The Archaeology of Castle Rock Pueblo

¹
Estimating the population size of a hamlet or village is important for understanding subsistence and sociopolitical organization, as well as regional population trends. Arriving at an accurate estimate of population, however, is tricky. Researchers have developed a number of methods for estimating population (Cook 1972*1; Hassan 1981*1), with varying success. These methods include estimating on the basis of floor area (LeBlanc 1971*1; Naroll 1962*1), number of households (Lightfoot 1994*1), number of rooms (Hill 1970*1), number of kivas, number of artifacts (Cook 1972*1:11–12; Hassan 1981*1:78–79), amount of food refuse (Cook 1972*1), area of roomblock rubble on the modern ground surface (Adler 1990*5; Schlanger 1987*1), hearth size (Ciolek-Torrello and Reid 1974*1), site size (Hack 1942*1), and number of burials or bones (Cook 1972*1). The success of these methods depends on many factors, and how well a specific method works for a particular site depends on the kind of site it is and on the types of data that were collected. The methods considered most useful for estimating the population of Castle Rock Pueblo include calculations based on total floor area or roofed area, the number of households, the number of rooms, and the quantity of human skeletal remains.¹

Population Based on Area

²
Studies in which population is estimated on the basis of architectural measures emphasize that usually not all structures in Puebloan sites were in use at one time. This requires a downward adjustment in the population estimate to allow for the abandoned structures. Castle Rock Pueblo, however, was occupied for a relatively short time, and it appears that only a few structures had fallen into disuse before the habitation of the village ended, sometime after A.D. 1274. Consequently, no adjustment was made to the following estimates to allow for abandoned rooms.

³
On the basis of his study of floor areas in 18 societies, Naroll (1962*1) suggested a formula of 10 m² (108 ft²) of floor area per person to estimate population. To use this well-known formula for Castle Rock Pueblo, I first summed the floor areas of all known structures, for a total of 396 m² (4,261 ft²). Dividing by 10 m² gives a population estimate of 39.6, or 40 people. This seems to be too few people to have constructed and occupied all of the structures known to have been present in the village. It is probably safe to assume, however, that we did not find all structures during our excavations. It is possible that a similar roomblock or at least a similar amount of floor space once existed in each residence cluster. Using the one mostly excavated roomblock (Roomblock 103) as a model, I projected the total floor area that would have been present in the village if all roomblocks were the same size. Roomblock 103 encompassed 31 m² (333 ft²) of floor area. If each of the 15 kivas of ordinary size was originally paired with 31 m² of surface rooms, then the total floor area of the village would have been 465 m² (5,003 ft²) of surface room area plus 188 m² (2,023 ft²) of kiva floor area. This would boost the total floor area for the village to 653 m² (7,026 ft²). Using Naroll's formula of 10 m² per person, I estimate a population of 65 people, which seems reasonable. But it is difficult to believe that we overlooked 257 m² (2,765 ft²) of surface rooms, and the use of Naroll's formula has resulted in seemingly low population estimates for researchers in the past (Hassan 1981*1:73; Hill 1970*1:76–77; Kuckelman 1977*1:82).

⁴
Dohm's (1990*1:Table 3) data on 22 historic-period Pueblo villages indicate an average of about 17 m² of total roofed area per person. Total known roofed area at Castle Rock (which, unlike floor area, includes benches and southern recesses of kivas) is 475 m² (5,111 ft²). Use of Dohm's formula gives a population estimate of 30 people for the village, which is even lower than the estimate derived using Naroll's formula. Using Roomblock 103 as the model roomblock for the site, as I did in applying Naroll's formula, I estimate the total projected roofed area of the village at 729 m² (7,844 ft²). Dividing this by 17 m² gives a population estimate of 43 people. This would mean an average of about three people per kiva, which still seems too low. To check this impression, I used the same formula on the area contained in the 103 residence cluster by itself (roomblock and kiva). In other words, I divided 50.3 m² (541 ft²) by 17 m² (183 ft²), which resulted in a population estimate of three people for that residence cluster. This is too few people to have constructed and used a kiva and six rooms, including four front rooms with hearths. I conclude that historic Pueblo households contain much more space per person than did Castle Rock Pueblo residence clusters.

⁵
Calculating the population of Broken K Pueblo in Arizona, Hill (1970*1) used a variety of methods to arrive at a formula of 4.55 m² (49 ft²) of area per person. Applying Hill's formula to the floor area actually found at Castle Rock Pueblo (396 m², or 4,261 ft²) yields a population estimate of 87 people. With 15 residence clusters (Structure 105 is believed to have been a communal structure and might not have been part of a residence cluster), Castle Rock Pueblo would have had an average of 5.8 people in each cluster, which is probably not unreasonable.

Population Based on Number of Households

⁶
Number of households can also be used to estimate population. Lightfoot (1994*1:148) concluded, on the basis of several cross-cultural studies of household size, that one could expect households in ancient pueblos to have ranged between two and 12 people and to have averaged between five and eight. The problem at Castle Rock is the same problem Lightfoot (1994*1:148–149) faced with the Pueblo I Duckfoot site, which was figuring out which physical remains represented a "household." Following Kroeber (1917*1:124), Beaglehole (1935*1:42) defined a household for modern Hopi villages as the house unit in which several people live around one hearth. This definition has been widely used by archaeologists in interpreting ancient pueblos (Cook 1972*1:16; Hill 1970*1:76; Kane 1986*1). If each of the four front rooms with hearths in Roomblock 103 at Castle Rock Pueblo represents a separate household, then this residence cluster housed between eight and 48 people. On the other hand, if the entire 103 residence cluster (kiva and rooms) represents one household, then the estimated number of residents falls to between two and 12 people. On the basis of Cook's (1972*1:Table 1) cross-cultural information on the floor areas of houses that probably each contained a single nuclear family, it appears that the front-and-back pairs of rooms in Roomblock 103 (for example, Structures 121 and 122) are very small to have housed separate nuclear-family households. Given the small sizes of the rooms, then, I propose that the 103 residence cluster consisted of one large household rather than four households. For the site as a whole, I believe that each of the 15 small kivas represents one household. Using the average household size of between five and eight people cited above, this results in a population estimate for the entire village of between 75 and 120 people.

⁷
In his calculations of numbers of people in households, Cook (1972*1:16) used a formula of 25 ft² (2.3 m²) of space per person for the first six persons in a household and allowed 100 ft² (9.29 m²) of space for each additional person. For Roomblock 103, this formula gives an estimate of eight people, or 10 people if the kiva is counted as domestic space. If eight to 10 people occupied each of the 15 known residence clusters at Castle Rock, then the village contained 120 to 150 people.

Population Based on Number of Rooms

⁸
Population estimates can also be based on the number of rooms in a village. In estimating the population of Broken K Pueblo, Hill (1970*1:75) used information from modern Hopi and Zuni pueblos to arrive at a formula of 2.8 people per room, based on an average room size of 9 m² (96.8 ft²). The average size of rooms that could be accurately measured at Castle Rock Pueblo was 9.15 m², so this formula should be usable for Castle Rock. At least 40 rooms are known to have been present at Castle Rock. If the population equaled 2.8 people per room, then 112 people lived in the village.

Population Based on Human Bones

⁹
The final method used to estimate the population of Castle Rock Pueblo was to calculate the number of people represented by the human bones found during excavations. It seems safe to assume that some people died during the 20-year occupation of the village, although no formal burials were encountered during the five seasons of excavation there. How many people died during the occupation of Castle Rock Pueblo? Schlanger (1992*1:15) estimated, on the basis of analyses of Southwestern burial populations, that four to six people out of every 100 would have died each year. The chronology for Castle Rock is not precise enough to enable calculation of the size of the village year by year, but if its population averaged 50 people for the last 10 years of occupation, then at least 20 to 30 Castle Rock inhabitants must have died during the occupation of the village. It is possible that, by chance, no test pits were excavated where any formal burials were located, or that all the dead were buried in a specific location that is as yet undiscovered (such as on the nearby floodplain of McElmo Creek). It is also possible, considering the widespread recreational digging that occurred in the area historically, that some burials were removed in historic times.

¹⁰
Although no formal burials were found, human bones were discovered at the site, all in locations and conditions indicating that these people died in the final days of occupation (see also "The Final Days of Castle Rock Pueblo"). The skeletal remains of some individuals were relatively complete and articulated, but the remains of most were disarticulated. The number of people represented by these remains must therefore be estimated. I estimate that at least 43 people died during the final days of the occupation. There is no direct evidence to indicate that all of the dead were residents of Castle Rock, but it can probably be assumed that the victorious attackers would not have left their dead in an enemy village. Various researchers have developed methods for projecting the number of inhabitants of a site on the basis of the number of human bones recovered, but many problems have arisen with these methods (see Cook 1972*1:4–7). Therefore I do not attempt to estimate the population of Castle Rock Pueblo on the basis of the number of human bones. At least 41 people are represented by the human remains documented at Castle Rock, and this number is offered as a minimum population for the village when the occupation ended.

Summary

¹¹
The population of Castle Rock Pueblo was estimated using various methods and formulas. Estimates based on floor area and total roofed area were as low as 43 people and as high as 87. Estimates based on the number of households present at the site tended to be higher—75 to 150 people. The number of rooms thought to have been in the village was used to estimate a population of 112 inhabitants. The human bones indicate that at least 41 people died in the village during the violent events that effectively ended the occupation.

¹²
Because above-ground rooms were difficult to find during the excavations (see "Architecture") and more emphasis was placed on finding kivas than on finding rooms during testing, I believe the most accurate population estimate is that based on the number of households in the village. The number of households is thought to be indicated by the number of kivas. Thus, I propose that the estimate of 75 to 150 people is the most accurate, although I favor the low end of this range because of the small number of rooms found.

¹ Analysis results for the human skeletal remains at Castle Rock Pueblo are not included in this publication. The complexities of the analysis and subsequent interpretations require treatment that is beyond the scope of this publication. An article is currently being written for submission to a professional journal.

References cited | To borrow, cite, or request permission
Copyright © 2000 by Crow Canyon Archaeological Center. All rights reserved.