclaimed land; the cleaning undertaken by CORDElCA revealed that three short branches join the main gallery in this stretch. The puquio then turns north, becomes an open trench, and

flows west again to its kocha. The kocha has been modified several times, most recently in 1992 when it was enlarged and its cement sides raised; this has caused the submergence of the lowest sec- tion of the trench. The point at which the gallery opens into trench has also been changed as houses and streets have encroached on the puquio.

Aja Alto is a short trench-type puquio. Aja was originally two separate puquios that flowed west into the same kocha. Sometime between 1944 and 1970 the course of the southern puquio was altered so that it now veers north and joins the northern one. The northern puquio comprises two open trench branches. The southern puquio begins as tunneled gallery; our measurements begin at the first visible ojo, but informants report that the gallery actually begins under the adjacent bed of the Aja River.

desert pampa
  I I
  0 5 km
XJS 9'    

Figure 5. Map of the Nasca Valley showing locations of puquios from Ocongalla to Soisongo: A Ocongalla; B Agua Santa; C Conventillo; D Soisonguito; E Soisongo. Dashed lines, in this and the following figures, indicate a puquio that has been destroyed or converted; if the puquio today functions as apozo-kocha, this portion of the feature is indicated

with a solid line.

Cuncumayo is a trench-type puquio whose kocha seems to have fallen out of use several years before it was observed in 1986. The puquio Curve begins as a deep open trench alongside a bedrock outcropping that forms a ridge extending into the valley. The trench flows south, along the side of the ridge, then turns west to its kocha, which was enlarged and lined with cement in 1987. Anklia begins as a short section of tunneled gallery, said by informants to begin under the bed of the Aja River, followed by a long section of filled-trench gallery; one short branch enters from the north. The open trench flows to a very large cement-lined kocha that was occupied until recently by a penguin.

Achako (Figure 7) is actually a group of two and perhaps three puquios that share a single kocha. The northern puquio is formed by the confluence of two open trench branches, whereas the southern puquio is a single long open trench. The puquios flow separately to the kocha, much as Aja did orig- inally. In the open land between the branches there is a short trench that now serves as a pozo-kocha, with a pumping station at its lower end.

On the south side of the Tierras Blancas River, the best known of all the puquios is Cantalloq. It was refurbished by CORDEICA in 1987, and some of the ojos were given spiral retaining walls so that tourists might more easily walk to the bot- tom and observe the water flowing through the gallery. Cantalloq has two major branches. The north branch is a tunneled gallery that begins at an unknown point and flows southwest under the Tierras Blancas riverbed. The southern branch parallels the riverbed and is also a tunneled gallery. At some time between 1970 and 1985, probably in 1978, two new ojos were excavated near the upper end of the branch. The stratigraphy exposed in the ojos reveals natural water-laid deposits, not artificial fill, and hence the gallery was without doubt built by tunneling. After the two branches join, a single gallery continues and then opens into a trench. The kocha is very large and cement-lined.

Cantalloq figures importantly in local legends about the puquios because of its proximity to Cerro Blanco, the enormous sand-covered sacred mountain that towers above the Nasca Valley. Local belief has it that the source of puquio water

puquio called Yapana. In 1942 he called it a lost puquio, but in 1977 he gave a description of it that sounds very much like Soisongo, which he omit- ted from the list (Rossel 1942:20 1, 1977: 169). It is nevertheless possible that a puquio once existed at Ayapana. Today there is an enormous pozokocha at exactly the location of what, on the 1944 photos, look like traces of a puquio.

Taruga Valley

The Taruga Valley is the smallest of the three val- leys, and has at present only two functioning puquios (Figure 9). In his 1926 field notes Kroeber (1 993) wrote that, "It was said that 28 puquios [sic; he means qjos]had been located and cleaned out in the past year (1925) on hacienda Taruga, upstream from it." Mejia (1939:562-563) mentioned only one puquio in this valley, the Pukyo de Taruga. In 1942 Rossel said there were three puquios, one on the right (north), two on the left (south); in 1977 he said there were two galleries, but on his map he showed three, arranged as described in 1942 (Rossel 19421202, 1977: 171, 194).

The floor of the Taruga Valley begins to widen just below the zone of infiltration. In the other valleys, the aquifer drops to about 10 m in the middle valley and then rises again to intersect the surface. In the TarugaValley, the water level drops gradually for about 5 km, but then drops off sud- denly to depths of 30 m or more. Such depths seem to have been beyond the technical capabili- ties of the puquios' builders. As a result, only a short stretch of this valley is suitable for puquio construction and use, and the maximum area watered was only about 5 km long.

Existing puquios. Located on the north side of the valley at the former Taruga hacienda is the puquio today called Santa Maria. This is the Pukyo de Taruga described by Mejia (1939), and the series of 28 ojos mentioned by Kroeber in 1926. The main branch of the puquio be&' 71ns as a tunneled gallery, and then is filled-trench gallery; a short filled-trench gallery joins it from the north. It continues as a filled-trench gallery, opens into an open trench, and empties into a large cement-walled reservoir.

On the south side of the valley is the puquio San Carlos, which begins as a tunneled gallery and then becomes a filled-trench gallery. In sev- eral areas the fill over the tunnel has been washed away, and the stone sides and roof of the tunnel are visible. A new kocha has been recently con- structed alongside an older one. The kocha and the lower end of the puquio were dry when we mapped it in October 1986, but the water was 60 cm deep in the gallery at the upper end.

Destroyed or Converted Puquio. About 1 km west of the Santa Maria puquio, also on the north side of the valley, are the remains of a third puquio, Camotal, which is no longer functioning. There is no kocha, and 1944 air photos show the trench emptying directly into the irrigation sys- tem. At present there is a wide, open trench, with a sizeable berm at its upper end. Above this point the land is slightly depressed and saturated with standing water, which suggests the presence of a subterranean gallery in which an obstruction is now causing the water to back up.

Probable Puquio. In 1993 a reliable informant reported that a fourth puquio once existed upstream of the others at a place called Travesia, but that it was buried by the river.

Las Trancas Valley

Today there are five functioning puquios in the valley, two of which have been substantially mod- ified in recent decades (Figure 10). In the past there may have been 1 1 or more. The Las Trancas Valley differs from the other two valleys in having a less steep gradient as the land drops in elevation from east to west. As a result, its puquios tend to be longer, on average, than those of the other valleys

In 1926 Kroeber wrote in his notes that, "Las Trancas is said to have a system of 80 [qjos],long in use" (Kroeber 1993). His investigations did not take him to Las Trancas, however, so he did not observe any of the puquios there. Mejia (1939) described and named six puquios that he observed in 1927, and provided a rough sketch map of the valley, but he indicated that some puquios had already been abandoned or destroyed. Mejia's data, although the most detailed of the published sources, are very ditricult to interpret. The names he gives for the puquios are either mistaken or have changed. We have been able to identify with certainty only three of the six he described; for the remaining three we can only determine the gen- eral area in which each puquio was located, and each of those locations includes more than one puquio. Rossel wrote in 1942 that there were

desert pampa


desert pampa I I I


Figure 9. Map of the Taruga Valley showing limits of arable Maria; B San Carlos; C Camotal.

eight puquios functioning; in 1977 he described six, said there were seven, but noted eight on his sketch map (Rossel 1942:20 1, 1977: 169, 171, 191194). He described only one in detail.

Existing puquios. The three puquios that still function, and that maintain traditional construc- tion, are located in the central portion of the dry middle valley. Situated on the south side of valley, the puquio Totoral (possibly Mejia's Pukyo B, La Marcha) comprises a gallery of indeterminate construction, then a filled-trench gallery. There are more ojos (more than 59) in this puquio than any other recorded. Below the gallery section a short open trench leads to the kocha.

The puquio Pampon (Mejia's Pukyo D, Totoral) is described in great detail by Rossel in both of his publications; he uses it as the "model" for the puquios of this valley, when in fact it is unique among them in several respects. It begins as two galleries, mostly of filled-trench construc- tion, on the south side of the valley. Unlike other galleries recorded, Branch A of Pampon begins at a single abrupt point, a circular stone-lined pit into which the subsurface water filters. It then flows a few meters to a large subterranean cham- ber that measures some 4 m across and 3 m deep, according to our informant, the man charged with its annual cleaning. From this point the gallery flows west along the valley edge until it meets the

Andean foothills


Andean foothills



5 km

KTS 95

land, the dry riverbed, and locations of puquios: A Santa

riverbed; within this stretch a short branch enters from the south. Branch B begins at a natural cul- de-sac in the adjacent hillside, angles northwest toward Branch A, passes beneath it, and then par- allels Branch A as they both approach the riverbed.

About one-fourth of the way across the riverbed, Branch B merges with Branch A. Rossel (1 942:20 1 ; 1977: 193) provides an eyewitness description of this confluence, and says that Branch B lies at a higher elevation than Branch A, and that the water cascades down a stone staircase where they join. On reaching the north bank of the river the puquio continues as a filled-trench gallery until it opens into a trench that in turn empties directly into the irrigation canal system. Because Pampon has no kochu, we defined its lower end as the point at which it crosses under the road running down the center of the valley and where it intersects several irrigation canals. From this point to the end of Branch A the distance is 1,488 m, the greatest we have recorded.

We find it especially interesting that, in his 1977 chapter, Rossel reproduces several of Gonzalez's plans and profiles of the puquios, including one of Pampon. It is the Pampon profile that leads us to believe that Gonzalez actually did study the puquios of the Las Trancas and Taruga valleys, but only published the data from Nasca.


--- - >----{fig-m::f<-&

Andean foothills

desert pampa

Kopara Las Trancas

Cerro Huayurf


  Andean foothills
  I I I I I I Andean foothills
  0         5km  
hJS 95              

Figure 10. Map of the Las Trancas Valley showing limits of arable land, the dry riverbed, and locations of puquios: A Totoral; B Pampon; C El Pino; D Kopara; E Chauchilla; F Huayuri; G Huaquilla Chica; H La Joya.

Located just west of Pampon is the puquio called El Pino (Mejia's Pukyo E). The uppermost segment of the north branch of the puquio has been converted into a pozo-kocha; the remainder of the branch is filled-trench gallery. The south branch is visible on the surface as filled-trench gallery that flows north from the edge of the riverbed; informants say that most water comes from this branch. The puquio continues as filled- trench gallery, then as open trench; a large berm suggests the presence of another branch entering from the south, as described by Mejia, but it has apparently fallen out of use. The puquio ends at a small kocha of recent construction.

Moving west, we find two puquios that have been extensively modified. The puquio Kopara (possibly Mejia's Pukyo F) is clearly visible as a large open trench in 1944 photos. Sometime before 1970 the base of the trench was lined with cement tubes so that the water could continue to flow, and the puquio was buried. At least six cement ojos were created for access into the tube. The upper en4 and hence the length, of this puquio is completely unknown; we were shown a well near El Pino and told it was an ojo of Kopara.

Farther downstream from Kopara is Chauchilla (also possibly Mejia's Pukyo F), which has also been entubed in cement. In the 1944 photo the puquio is visible as a large open trench, with a possible gallery at the upper end where it begins at the edge of the river. In the 1970 photo 24 cement tube "wells" are visible that follow exactly the trajectory of the old trench, and provide access to the now-buried puquio. There is a large cement reservoir at the lower end of the line of wells, approximately where the kocha of the old puquio was situated.

Destroyed or Converted Puquio.~. Three other puquios were clearly functioning in 1944; two have since been converted into pozo-kochas, and the third destroyed. One was located at the upper end of the middle valley on the north side, at the base of Cerro Huayuri. It is clearly visible in the 1944 photos, and was apparently functioning at that time. The upper portion was a gallery with at least four small square ojos, the lower portion was an open trench, and there was no kocha. This is very probably Mejia's (1939:560-561) Pukyo A, which he called the Pukyo Perdido, and Rossel's (1977:192) Wayuri. By 1970 it had been converted into a pozo-kocha, today called El Limon.

In the 1944 photos another puquio is clearly visible in the Las Trancas region; it was an open trench that began northeast of the mound called Huaquilla Chica. It followed along the northern perimeter of the mound, and emptied directly into irrigation canals to the west. Along the upper end there was a large mound of soil, probably excavated

from the trench. By 1970 the uppermost end had been converted into a pozo-kocha, and the rest, including the mounded soil, had been entirely obliterated. This is probably Mejia's Pukyo C, which he called El Pampon.

Another puquio, La Joya (also possibly Mejia's Pukyo F), existed near Kopara. Located just south of the river, it began as a gallery with six large conical ojos visible in the 1944 photo. It then opened into a trench that led to a small reservoir. There is no trace of this puquio on the ground today other than a higher-than-usual density of stone cobbles in a linear path through the modern fields. Informants said that La Joya was very defi- cient in water and was filled in sometime in the 1950s by the hacienda owner.

Possible Puquio.~. In addition to the three for- mer puquios, others may have existed that are now destroyed or converted. At the upper end of valley on the south side at La Marcha, opposite the Huayuri puquio, there are traces of a possible destroyed puquio visible in the 1944 photo. Mejia's (1939:561) Pukyo B may refer to this puquio, or to Totoral. On the north side of the val- ley, near the small settlement at Santa Luisa, a short length of depressed land between two mounds is visible in the 1944 air photo and looks like the remnants of the upper end of a puquio. It appears that there was a square hole or well dug into the depression, which would support the notion that there was a filled-trench gallery below the surface. At present one of the mounds has been leveled, and the depression filled.

Below Chauchilla the valley floor narrows and there are no settlements, historic or prehistoric, for about 5 km. However, in the vicinity of Poroma there is a sizeable Late Intermediate Period occupation, and traces of old agricultural fields are vis- ible in the 1944 photos. We suspect there may have been a puquio in this area as well.

The Puquios in Archaeological Context

Now that we have established the present condi- tion and attributes of the puquios, the most intrigu- ing question facing us is, "When were the puquios built?" Few people familiar with them lack an opinion on the subject. There is a strong local belief that they were built by the Nasca Culture of the Early Intermediate Period (Table 2), although some residents of Nasca maintain that it is more likely that they were built by the Inkas. Among the published sources that focus generally on Nasca or specifically on the puquios there is little agree- ment on the subject. Gonzalez (1934) identifies them as Inkaic, whereas Regal (1942) and Rossel (1 977: 193-1 94) attribute their construction to the Nasca culture of the Early lntermediate Period. In a footnote to the reprinted version of Gonzalez's article (Gonzalez 1978: 130), Ravines states that there is a general consensus that the puquios cor- respond to the Early lntermediate Period Nasca Culture. Peterson (1980:21) suggests that the sys- tem was begun in Middle Nasca (Nasca 5, in our terminology) times and finished in Late Nasca times. Barnes and Fleming ( 1991) argue that the puquios are not prehistoric at all, but rather were built by the Spanish in the Colonial period. Although many of these opinions are valid, given the bases upon which they were formed none of the writers had access to the archaeological data that are now available as the result of the recent settlement surveys of the region. We attempt here to rectify this situation.

Historic Documents: Absence qf' Evidence vs. Evidence of Absence

One might logically begin with Spanish historic records, seeking mention of the puquios-refer- ences to their existence, if they were Prehispanic, or to their construction, if they were built by the Spanish. Barnes and Fleming (1 99 1 :57) have argued that the lack of mention of the puquios in early Spanish documents indicates that the puquios did not exist prior to the seventeenth cen- tury. But neither are there any descriptions of Spanish construction of puquios. The fact that puquios were not mentioned by Cieza de Leon (1984:221) is cited by Barnes and Fleming as strong evidence that the puquios did not exist in the mid-1500s. However, they fail to take into account the fact that Cieza never actually traveled along the coast of Peru south of Lima (Estrada

1987). He was never in Nasca and could not have had the opportunity to observe the puquios, so his oversight should not surprise us. Four of the other five sources cited by Barnes and Fleming (1991:57) do not refer specifically to Nasca at all; indeed, most refer to Chile. The fifth, that of Reginaldo Lizarraga (1 909), merits closer atten- tion. Written in 1605, it includes the earliest

description of the puquios of which we are presently aware.

Lizirraga traveled through much of South America, including the south coast of Peru. Like most Spaniards he was impressed with the dry- ness of the coast, and the general lack of water, and was careful to mention available water resources in each valley. He specifically noted where rivers were adequate or not, and in some cases gave details about irrigation. In specific cases he pointed out who built andor used partic- ular features, and distinguished between the Spanish, the Inkas, and the indios (natives), always noting that the latter were very few in number owing to the massive post-conquest depopulations. He has the following to say about water in Nasca:

[The Nasca river] is lacking of water in the winter, which is the time that in the Sierra it does not rain, and here [on the coast] is the time of the garuas. But in the summer, which is the time of the rains in the sierra, it is a large and even dangerous river. . . . The Indians, in the dry season, make use of wells made by hand, at intervals, and in high places, as reservoirs of water from which they take ucequius to begin the planting and to sustain themselves until the river comes [Lizarraga 1909:522; author's translation].

There are three points to be noted in this pas- sage. First, Lizarraga points out the utter lack of water in the Nasca River during the dry season, and the apparent need to use groundwater when the river was dry; he says this about no other coastal valley. (He seems to assume that the river flows every wet season, which is not the case.) Second, he describes the alternative source of water as wells built at intervals on high ground. This is undoubtedly a description of the puquios: a sequence of ojos andor kochas, on high ground above the level of the riverbed, providing water for irrigation. Third, he is specific that these were used by the Indians. He does not attribute them to the Inkas, as he does certain irrigation features in Cafiete (Lizarraga 1909:5 19), or to the Spanish.

Although this source implies prehistoric puquio construction, in general we find the historic docu- ments to be rather ambiguous for our purposes. Let us turn instead to the archaeological data and con- sider several lines of evidence that might be useful in estimating the date of construction of the puquios.

Table 2. Peruvian South Coast Chronology

Numerical Nasca Valley Standard Uncalibrated Dates Chronology Chronology Dates

1532 Late Horizon Late Horizon

Late Late

lntermediate lntermediate


A.D. 1000 4

  Middle 3  
  Middle     A.D. 800
    Horizon 2  
    A.D. 650
A.D. 750    
  Late Nasca  
    Early 6
  Nasca 5 Intermediate 5
    A.D. 450
A.D. 400    
  Early Nasca 3
A.D. 1    
  Montana 100 B.C.
    Early 9
    Horizon 8
  La Puntilla  

Absolute Dating, Stratigraphy, and Direct Associations

An obvious solution to the problem might be to obtain radiocarbon assays on the wooden beams that form the ceilings of the galleries. Unfortunately, because wood does not endure more than a few decades, the beams are regularly replaced, and hence dating them is not likely to indicate the date of construction of the puquio.

Wooden lintels from the Cantalloq and Majoro puquios collected by Hermann Trimborn yielded dates of 110 * 100 B.P. and 140 * 100 B.P. (see Scharpenseel and Pietig 1974). Another sample was collected from a beam of the puquio Visambra in 1986, during the CORDEICA-sponsored clean- ing. The worker who retrieved the beam said it was the oldest piece of wood in the puquio, and that it was certainly put there by the Inkas when they built the puquio. However, like Trimborn's samples, the beam yielded a date of only 124 * 65 B.P. (SMU 2237; wood; "C = -26.8), somewhere in the mid-nineteenth century A.D. Either the sample was contaminated, or, more likely, the beam was a replacement, installed during a cleaning of the puquio in the last century.

A potentially more fruitful approach is that of Clarkson and Dorn (1 991, 1995; Dorn et al. 1992). Dorn observed the formation of desert varnish on stone lintels inside the Cantalloq and Orcona puquios; microscopic analysis indicates that this type of varnish forms only in damp or wet envi- ronments. The AMS assays of the samples were Orcona: 1460 * 50 B.P., cal A.D. 552-644 (1sigma error); and Cantalloq: 1430 * 60 B.P., cal

A.D. 591-658 (I-sigma error) (Clarkson and Dorn 1995). These are minimum ages, because the var- nish begins to form after the stones were cut. The results suggest, therefore, that the stones were cut in roughly the Early Intermediate Period.

Turning to stratigraphic associations, if puquio trenches were dug through existing (or aban- doned) settlements, then the dates of those sites should provide a lower-limit date for puquio con- struction. Unfortunately, we have never observed evidence of archaeological remains in the side walls of puquio trenches or qjos. It is apparent that prehistoric people did not live on the arable valley bottom, and hence puquios were not exca- vated through extant sites.

Excavation of the puquios themselves (cf. Barnes 1992) is liable to yield little or, worse, mis- leading evidence. It is obviously not possible to excavate open trench segments or tunnels because they are excavations themselves. The only seg- ments that might yield cultural remains are the filled-trench galleries. Artifacts in the fill should provide a lower-limit date for puquio construction if, and only if, the filling took place at the time of the initial construction. Unfortunately if, as we suspect, filling took place long after the puquios were built, possibly within the last century, then any artifacts found in trench fill might pertain to periods long after initial puquio construction.

A more productive line of investigation might be to consider archaeological sites in direct asso- ciation with puquios. Although people did not live on the arable valley bottom, the excavation of puquio trenches created large elevated berms that were unsuitable for cultivation, but that might be suitable for limited habitation or other activities. Sites found on the berms must necessarily post- date puquio construction. Middle Horizon and Late Intermediate Period artifacts have been observed on the berms of the puquios Anklia, Achako, Agua Santa, Soisonguito, Pangaravi, and Santa Maria; and there is a scatter of Nasca phase 5 and Late Nasca artifacts at Pangaravi as well. While the presence of these artifacts might result from the excavation of a puquio trench through a pr-e-existing site, we sought but found no evidence of such sites. We suspect that numerous other small occupations may be located on puquio berms, but are concealed by modern houses.

The construction of puquios also created barri- ers to the distribution of irrigation water, and in the case of the puquio Totoral a small parcel of land was rendered unusable for agriculture. The fact that a Nasca 5 village was established on this parcel of land suggests that the puquio existed at that time.

In sum, sites directly associated with puquios are not uncommon, and the few cases noted indi- cate that the puquios probably existed in the Middle Horizon and the Late Intermediate Period, and may have existed as early as Nasca 5 of the Early Intermediate Period.

Settlement Putterns

Settlement-pattern data, specifically the distribu- tion of habitation sites across the landscape, pro- vide the most secure evidence presently available for dating the construction and use of the puquios. Recalling our earlier discussion of the availability of surface water in different parts of the valley (Figure 2), we make the following assumptions. Prior to the time of construction and use of the puquios, permanent habitation sites should be rare to nonexistent in the dry mid- dle valleys. Although the land might be irrigated and cultivated during years in which the river flowed most years saw little water in this portion of the valleys, and certainly not a drop during the dry season. It is therefore extremely unlikely that permanent habitation sites were established in the dry portions of the valleys in the absence of puquios. This does not, of course, preclude the establishment of cemeteries in such areas. Perennial water is available in the lower and upper valleys, and usually in the zone of infiltra- tion, and hence it is in those locations that we should expect to see permanent villages prior to puquio construction. The lower valley is rather inhospitable due to high temperatures and terri- ble wind storms, so we should anticipate finding most habitation in the zone of infiltration and in the upper valley.

After puquio construction, the middle valley was made habitable year-round and we should therefore see a shift in site locations into this region correlated with the use of the puquios. In fact, given that there is much more arable land in the middle valley, and that puquios are more reliable than river water, we might expect to see movement of the majority of the population into the middle valley as the use of the puquios increased.

Archaeological survey (Figure 11) was directed by Schreiber and conducted on foot by a crew of three to four persons. The region com- prises four zones: valley edges, ridges and hill- tops, valley bottoms, and desert panzpu. Because nearly all sites are located on the valley edges, immediately adjacent to the arable valley bottom, these areas were surveyed completely. Ridges and hill tops were surveyed only in the lowest portion of the Andean foothills, up to 250 m above the valley floor. Sites in the valley bottom are located only on low rises of land where irrigation water cannot reach; these include natural and man-made mounds and puquio berms. All such areas were surveyed except where modern settlements (i.e., the dogs at them) caused us to skirt some pieces of land. In addition to these parcels, we surveyed up to 50 percent of the valley bottom in each val- ley, but thus far we have not located any prehis- toric settlements on potentially arable land. Survey of the valley bottom did allow us to define the limits of arable land established by traditional technology; the use of motor-driven pumps has expanded cultivation significantly in some parts of the valleys. Survey did not extend away from the valleys onto the desert plain or up alluvial fans of side valleys; these areas are covered with geo- glyphs, and it is prohibited to walk on or near them without special permission. Because the geoglyphs were not a primary focus of our study, we did not seek such permission.

We have completed systematic survey in the Aja and Tierras Blancas valleys, beginning at an elevation of 1,200 m as1 in each, and continuing down through the Nasca Valley to an elevation of 275 m asl, a distance of about 50 km. The survey area includes all four hydrologic sectors: upper valley, zone of infiltration, and middle and lower valley. We have also extended limited reconnais- sance up to 1,600 m as1 in the Tierras Blancas Valley. In the Taruga and Las Trancas valleys we have completed systematic survey only in and around the dry middle portions of the valleys, with some reconnaissance of areas above and below, and the survey has not yet been extended through the valleys' upper and lower reaches. Thus, the Nasca Valley and its tributaries should provide both the expected "before" and "after" patterns, if they exist. The Taruga and Las Trancas valleys should show evidence of the "after" pat- tern, but we must await further research to provide the "before" picture.

When we compare settlement patterns of the Early Nasca period with those of the Late Intermediate Period in the Nasca Valley (Figures 12 and 13), we clearly see our expected "before" and "after" patterns. In Early Nasca times sites were distributed in the lower valley, and in the zone of filtration and the upper valley. In the lower valley Cahuachi, a major ceremonial center (Silverman 1993), occurs along with numerous cemeteries (Ogburn 1993). It is no surprise to us that the region was sacred to the prehistoric inhab- itants; the emergence of the river at Las Caiias, in the midst of some of the driest territory on the coast of Peru, must have had great religious sig- nificance. Permanent habitation sites seem to have been lacking in the lower valley in Early Nasca times, probably owing to the heat and wind.

In contrast, numerous habitation sites, mostly small villages, were located in the zones of infil- tration and upper portions of the Aja and Tierras Blancas valleys. The extensive occupation of the zone of infiltration suggests that water was rela- tively plentiful in Early Nasca times, and that

limits of systematic survey areas of limited reconnaissance

Figure 11. Boundaries of the areas systematically surveyed, 19861994.

there were no extended periods of drought that would cause people to move to higher elevations. Not a single habitation site is found in the dry middle valley, although numerous cemeteries and two small ceremonial centers (Pueblo Viejo and Cantalloq) are present. Likewise, in the Taruga and Las Trancas valleys, no habitation sites are found in the middle valleys.

In sum, settlement patterns of the Early Nasca period match our expectations regarding patterns prior to the construction of the puquios. The lack of permanent occupation of the middle valleys indicates to us that the puquios were not yet in use.

This pattern changed completely by the Late lntermediate Period. In the Nasca Valley settle- ments are found throughout the dry middle valley. The sites in the zone of infiltration and upper val- ley are smaller and fewer in number than those in the middle valley, an indication that the majority of the population lived in the middle valley. In both the Taruga and Las Trancas valleys there are likewise maJor settlements in the middle valley in the Late Intermediate Period. These data indicate to us that puquios were in use by the Late lntermediate Period and long before the arrival of the Spanish. Furthermore, in the Late Horizon immediately after, the Inkas established a major center at Paredones, in the heart of the dry zone (Figure 13). Water must have been available at or near the site; at present the land in front of Paredones is watered by the La Gobernadora puquio, and the Pangaravi puquio is nearby. It is probably no accident that the Spanish also chose this portion of the valley for their town; in fact, they located Nasca on a tract of land watered by the most productive puquio of all, Visambra.

The differences in settlement distribution demonstrate to us that the puquios were built after roughly A.D.450, but well before the Inka arrival in A.D. 1476. Can we be any more precise regard- ing the period of construction within that millen- nium? We should be able to determine the period of initial use by identifying the period in which people first established major permanent settle- ments in the dry middle valley. The data from all three valleys suggest that initial use of puquios began by Nasca 5 (Figure 14). In the Nasca Valley, settlements were first established in the middle valley in Nasca 5. The middle Taruga Valley includes a very large Nasca 5 site, and sev- eral maJor Nasca 5 sites were established in the middle Las Trancas Valley as well. These data suggest an initial period of use of at least some of the puquios in Nasca 5 times. The association of Nasca 5 sites with the Pangaravi and Totoral puquios also supports this interpretation.

large village

o small v~llage

Figure 12. Settlement patterns of the Early Nasca period. Habitation sites are located in the upper valley where peren-

It is interesting that while people were moving down into the middle valley, others were moving farther up-valley in Nasca 5. New and larger settlements were established at elevations above 1,050 m as1 in the Aja Valley, and above 1,150 m as1 in the Tierras Blancas Valley.Movement out of the zone of infiltration and into the upper valley is exactly the expected pattern during periods of prolonged drought, which might also account for the need to develop new sources of water, such as puquios. Prehistoric precipitation in the south highlands of Peru has been documented through the study of ice cores by Thompson et al. (1985). The data do not yet extend to periods prior to A.D. 470, but they do indicate a prolonged drought between A.D. 540 and 560, and an especially severe drought between A.D. 570 and 610. There is evidence in the puquio constructions themselves that suggests they were built in times of severe drought; the fact that the galleries are longer than they need to be today and are no longer cleaned to their original end suggests that the aquifer is slightly higher today than it was at the time of puquio construction.

The shift in settlement locations is even clearer by Late Nasca times (Figure 15). There was a complete change from the pattern of numerous small villages seen in Early Nasca times to a pattern characterized by a limited number of very large towns. In the middle Nasca Valley there is a cluster of one large and two small sites; in the upper Tierras Blancas there is a similar cluster. In the middle Taruga Valley growth of the large Nasca 5 site to cover some 16 hectares made it the largest site in the region in Late Nasca times. In the middle Las Trancas Valley several large villages were occupied in the Late Nasca period. Our data indicate that the Late Nasca period was a time of population aggregation and increased sociopolitical complexity. More to the point, the location of major sites in the dry middle valleys indicates very strongly a reliance on puquios in Late Nasca times.

Location key



desert pampa

6 -----'----{fig-~:1p3-,


Today the puquios provide the people of the Nasca, Taruga, and Las Trancas valleys with water year-round for both irrigation and domestic use. We have attempted to provide a brief, updated description of the system of puquios, a discussion of how they function and how they were built, and the condition of each one of them today. We have considered the archaeological data that suggest a prehistoric date of construction,

o small village

small town

Figure 13. Settlement patterns of the Late Intermediate Period. The majority of the population lived adjacent to lands watered by puquios, which suggests that the puquios existed at this time. The location of the lnka site, Paredones, which was established at the beginning of the Late Horizon, is also noted.

large vlllage

o small village

Figure 14. Settlement patterns of Nasca 5. Habitation sites are located at higher elevations than in Early Nasca, suggesting drought conditions. For the first time sites are located adjacent to lands watered by puquios, which suggests that the puquios were first used at this time.

large village

o small village

Figure 15. Settlement patterns of the Late Nasca period. Population aggregation into a small number of very large sites suggests changes in sociopolitical organization. The presence of large sites adjacent to lands watered by puquios sug

gests that puquios were in use at this time.

and our best estimate for the date of initial construction and use of the puquios is Nasca 5, the

period between the and Late Nasca phases Of the Intermediate Period. This is not to say that all the puquios were built in Nasca 5 times. It is not unlikely that puquio

construction and modification continued over the centuries' as grew' and the need for water grew ever greater. It is that puquios are constantly undergoing structural modifications even now. It is suite likelv that modifications in the centuries the spanish

caused them to ble the qanats Of the Old and have given rise to some confusion regarding their authorship.

The archaeological evidence indicates that the usmiling so charmingly described by Markham was created well over a millennium ago by the Nasca people of the Early Intermediate Period. The puquios are perhaps the most enduring legacy of the ancient Nasca Culture.

'Onquest have

Ackizowledgments. We wish to thank our many colleagues who helped collect the data, and whose ideas and arguments eontributed greatly to the clarification of our own. Funding for the fieldwork was provided by the Consejo Naeional de Cieneia y

Tecnologia (CONCYTEC), the University Research Expeditions Program of the University of California, the Academic Senate of the University of California at Santa Barbara. and the National Geographic Society; we gratefully

-. -

acknowledge the support of these agencies. In Nasca we thank Aroldo Corzo, Julio Garcia, Miguel Pazos, Otto Pflucker, Ol~viaSejuro, Felix Solar, and the innumerable users of the puquios who were only too happy to share thelr knowledge with us. We offer profound thanks to the medlcal staff of the

U.S. Embassy in Lima for their kindness in providing anti


rabies serum and vaccine. We extend our appreciation to our colleagues who read and commented on an earlier draft of this paper: Wanvick Bray, Patrick Carmiehael, LUISJaime Castillo, Persis Clarkson, An~taCook, John H. Rowe, and Helaine Silverman; and we thank Kirk Frye, who assisted in the ereation of computer graphics. We alone are responsible for our interpretations. Because this study is ongoing, we shall endeavor to continue to increase our store of knowledge of the puquio~and to correct any inadvertent errors of fact through future investigation in Nasea. After all, we have both drunk the waters of "isambra. Salud!

References Cited

Barnes, M. 1992 Dating ofNazca Aqueducts. Scientific Correspondence, Nature 359:1 1 1. Barnes, M., and D. Fleming 1991 Filtration-Gallery Irrigation in the Spanish New World. Latin American Antiy uity 2:48-68.

Bray, W. 1992 Under the Skin of Nazca. Nuturc. 358: 19. dc Cieza de Leon, I? 1984[1553] CrAnicu del Peni. Primeru Purte. Pontificia Universidad Catolica del Peni, Lima. Clarkson, P. B., and R. I. Dorn

1091 Nuevos Datos Relativos a la Antigucdad de los Geoglifos y Pukios de Nazca, Peni. Boletin dc. Limu 78:3345. Lima.

1995 New Chronometric Dates for the Puquios of Nasca, Peru. Lutin Americun Antiquitv 6:5&69. CORDEICA (Corporacion de Desarrollo de Ica)

1987 Proyccto Rchabilitacion de Acueductos Arqueologicos de Nasca. Corporacion dc Desarrollo de Ica, Consejo de Nasca, y lnstituto Nacional de Cultura Direccion General del Patrimonio Cultural y Monumental. Lima.

Dorn, R. I., P. 9.Clarkson, M. F. Nobbs, L. L. Locndorf, and D.

S. Whitley 1992 New Approach to the Radiocarbon Dating of Rock Varnish, with Examples from Drylands. Annals of the Association qfAmericun Geographers 82(1): 136 15 1.

Estrada Ycaza, Julio 1987 Andanzas de Cieza por Tierras Americanas. Banco Central del Ecuador, Sucursal Mayor de Guayaquil. Archivo Histhrico del Guayas, Coleccibn Monografica, Publicacion no. 14. Guayaquil, Ecuador.

Gonzilez Garcia, M. F. 1934 Los Acueductos lncaicos de Nazca. A~wasr lrrigcrcihn 2(2):207-222. Lima. 1078 Los Acueductos Incaicos de Nazca. In Trcnologiu Andinu, edited by R. Ravines, pp. 128-156. lnstituto de Estudios Peruanos, Lima. (Reprint of 1935 article with annotations by Ravines.)

Kobori, I. 1960 Human Geography of Methods of Irrigation in the Central Andes. Andesu, Andes, the Report ($the Univr2rsit). of 7hkyo Scientific Expedition to the Andes in 1958, pp. 7&97,417 420. Bijitsu Shuppan sha, Tokyo.

Kroeber, A. L. 1996 [I9261 Archac~ologicul E.rplorutions in Nazcu, Peru: The I926 Marshall Field Expedition. Edited by P. H. Carmichael. Fieldiana. Field Museum of Natural History, Chicago, in press.

Lancho Rojas, J. 1986 Descripcicjn y Proh1ema.s de Mantmimiento J, Rehuhilitucicin-dc, 10s Aclleducto.~ de Na.~ca. Report to CONCYTEC (Consejo Nacional de Ciencia y Tecnologia).

Lizirraga, R. de 1909 [I 6051 Description Breve de Toda la Tierra del Peru, Tucumin, Rio de la Plata y Chile. Nucvu Bihlioteca de Autore\. Espaiiolt,.~ 15:485460. Bailly-Bailliere e Hijos Editores, Madrid.

Markham, C. R. 1991 [1908-19121 Markham in Peru: The Travels of Clements R. Markham, 1852-1853. Edited and with an introduction by P Blanchard. University of Texas Press, Austin.

Mejia Xesspe, M. T. 1939 Acueductos y Caminos Antiguos de la Hoya del Rio Grande de Nasca. Actus y Truhujo.~ C'ientificos del XXVII (hngreso Internacionul dr Am~~ricanistas 1559 -569. Lima.

Ogburn, D. E.

1993 The (Lmeterir:c. ofNusca. Unpublished Master's paper, Department of Anthropology, University of California, Santa Barbara.

ONERN (Oficina Nacional de Evaluacion dc Recursos Naturalcs) 1971 lnventurio, Evuluacitin y Uso Rucionul de 10,s Recursos Nuturules de lu Ci)stu: ('uenca del Rio Grunde (Nuzcu).

1973 Inventario, E~~uluucibn

y Rucionul dc. los Recursos Nuturule, dc. lu C'ostu: Cuencu del Rio ('hicarnu. Lima. Peterson G., G.

1980 Evolucitin y Ilesui~uricion dc. lus Altus ('ulturus Purucas-Chhuachi (Nuscu). Univcrsidad Nacional Federico Villareal, Lima.

Regal, A. 1943 Los Acueductos Precolombinos de Nasca. Revi.stu de lu Univer.sidud C'uttilicu del Peru 1 1 (&5):210-213. Lima. Reinhard, J. 1986 The Nuzca Lines: A Ne~s Perspective on their Origin und Meaning. Editorial Los Pinos E.I.R.L., Lima. Rossel Castro, P A. 1942 Sistema de lrrigacihn Antigua de Rio Grande de Nasca. Revista drl Museo Nucionul 1 l(2): 196 202. Lima. 1977 Arqlleologia Sur drl Perzi. Editorial Universo S.A., Lima. Scharpenseel, H. W., and F. Pietig 1974 University of Bonn Natural Radiocarbon Measurements VII. Rudiocarhon 16(2): 143-165. Schreiber, K. J. 1987 Pro,vecto Arqurologico 10s Pukios dr Nu.scu. Report submitted to the lnstituto Nacional de Cultura, Lima. Schreiber, K. J., and J. 1,ancho Rojas 1988 Los Puquios de Nasca: Un Sistema de Galerias Filtrantes. Boktin de Lima 595 1-62, Silverman, H. 1993 (hhuuchi in the Ancient Nasca World. University of

Iowa Press, Iowa City. Thompson, L. G., E. Mosley-Thompson, J. F. Bolzan, and B. R. Koci

1985 A 1500-Year Record of Tropical Precipitation in Ice Cores from the Quelccaya Ice Cap, Peru. Science 229971-973.

Zuidema, R. T.

1986 lnka Dynasty and Irrigation: Another Look at Andean Concepts of History. In Anthropological Hi.stor)j ofAndeun Polities, edited by J. V Murra, N. Wachtel, and J. Revel, pp. 177-200. Cambridge University Press, Cambridge.


1. The word puquio is Quechua, and it usually refers to a nat- ural spring. However, its meaning can also be quite broad and may refer to natural water sources modified by human action (Zuidema 1986: 199 [endnote 51). In Nasca the term acurdllcto has received increasing usage in recent decades, in part to attract tourists to see what the local people consider to be some of their greatest prehistoric accomplishments, and in part because European terms are seen to be of higher status than indigenous ones. However, because the term used most widely in Nasca by the people who actually use this irrigation system is puquio, we use it herein.

  • Recommend Us