Journal of Anthropological Archaeology 18, 201–239 (1999) Article ID jaar.1999.0336, available online at http://www.idealibrary.com on

Room Size, Organization of Construction, and Archaeological Interpretation in the Puebloan Southwest

Catherine M. Cameron

Department of Anthropology, Campus Box 233, University of Colorado, Boulder, Colorado 80309-0233

E-mail: [email protected]

Received April 28, 1998; revision received November 13, 1998; accepted December 20, 1998

The size of architectural space (floor area) is a variable that is readily preserved in the archaeological record, is easy to measure, and has been used in a variety of types of archaeo- logical interpretations, from determinations of room function to reconstructions of social orga- nization. The use of room size in the interpretation of Southwestern pueblos is reviewed here. Assumptions about the meaning of room sizes are explored using data on rooms, houses, and households from the historic Hopi Pueblo of Orayvi. Finally, room size is used, along with other variables, to examine the organization of construction at large, late prehistoric pueblos. Plan- ning and coordination of construction identified at some of these sites suggest a more central- ized social system like those of the historic and modern Eastern Pueblos. © 1999 Academic Press

Architecture communicates an abun- from the American Southwest, this paper dance of cultural information to archaeol- evaluates the use of architectural space in ogists reconstructing prehistoric social archaeological interpretation and demon- systems. Room size (floor area) is the ar- strates that the size of architectural space chitectural variable that is most likely to can be dependent on technological and survive in the archaeological record and is cultural factors, including group specific most easily measured. A basic assumption approaches to the organization of con- that underlies the use of domestic floor struction. area in archaeological interpretation is In the northern part of the American that people create architectural spaces of Southwest (Fig. 1) puebloan structures built an appropriate size for specific activities of stone or mud represent a crucial part of and for the numbers of people who will the archaeological record after A.D. 700. Ar- use those spaces. Technological factors af- chaeological interpretation of these struc- fecting room size and culture specific at- tures has been based largely on ethno- titudes toward the use of space are impor- graphic models developed from the study tant but often are not considered. of historic and modern Western Pueblos, The size of architectural space has been especially the work of Victor Mindeleff used as an index of population (Naroll (1989) at Hopi and Zuni during the 19th 1962), as one of a suite of variables that century (see Ciolek-Torrello 1985 for a dis- inform on the function of prehistoric ar- cussion of the ethnographic model of chitectural space (Hill 1970), as an indica- Puebloan room function). In this study, his- toric data from the Hopi (Western) Pueblo tor of changes in social organization 1, (Crown and Kohler 1994; Martin and of Orayvi * (Fig. 2) are presented in order to Rinaldo 1950), as a measure of social sta- examine archaeological expectations about tus (Wilk 1983), and, rarely, as an ethnic * See Notes section at end of paper for all foot- group marker (Baldwin 1987). Using data notes. 201 0278-4165/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved. 202 CATHERINE M. CAMERON

FIG. 1. The northern Southwest showing the locations of archaeological sites and regions, modern pueblos, and modern towns mentioned in the text. the meaning of differential room and house When room sizes at Orayvi are com- size. Surprisingly, the Orayvi data do not pared with those of large, late prehistoric closely fit the ethnographically based ex- pueblos an intriguing contrast is appar- pectations commonly used by Southwest- ent: room sizes at some pueblos, espe- ern archaeologists. Room size and function cially Eastern Pueblos, are very standard are only generally correlated. Families of in size, while others, especially Western different sizes and organization occupy Pueblos, are variable. Underlying the dif- houses of similar sizes, and the average ferences in room size among prehistoric number of square meters of space occupied pueblos is an apparent fundamental dif- by individuals is quite different from that ference in how construction (and presum- reported in cross-cultural studies. A num- ably other aspects of society) is organized. ber of reasons for these deviations from Standard sized rooms are apparently the expected patterns are suggested, both tech- result of construction techniques that al- nological and social, including the develop- low many rooms to be built as a single mental cycle of the domestic group, the ag- event (ladder-type construction). A dis- gregated and terraced nature of pueblo tinctive site layout is produced: rows of buildings, the status of the homeowner, and rooms outlining plazas of similar size. the availability of building materials and la- Construction of these sites was apparently bor. undertaken by a large number of people ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 203

FIG. 2. The Hopi Mesas in northeastern Arizona showing the location of Orayvi and other Hopi towns. and coordinated at a level above the cal expectations. Room size patterns found household. Such large-scale, coordinated at Orayvi are then compared to a sample of construction of houses has not been doc- prehistoric Eastern and Western Pueblos. umented at historic Western Pueblos Differences in the organization of construc- where home building was accomplished tion among prehistoric pueblos are dis- at the household level by women. In mod- cussed. The coordinated construction of a ern and historic Eastern Pueblos, house large number of houses in a single construc- construction undertaken as a group project tion event suggests that during the unset- by men has been documented and sup- tled last centuries of the prehistoric period ports suggestions (Ware and Blinman, in (A.D. 1250–1500) when abandonment and preparation) that these groups have a population aggregation are common pat- more highly integrated and centralized terns, people were able to move in relatively social system than the Western Pueblos. large, coherent communities. Finally, sug- The paper begins with discussion of the gestions about the historic connections technology of pueblo construction, espe- among prehistoric and modern social and cially technological constraints on room and ethnic groups are made. house size, and a review of the use of archi- tectural space in archaeological interpreta- THE TECHNOLOGY OF PUEBLO tion in the Southwest. The Orayvi data set is CONSTRUCTION introduced and the causes of variability in room and house sizes at Orayvi are exam- An understanding of the parameters ined within the framework of archaeologi- conditioning puebloan room size requires 204 CATHERINE M. CAMERON a review of construction methods. Puebloan the 10–12th century Chacoan Great Houses buildings typically consist of a set of at- where wide core-and-veneer walls were tached, dwellings called a roomblock. built with closely fit stones. The spaces Dwellings, which are occupied by a single between the stones were often filled with household (a task-oriented residence small sandstone spalls (“chinking stones”) group; Netting et al. 1984:xx), usually in- leaving almost no exposed mortar, pro- clude several rooms, are often multi- ducing remarkably stable, long-lasting storied, and built in a terraced fashion. walls (Lekson 1986). Chacoan Great Houses Pueblos are made of stone, earth, and are an unusual case, however. At other wood and are the product of many centu- sites in the northern Southwest, stone ries of architectural development. In most walls lack the elaborate masonry tech- parts of the northern Southwest, the move niques practiced by Chacoan builders. to above-ground dwellings from pit struc- Walls may be only one or two stones wide, tures occurred after A.D. 700. Large, con- although the compound masonry of the tiguous, multi-storied domestic buildings, Mesa Verde area is wider (c.f. Rohn 1977: like those that the Spanish encountered 241) At Hopi and Zuni in the late 19th when they first entered the Southwest in century, Mindeleff (1989:140) comments the 16th century, become common after 2 that masonry walls had “. . . been pushed A.D. 1200. It is these large, late pueblos to the limit of thinness” and were some- that this paper addresses. times only barely capable of holding the weight of several stories. Walls Coursed adobe walls are found exten- sively in the northern Rio Grande and also Pueblo walls are of stone mortared with in the Little Colorado region of the north- mud or of earth built up in courses by 4 hand (“coursed adobe”); the prehistoric ern Southwest but are rare in the Four- use of adobe brick has also been noted. Corners area. The construction method Wall strength and sturdiness can deter- uses stiff mud that is built up in courses; mine room size because walls usually bear each course is allowed to dry before the the weight of heavy roofs—the larger the next course is added (Stubbs and Stallings room, the heavier the roof. With multisto- [1953:26] provide an excellent description ried construction, lower story walls must of the technique). This process can pro- also support upper story walls. Upper duce a very sturdy wall that is functionally story walls are built directly on top of equivalent to stone construction for load- lower story walls, even when the upper bearing walls. Like masonry buildings, story is a later addition. Room size for coursed adobe structures can be multisto- upper and lower stories will, therefore, ried and terraced (Cameron 1998). E. almost always be identical, even though Charles Adams (personal communication, they were typically (based on ethno- 1997) suggests, however, that differences graphic information) used for different in load-bearing strength between adobe purposes (Ciolek-Torrello 1985).3 and masonry walls may result in smaller Stone construction is the most common adobe rooms. Recently, the use of adobe wall fabric in the northern Southwest. The brick has been recognized at a few prehis- strength and stability of stone walls de- toric sites in the northern Southwest pends on a number of factors including (Johnson 1992, Gann 1996). the ratio of stone to mortar, wall width, Adobe walls, whether coursed adobe or footings, and exposure to the elements. adobe brick, were thin. At Arroyo Hondo The apogee of stone wall construction in Pueblo and other northern Rio Grande the prehistoric northern Southwest were sites, walls ranged from 20 to 35 cm in ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 205 width. At Arroyo Hondo, lower courses cies of cedar whose odor and dense mor- were widest and width decreased with up- phology tends to retard decay. per courses (Creamer et al. 1993:16). At Methods of roof support are important Homol’ovi III in the Little Colorado re- determinants of room size as demon- gion, Gann (1996:95) found that adobe strated in James’ (1997) recent study of bricks were about 31 cm wide but de- historic Hopi and Zuni houses. He sug- creased in height in upper portions of gests that new methods of roof support walls. Thin adobe walls were apparently were developed after Euroamerican con- still the norm in late 19th century Zuni. tact that allowed larger rooms to be built. Mindeleff (1989:140) noted that “The Most descriptions of historic or prehistoric adobe walls are built only as thick as is puebloan architecture report that walls absolutely necessary, few of them being support the heavy roofs. At Hopi, how- more than a foot in thickness.” He ob- ever, James found historic illustrations served several collapsed rooms and com- and documents showing the use of inte- mented on “. . . the insufficiency of the rior posts and masonry “buttresses” or thin walls to sustain the weight of several piers that supported major roof beams stories.” across the width of a room. This technique took some of the weight of the roof off the Roofs walls and allowed two adjoining rooms to be combined into a single large room. At Zuni, larger rooms could be constructed Puebloan roofs are constructed of large by placing massive roof beams across the beams (vigas), generally laid parallel to the length of a room and supporting them on long axis of the room and socketed into very thick walls. Although a number of the walls (Ahlstrom, Dean, and Robinson 1991:629–630; Lange et al. 1993). Smaller scholars have documented room size in- secondary beams are laid perpendicular creases (among other architectural changes) to the primaries. This lattice is covered at Hopi and Zuni beginning in the last two with brush, twig, or reed “closing mate- decades of the 19th century and the early rial” and then covered with a thick cap of 20th century (Adams 1982; Ahlstrom et al. earth. The size of timbers available for 1978; Cameron 1999; Ferguson and Mills roofing may have been an important fac- 1987) James believes that room sizes in- tor imposing a limit on room size (Adams, creased much earlier. He dates the use of personal communication, 1997; Ciolek- both interior posts and larger roof beams Torrello 1985). At many historic and pre- to at least the mid-19th century. historic pueblos, Douglas fir and pon- Interestingly, prehistoric builders proba- derosa pine were preferred because these bly had knowledge of the techniques James trees produce long, relatively straight (1997) describes. Interior posts roof supports timbers useful for large roof beams were found prehistorically in Basketmaker (Ahlstrom, Dean, and Robinson 1978:29; pit structures (Cordell 1997), at Great Creamer et al. 1993). Pinyon pine and ju- Houses in Chaco Canyon (Windes 1987: niper produce shorter, smaller, more 288), in rooms at Pot Creek Pueblo the Taos crooked beams, although the decay- area (Wetherington 1968:26–33), and at Pe- resistant properties of these species may cos Pueblo (Kidder 1958:81; Fig. 29, 32). have led to their use at some sites. Sulli- Wetherington (1968:32) lists other archaeo- van (1974) has suggested that juniper may logical examples and suggests that upright have been preferred for roof construction post roof supports usually occur in late pe- in storage rooms at Grasshopper Pueblo riod pueblos in the Northern Rio Grande, in east-central Arizona because it is a spe- but also reports their occurrence at Pueblo I 206 CATHERINE M. CAMERON sites in the Mesa Verde region. These ar- century site in the northern Rio Grande chaeological examples suggest that Puebloan region, Creamer and her colleagues (1993: builders had the technology to build larger 139) suggest that the large Component I rooms, but chose not to. James (1997) sug- occupation may have depleted surround- gests that the introduction of Euroamerican ing timber forcing the later Component II tools, domestic animals, transportation occupants to cut younger, smaller trees methods, and Spanish-introduced concepts and in some cases to forgo vigas alto- of the use of space combined to make larger gether. Upper stories were common in rooms appealing to residents of Hopi and Component I but absent in Component II, Zuni. possibly because adequate roof supports could not be found. Materials and Labor At the Hopi village of Walpi different species of trees were used for different The availability of building materials for parts of the roof. Primary beams are of construction was almost certainly a factor pine, fir, or cottonwood, secondary beams conditioning room size in the prehistoric are of juniper and closing material is gen- Southwest. Good quality stone and mud erally willow or reed (Adams 1982:51; Ahl- for mortar or coursed adobe construction strom, Dean, and Robinson 1978). Pine were necessary for sturdy walls. At Zuni, and fir were the most difficult to procure. Mindeleff (1989:139) noted that usable At Orayvi, the University of Arizona’s outcrops of stone were some distance Laboratory of Tree-Ring Research found away and masons used stone sparingly. At that sources of preferred timber near the both Hopi and Zuni, stone was frequently village—pine or Douglas fir—were appar- recycled from earlier buildings (Cameron ently exhausted by the late 1700s (Doug- 1999; Ferguson and Mills 1987:253; Fergu- son et al. 1990:110; Krober 1917:195). At lass 1929:754). By the late 19th and early prehistoric Arroyo Hondo pueblo, Creamer 20th centuries, however, wagons and draft and her colleagues (1993:14–15) suspect animals had become common at Hopi. As that available outcrops of stone were ex- a result, larger beams from greater dis- hausted after the first rooms were built, so tances could be procured and the size of later rooms were constructed of coursed rooms increased (Cameron, in press). Like adobe. The construction of coursed adobe stone and adobe, roof beams were fre- walls requires quantities of good quality quently reused (e.g., Ahlstrom, Dean, and mud and abundant water. Abundant wa- Robinson 1978:37–39). ter is also necessary for mortar in stone The organization of housebuilding ac- walls and house-building was typically tivities (procuring timbers and stone, pro- undertaken when rainwater was likely to curing and mixing adobe or mortar, con- be available. structing walls and roofs) may have been Perhaps most costly to procure were an important determinant of puebloan roof timbers, especially the large vigas room size. Most ethnographic accounts which had to be cut, seasoned, and hauled report Pueblo housebuilding was under- to the building site. Access to stands of taken at the household level by male kin timber may have an important factor con- (Jorgenson 1980:152, Cu-104/V-229). This ditioning room size. At prehistoric Grass- was apparently not the case among the hopper Pueblo in northeastern Arizona matrilineal Hopi where women owned the (discussed below) Ciolek-Torrello (1985) houses. Mindeleff (1989:101) reported that believes that rooms are large partly be- Hopi men said they built the houses while cause of the ready availability of large women only plastered them. However, he trees. At Arroyo Hondo Pueblo, a 14th observed that a woman and her female ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 207 kin actually accomplished much of the particular patterns seem to be associated house-building with only occasional male with certain room size characteristics. Ad- assistance. Cameron (1999) shows a house ams (1991) proposes that settlement lay- under construction at Orayvi with two out for large, late pueblos is of two main women laying stones and four to six types: enclosed plaza and street oriented women mixing mortar. Only one figure (see also Reed 1956). In enclosed plaza may be male—an individual bent over the pueblos, roomblocks are built up around roof. James (1997:435) also presents evi- one or more plazas which were used for dence that women at both Hopi and Zuni domestic activities and religious ceremo- built and owned their homes. nies. In street-oriented pueblos, like Among modern Eastern Pueblos, where Orayvi, plazas are the areas (“streets”) be- residence is generally patrilocal or neolocal, tween linear roomblocks, and were also men often own the houses (Hill 1982:20; Or- used for domestic and ceremonial pur- tiz 1979:290) although in some cases either poses. Roomblocks in street-oriented sex may own and inherit a dwelling (Hoebel pueblos tend to face southeast, a direction 1979:411; White 1942:196). Where house- that ensures maximum solar efficiency building is discussed, construction is often and is an important ritual direction for reported to be by male relatives or is a co- many Pueblo people. Terracing in both operative project assigned to males. For ex- types of settlements allowed upper story ample, at Santa Clara Pueblo, Hill (1982:73– living rooms to receive maximum light 74) notes that houses are built by a father and ventilation, while occupying a defen- and his sons, a group of brothers, or the sive position atop a platform of lower affinal relatives of a husband and wife (see story rooms. also Ortiz 1979:191). Parsons (1936:52) re- In this paper, a dichotomy is recognized ports that Northern Tiwa house construc- within prehistoric enclosed plaza pueblos. tion is a cooperative project of male family “Linear plaza” pueblos are characterized members. House building was not exclu- by a few rows of rooms which define sev- sively a household activity however. Adolph eral, relatively large, generally enclosed Bandelier reported that house construction plazas. “Agglomerative” layouts (Mills was organized at the village level (Lang 1998) consist of dense clusters of rooms and Rilley 1966:97). At Santo Domingo he surrounding small plazas. Some linear was told that “When a house is built, the plaza pueblos were built using a tech- Pueblo details ten men to build the walls nique that has been called “ladder type” and also ten to construct the roof (Lange and Rilley 1966:97).” At Santa Clara Pueblo, construction (Cordell 1996; also “aggre- Hill (1982:73) reports that a cacique might gate construction,” Creamer et al. 1993: assign a house-building task (in hardship 16). At Arroyo Hondo Pueblo near Santa cases) to the male membership of a moi- Fe, this method was used to build many ety or a governor might ask the village to rooms at the same time by constructing cooperate in such a project. The rebuild- two long walls first and then subdividing ing of Acoma in the 17th century seemed the space between them with cross walls to have been a planned endeavor in which (Creamer et al. 1993:16). This technique much of the village probably took part creates a number of rooms in one con- (Robinson 1990). struction episode and rooms tend to be very similar in size because at least one The Organization of Settlements dimension (width) will be the same for each room. Rooms constructed in pueblos Settlement layout is important to an ex- with agglomerative layouts seem to have amination of puebloan room sizes because been individually built, although, of course, 208 CATHERINE M. CAMERON many rooms could have been under con- important for determining household size. struction at the same time. Although house size might seem to be Ladder type construction suggests that directly related to social needs (the addi- house building was an activity undertaken tion of a room to accommodate a growing by a large group of people, rather than an family), there are technological limits on individual family. Ladder type construc- Puebloan house size. As noted above, the tion creates a single row of rooms and closely built houses typical of both en- linear plaza pueblos generally consist of closed plaza and street-oriented pueblos several rows of rooms surrounding a limit a homeowner’s ability to add rooms plaza. Ample archaeological evidence horizontally. Additional rooms could be suggests that dwellings cross-cut these added vertically, but there are limits to the rows, so that families have a plaza facing number of stories that can be accommo- room(s), as well as interior and rear dated using puebloan construction tech- rooms. Like the construction of a modern nology (Cameron 1999). Of course, the apartment building, ladder type construc- construction of a new roomblock avoids tion can create many houses through the these difficulties. Throughout this paper coordinated construction of many cells. both room size and house size are dis- Agglomerative construction does not re- cussed. Both are important, related pa- quire a similar level of coordination in rameters of puebloan use of space. room construction; construction can be accomplished room by room. ROOM SIZE AND The linear plaza settlement form also ARCHAEOLOGICAL suggests coordination above the level of INTERPRETATION IN THE the household (see Cordell 1998:27 and SOUTHWEST Kidder 1958:63). Unlike agglomerative pueblos, the closely packed dwellings of Even though there are clear technolog- this enclosed plaza design limits the easy ical constraints on puebloan room size, addition of individual houses. Instead, ex- technology has generally been given little pansion consisted of using roomblocks to consideration in archaeological studies of create a new plaza or, occasionally, adding room size (but see Ciolek-Torrello 1985). another row of rooms on the interior Instead, the size (floor area) of Pueblo (plaza-facing) side of the roomblock (e.g., rooms has been assumed to reflect social Creamer et al. 1993, Figs. 7.5–7.7). This requirements. As a result, room size has pattern of growth suggests that commu- been used in two basic, but intertwined nities that build these sites share an un- types of archaeological interpretation: es- derstanding about the size of plaza space timates of prehistoric population and re- that would be surrounded by roomblocks construction of prehistoric social groups. (Cameron 1999). A coordinated building For both, the identification of room func- effort using the ladder construction tech- tion—with room size usually an important nique would be a likely approach to con- variable—is a first step. struction. Once rooms are assigned to a functional type (generally using other indicators as Rooms and Houses well, such as floor features and artifacts) rooms of different types can be combined A final technological consideration in- into houses. The identification of houses volves house size. In the Southwest and permits population estimates because the elsewhere, house size is reported either as number of houses can be multiplied by an numbers of rooms or the combined floor average household size. Population can area of rooms. House size is especially also be estimated directly, dividing house ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 209 size by a constant number of square tion through an analysis of artifacts and meters per person (Naroll 1962; Brown ecofacts present in rooms. The methods 1987). House size has been used to study used in Hill’s study have been heavily of the make-up of the household, changes criticized (especially his lack of consider- in household composition over time, and ation for the processes that result in the differences in household wealth and/or deposition of artifacts on floors), but not status (although seldom in the South- his equation of room size with room func- west). Room size has also been used, tion (except Ciolek-Torrello 1985). rarely, in the identification of prehistoric E. Charles Adams (1978, 1983) also tested ethnic groups. Each of these uses of room the ethnographic model relating room size size are discussed below. to room function. He determined room function for abandoned rooms at the Hopi Room Size and Room Function Pueblo of Walpi through interviews with informants and then collected architectural Early scholars noted that very small and artifactual data from these same rooms. rooms were appropriate for only a limited These data were used to define several cat- number of activities (primarily storage) egories of room use. Rooms of different uses and room size eventually became one of a at Walpi were found to differ significantly in group of attributes, including floor fea- size. Rooms from two prehistoric sites, in tures and artifacts, that were used by ar- the Kayenta area (Betatakin) and on Mesa chaeologists to define room function. Verde (Badger House), were categorized by Drawing heavily on Cosmos and Victor room function using the architectural at- Mindeleff’s architectural studies among tributes developed at Walpi (room story lo- the Hopi (Mindeleff 1900; Mindeleff 1989), cation, number of interior wall doors, pres- most archaeologists, especially those studying Western Pueblos, have assumed ence or absence of doors in exterior walls). a close relationship between room size Adams found significant differences in size and room function. At least one study of between rooms of different functions. Other room function at a prehistoric Western archaeologists working in the Western pueblo has questioned this relationship, Pueblo area have continued to find room however (Ciolek-Torrello 1985). size an important indicator of room function Hill’s (1968, 1970) study of Broken K (Jorgensen 1975; Lowell 1991; Hansen and Pueblo, a 12th–13th century site in east- Schiffer 1975; Sullivan 1974), often identify- central Arizona was the first archaeologi- ing a bi- or tri-modal distribution of room cal study to test the assumption that floor sizes with the largest rooms assumed to area (among other variables) was an indi- have been used for ceremonial activities cator of room function. He articulated the and the smallest for storage. commonly held archaeological assump- Not all archaeologists are convinced tion that “...a large room containing a that room size is a good indicator of room firepit and mealing bin is called a living function, however. In his comprehensive room or habitation room, while a small study of room function at Grasshopper room without a firepit is called a storage Pueblo, Ciolek-Torrello (1978) did not find room (Hill 1970:37).” The largest rooms at floor area to be useful. Rooms at Grass- Broken K Pueblo were assumed to have hopper are relatively large (16.3 sq m on been used for ceremonial purposes. Hill average) in comparison to other sites such assigned rooms to different types based as Broken K Pueblo.5 Ciolek-Torrello at- on their size and the presence certain dis- tributes this to the ready availability of tinctive floor or wall features, then con- large trees in the Grasshopper area for use firmed the architecturally assigned func- as roof beams. Instead of a bi- or tri-modal 210 CATHERINE M. CAMERON distribution of room sizes, Ciolek-Torrello household to arrive at site population. Al- found that room floor areas at Grasshop- ternatively, the number of living rooms per were normally distributed about the can simply be counted, assuming each mean (1985: Fig. 3). While Ciolek-Torrello family occupies a one living room (Sch- observes that, in general, storage rooms langer 1986:570). Room size is an impor- are smaller than habitation rooms, the dif- tant variable in this equation because it is ference in size is slight and range of sizes so frequently used to define room func- within types is great. He explains the lack tion which is essential to both methods of of correlation between room size and estimating population. room function partly as a result of con- In 1962, Naroll proposed a constant, stant remodeling at Grasshopper with a cross-cultural relationship between quan- resulting change in room function (a con- tity of roofed space and population. clusion also reached in Sullivan’s 1974 “Naroll’s Constant” resulted, in the study). He also cites a technological con- Southwest, in the occasional use of room sideration: Grasshopper is a multistoried size as a means of calculating population pueblo where upper and lower stories (Drager 1976; Snow 1976:A224, see also were the same size even though (based on Bullard 1962:123). Naroll’s original pro- ethnographic analogy) they should have posal was refined (Brown 1987; Leblanc been used for different purposes (habita- 1971) and new calculations were sug- tion and storage, respectively). gested for use in the Southwest (Clarke While the constant rebuilding and re- 1974) and elsewhere (Castleberry 1974; modeling observable in prehistoric and Kolb 1985; Webel 1979) but it is still not a historic pueblos might obscure relation- commonly used index of population size ships that once existed between room except as a general measure for unexca- function and room size, ethnographic data vated pueblos (e.g., Wilshusen 1995:70; from historic pueblos, as well as pueblo- see also Schlanger 1986:579 who develops like structures outside the Southwest, a project-specific constant). At Orayvi suggest that this is not always the case (discussed below), pueblo house size ap- (Cameron 1999). These data indicate that pears to vary widely from the constants habitation rooms tend to be modified proposed by Naroll and others. This may more frequently than rooms of other types be a result of technological rather than (but see Lowell 1991:36–38) because they social factors, however. Because not all are used more intensively and deteriorate space in a pueblo house is active use, the faster. Most often they continue to be used fit may be closer than it seems. as habitation rooms, however. In fact, change in structure function is commonly Room Size, House Size, and Social a result of structural deterioration and de- Organization cisions not to rebuild (i.e., a dilapidated living room used for storage). Several studies have attempted to ac- count for changes in room or house size at Room Size and Population Estimates pueblos by relating these measures to economic activities, the size and organiza- The most common method of estimat- tion of households, to community organi- ing population at excavated prehistoric zation, or social interaction. In the Cochiti pueblos involves assigning rooms to func- area near Santa Fe, New Mexico, Hunter- tions and grouping them into “suites” that Anderson (1979:182) found greater vari- are assumed to have sheltered a house- ability in room sizes at larger sites (dating hold. The number of suites is multiplied to the Pueblo IV period) and suggests that by a constant number of individuals per these sites would be used for a wider va- ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 211 riety of functions. At small sites (Pueblo III Creek Pueblo). They observed a change and Pueblo IV farmsteads), Hunter- through time in the frequency of these Anderson suggests that smaller room household types at Grasshopper that sug- sizes may be related to a decrease in the gested a change in both household size number of people or amount of time spent and use of the pueblo. Households occu- at farmsteads, or a lack of certainty about pying multiple-room houses were larger whether the sites would be reused. and more established; these houses were A change in economic activities caused found in the main part of the pueblo. changes in room and house sizes at Pot Families occupying only a single, multi- Creek Pueblo, a site located near Taos functional rooms were typically located in New Mexico, that dates primarily to the outlying roomblocks. Single room houses late 13th and early 14th centuries. Crown appear late in the history of the pueblo and Kohler (1994:111) found an increase in and may have been used only part-time size for both rooms and residence units at during the last occupation episode. The the Pot Creek through time. They at- change in household size over time is sup- tribute this to an increase in household ported by a decrease in cooking hearth size brought on by population aggrega- size that would have serviced a smaller tion (see also Holschlag 1975). Population family (Reid and Whittlesey 1982:696; see aggregation resulted in an expanded de- also Ciolek-Torrello and Reid 1974). pendence on agricultural products but an In a comprehensive study of prehistoric increase in distance to fields that may Pueblo households in the Southwest, have caused scheduling problems related James (1994) has linked room size with to agricultural work versus child care. patterns of residence and descent. Using This conflict could be solved by a shift to measurements from almost 9000 rooms at extended families. Extended families, they 350 sites in 11 areas of the Southwest, he believe, would necessitate larger houses associates matrilocal residence and matri- with larger rooms. Kulishek and others lineal descent with Puebloan groups on (1994) note that Crown and Kohler rely the Colorado Plateau and adjacent re- heavily on contemporaneity of construc- gions, including the northern Rio Grande. tion in defining residence units. While not These groups, prehistorically, tended to questioning Crown and Kohler’s postu- have rooms that were under 10 sq m in lated increase in household size, Kul- size. Below the Mogollon Rim, groups in isheck and his colleagues point out that the southern Southwest had very large simply because a group of rooms were rooms, more than 21 sq m, that James built as a unit does not mean that those associates with patrilocal residence and rooms were occupied by a single family. patrilineal descent. He associates resi- In other words, they suggest that Pot dence and descent patterns with differ- Creek experienced a change in building ences in subsistence practices, especially technology, rather than (or in addition to) corn storage, but also food preparation. a change in social organization.6 Two studies see changes in room or Change in household type was also house size as reflecting changes in com- identified at Grasshopper Pueblo and was munity organization or social interaction, associated with changes in how the rather than modal family size. Dohm pueblo was used. Here, Reid and Whit- (1990) proposes that the spatial proximity tlesey (1982) identified two household of houses is a determinant of house size. types: those occupying multiple rooms Using data from modern and historic and those occupying single rooms (see Eastern and Western Pueblos, she found Lowell 1988 for a similar study at Turkey that with increasingly aggregated popula- 212 CATHERINE M. CAMERON tions, house size (measured both by num- household. Where neolocal households ber of rooms and quantity of roofed area) predominated, homeowners were more in pueblos increases. Dohm ties this likely to use the size and decoration of change to an increased need for privacy as their houses to express their status within people aggregate into close-packed, con- the community. Other studies show that tiguous pueblos. Privacy is sometime in village-level, egalitarian societies, achieved, she suggests, by subdividing house size can be an indicator of differen- rooms, thereby increasing the number of tial status or wealth, even when the dis- rooms in a house and decreasing their tinction is not consciously made. Wilk’s size. (1983) study of the Kekchi Maya found In a more detailed study of social inter- that individuals with greater social status actions, Ferguson (1996:132–135) com- had houses of larger size, although the pared room sizes at modern Zuni Pueblo Kekchi “. . . denied vehemently that the for vernacular (user built) houses and size of a house is any reflection of a house- houses built by the Federal government’s hold’s wealth, status, or power (1983:103).” Department of Housing and Urban Devel- In Iranian Kurdistan, Kramer (1980) found opment (HUD). He found that both house a positive correlation between compound types were of similar size (about 120 sq. size and wealth. In the Ecuadorian Ama- m), but that vernacular houses had fewer, zon, Brenda Bowser’s study of the small- larger rooms, while HUD houses had a scale, segmental, egalitarian village of greater number of smaller rooms. These Conambo found that house size (floor differences reflect and contribute to a re- area) is strongly correlated with social sta- formulation of the norms of Zuni society. tus (personal communication, 1997). High For example, because kitchens in HUD status households in Conambo are also houses are too small to hold the large beginning to roof their houses with tin, an numbers of people that traditionally gath- expensive and ineffective material, but ered in these spaces, Ferguson believes one that is an indicator of success in con- that the role of women is changing in Zuni tact with the outside economic world. society. He observes that the small HUD In the Southwest only a handful of stud- kitchens “. . . weakens the social role of ies explore the relationship between social women by spatially constraining their so- status and pueblo house sizes. At Grass- cial interaction during family gather- hopper Pueblo, Reid and Whittlesey ings....The role of women shifts from (1982:696) suggest that multiroom houses being the social interactional focus of so- cial activity to a role more characteristic of may have been occupied by older, more a servant who caters to the needs of other established, “wealthier” families. Simi- groups of people in the house (1996:142).” larly, at Turkey Creek, Lowell (1988:91) believes that families occupying only a House Size and Social Status single room might have been poorer than families with multiroom houses. In the Blanton’s (1994) recent cross-cultural Mogollon area, Lightfoot and Feinman study of houses and households in peas- (1982) suggest that large, early pit struc- ant communities found that house size tures found in some villages, housed vil- was one element in a complex system of lage leaders with greater access to exotic architectural communication. In cultures goods. where large extended family households In most Southwestern studies, however, were typical, house layout and decoration Southwestern archaeologists link large was directed toward pressuring younger rooms or houses with the location of cer- members to sacrifice for the good of the emonial activities, not with the homes of ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 213 high status individuals—even though in archaeologists have begun to recognize an most middle-range societies, social rank is association between room size and ethnic closely tied to roles within a ceremonial or or cultural affiliation. For example, James’ religious hierarchy. For example, in (1994) pan-Southwestern study (described Puebloan society, high status individuals above) found that room sizes are associ- are usually religious leaders and at the ated with different patterns of residence Western Pueblos, these individuals would and descent. He associates these patterns be clan leaders—especially the leaders of with differences in the organization of the highest ranking clans. In his study of room space for corn storage and process- Walpi Pueblo, Adams (1983) included clan ing among communities above and below houses—the habitation room of a clan ma- the Mogollon Rim. James believes that triarch—with “religious rooms” and ob- patterns of residence and descent, and served that religious rooms were larger their associated architectural correlates, than rooms of other types. Adams did not become established early and are quite associate these differences in room size stable and recognizable for cultural groups with status differences however, but with in different parts of the Southwest. In fact, religious use. he questions a often-cited migration from Similarly, during the Chacoan era in the the Kayenta area to the 13th century site of northern Southwest, there is long-standing Point of Pines, Arizona (Haury 1958) be- recognition that rooms in Chacoan Great cause room sizes of the immigrant group Houses are much larger than rooms in con- are not similar to those in the Kayenta temporary unit pueblos. Many Chacoan area (1994:161–162). rooms lack evidence of domestic activity, In a similar but less comprehensive study however. Great Houses have most fre- Baldwin (1987) also uses room size to define quently been interpreted as ceremonial or prehistoric Southwestern cultures. Building community structures or even storehouses, from classic studies of proxemics and archi- not primarily as the residences of high sta- tecture (Hall 1966, 1968) and the theory of tus individuals (Lekson et al. 1988). Finally, architectural design offered by McGuire Lightfoot and Feinman’s (1982) suggestion and Schiffer (1983), he suggests, with regard that large early large pit structures in the to pueblo buildings, that “. . . the largely Mogollon area, housed village leaders with subconscious, culture-specific proxemic greater access to exotic goods was heavily system of the builder is a major determinant criticized for assuming that they had cor- of some properties of rooms, particularly rectly identified the archaeological corre- lates of status (Cordell 1984:319–320). Large the size and shape of dwelling spaces (1987: structures in the Mogollon area are, instead, 166).” He uses cumulative frequency curves believed to reflect the location of ceremonial (ogives) of room sizes to distinguish popu- activities. lations with culturally specific building pat- terns at sites in Chaco Canyon, the Mesa Room Size and Cultural Identity Verde area, and the northern Rio Grande. Where cumulative frequency curves over- Because the types of indoor activities lap, Baldwin (1987:169,173) postulates immi- undertaken by village-level horticultural- gration and cultural mixing. ists practicing corn agriculture are broadly In the Western Pueblo area, during the similar (food processing, cooking, storage, Pueblo IV period (A.D. 1275–1550), ar- tool manufacturing and maintenance, eat- chaeologists have begun to link changes ing, sleeping) the size and shape interior in room sizes to the arrival of immigrants space in Southwestern pueblos might be to the area and to internal population assumed to be similar. Recently, however, movements. In the Silver Creek area of 214 CATHERINE M. CAMERON east–central Arizona, Mills (in press) finds 2). It is an ancient pueblo, perhaps first a decrease in room size after A.D. 1325 occupied in the 1100s. The Hopi Pueblos that she associates with a population in- had little contact with Euroamericans un- flux. At the Homolovi pueblos, located on til the late 1800s, which is why they are the middle Little Colorado River near favored by archaeologists as models for Winslow, Arizona, Adams (1998; see also prehistoric pueblos (but see James 1997). Adams 1996), has found that immigrants Orayvi was the largest of the Hopi villages converged on this area during the early in the late 1800s and reportedly the most Pueblo IV from both the north and the conservative. In 1906, the pueblo was the south. Room sizes at the pueblos occupied scene of a dramatic factional split that re- by each of these groups were different. sulted in the departure of half of the res- Both Homolovi and the Silver Creek cases idents. These emigrants founded two new are discussed further below. villages nearby. James (1994) is the only scholar to pro- The split and other aspects of Orayvi pose material causes for room size differ- social organization have been studied by ences among groups in the Southwest: numerous scholars. One of the first and agricultural practices that result in differ- most comprehensive studies of Orayvi was ences in patterns of residence and de- undertaken by Mischa Titiev in the 1930s scent. Nevertheless, the examples pre- (Titiev 1944). His work, a study of social sented here indicate that a number of organization at Orayvi, has served as a Southwestern scholars are beginning to baseline for virtually all other studies of explore the importance of culturally based the village, including this one. Other schol- conceptions of the use of space in deter- ars have focused more directly on the mining room size and, presumably, house causes of the split, attributing it to a vari- size. ety of factors ranging from acculturative pressure on the Hopi by the U.S. Govern- THE ORAYVI DATA SET ment (Clemmer 1978:58, 76) to an inten- Historic records from Orayvi, including tional effort by Orayvi’s religious leaders maps and census materials provide an un- to overturn Orayvi’s politico-religious or- paralleled set of data for examining the der (Whiteley 1988). Most recently, in a historic use and modification of pueblo detailed study of social stratification at structures (Cameron 1999). For the Orayvi, Jerrold Levy determined that eco- present study, the most important aspect nomic distress, caused by a deteriorating of these data is that they provide a link environment, was the ultimate cause of between the pueblo house—a set of rooms the split (Levy 1992). Levy found that the occupied by a household—and the in- families who were forced out of the village dividuals who made up the household. during the split were almost all from land- The exact articulation of house and house- less and low status lineages. Because of hold is not, to my knowledge, available for the work of Titiev and other scholars, a any other 19th century Southwestern wealth of architectural and census data pueblo. These data provide an important are available for Orayvi. These data are foundation for evaluating commonly held derived from a number of sources. A archaeological assumptions about the highly accurate plan of Orayvi was pro- meaning of pueblo room sizes and inter- duced by Victor Mindeleff in 1887 (Mind- pretations of temporal and spatial differ- eleff 1989); this map was used to calculate ences in room sizes. floor area for rooms and houses in this Orayvi is located on Third Mesa, the study. During the early 1930s, Titiev westernmost of the three Hopi Mesas (Fig. worked with elderly informants at Orayvi, ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 215 to produce a reconstructed “census” of ful set of data for examining the relation- the pueblo for the period around 1900, just ship between the Orayvi house and before the split. He used a tracing of Min- household at this historic pueblo. deleff’s map and recorded the members of each household and the rooms they had Orayvi Room and House Size occupied. Titiev used the census in his well-known monograph on Orayvi (1944), In 1887 when Mindeleff mapped Orayvi but his detailed census notes were never it was a large settlement of about 1100 published. rooms arranged in long, terraced room- In the 1970s, Levy discovered Titiev’s blocks of attached houses that generally census notes at the Museum of Northern faced southeast (Fig. 3). This “street-ori- Arizona (Titiev n.d.) and undertook a new ented” layout (Adams 1991:103) was com- study of social organization at Orayvi and mon during the late prehistoric and his- the causes of the split (Levy 1992). Levy toric time periods. Most houses were only one room wide, with approximately three constructed a computer file that included ground floor rooms, two second story all of Titiev’s census information, but also rooms, often a third story room, and occa- included data from the 1900 U.S. census. sionally a fourth story room (Fig. 4; Cam- Titiev’s census was generally accurate, but eron 1996a). The areas between room- it occasionally included in a household blocks were used as plazas for domestic individuals who were recently deceased activities. The Main Plaza was the scene of in 1900 or were not yet born. The 1900 U.S. much ceremonial activity, as, occasionally, census listed the individuals in each was the Snake Dance Plaza (Fig. 3). household at a single point in time. Levy’s Titiev (n.d.) recorded 188 households at computer census files included the name Orayvi in 1900 and the room numbers that of the female household head, her age in made up the houses they occupied (Fig. 5), 1900, and the number and relationship of but I was not able to use quite all of these other people living in her house (as well households to examine the relationship as other information). between house and household size. Be- Levy generously permitted me to use tween the time Mindeleff mapped Orayvi his computer census in a study of archi- and 1900 (the date of Titiev’s recon- tectural change at Orayvi (Cameron 1991, structed census), 34 new houses had been 1996a, 1999). These data allowed me to added to the village. Titiev sketched the link the spaces on Mindeleff’s map that approximate location of new houses on Titiev had defined as the lodgings of a his copy of Mindeleff’s map, but did not single household with the individuals show number of rooms or stories accu- who lived in those rooms in 1900. Titiev rately. When these new houses could be also recorded households living in new seen in photos, the number of rooms houses that had been built during the 13 could sometimes be counted (Cameron years between Mindeleff’s 1887 map and 1999). Floor area for rooms and houses the 1900 census. I used a series of historic that are not on Mindeleff’s map could not, photographs collected from numerous of course, be calculated. museum and university archives to chron- The number of rooms for each house at icle change to the houses over time, doc- Orayvi was calculated by adding together umenting architectural change at the all of the room numbers listed by Titiev as pueblo over a period of more than 70 the residence of a single household, then, years, from 1872 to 1948. Mindeleff’s map, for each room, adding the number of sto- Levy’s census, and my study of architec- ries recorded on Mindeleff’s map. Adjust- tural change combined provide a power- ments were made for architectural change 216 CATHERINE M. CAMERON

FIG. 3. Map of Orayvi redrawn from the original produced by Victor Mindeleff in 1887 (Mind- eleff 1891). The named structures are kivas. Courtesy Smithsonian Institution, National Anthropo- logical Archives. observed in photographs for those houses room size at pueblos in the northern that could be seen in photos—deleting or Southwest was below 10 sq m prior to adding rooms that had been abandoned, Spanish contact. He suggests that even demolished, or newly built between 1887 though the Hopi and Zuni were not in and 1900.7 direct contact with Euroamericans from Floor area measurements were ob- the late 16th century to the late 19th cen- tained for 1001 rooms in 154 houses (Table tury, their architecture, including room 1). Rooms at Orayvi ranged in size from size, was affected by Euro-American pres- about 1 sq m to almost 60 sq m, averaging ence in the Southwest as early as the mid- almost 16 sq m. James (1997) has argued, 19th century, if not before. As discussed using abundant archaeological data, that above, he documents the use of new ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 217

FIG. 4. A cross-section of an Orayvi house. Most houses at Orayvi were only one room wide. Rear, lower story rooms may have been little used except as dead storage or an architectural platform. Modified from Adams (1982, Fig. 18). methods of roof support to combine one of Titiev’s households) ranged in size rooms into a single large space. He notes from less than 6 sq m (a small, one room that these changes were much greater at house that may have been undergoing ren- Zuni than at Hopi because the Hopi were ovation) to over 300 sq m. Average house more conservative and had restricted ac- size was more than 100 sq m, approximately cess to large roof beams. the size of a typical 1950s American ranch- Average room size at Orayvi does sug- style home. These figures conform with gest that by 1887 rooms at Orayvi were cross-cultural studies that suggest that in larger than would have been the case pre- societies with matrilocal residence, house historically. The degree of change is diffi- size can be expected to range between 80 cult to assess because of the method used and 270 sq m (Divale 1977 cited in James to measure room sizes: using the roofed 1994:19; patrilocal societies reportedly have area on Mindeleff’s map means that inte- smaller houses: 15 to 43 sq m). rior walls that do not protrude above the highest roof cannot be seen. While James The Fit between Houses and Households (1997) suggests that large rooms were cre- ated by removing interior cross walls, it is More than two-thirds of the households also possible that some of the largest, on Titiev’s census could be identified in lower story rooms recorded at Orayvi the 1900 U.S. Census (n ϭ 115), allowing were subdivided by cross-walls that could an exact count of the individuals living in not be seen on Mindeleff’s map (see Note particular houses. In 1900, the number of 5). Even if the amount of domestic space at people living in Orayvi houses ranged historic Orayvi was somewhat greater from one to 12 with an average of between than at prehistoric pueblos, however, the 5 and 6 people (Table 1; Cameron 1996a). relations between people and space dis- In spite of ethnographic accounts (for ex- cussed here are still pertinent to the inter- ample, Eggan 1950) and large house size, pretation of archaeological cases. Orayvi households were not primarily Houses had between 6 and 7 rooms. The matrilineal extended families (Table 2). floor area of houses (the combined area of More than half the households at Orayvi each room and each story associated with in 1900 were nuclear families (mother, fa- 218 CATHERINE M. CAMERON

FIG. 5. Mindeleff’s 1887 map of Oraibi annotated by Mischa Titiev for the period around 1900. Room numbers were assigned by the Laboratory of Tree-Ring Research during the Second Beam Expedition. The map shows new rooms and new roomblocks built after 1887. ther, and unmarried children); extended common activities at the Hopi villages, but families only made up 22% of the house- in the late 19th century the introduction of holds (Cameron 1996a, 1999; Levy 1992; Euroamerican technology and architectural Whiteley 1988). Remaining families were style and apparent population growth at couples without resident children or a va- Orayvi may have increased the pace of riety of other household types. building and rebuilding. Two-thirds of the Examination of the fit between houses 59 houses I observed in my study of archi- and households at Orayvi is complicated by tectural change were modified between evidence that many houses were being 1887 and 1900; rooms were rebuilt, aban- modified during the 13 years between Min- doned, or demolished and new rooms were deleff’s map and the 1900 census. Recon- built (Cameron 1999). struction and remodeling had always been It is likely that even if room size in- ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 219

TABLE 1 Orayvi House and Household Size

Mean Minimum Maximum Standard deviation Number

Room size (sq m) 15.6 1.2 58.6 9.4 1001 House size (Area in sq m) 103.1 5.7 306.4 59.0 154 House size (# rooms) 6.5 18 1 3.4 160 Household size 5.5 1 12 2.2 110 creased during historic times, as sug- ilies did not have significantly larger houses gested by James (1997) house size did not than nuclear families at Orayvi (Table 3). increase. James (1997:435–436) proposes Extended families were significantly larger that room size increased through the use than nuclear families (t value ϭ 3.941, p Ͻ of internal roof supports that allowed two .0007), but they did not have significantly rooms to be combined into one (an inter- larger houses whether measured by num- nal wall could be demolished). Combined bers of rooms or as total house area (num- rooms were observed at Orayvi during the ber of rooms: t value 1.045, p Ͼ .0383; mean late 19th century (Cameron 1999). This area of house: t value .19, p Ͼ .847). Al- process would create a larger room, but though the sample size is small, couples not a larger house. In fact at Orayvi, al- actually had the largest houses at Orayvi though some rooms increased in size dur- and “other” household types had the small- ing the years between 1887 and 1990, and est houses (Table 3). living rooms were moved from upper to The lack of fit between house and house- lower stories, I found evidence that hold at Orayvi may be attributed both to houses generally stayed within the same historical factors and the developmental cy- architectural footprint (except for evi- cle of the domestic group.9 Houses at dence of expansion to the southeast; see Orayvi are highly variable in size (Table 3). below). For purposes of floor area calcula- They have long use-lives, but in spite of tions discussed below, I argue that house evidence of frequent remodeling and re- size (floor area) did not change substan- 8 building, family configuration almost cer- tially between 1887 and 1900; for those tainly changed more frequently than house houses visible in photos, the numbers of size. For example, Table 3 demonstrates that rooms were adjusted to account for archi- couples lived in large houses with many tectural modification. rooms that were clearly built for, and likely In contrast to Crown and Kohler’s expec- once housed, a bigger family. tations for Pot Creek Pueblo, extended fam- Titiev reported that, among the Hopi, daughters continued to reside with their TABLE 2 mothers after marriage, their husbands Orayvi Household Type joining them, as they raised their children in a matrilineal extended family (1944:46). Household type Number Percent There is evidence, however, that the ex- Nuclear 54 49.5 tended family was only a temporary con- Extended 24 22.0 figuration until a young couple became Couple 9 8.3 settled and could establish an indepen- Other households 22 20.2 dent household (Emory Sekaquaptewa, 220 CATHERINE M. CAMERON

TABLE 3 House Size by Household Type

Mean Mean no. of Mean floor area Household type household size rooms per house (sq m) per house

Nuclear 5.4 6.2 97.0 SD 1.7 3.3 45.3 No. Hhd./houses 54 46 45 Extended 7.0 6.8 100.7 SD 2.0 2.8 55.0 No. Hhd./houses 24 21 21 Couple 2.0 6.5 108.4 SD 1.9 46.0 No. Hhd./houses 9 5 5 Other 5.2 6.7 90.2 Households 1.9 3.5 56.8 No. Hhd./Houses 22 20 20 All households (incl. unk. hhd type) 5.5 6.4 103.1 Number Hhd./house 109 160 154

Note. Fewer houses were included in the calculation of floor area than had been determined for numbers of rooms because floor area could not be calculated for any of the houses added to Mindeleff’s map by Titiev. Where these houses could be seen in photos, rooms could sometimes be counted. personal communication 1990; Levy 1992: to inherit the natal house, but as soon as 24). Census figures, discussed above, bear her parents died, the house would, once this out (Cameron 1996a; Whiteley 1988). again, be occupied by a nuclear family. The matrifocal extended family may have been the ideal, but it was not the most Floor Area per Person common household type at Orayvi. Orayvi in 1900 had a variety of household types and this was probably always the case. If Naroll (1962) found, in a cross-cultural certain household configurations, such as study, that dividing roofed floor space by extended families or elderly couples, were population resulted in an average of 10 sq expected to be temporary, there would be m of roofed floor space per person. little incentive to frequently adjust house “Naroll’s Constant” has been repeatedly size. questioned and refined (LeBlanc 1971; Furthermore, the architectural style of Clarke 1974; Kolb 1985). Castleberry (1974) Orayvi hinders the expansion of houses proposed a constant of 6 sq m per person, and, given an expected relationship be- a figure supported by Brown (1987) and tween house and household size, proba- Kolb (1985), although with caveats. bly always restricted the development of At Orayvi, calculating floor area per extended families. Orayvi houses, as in all person resulted in a surprisingly high fig- street-oriented pueblos, are joined in long ure, 21.34 sq m per person (Table 4), more rows. Houses can expand vertically, but than twice that originally proposed by not horizontally without impinging on Naroll and more than 3 times the figure property held by someone else. The estab- estimated by Castleberry (1974) and lishment of new, noncontiguous houses Brown (1987). The mean conceals a high would be a necessity if many daughters degree of variability, however. Calcula- married and started new families (Cam- tions of floor area per person could be eron 1996a). One daughter would remain made for 92 households (Table 4), and ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 221

TABLE 4 “creep” of houses toward the southeast Floor Area per Person at Orayvi for Different (Cameron 1996a,b). As a result of this pro- Household Types cess, even though households had a num- Household type Mean Standard Deviation ber of rooms and apparently a considerable and number (sq m) (sq m) amount of floor area per person, much of this floor area was not usable space. Based Nuclear 20.6 10.82 on studies at both prehistoric and historic no. 45 Extended 16.1 8.55 pueblos, some of these unused rooms be- no. 21 came garbage dumps (Kidder 1958; Adams Couple 54.2 23.0 1983), although they continued to function no. 5 as a defensive terrace upon which dwellings Other 17.0 12.2 were built. If only rooms in active use were no. 18 All houses 21.3 15.5 counted, floor area per person at Orayvi no. 92a might approach constants derived by Naroll and others. Because the Orayvi data were a Includes houses for which household type could exclusively external and room use could not not be determined. be directly determined, such calculations are beyond the scope of this study. these households showed a range of 1 to 90 sq m per person. Room Size and Social Status at Orayvi The wide deviation from proposed uni- versal floor area requirements may be the Both ethnologists and archaeologists result of patterns in the way multistoried have traditionally considered Pueblo peo- pueblos grow. Although Dohm (1990) pro- ple to be largely egalitarian, although the poses that increased crowding at large ag- point has been argued in both subdisci- gregated pueblos like Orayvi increases plines, especially during the past few de- the need for privacy and results in an in- cades (Lightfoot and Upham 1989 pro- crease in house size, pueblo growth pat- vide a summary of this controversy). Clearly, terns may be equally important. In the social stratification is not marked among 1800s, much of Orayvi was three stories prehistoric and historic pueblos and is tall; parts of the village reached four sto- only proposed consistently for certain ar- ries. The terraced structures faced south- chaeological cases, such as the Chacoan east both for maximum solar efficiency Regional System. However, even suppos- and because this is an important ceremo- edly “egalitarian” societies exhibit differ- nial direction for the Hopi. Front, upper ences in wealth and status that may have story rooms were living rooms; lower archaeological correlates. At Orayvi, the story rooms were used for storage. Ter- relationship between house size, wealth, raced structures were defensive; lower and social rank is not strong, but there is story rooms had no doors and upper sto- some evidence that clan houses, the cere- ries could be reached only by ladders that monial home of the Clan Mother and the could be pulled up in case of attack. place where religious paraphernalia was Photographic and documentary evidence kept, may have been larger than other shows that houses expanded as new rooms houses. were built onto the front of the dwelling In his exploration of social stratification (Cameron 1996a,1996b). As new, front at Orayvi, Levy (1992) found that the ma- rooms are built, rear rooms are blocked and trilineal clans of which Hopi society is became increasingly dark, unventilated, comprised could be assigned to one of and difficult to access. Eventually, they are three ranks based on ownership of cere- abandoned. This process resulted the slow monies, ownership of major political of- 222 CATHERINE M. CAMERON

TABLE 5 House Size and Social Status at Orayvia

Lineage status Mean house size SD Numberb

Prime 108.7 60.2 57 Marginal 100.9 55.0 60

Minimum Maximum Clan rank house area house area Mean area No. housesc (sq m) (sq m) (sq m) houses

3 25.11 232.33 98.6 24 2 60.9 306.4 102.74 62 1 9.1 239.8 107.7 40

a t value ϭ .74, significance ϭ .463. b Lineage status was unknown for 37 houses. c Clan rank from Levy (1992) Table 3.1 (Rank 3 ϭ highest).

fices, and access to good quality agricul- lowest ranking clans: in other words, the tural land. During the 1906 split, Levy sets of households which should have had found that high ranking clans and prime the highest and lowest status at Orayvi. and alternate lineages of middle ranking The mean area for houses of prime lin- clans tended to remain at Orayvi while eages of the five highest ranking clans is landless clans and marginal lineages of compared to average house size in Table middle-ranking clans left. Levy’s ranking 6. The Bear Clan is perhaps the highest system provides a method of linking ranking clan at Orayvi. In Orayvi oral his- house size with social status and wealth. tory, it is the first clan to arrive at Orayvi As a first attempt to discern a relationship and is the clan from which the village between house size and social status, chief is always selected (Levy 1992:25,33). houses at Orayvi were divided into two The three prime lineages of this high groups: those in which the female house- ranking clan had houses with floor areas hold head represented a prime lineage and that were considerably larger than the av- those in which the female household head erage house size at Orayvi. The 11 other represented a marginal lineage (Table 5). houses owned by prime lineages of high Houses of prime lineages were slightly ranking clans, however, had houses that larger than those of marginal lineages, but were quite variable in size. Interestingly, the difference was not significant. Next, of the seven houses that were below av- Levy’s (1992: Table 3.1) three clan ranks erage in size, most may have been in the were used to order houses by area (Table 5). process of expansion. Two were under Minimum and maximum house sizes for construction when Mindeleff mapped each of the three ranks were similar. The Orayvi and four were located in the small three largest houses at Orayvi belonged to roomblocks at the edges of the village, women of the middle rank. Of the three where new construction was taking place. smallest houses, one belonged to a women High ranking clans may have been better of rank 1, one to a woman of rank 2, and the able to afford the costs of new construc- third was unknown. tion and gather the materials and labor Finally, house sizes were examined for necessary to build the new structure. the prime lineages of the highest ranking Archaeologists have often assumed that clans and the marginal lineages of the clan houses would be larger than other ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 223

TABLE 6 House Size for Households of Different Statusa

Houses of the prime lineages of the five highest ranking clans HHd No. House size (sq m)

Bear Clan 23 113.7 161d 152.8 162 121.1 Mean house area, Bear Clan: 129.2, SD 20.8 Spider Clan 13 (no area) 110b 76.1 134 112.0 (faces Main Plaza) 166b 76.9 Mean house area, Spider Clan: 88.31, SD 20.54 Parrot Clan 9c 55.1 73c 47.4 Mean house area, Parrot Clan: 51.2, SD 5.4 Bow Clan 1c 28.1 131 127.8 (faces Main Plaza) 165d 213.9 Mean house area, Bow Clan: 123.3, SD 93.0 Patki Clan 5c 30.6 36c 84.1 61d 232.3 Mean house area, Patki Clan: 115.7, SD 104.5

House size for the marginal lineages of the lowest ranking clans

Mean house size: 103.8 Standard deviation: 52.3 Largest house: 239.8 (household 137) Smallest house: 36.3 (household 141) Number: 17 (for which floor area could be measured)

Note. Mean house area for all Orayvi houses (for which floor area could be measured) ϭ 103.1 m2, SD 58.95, n ϭ 154. a Clan status and prime and marginal lineage status from Levy 1992. b Houses under construction in 1887. c Houses at the edge of the village. d Possible clan houses. houses (Creamer 1993:117) and Adams clan houses (Levy, personal communication (1983) demonstrated this relationship with 1997) the clan house would have been lo- data from Walpi Pueblo. Titiev (1944:47) re- cated in the home of a woman from one of ported the great concern that people at the clan’s prime lineages (Levy 1992:24; Orayvi had for clan houses and even today Titiev 1944:47). Therefore, one of the homes extraordinary efforts are made to preserve of each of the five high ranking clans on these special houses in Hopi villages. While Table 6 must have been the clan house. For it was not possible to determine from each of these clans, except the Parrot Clan, Titiev’s notes exactly which houses were one house is considerably larger than the 224 CATHERINE M. CAMERON others. These larger houses may have been three sizes (small storage rooms, large clan houses and the home of the Clan habitation rooms, and very large ceremo- Mother. nial rooms), then room sizes at Orayvi While most prime lineages of high should show a tri- or at least bi-modal ranking clans do not tend to have signifi- distribution. This is not the case, however. cantly larger houses that average, neither A histogram of Orayvi room sizes shows a do marginal lineages of low ranking clans normal distribution across a wide range of have very small houses. House size for room sizes (Fig. 6); there is no patterning marginal lineages of the lowest ranking that would suggest that room sizes fall clans is presented in Table 6. Average size neatly into two or more size classes. for the 17 houses for which data were Another measure of room function that available matched exactly the average size was observable at Orayvi is the level or for all Orayvi houses (103 sq m). With the story at which a room is located. Adams possible exception of clan houses, house (1983) found that at the Hopi Pueblo of size and social status do not seem to be Walpi, storage and granary rooms tended correlated at Orayvi. to be on ground floors. Living rooms were While house size and status seem to be in upper stories. Using the story of rooms poorly correlated at Orayvi, the location of at Orayvi as an indicator of room function, houses is another potential status indicator. the size of rooms of different stories was The most important house sites at Orayvi calculated. (Measurements are of rooms were likely those surrounding the Main sizes in 1887 before Euro-American influ- Plaza or the Snake Dance Plaza where most ence changed significantly the location of ceremonial activity takes place. For exam- Hopi storage and living rooms). Because ple, when Bacavi was built after the split by of the assumption used in this study that emigrants from Orayvi, the village was each story of a multistoried room is the planned so that the houses of the most same size, only the top floors of each room prominent families (especially the Bear were used in this calculation. In fact, then, clan) were located on the plaza (Whiteley the comparison is between different ter- 1988:125). At Orayvi before the split, how- races rather than different stories: front, ever, this pattern was not apparent: only single-story terraces, middle, two-story two of the 14 houses of the prime lineages of terraces, and rear, three-story terraces the highest ranking clans were located (Fig. 4). The assumption that front, lower around the plaza (Table 6). House location story rooms will be storage rooms and did not seem to be a status indicator. recessed upper stories, living rooms, is Room Function and Room Size at Orayvi still appropriate. There is a significant dif- ference in room size for rooms in different Both Hill (1970) and Adams (1978, 1983) stories (Table 7). Front, first floor rooms, proposed multiple indicators of room most likely to be storage, have the lowest function, including floor features, such as mean area, and rooms increase in size as hearths or mealing bins, ventilators, wall the story increases. niches, doorways, and story in the village Although room sizes at Orayvi are not structure. Both found room size to be a clearly bi- or tri-modal, based room posi- strong indicator of room use. Because data tion, rooms of different function seem to be from Orayvi are based largely on maps differently sized. Ciolek-Torrello suggests and photos of house exteriors, there is that at Grasshopper (1985) the frequent re- little direct evidence concerning room use modeling that characterizes Hopi architec- for this pueblo. But if, as Hill and others ture may have blurred a once more clear- have suggested, pueblo rooms come in cut relationship between room size and ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 225

FIG. 6. Histogram of Orayvi room sizes showing a relatively continuous distribution of sizes. room function. But it is also likely that the ROOM SIZE AND THE household-based construction evident at ORGANIZATION OF Hopi did not produce rooms in standard- CONSTRUCTION IN THE ized sizes. Factors such as the size of avail- PREHISTORIC SOUTHWEST able roof timbers and the number of people who could be persuaded to work on a build- A comparison of patterns of room sizes ing project may have been important deter- at Orayvi with those found at prehistoric minants of room size. This combination of sites highlights apparent differences in factors would produce the variety of room the organization of construction through- sizes apparent at Orayvi. out the northern Southwest. Rooms at Orayvi were variable in size, but at some TABLE 7 prehistoric pueblos, especially certain Orayvi Room Size by Storya Eastern Pueblos, rooms size is quite uni- form. This is also true for a few Western Story Mean SD Number Pueblos. Ladder type construction (de- First 12.3 6.7 207 scribed above), may account for some of Second 14.4 8.9 202 the modularity and may result in sites Third 18.1 10.4 171 with characteristic layouts: linear plaza Fourth 21.5 10.0 12 pueblos with relatively straight rows of a Only the top story for each “story” type is used in rooms outlining one or more plazas. Uni- calculating the mean as each lower story is assumed form room sizes, ladder type construction, to be the same size as the top story. and linear plaza pueblos may be the result 226 CATHERINE M. CAMERON of a coordinated construction effort in- techniques at these pueblos that is lacking volving a large portion of a community. at other pueblos. Where these characteristics are not ob- Ladder-type construction is just begin- served, construction may be organized at ning to be recognized and reported by the household level. A few suggestions Southwestern archaeologists. Although concerning the origin of these different most of the Eastern Pueblos on Table 8 organizational types is made here. were excavated at the turn of the century Table 8 shows room sizes for a number before much attention was paid to con- of large, late prehistoric (post A.D. 1200) struction methods, ladder type construc- Western and Eastern Pueblos. Compari- tion has been noted at some Eastern son with Orayvi required large, exten- Pueblos (Creamer et al. 1993, Cordell 1996, sively excavated pueblos with room sizes Roney 1996). There was no evidence of reported or mapped at a scale that would ladder type construction at Orayvi al- allow for measurement of rooms. Eastern though it has been reported at a few pre- Pueblos for which individual room size historic sites in northeastern Arizona data could be obtained are from a re- (Dean 1996,10 Gilpin 1989), including sev- stricted region of the northern Rio Grande, eral of the Homolovi pueblos (Adams per- south of Santa Fe: the Galisteo Basin and sonal communication 1997). surrounding areas. A majority of these Ladder type construction results in a sites were excavated by Nels Nelson characteristic, linear site layout. For exam- (1914) early in this century and he pro- ple, most of the Eastern Pueblos on Table vided a table of room dimensions for the 8 are characterized by a few rows of rooms rooms that he excavated. Room sizes for which define several, relatively large, gen- Western Pueblos were more difficult to erally enclosed plazas (Fig. 7). The same is obtain and as a result, the sample is true at Homolovi II. At Fourmile Ruin, the smaller and more widespread, including latest part of the site (including Room- sites both above and below the Mogollon block D) is composed of linear room- Rim (Fig. 1). blocks, although the plaza they define is Like Orayvi, several of the prehistoric not fully enclosed. These linear plaza lay- Western Pueblos have rooms of variable outs likely result, at least partly, from the size, ranging from a little more than 1 sq construction of long continuous walls re- m to more than 35 sq m. Standard devia- quired for ladder type construction. In tion of room sizes for these pueblos is contrast, at a number of Western Pueblos smaller than Orayvi, and Orayvi has a on Table 8, especially those below the number of very large rooms; however, all Mogollon Rim, plazas are present, but the of the prehistoric sites have far fewer total organization is much less formal. For ex- rooms than Orayvi, so some of Orayvi’s ample at Turkey Creek and Grasshopper variability may be due to sample size. Pueblo, dense clusters of rooms surround Rooms at the Eastern Pueblos were of small plazas (Fig. 8). This layout, which very standard size: mean areas were be- may have begun at the large Classic sites tween 6–7 sq m and most had standard of the Mimbres region, such as Galaz and deviations of 2 sq m or less from the mean. Swartz, has been termed “agglomerative” Several Western Pueblos also show signif- (Mills 1998). icant regularity in room sizes: the Ho- These two different settlement patterns molovi Pueblos and Roomblock D at Four- may reflect differences in the organization mile Ruin. The remarkable regularity in of groups of people who moved about the the sizes rooms at some prehistoric pueb- Southwest during the centuries just be- los suggests a standardization of building fore Spanish contact—a time of dramatic, ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 227

TABLE 8 Room Sizes at Prehistoric Pueblos

Largest Smallest Mean No. room room area SD No. rms Source of room Site Time period stories Total rooms (sq m) (sq m) (sq m) (sq m) in mean size data

Eastern Pueblos

Arroyo Hondo 1300–1330 2 1000 9.3 3.8 6.3 1.1 62 Creamer et al. (1993) Comp. I Pueblo Largo 1275–1475 1–2? 480ϩ ground 10.0 5.0 7.1 1.6 13 Nelson (1914) (Galisteo Basin) fl. Pueblo San Lazaro 1275–1675 2 488 ground 11.1 3.7 6.9 1.7 44 Nelson (1914) (Galisteo Basin) fl. prehis. Pueblo San Lazaro 1275–1675 2 488 ground 15.0 3.8 8.1 1.9 80 John Ware, (personal (Galisteo Basin) fl. prehis. comm. 1997) (measured from map) Pueblo Galisteo 1275–1700 2 567 ground 9.7 5.5 7.0 1.1 23 Nelson (1914) (Galisteo Basin) fl. San Cristobal 1375–1675 1–2? 1645 19.7 2.7 7.1 2.0 202 Nelson (1914) (one very lg. (Galisteo Basin) room; cross walls probably removed) Pueblo She 1300–1600 1–2 1543 9.6 4.9 7.1 1.3 27 Nelson (1914) (Galisteo Basin) Pueblo Colorado 1275–1600 2 881? 9.0 4.4 6.6 1.1 40 Nelson (1914) (Galisteo Basin) Pueblo Blanco 1400–1600 2 1400–2000 12.3 4.0 7.4 1.9 34 Nelson (1914) (Galisteo Basin) Paako (Prehist.) 1300–1425 2 253? 14.3 1.6 6.3 1.9 122 Laboratory of Anthropology 1936 WPA excavations Paako (Histor.) 1500s 1 ? 14.6 1.4 7.4 2.4 89 Laboratory of Anthropology 1936 WPA excavations Pueblo del 1350–1550 1–3 198 ground 13.7 2.2 6.9 2.1 141 Snow (1976) Encierro fl. rms Rowe Ruin 1240–1425 2 ? 10.6 3.6 6.9 2.0 21 Cordell (1998) Tables2&3

Western Pueblos

Carter Ranch 1100–1225/1250 1 39 19.0 3.5 10.1 4.5 23 Longacre (1970) (measured from map) Broken K 1150–1280 1 92 30.4 2.5 9.4 5.2 42 Hill (1970; Table 6) Turkey Creek 1240–1300 1 335 32.8 1.2 8.2 5.4 289 Lowell (1986) Grasshopper 1300–1400 2 500ϩ 29.2 1.9 15.5 4.9 86 Ciolek-Torrello (personal communication, 1998) Mariana Mesa 1100–1300 2 500–600 13.6 3.7 6.2 2.4 27 McGimsey (1980) (2 very Site 616 (8.5) (5.6) (1.3) (25) large rooms may lack cross walls; figures in () are w/out these room) Homolovi IV 1260–1285 1? 150 12.5 1.6 5.5 2.2 79 Adams (personal communication, 1998) Homolovi III 1280–1300 1? 40 12.2 6.3 8.6 2.4 8 Adams (personal (early) communication, 1998) Homolovi III (late) 1325–1375 1? 10 11.0 8.6 10.2 1.3 3 Adams (personal communication, 1998) Homolovi II 1330–1400 2 1200 12.8 2.3 7.2 1.8 61 Adams (personal communication, 1998) Homolovi I 1280–1390 2 800 14.7 1.8 5.7 2.3 178 Adams (personal communication, 1998) Fourmile Ruin, 1350–1400 ? 50 (Rmblk D) 11.5 3.9 6.7 1.5 39 Johnson (1993) (measured Roomblk D from map)

large-scale abandonments and aggrega- have been designed to accommodate the tion of population into large villages immigration of large, relatively well- (Adler 1996). Linear plaza pueblos may organized groups of people; agglomera- 228 CATHERINE M. CAMERON

FIG. 7. Arroyo Hondo Pueblo (Component 1), located near Santa Fe, New Mexico, was built and occupied during the early 13th century. It has a linear, plaza-oriented settlement layout typical of many Eastern Pueblos. Reprinted, by permission, from The Architecture of Arroyo Hondo Pueblo, New Mexico by Winifred Creamer. Drawing by Katrina Lasko. © 1993 by the School of American Research, Santa Fe. ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 229

FIG. 8. Turkey Creek Pueblo, located in east-central Arizona, shows an agglomerative layout with rooms massed around plazas, which characterizes some Western Pueblos. Reprinted, by permission, from Prehistoric Households at Turkey Creek Pueblo, Arizona by Julie Lowell, Copyright 1991 The University of Arizona Press. 230 CATHERINE M. CAMERON tive pueblos could more easily accept the Robinson found room size, house plan occasional household (Cameron 1999; but and the number and size of vigas was see Cordell 1996:235). At agglomerative highly variable. Even though these two pueblos, a new houses could simply be pueblos have been occupied for approxi- tacked to the outside of the existing struc- mately the same length of time, a planned ture. This design would allow individual core was visible at Acoma, but not at families to join the community at any time Walpi. and the pattern of site growth would be Robinson (1990) attributes the differ- accretional. Linear plaza pueblos could ences in construction between these two not be as easily expanded. New room- pueblos to a higher degree of social inte- blocks would have to be built surrounding gration at Acoma, although he cannot rule new plazas and the builders would have out the possible influence of Spanish to agree on the size of plaza space to out- priests there. Robinson notes that Acoma line. Construction would be organized at a has existed as a community for over 600 level above the household. years and suggests that its leaders may have been able to exert considerable con- A Consideration of Differences in Length of trol over construction. In other words, he Settlement Occupation finds greater social integration at Acoma that is expressed architecturally. Standardized room sizes, ladder type construction, and linear plaza site layouts Organization of Construction among all suggest that construction at some pre- Prehistoric Pueblos historic pueblos was coordinated above the household level. At other sites, these If length of settlement occupation can features are not observed suggesting that be ruled out as the primary cause of dif- construction was more likely undertaken ferences in room sizes, settlement layout, by individual families. Differences in the and the prevalence of ladder-type con- length of settlement occupation must be struction, then social or cultural differ- considered, however. Historic pueblos, ences that affect the organization of con- like Orayvi, have been occupied for hun- struction may be investigated as a factor. dreds of years. It is possible that the ear- Although no one has directly attributed liest construction at these villages may variation in this suite of attributes to par- have been an organized endeavor by a ticular prehistoric social or ethnic groups, number of families resulting in a “planned two studies from the Western Pueblo area core” of buildings that have been obscured show that archaeologists are beginning to by centuries of rebuilding. find room sizes, construction methods, A comparison of the Hopi Pueblo of and site layout key pieces of evidence for Walpi and the Pueblo of Acoma, both of exploring ethnicity and tracking the mas- which were built about 300 years ago, sug- sive population movements that charac- gests that this is not necessarily the case terize the late 13th and 14th centuries and supports a fundamental organiza- (Adler 1996, Cameron 1995; see Reid 1998 tional difference in construction methods for a recent overview of population move- among prehistoric pueblos (Fig. 1).11 Rob- ment in the Southwest). inson (1990) and Earls (1988) found that In the Silver Creek drainage of east- rooms at Acoma were very standardized central Arizona, Mills (in press) has ob- in size, houses had remarkably similar served differences in room size and site plans, and there was even similarity in the layout and attributed them to ethnic dif- numbers and sizes of vigas used in roofing ferences. During the period from A.D. (Robins 1990:104). At Walpi, however, 1275–1325, sites in the Silver Creek drain- ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 231 age were built of “massive roomblocks en- pertinent to understanding differences in closing central plazas...(Mills 1998).” the organization of construction among Rooms at these sites are very large, about prehistoric pueblos in both areas. Ware 15 square meters, comparable to rooms at and Blinman (in prep), in a recent paper, Grasshopper Pueblo, 30 miles to the south have emphasized the importance to ar- (Table 8). Between 1325 and 1390, Mills chaeology of ethnographic reports of believes that architectural and ceramic highly integrated social systems among change suggest immigration into the area. the Eastern Pueblos. They show that so- She observes that at least part of Fourmile dalities operated as centralized polities Ruin (dated to the mid- to late 14th cen- and they cite Jorgensen (1980:222) who tury) has an “open plaza” layout (equiva- finds the Eastern Keres and Tanoan Pueb- lent to linear plaza) and Roomblock D los villages “. . . more integrated and more shows evidence of rapid construction.12 centrally controlled in economic, social, The rooms at Fourmile are also much ceremonial, and warfare matters than all smaller—about 10 square meters. Mills other political communities in western notes, however, that bond and abut se- North America.” quences indicate that Roomblock D was A review of Hill’s (1982) comprehensive not built in one construction episode (i.e., ethnography of Santa Clara Pueblo illus- no ladder construction). In spite of the trates the centralization of powers in the linear form of late Pueblo IV sites in the two caciques, who headed the religious hi- Silver Creek drainage, the lack of coordi- erarchy for each of the two moieties. Hill nated (ladder type) construction at these describes the functions of the caciques at sites leads Mills to suggest that construc- Santa Clara as “. . . extensive. They were tion was organized at the household level. charged with responsibility for the inter- At the Homolovi Pueblos in the Middle nal well-being of the village...(includ- Little Colorado area, Adams (1998; see ing) the guidance of political and secular also Adams 1996) finds settlers from both affairs (1982:185).” During his stays among the Hopi Mesas and the Silver Creek area the Eastern Pueblos, Adolph Bandelier occupying several large pueblos with no- noted “much communism” among the ticeable architectural differences. For ex- Eastern Pueblos in the accomplishment of ample, Homolovi IV had relatively small public works such as bridge building and rooms (Table 8) and a clustered layout care of the irrigation systems (1966:111– built on the top and sides of high butte. In 114, 196). As described above, both Hill contrast, Homolovi III was a linear room- and Bandelier found that house construc- block with rooms that at first averaged tion was accomplished, at least part of the more than 11 sq m and then decreased, time, by village-directed parties of male with later rooms averaging 7.5 sq m (Table workers. 8).13 Based on ceramic and other evidence, In contrast to the highly organized East- Adams shows that the inhabitants of Ho- ern Pueblos, Ware and Blinman point out molovi III originated in the Silver Creek that the Hopi and other Western Pueblos area, while the Homolovi IV populations were dominated by kinship groups (ma- derived from the Hopi Mesas. trilineal clans) and that “. . . the weakest It seems apparent that there were social sodality expressions in western North organizational differences in among large, America developed among groups with late pueblos that are reflected in prehis- the strongest kinship-based polities....” toric architecture. Similarly, there are or- In other words, the strong clan system in ganizational differences between modern the Western Pueblos, like Orayvi, pre- Eastern and Western pueblos that may be vented the kind of centralization evident 232 CATHERINE M. CAMERON among the historic and prehistoric East- terial) nor were they built at the same ern Pueblos. Ethnographic and archaeo- scale as Chacoan Great Houses. The plan- logical evidence suggest that construction ning and organization that went into con- at some Western Pueblos was not orga- struction at Eastern and some Western nized above the household level. Mind- Pueblos, however, might be seen as a eleff described much of the work of house muted echo of Chacoan building tech- construction at Hopi as being carried out niques. Aspects of the highly organized by a woman and her female relatives building traditions developed in Chaco— (1989:101). At Turkey Creek Pueblo, Low- including coordinated construction of ell (1989:193) found evidence for house- many rooms in one building event—may hold autonomy in a number of types of have been applied by post-Chacoan peo- activities, including food production and ples in the construction of later Pueblos. food storage. Although Lowell (1989:188) Differences in the gender of house did find groups of rooms defined by un- builders may be an important aspect of broken walls which she defined as “supra- the organization of pueblo construction. households” (groups of households that In his study of patterns of residence and may have shared some activities), room descent among prehistoric Southwestern size and overall site layout suggests that peoples, Steven James (1994) argues that house construction was generally under- in matrilineal societies women built and taken at the household level. owned houses and rooms were small (he Ware and Blinman’s (in prep) most in- notes that men may procure building ma- triguing suggestion is that the highly terials). In patrilineal societies, men built structured Eastern Pueblos may represent and owned houses and rooms were large. an historic continuity with the 10th–12th Based on room size, James classifies most century Chaco Regional System, the most of the prehistoric pueblos in the northern readily acknowledged complex social sys- Southwest as the product of societies with tem in the Southwest. Ware and Blinman matrilineal residence and descent where believe that the highly centralized sodality women build houses. He makes an excep- system evident in modern and historic tion for Chacoan Great Houses. “Given Eastern Pueblos may echo the even more the monumental architecture at Chacoan centralized and complex Chacoan politi- Great Houses, the majority of room con- cal system. Other scholars have also found struction probably would have been per- connections between Chaco and modern formed by men in the society (James 1994: Puebloan peoples living along the north- 188).” ern Rio Grande (Ford, Schroder, and Gender issues in the organization of Peckham 1972). pueblo construction deserve far more The distinguishing characteristic of study than they can be given here. There Chacoan Great Houses, besides their wide seems, however, to be a dramatic organi- walls, is the extraordinary degree of plan- zational contrast between the construction ning that went into their construction (Lek- of Chacoan Great Houses by groups of son 1986). Great Houses were obviously highly skilled male (?) masons and the the result of a highly coordinated effort by construction of individual houses by a number of skilled masons who built 100 women as observed at modern Western or more rooms in one construction event Pueblos like Orayvi and inferred for many following a preconceived plan. None of prehistoric pueblos. James believes that the pueblos on Table 8 were made with the Chacoan building tradition ended in elaborate Chacoan masonry techniques A.D. 1150 (1994:201). It is possible, how- (in fact, coursed adobe is a common ma- ever, that some highly organized settle- ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 233 ments in the 13th, 14th, and 15th centuries supported by the Orayvi data. Closely reflect a continuation of pueblo building packed puebloan houses cannot be easily as an hierarchically directed male activity. expanded which places a limit household Of course, the relationship between Cha- size. Furthermore, frequent changes in coan peoples and later Southwestern pop- household configuration mean that there is ulations remains to be discovered. not a close link between house and house- hold size at Orayvi. Such shifting domestic CONCLUSIONS arrangement were likely always a feature of puebloan society making it difficult to use This paper demonstrates that the size of changes in prehistoric house size as an in- architectural space has important implica- dicator of changes in household organiza- tions for the interpretation of prehistoric tion. Furthermore, house size does not seem social systems. It is one aspect of material to be determined by social status in Pueblo culture that is readily preserved, easy to society as it sometimes is in other “egalitar- measure, and is conditioned by a variety ian” societies. Clan houses of high ranking of technological and social behaviors that clans, which would house individuals who can be investigated in archaeological cases. control important religious ceremonies and Although archaeologists tend to see social their associated paraphernalia, may be requirements as the most important de- larger than other houses, but, in general, the terminants of room size, the Orayvi data relationship between status and house size and comparison of Orayvi with prehistoric was not strong. Eastern and Western Pueblos highlighted A comparison of room sizes, site layout, two other important determinants: the and the prevalence of ladder type con- technology of pueblo construction and the struction at Orayvi and at prehistoric East- organization of construction. Room sizes at Orayvi are quite variable ern and Western Pueblos suggests differ- and while room function may account for ences in the organization of construction some of this variability, the technology of across the prehistoric Southwest. Planning puebloan construction may be a more im- and coordination of construction exhib- portant factor. This may also be true at ited by some large, late Pueblos, especially other large, multistoried pueblos. Avail- Eastern Pueblos, suggests organization able building materials, especially the size above the household level and possibly of roof beams, limits room size. In multi- construction by males. It implies that pop- storied pueblos, because of the need for ulation movement and the establishment wall support, the size of lower story rooms of towns may have been accomplished by generally determines the size of upper relatively large, coherent groups. Planned stories (but see James 1994, 1997). The na- and coordinated building construction ture of pueblo construction also affects may even reflect historical connections presumed constant relationships between with the complex Chacoan Regional sys- numbers of individuals and quantities of tem. On the other hand, construction or- roofed space. While there appears to be a ganized at the household level, likely un- great deal of space for each individual in dertaken by women, seems to characterize Orayvi house, much of that space may some of the Western Pueblo area and sug- function only as an architectural platform gests the aggregation of much more or a trash disposal area; average active loosely knit groups of households with space per individual may actually be sim- much weaker village-wide ties. ilar to that found in cross-cultural studies. Room size is an important indicator of Certain archaeological expectations link- the organization of pueblo construction ing house size and social variables were not and may provide an avenue for exploring 234 CATHERINE M. CAMERON social boundaries and, potentially, migra- double walls separating the rooms used by one tion paths in the prehistoric Puebloan household from those of another, allowing easy Southwest. The examination of a far larger identification of houses. Neither Crown and Kohler (1994) nor Kulisheck et al. (1994) mention this con- sample of sites and exploration of techno- struction feature, nor does it seem to have been logical and social variables other than reported elsewhere in the Southwest. room size will almost certainly deepen our 7 Mindeleff’s map was used to calculate floor area for understanding of organizational differ- rooms and houses at Orayvi. Measurements were ences among prehistoric pueblos as well taken from of Mindeleff’s five original field maps, pro- vided to me by the Smithsonian Institution’s National as the origins of these differences. Anthropological Archives as photographs reduced to 8 by 10 inches. I enlarged the photographs to double ACKNOWLEDGMENTS Mindeleff’s original field scale (1 inch ϭ 20 feet) to allow rooms to be easily measured. Because of the probable distortion involved in enlarging and reducing Several people provided helpful comments on drafts the maps, length and width were measured only to the of this paper: Chuck Adams, Doug Bamforth, Brenda nearest 1/2 foot. Measurements were taken from the Bowser, Linda Cordell, Steve Lekson, Barbara Mills, interior of one wall to the interior of the opposite wall. and two anonymous reviewers. Chuck Adams, Richard Length and width measurements for rooms were then Ciolek-Torrello, and John Ware generously provided converted to meters. Measurements were taken of up- me with room size data from their projects. Jerrold per story rooms visible on Mindeleff’s plan view and Levy not only allowed me to use his computerized each lower story was assumed to be the same size as Orayvi census data for this and other projects, but the top story. This is a reasonable assumption given the compiled data on household ranking which I used to difficulty of building an upper story wall without the assess the relationship between house size and social support of a lower story wall. Bottom story rooms status at Orayvi. Steve James provided me with copies might, of course, be divided which would increase of his recent work which were extremely helpful in lower story room counts and decrease room sizes, but revising the manuscript. Andrew Sell carefully re- not affect floor area for houses. corded the size of Orayvi rooms using Mindeleff’s 8 map. The paper profited especially from my conversa- At the Hopi Pueblo of Walpi, house size did tions with Chuck Adams, Eric Blinman, Brenda increase as many areas of the pueblo were aban- Bowser, Barbara Mills, and John Ware each of whom doned during the 20th century; the number of rooms was insightful, encouraging, and generous with their per house decreased and house size increased (E. time. Thanks to all for your help! None of these folks, of Charles Adams, personal communication 1997). The course, are responsible for errors in the final product. figures for Orayvi are from before the split and the abandonment of much of the village. 9 Ferguson (1996:125–126) found a similar wide NOTES variability in the number of square meters per per- son in houses at modern Zuni Pueblo. He attributes 1 The spelling of Orayvi follows the Hopi Dictio- the variability both to the developmental cycle of the nary (1998). domestic group and to the occurrence of a number of 2 Chacoan Great Houses were large, aggregated, different household types at Zuni. 10 and multi-storied beginning in A.D. 900, but they Dean (1996) found “spinal” roomblocks in Kay- were not primarily domestic structures. enta area which he notes are similar to “ladder type” 3 Lower story rooms could, of course, be subdi- construction found in the Eastern San Juan Basin, vided. The frequency with which this occurs in either but notes “we currently lack the data to assess func- prehistoric or historic pueblos is unknown. tional similarities between the two architectural 4 Coursed adobe is common in the southern forms (1996:38).” Southwest, for example at Classic Period Hohokam 11 Note that Eggan (1950) considered Acoma a sites in southern Arizona and in the El Paso/Casas Western Pueblo. However, it is geographically close Grandes area of southern New Mexico and northern to the Rio Grande (Eastern) Pueblos and residents Chihuahua (Cameron 1998). speak Keres (Fox 1967). 5 Similar large rooms have been reported by Mills 12 The difference in the average room size shown (1998) at Bailey Ruin. E. Charles Adams (personal on Table 8 (6.7 sq m) and reported by Mills (1998; 10 communication 1997) emphasizes that the large, late sq m) results from differences in how room measure- pueblos below the Mogollon Rim (such as Grasshop- ment were made (Mills, personal communication per Pueblo) have much larger, more multifunctional 1998). rooms than do pueblos above the Mogollon Rim. 13 The availability of driftwood used in roof con- 6 Holschlag (1975) reports that houses at Taos have struction is also a factor in room size in the middle ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 235

Little Colorado area; the area is a considerable dis- Bullard, William R. Jr. tance from forests. 1962 The Cerro Colorado site and pithouse archi- tecture in the Southwestern United States prior to A.D. 900. Papers of the Peabody Mu- REFERENCES seum of Archaeology and Ethnology, Harvard University XLIV(2). Peabody Museum, Cam- Adams, E. Charles bridge, Massachusetts. 1978 Room use as interpreted from room size. Paper Cameron, Catherine M. presented at the 43rd Annual Meeting of the 1991 Architectural change at a southwestern pueblo. Society for American Archaeology, Tucson. Ph.D. dissertation, University of Arizona, 1982a Walpi archaeological project, synthesis and In- Tucson. terpretations. Ms on file, Museum of North- 1996a Observations on the Pueblo house and ern Arizona, Flagstaff. household. In People who lived in big houses: 1982b The architectural analogue to Hopi social Archaeological perspectives on large domestic organization and room use and implications structures, edited by Gary Coupland and E. B. for prehistoric Southwestern culture. Amer- Banning, pp. 71–88. Prehistory press, Mono- ican Antiquity 48(1):44–61. graphs in World Archaeology No. 27, Mad- 1991 The origin and development of the Pueblo ison, Wisconsin. Katsina cult. University of Arizona Press, 1996b Multi-story construction in Southwestern Tucson. Pueblo architecture. In Interpreting South- 1996 River of change: Prehistory of the Middle Little western diversity: Underlying principles and Colorado River Valley, Arizona. Arizona State overarching patterns, edited by J. J. Reid and Museum Archaeological Series 195. Arizona P. R. Fish, Arizona State University Anthro- State Museum, Tucson. pological Papers, Tempe. 1998 Late prehistory in the Middle Little Colo- 1995 Migration and the movement of Southwest- rado River Area, a regional perspective. In ern People Journal of Anthropological Archae- Migration and reorganization of the Pueblo IV ology 14(2):104–124. Period in the American Southwest, edited by 1999 Hopi dwellings: Architectural change at Orayvi. Katherine Spielmann, Anthropological Pa- University of Arizona Press, Tucson. pers of Arizona State University, No. 51, Castleberry, Samuel E. Tempe. 1974 Further refinements of formulae for deter- Adler, Michael A. mining population from floor area. World 1996 The prehistoric Pueblo world, A.D. 1150–1350. Archaeology 6:118–122. The University of Arizona Press, Tucson. Ciolek-Torrello, Richard 1978 A statistical analysis of activity organization: Alstrom, Richard V. N., Jeffrey S. Dean, and William Grasshopper Pueblo, Arizona. Ph.D. disserta- Robinson tion, Department of Anthropology, Univer- 1991 Evaluating tree-ring interpretations at sity of Arizona, Tucson. Walpi Pueblo, Arizona. American Antiquity 1985 A typology of room function at Grasshopper 56(4):628–644. Pueblo, Arizona. Journal of Field Archaeology Baldwin, Stuart J. 12:41–63. 1987 Roomsize patterns: A quantitative method Ciolek-Torrello, Richard and J. Jefferson Reid for approaching ethnic identification in ar- 1974 Change in house size at grasshopper. The chitecture. In Ethnicity and culture, Proceed- Kiva 40(1–2):39–48. ings of the Eighteenth Annual Conference Clarke, Steven K. of the Archaeological Association of the 1974 A method for the estimation of prehistoric University of Calgary, edited by Reginald Pueblo populations. The Kiva 39(3–4):283– Auger, Margaret F. Glass, Scott MacEach- 287. ern, and Peter H. McCartney, pp. 163–174. University of Calgary. Clemmer, Richard O. 1978 Continuities of Hopi culture change. Acoma Blanton, Richard E. Books, Ramona, CA. 1994 Houses and households, a comparative study. Cordell, Linda S Plenum Press, New York. 1984 Southwestern archaeology. Annual Review of Brown, Barton McCaul Anthropology 13:301–332. 1987 Population estimation from floor area: A re- 1986 Big sites, big questions: Pueblos in transi- study of “Naroll’s Constant.” Behavior Sci- tion. In The prehistoric pueblo world A.D. ence Research 21(1):1–49. 1150–1350, edited by Michael A. Adler, pp. 236 CATHERINE M. CAMERON

228–240. University of Arizona Press, Tuc- Ferguson, T. J. son. 1996 Historic Zuni architecture and society, an ar- 1997 Archaeology of the Southwest. Academis Press, chaeological application of space syntax. An- San Diego. thropological Papers of the University of Ar- 1998 Before Pecos: Settlement aggregation at Rowe, izona, No. 60. University of Arizona Press, New Mexico. Maxwell Museum of Anthro- Tucson. pology Anthropologica Papers No. 6, Albu- Ferguson, T. J., and Barbara J. Mills querque. 1987 Settlement and growth of Zuni Pueblo: An Creamer, Winifred, with Catherine M. Cameron and architectural history. The Kiva 52:243–266. John D. Beal Ferguson, T. J., Barbara J. Mills, and Calbert Seciwa 1993 The architecture of Arroyo Hondo Pueblo, New 1990 Contemporary Zuni architecture and soci- Mexico. School of American Research Press, ety. In Pueblo style and regional architecture, Arroyo Hondo Archaeological Series, Vol. 7, edited by Nicholas C. Markovich, Wolfgang Santa Fe. F. E. Preiser, and Fred G. Sturm, pp. 103–121. Crown, Patricia L., and Timothy A. Kohler Van Nostrand Reinhold, New York. 1994 Community dynamics, site structure, and Ford, Richard I., Albert H. Schroeder, and Stewart L. aggregation in the Northern Rio Grande. In Peckham The ancient Southwestern community, models 1972 Three perspectives on Puebloan prehistory. and methods for the study of prehistoric social In New perspectives on the Pueblos, edited by organization. edited by W. H. Wills and Rob- Alfonso Ortiz, pp. 19–40. University of New ert D. Leonard, pp. 103–118. University of Mexico Press, Albuquerque. New Mexico Press, Albuquerque. Fox, Robin Dean, Jeffrey S. 1967 The Keresan Bridge: A problem in Pueblo eth- 1996 Kayenta Anasazi settlement transforma- nology. Athlone Press, London. tions in Northeastern Arizona: A.D. 1150– 1350. In The prehistoric Pueblo world A.D. Gann, Douglas W. 1150–1350, edited by Michael A. Adler, pp. 1996 The use of adobe architecture in the 29–47. University of Arizona Press, Tucson. Homol’ovi region. In River of change: Prehis- Divale, William T. tory of the Middle Little Colorado River Valley, 1977 Living floor area and marital residence: A Arizona, edited by E. Charles Adams, pp. replication. Behavior Science Research 12:109– 93–106. Arizona State Museum Archaeolog- 115. ical Series 195. Arizona State Museum, Tuc- son. Dohm, Karen 1990 Effect of population nucleation on house Gilpin, Dennis size for Pueblos in the American Southwest. 1989 Great houses and pueblos in Northeastern Ari- Journal of Anthropological Archaeology 9:201– zona. Paper presented at the 1989 Pecos 239. Conference, Bandelier National Monument, New Mexico. Douglass, Andrew E. 1929 The secret of the Southwest solved by talk- Hall, Edward T. ative tree rings. National Geographic Society 1966 The hidden dimension. Doubleday, New York. Magazine, Vol. 56, No. 6, pp. 736–770. Wash- 1968 Proxemics. Current Anthropology 9:83–103. ington, DC. Hansen, John A., and Michael B. Schiffer Drager, Dwight L. 1975 The Joint Site—A preliminary report. In 1976 Anasazi populations estimates with the aid chapters in the prehistory of Eastern Ari- of data derived from photogrammetric zona, IV. Fieldiana: Anthropology 65:47–91. maps. In Remote sensing experiments in cul- Haury, Emil tural resource studies: Non-destructive methods 1958 Evidence at Point of Pines for a prehistoric of archaeological exploration, survey, and anal- migration from Northern Arizona. In Migra- ysis, edited by Thomas R. Lyons, pp. 157– tions in New World culture history, edited by 171. National Park Service. Raymond Thompson, pp. 1–6. University of Earls, Amy C. Arizona Bulletin No. 29. University of Ari- 1988 History of room use in area H of Acoma Pueblo. zona Press, Tucson. Cibola County, New Mexico. Hill, James N. Eggan, Frederick R. 1968 Broken K. Pueblo: Patterns of form and 1950 Social organization of the western Pueblos. Uni- function. In New perspectives in archaeology, versity of Chicago Press, Chicago. edited by Sally R. Binford and Lewis R. Bin- ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 237

ford, pp. 103–142. Aldine Publishing Com- Kidder, Alfred Vincent pany, Chicago. 1958 Pecos, New Mexico: Archaeological notes. 1970 Broken K. Puebo: Prehistoric social organi- Papers of the Peabody Foundation for Archaeol- zation in the American Southwest. Univer- ogy 5. Phillips Academy, Andover. sity of Arizona Anthropological Papers, No. 18. Kolb, Charles C. Hill, W. W., edited and annotated by Charles H. 1985 Demographic estimates in archaeology: Lange Contributions from ethnoarchaeology on 1982 An Ethnography of Santa Clara Pueblo, New Mesoamerican Peasants. Current Anthropol- Mexico. University of New Mexico, Albu- ogy 26(5):581–599. querque. Kramer, Carol 1980 Estimating prehistoric populations: An eth- Hoebel, E. Adamson noarchaeological approach. In L’Archeologie 1979 Zia Pueblo. In Handbook of North American de ’Iraq, edited by M-T. Barrelet, pp. 315–334. Indians, volume 9, Southwest, edited by Al- Center National de la Recherche Scienti- fonso Ortiz. Smithsonian Institution, Wash- fique. ington, DC. Kroeber, Alfred L. Holschlag, Stephanie L. 1917 Zuni Kin and Clan. Anthropological Papers of 1975 Pot Creek Pueblo and the Question of Prehistoric the American Museum of Natural History, 18(2): Northern Tiwa Household Configuration. Ph.D. 39–204. dissertation, Washington State University, Kulisheck, Jeremy, Michael Adler, and John Pullman. Hufnagle 1998 Hopi Dictionary/Hopiikwa Lava`ytutuveni. Uni- 1994 Diversity and continuity in classic and protohis- versity of Arizona Press, Tucson. toric villages of the Taos district, Northern New Hunter-Anderson, Rosalind Mexico. Paper presented to the 59th Annual 1979 Observations on the Changing Role of Small Meeting of the Society for American Ar- Structural Sites in the Northern Rio Grande. chaeology, Anaheim, CA. In Archaeological investigations in Chchiti Res- Lange, Richard C., Craig P. Howe, and Barbara A. ervoir, New Mexico, edited by Jan V. Biella Murphy and Richard C. Chapman, pp. 177–186. Of- 1993 A study of prehistoric roofing systems in fice of Contract Archaeology, University of Arizona cliff dwellings. Journal of Field Ar- New Mexico. chaeology, 20:485–498. LeBlanc, Steven James, Steven R. 1971 An addition to Naroll’s suggested floor area 1994 Regional variation in prehistoric Pueblo house- and settlement population relationship. holds and social organization: A quantitative American Antiquity 36(2):210–211. approach. Ph.D. dissertation, Arizona State University, Tempe. Lekson, Stephen H. 1997 Change and continuity in Western Pueblo 1986 Great Pueblo architecture of Chaco Canyon. households during the historic period in the University of New Mexico Press, Albuquer- American Southwest. World Archaeology que. 28(3):429–456. Lekson, Stephen H., Thomas C. Windes, John R. Stein, and W. James Judge Johnson, Douglas A. 1988 The Chaco Canyon community. Scientific 1992 Adobe brick architecture and Salado ceram- American 259(1):100–109. ics at Fourmile Ruin. In Proceedings of the Second Salado Conference, Globe, Arizona, ed- Levy, Jerrold E. ited by Richard C. Lange. Arizona Archaeo- 1992 Orayvi revisited: Social stratification in an logical Society, Phoenix. “egalitarian” society. School of American Re- search, Santa Fe, NM. Jorgensen, Joseph G. Lightfoot, Kent, G., and Gary M. Feinman 1980 Western Indians: Comparative environments, 1982 Social differentiation and leadership devel- languages, and cultures of 172 Western Ameri- opment in early Pithouse villages in the can Indian tribes. W. H. Freeman, San Fran- Mogollon region of the America Southwest. cisco. American Antiquity 47:64–86. Jorgenson, Julia Lightfoot, Kent G., and Steadman Upham 1975 A room use analysis of Table Rock Pueblo, 1989 Complex societies in the prehistoric Ameri- Arizona. Journal of Anthropological Research can Southwest: A consideration of the con- 31:149–161. troversy. In The Sociopolitical Structure of Pre- 238 CATHERINE M. CAMERON

historic Southwestern Societies, edited by Ortiz, Alfonso Steadman Upham, Kent G. Lightfoot, and 1969 The Tewa world. The University of Chicago Roberta A. Jewett, pp. 3–32. Westview Press, Press, Chicago. Boulder. Parsons, Elsie C. Lowell, Julie Carol 1936 Taos Pueblo. General Series in Anthropology, 1989 Flexible social units in the changing com- Vol. 2. munities of Point of Pines, Arizona. In Reid, J. Jefferson Households and communities, edited by A. S. 1998 Return to migration, population movement, MacEachern, David Archer, and Richard and ethnic identity in the American South- Garvin, Proceedings of the Twenty-First west: A peer reviewer’s thoughts on archae- Chacmol Conference, University of Calgary, ological inference. In Vanishing river, land- Department of Archaeology, Calgary. scapes and lives of the Lower Verde Valley, 1988 The social use of space at Turkey Creek edited by Stephanie M. Whittlesey, Richard Pueblo: An architectural analysis. The Kiva Ciolek-Torrello, and Jeffrey H. Altschul, pp. 53(2):85—100. 629–638. SRI Press, Tucson. 1991 Prehistoric households at Turkey Creek Pueblo, Reid, J. Jefferson Reid and Stephanie M. Whittlesey Arizona. Anthropological Papers of the Uni- 1982 Households at Grasshopper Pueblo. Ameri- versity of Arizona, No. 54. University of Ar- can Behavioral Scientist 25(6):687–703. izona Press, Tucson. Robinson, William J. McGuire, Randall, and Michael B. Schiffer 1990 Tree-ring studies of the Pueblo de Acoma. 1983 A theory of architectural design. Journal of Historical Archaeology, Vol. 24. Anthropological Archaeology 2:277–303. Martin, Paul S., and John B. Rinaldo Rohn, Arthur H. 1950 Sites of the Reserve Phase, Pine Lawn Val- 1977 Cultural change and continuity on Chapin ley, Western New Mexico. Fieldiana: Anthro- Mesa. The Regents Press of Kansas, Law- pology 38(3). Chicago Natural History Mu- rence. seum, Chicago. Roney, John R. Mills, Barbara J. 1996 The Pueblo III period in the Eastern San 1998 Migration and Pueblo IV community reor- Juan Basin and Acoma–Laguna Areas. In ganization in the Silver Creek area, East- The prehistoric Pueblo world A.D. 1150–1350, Central Arizona. In Migration and reorganiza- edited by Michael A. Adler, pp. 145–169. tion of the Pueblo IV period in the American University of Arizona Press, Tucson. Southwest, edited by Katerine Speilman. Ar- Schiffer, Michael B. izona State University Anthropological Pa- 1983 Toward the identification of formation pro- pers No. 51, Tempe. cesses. American Antiquity 48:675–706. Mindeleff, Cosmos Schlanger, Sarah 1900 Localization of Tusayan Clans. Bureau of Eth- 1986 Population studies. In Dolores archaeological nology Nineteenth Annual Report, pp. 635–653. program: Final synthetic report, compiled by Mindeleff, Victor David A. Breternitz, Christine K. Robinson, 1989 A study of Pueblo Architecture, Tusayan and Timothy Gross, pp. 493–524. Bureau of and Cibola. In Eighth annual report of the Reclamation, Engineering and Research Bureau of American Ethnology 1886–’87, pp. Center, Denver. 13–228. Government Printing Office, Wash- Snow, David H. ington. Reprinted, 1989 Smithosonian Insti- 1976 Archaeological excavations at Pueblo Del tution Press, Washington, DC. Encierro, LA 70, Cochiti Dam Salvage Naroll, Raoul Project, Cochiti, New Mexico. Final report: 1962 Floor area and settlement population. Amer- 1964–1965 field seasons. Museum of New ican Antiquity 27(4):587–589. Mexico, Laboratory of Anthropology Notes No. Nelson, Nels C. 78. 1914 Pueblo ruins of the Galisteo Basin, New Stubbs, Stanely, and W. S. Stallings Mexico. Anthropological papers of the Ameri- 1953 The excavation of Pindi Pueblo, New Mexico. can Museum of Natural History XV(I). Ameri- Monographs of the School for American Re- can Museum of Natural History, New York. search, No. 18. University of New Mexico Netting, Robert McC., Richard R. Wilk, and Eric J. Press, Albuquerque. Arnould Sullivan, Alan P. III 1984 Introduction. In Households. University of 1974 Problems in the estimation of Original California Press, Los Angeles. Room function: A tentative solution from ROOM SIZES AND ARCHAEOLOGICAL INTERPRETATION 239

the Grasshopper Ruin. The Kiva 40(1–2):93– White, Leslie A. 100. 1982 The Pueblo of Santa Ana, New Mexico. Titiev, Mischa Memoirs of the American Anthropological Asso- 1944 Old Oraibi: A study of the Hopi Indians of Third ciation, Vol. 60. Menasha, WI. Mesa. Peabody Museum of American Ar- Whiteley, Peter M. chaeology and Ethnology Papers, Vol 22, 1988 Deliberate acts. University of Arizona Press, No. 1. Harvard University Press, Cam- Tucson. bridge. n.d. Household census of Oraibi done in 1933–34. Wilk, Richard R. Transcription and computerized data base in 1983 Little house in the jungle: The causes of possession of Dr. Jerrold Levy, Department variation in house size among modern Kek- of Anthropology, University of Arizona, Tuc- chi Maya. Journal of Anthropological Archae- son. ology 2:99–116. U.S. Census of Population 1900 Population census schedules: Moqui Indian res- Wilshusen, Richard H. ervation. National Archives of the United 1995 The Cedar Hill Special Treatment Project: Late States. Microfilm Roll T623, Tape 48. Pueblo I, Early Navajo, and historic occupations in Northwestern New Mexico. Compiled by Ware, John A. and Eric Blinman Richard H Wilshusen, La Plata Archaeolog- 1998 Archaeology, Historical ethnography and Pueblo social history. Unpublished manuscript. ical Consultants, Research Papers, No. 1. Wetherington, Ronald K. Windes, Thomas C. 1968 Excavations at Pot Creek Pueblo. Fort Burgwin 1987 Excavations at the Pueblo Alto Complex, Chaco Research Center, No. 6. Taos, New Mexico. Canyon New Mexico 1975–1979. Volume I. Wedel, Waldo Summary of Tests and Excavations at the Pueblo 1979 House floors and native settlement popula- Alto Community. Publications in Archaeol- tions in the Central Plains. Plains Anthropol- ogy 18F, Chaco Canyon Studies, National ogist 24(84), Part 1:85—98. Park Service, Santa Fe.