More than three thousand years ago, a great tradition of stone sculpture and mega­lithic architecture emerged in Peru’s north highland valleys of which the Early Horizon (1200-200 B.C.) temple complex at Chavin de Hunter is a spectacular early example. In the subsequent Early Intermediate Period (200 B.C.-A.D. 550), the Callejón de Huaylas (or upper Santa River Valley; Fig. 2) appears to have become the center of this stone-working tradi­tion. We know this, however, not from temple complexes like Chavin, but from a collection of clearly local but largely unprovenienced statues and decorated stone lintels now housed in the museum of the modern city of Huaraz. This stonework is superb, yet most of the settlements of the period are unimpressive, with monumental stone constructions lim­ited to subterranean tombs and mas­sive retaining walls.

The contradiction between such an elaborate stone sculptural tradi­tion and such modest settlements is puzzling. Have archaeologists some­how missed the monumental capitals of the Early Intermediate Period in the Callejón de Huaylas? Or were there no such capitals, and were these impressive sculptures produced in­ stead by autonomous communities of peoples whose stone architecture, and perhaps their statues as well, were primarily mortuary in function?

The subsequent Middle Horizon Period (A.D. 550-1000) brings with it another set of questions about the nature of settlement in the Callejón de Huaylas. This was a time when the styles of the city of Huari, 550 km to the south, spread throughout much of highland and coastal Peru, bring-in to an end the Early Intermediate Period. What effects did this Huari expansion have on the Callejón de Huaylas, with its striking local tradi­tion of stone working and architec­ture? And what in turn might local settlements in the Callejón& de Huaylas tell us about the organization and administration of the important pre-Incan polity of Huari? The site of Honcopampa seemed likely to hold some answers.


Honcopampa is the ruin of a small city 11,400 feet above sea level on the western flank of the glaciated Cor­dillera Blanca (Fig. 1). It is char­acterized by a large concentration of impressive, multistoried buildings often called chullpas (see below). Many archaeologists date these stnic-tures to the Middle Horizon and interpret Honeopampa as a provin­cial capital built by Huari rulers to govern the subjugated peoples of the Callejón de Huaylas. Others disagree and suggest that the buildings were instead constructed during the Early Intermediate Period by leaders of a complex local polity called Recuay.

The site was briefly investigated in 1981 by American archaeologist Gary Vescelius and Peruvian Hernán Amat. They reported that Hancopampa was a Middle Horizon center, but their findings were never published and remain controversial.

Establishing the site’s dates would be the first step in providing a proper historical context. Was Honcopampa a “missing” Early Intermediate Period capital of the Recuay polity, or was it the provincial administrative center of an expansive Middle Horizon Huari state? These questions moti­vated my investigations in the sum­mer of 1987.

Architecture at Honcopampa

Honcopampa’s archaeological zone surrounds a shallow bowl-like depres­sion high on the eastern side of the steep Callejdn de Huaylas. The de­pression is about a kilometer across and comprises pasture and marsh areas; modern inhabitants state that the name Honcopampa (perhaps more properly spelled Honcopampa) derives from a descriptive term in the local Quechua language for “wet land where people sink into the ground.”

Three large concentrations of well-preserved buildings are located on high ground at the northern edge of the marsh, but most of the well-drained slopes surrounding the de­pression are dotted with various align­ments of boulders—straight lines, circles, V shapes, grids, terraces, and perhaps even plazas and courts (Fig. 3). At least some of these boulder alignments were once monumental walls, constructed of massive stones with infill of smaller rocks between them (see Fig. 4).


Small rectangular buildings are also scattered across the hills sur­rounding the depression. Founda­tions of these buildings were dis­covered on cultivated slopes near modern homesteads, and also on steep and windy hilltops (Fig. 9). Called chullpas by the natives, the buildings range in size from only a couple of meters square to at least 4 or 5 meters on a side. They some­times occur in groups of two or three.

Chullpas were constructed not of boulders with smaller infill stones but in a technique called block and spall. The masonry of the walls consists of more or less rectangular stone blocks carefully and attractively combined with small, flattish rock slabs or spafls set in strong clay mortar. The stout little structures were roofed with huge stone slabs, had large door jambs and megalithic lintels over tiny doorways (Fig. 10).

Today, most of these scattered chullpas are so severely damaged that their remains are often difficult to detect. They have been the targets of intensive looting as well as de­liberate dismantlement for their megalithic roof stones and door lintels, which are reused in bridges as well as for door and window jambs in modern buildings. Their original con­tents are missing or in disarray, mak­ing functional interpretations diffi­cult. Most archaeologists accept them as mortuary monuments, but while human bones have been found in some of their chambers, it has yet to be demonstrated that the chullpas all belong to a single category of build­ings intended for the burial of human remains. Chullpas are also found in each of the three architectural concentrations on the northern side of the marshy depression (Fig. 5). Each concentra­tion has its own characteristic set of building forms, but masonry tech­niques and the general chullpa build­ing type are similar in all three.

Alma Funcu is a long, low hill with the remains of nine or more chullpas. Most of these chullpas are much bigger than the ones dispersed on the surrounding hills, but they share the same masonry. Many are two stories tall. Seven or more of the chullpas are arranged around three sides of a courtyard. They enclose a U-shaped area that is open to the north (see cover).

In the southwest corner of the ‘U’ is Honcopampa largest chullpa, meas­uring 12 by 16 meters (Fig. 6). At ground level, this building has three doorways on the north side that open into the U-shaped court, and a single doorway on each of the other sides. The interior of the building is divided into 20 chambers that are arranged in interconnected groups of 4 or 4 rooms (Fig. 7). The second floor of the chullpa has a single entrance to the north and contains 6 intercon­nected chambers, while an old report claims that what may have been a third story was visible in the past. Hernán Amat, who excavated in the chambers of this chullpa with Gary Vescelius in 1981, reports finding human bones as well as pottery in the Viñaque style (a style diagnostic of Huari during Middle Horizon 2, about A.D. 700-850).

Five other chullpas in the AU’ also have two stories, multiple doorways, and numerous internal chambers. Many, but not all, of the chambers are tall enough to stand in comfort­ably and large enough for several adults to sleep on the floor. Some of the ceilings are stained from smoke, and occupation debris can be found on the floors. But the chambers have been used as camp sites by travelers and trekkers for many years, and the multiple reoccupations make it very difficult to determine how the cham­bers were used by their original builders.

Unfortunately, many of Ama Puncu’s chullpas are in such poor repair that they cannot be mapped accurately without clearing and excavation. Two that are in extremely bad condi­tion were identified less than 100 meters west of the U-shaped group, but still on the Ama Puncu hill. There are also traces of stone walls that appear to have belonged to construc­tions other than chullpas. Perhaps at some time the Ama Puncu hill had a significant diversity of architectural forms, as well as corresponding cul­tural activities. However, habitation refuse is very scarce among current surface remains, and it seems un­likely that Ama Puncu was ever an important residential area.

Chucara Ama is located about a kilometer northeast of Ama Puncu. A large chullpa occupies an extensive rectangular platform. Remains of several smaller chullpas are also in evidence, as well as many walls that probably belonged to terraced plazas. An outstanding and unique feature at chucara Ama is a red rock that looks like a bedrock outcropping, surround­ed by well-constructed stone walls. As in the case of Alma Puncu, surface refuse is scarce at Chucara Ama.

H-shaped Buildings and Patio Groups on the Purushmonte Hill

Purushmonte is a hill with the larg­est area of architectural remains at Honcopampa. It is covered by very dense brush, giving the initial impres­sion that the ruins are limited to big terrace walls. However, clearing the vegetation revealed lovely buildings with great doorways (Fig. 14). Many walls were standing 2 meters tall. Periodic cultivation had destroyed other walls to the bottom of the plow zone, but ca. 30 centimeters below the surface they were relatively well preserved. Once the vegetation had been cut down, shallow trenches revealed nearly complete building plans without disturbing the deeper occupation strata that will be so im­portant for more detailed studies of Honcopampa in the future.

In spite of dense vegetation, abun­dant occupation refuse was recover­ed, including large grinding stones found in many of the buildings. Some small, poorly preserved chullpas were also recorded. Several large retaining walls on the Punishmonte hill are of boulder and infill con­struction, but most of the walls, ex­cept for a few examples that may be technologically intermediate, are of block and spall construction.

Two additional building forms were identified at Purushmonte. Both are complex, multi-roomed com­pounds, but the first surrounds a central D-shaped room or tower, while the second surrounds a rectan­gular patio. Surface remains suggest that there are somewhere between 6 and 12 hectares of such compounds. The well-preserved architecture clear­ed in a 2.5 hectare study area may all belong to one or the other of these two compound classes (Fig. 8).

D-shaped building complexes are less frequent than patio groups. Only two examples were found, both lo­cated in the southern or lower part of the Purushmonte study area. How­ever, since the lower part of the slope has been much more disturbed by modern and historic activity, the scar­city of D-shaped buildings may not be indicative of their original num­bers.

The larger D-shaped building, Ac-14, has a doorway in the flattened side that faces south, and another feature that may be a door in the northwestern section of the wall, at a significantly higher level. One part of the D-shaped wall stands over 5 meters tall suggesting that in its orig­inal condition the building was a tower-like construction that may have contained multiple floors. The smaller D-shaped building, Ac-14, just north of AC-14, has several large niches in the interior of its curved wall. Since both of the D-shaped buildings have rooms abutted onto their exteriors, they appear to have been the foci of elaborate architec­tural complexes. One room on the east side of AC-14 has a row of stones projecting from the inner face of the wall, perhaps corbels to support a second floor. However, they are only a few centimeters above the level of the modern surface of the ground.

No excavations were conducted in the D-shaped buildings. An old trench in Ac-13, probably excavated in 1961 by Amat and Vescelius, was cleaned and inspected. Its walls revealed ashy zones but few artifacts, and it seems likely that AC-13 was either cleaned regularly and carefully, or it was not a residential structure. Since D-shaped buildings were identified in only the lower part of Punishmonte, spatial separation may have correlat­ed with functional or ethnic differ­ences within the city of Honcopampa.

The rectangular patio group com­plex is the most frequent building form in the sample study area, and all the examples are located higher on the hill than D-shaped buildings, Patio group masonry, like that of chullpas and D-shaped buildings, is of the block and spall type. How­ever, the quality of patio group ma­sonry varies, depending on its loca­tion (Fig. 11).

The patio groups consist of four elongated halls around a rectangular court or patio, creating an enclosed compound. The idealized Honcapampa form is oriented more or less to cardinal directions, and has a main entrance in the center of the east side of the compound. The entranceway is a corridor with megalithic door jambs and great stone lintels. It pro­vides access to the central patio, and sometimes to one of the adjacent perimeter halls. Each long perimeter hail is divided into rooms, usually three. One perimeter hall, usually on the west side, is wider than the rest; it too is divided into rooms (Fig. 12). Each room has a doorway with large stone jambs and lintels connecting it with the patio, but direct access be­tween rooms is rare. The rooms were probably one story high. The largest lintel in each patio group is located over the central doorway into the wide hall, which also has the finest masonry. These lintels, ranging from 3 meters to more than 4 meter. in length, are extremely impressive.

Occasionally patio groups have a secondary entrance from the outside: a small corridor into the patio from one side of the rectangular compound. More rarely, there is a doorway through the outer wall directly into one of the hall rooms. Patio groups also appear to have had broad benches about 20 to 30 centimeters high that encircled the entire edge of the central enclosure. It is within the patios that large grinding stones are typically found.

Compounds Ac-2, AC-4, Ac -4, AC-5, AC-6, Ac-9, and AC-11 are especially good examples of the rec­tangular patio group complex, even though few of them meet all of the ideal criteria discussed above. Other buildings at Purushmonte also appear to have been constructed with the patio group model in mind. Future research will clarify the formal aspects of these buildings and reveal whether they belong to subclasses of the patio group complex, or perhaps to another formal class that is yet to be defined.

One rectangular patio group de­serves special note. Compound AC-1 lacks a central entrance with cor­ridor. It has no large lintels, and all of its perimeter halls are of approxi­mately the same width. Significantly, there is is a row of projecting stones, or corbels, preserved in two of the perimeter hall rooms. This suggests that the peripheral halls of this build­ing may have been multi-storied.

Four small stratigraphic test pits were excavated in hall rooms of four rectangular patio groups. In three cases, AC-2, Ac-5 and Ac-9, resi­dential debris was abundant. Only the excavation in Ac-1 yielded little more than sterile soil. In view of these results, as well as the general abun­dance of occupation refuse on the surface (including numerous grind­ing stones), I conclude that Honco­pampa’s rectangular patio groups were primarily residential in function.

Several of the rectangular patio groups revealed alterations that re­present significant remodeling. Doorways were blocked or added, rooms reshaped, walls removed, and new walls, some of them curved, added. AC-18 appears to have suf­fered only minor modifications, but AC-4 was dramatically altered. AC-18 has been so extensively rebuilt that its original form is no longer clear, and it may or may not have been a rectangular patio group. I infer that these remodelings represent a final phase of occupation at Honcopampa, when ideas about the shape and use of buildings had changed.

The Dating of Honcopampa

With the exception of a handful of Early Intermediate Period sherds not found in architectural contexts, all the ceramics from Honcopampa date to the Middle Horizon or later (see box on dating). Sherds of the Middle Horizon are associated with unmodified patio groups, D-shaped buildings, and chullpas, as are cor­roborating radiocarbon dates. Sherds dating to the end of the Middle Horizon and perhaps on into the Late Intermediate Period are associated with plough zone strata and the re­modeled patio groups that mark the end of Honcopampa’s prehistoric occupation.

The masonry of the patio groups, D-shaped buildings, and chullpas is stylistically united by the block and spall construction technique. While late modifications of several patio groups also employ block and spall construction, it is of the lowest quality; seemingly earlier construc­tion, although not presently datable, is easily differentiated by its boulder and refill technique. Consequently, it appears that the block and spall ma­sonry style as well as the patio group, D-shaped building, and chullpa forms belong to one time period. Honco­pampa is best interpreted as a Middle Horizon center.

Architecture, Honcopampa, and Huari

In many ways, Honcopampa’s archi­tecture resembles that of Huari. The patio group was a standard building form at Huari and it is the form most commonly associated with what have been interpreted as Heart’s pro­vincial administrative centers during the Middle Horizon Period. We also know that D-shaped buildings are important architectural features at Huari, probably dating to the end of the Early Intermediate Period and the beginning of the Middle Horizon. Only one of Honcopampa building forms is not typical of Huari and that is the chullpa. Although chullpas may have analogies at Huari in megalithic, dressed stone chambers that often included several rooms and even two or three floors, these seem to have been subterranean, or surrounded by rough stone walls that retained a covering of earth.

However, while the principal con­struction phase at Honcopampa is Middle Horizon in date and related to Huari, the architecture is not simply a Huari import. Honcopampa responded to the Middle Horizon with a strategy that incorporated some Huari building traits while re­jecting others in favor of a continuing affiliation with an old northern tradition of monumental masonry archi­tecture. Two of Honcopampa build­ing forms are shared with Huari, but masonry of the block and spall style that characterizes the patio groups, D-shaped buildings and chullpas is not. Huari’s buildings were covered with clay and lime plaster, giving them a brilliant white finish. Unlike Honcopampa careful masonry, their rough stone construction is durable but not aesthetically patterned with stones of selected size and shape.

Huari architecture also avoids the use of large stone lintels or door jambs, preferring to de-emphasize doorways. Consequently, Honcopampa explicitly emphasized spaces, marked by block and pall masonry walls of differing quality, by larger entrance lintels, and by more massive door jambs, must be considered part of a northern architectural tradition foreign to Huari.

Indeed, block and pall construc­tion characterizes Early Intermediate Period architecture at Pashash and in Huamachuco and Cajamarca, basins farther north than the Callejón de Huaylas. Emphasis on megalithic lintels is another feature widely spread in the north. There are mega­lithic towers in the northern basin of Chota, tall, narrow buildings of four floors whose masonry includes dress­ed block construction. These may belong to the Early Intermediate Period, and perhaps are related to, or even antecedents of, the chullpas at Honcopampa and elsewhere in Peru’s highlands. Furthermore, among the architectural remains described for Huamachuco are Early Intermediate Period buildings whose forms are similar to patio groups.

With the field work of 1987, a major question about Honcopampa has been answered. The little city was a Middle Horizon center with buildings characteristic of Huari. But the enigma of Early Intermediate Period settlements and political or­ganization in the Callejón de Huaylas remains unresolved. And now there is more than the sculpted statues in the Huaraz museum to be explained. Honcopampa architecture is a blend of Huari building traits with a little understood but obviously powerful northern tradition, in which we must find the antecedents of block and spall masonry as well as architectural spaces marked with megalithic lin­tels and great door jambs. Do these techniques derive from boulder and infill constructions, and were there boulder and infill buildings at Honcopampa before the patio groups, Dshaped compounds, and chullpas were built? Finally, who occupied Honcopampa, and by what authority did they draft the labor to build the city? It seems most likely that northerners, not Huaris, ordered the patio groups redesigned to include so many northern conventions. Only the Ac-1 patio group is true to Huari standards, but it yielded no occupa­tion debris.

The architectural ramifications of Honcopampa indicate a complex cul­tural history. They suggest that con­tinued research will reveal Honcom­pampa to have been a dynamic city where southern Huaris and native northerners worked out conflicts and accommodations in the construction of a new system of regional power and control.