This report discusses all animal bones collected during excavations by the Crow Canyon Archaeological Center at Castle
Rock Pueblo. The total site assemblage reported here is a mixture of specimens derived from a probabilistic sampling
design and from judgmental placement of excavation units.
Most zooarchaeological interpretation requires comparison between sites or comparison of archaeological assemblages
with models of animal bone frequencies derived from ethnographic or experimental studies. For example, the relative
frequencies of different species at Castle Rock Pueblo must be interpreted by reference to other sites in the region. Analysis
of the frequency of different parts of the skeleton has to consider models of the survival of bones of different densities and
models of the effects of human behavior on the transport of different parts
of a prey carcass (Lyman 1994*1).
The zooarchaeological data from Castle Rock Pueblo have already been compared with data from other Pueblo III
assemblages in the Sand Canyon locality (Driver 1996*1, 1997*1; Driver et al. 1999*1). The data have also been
incorporated into a larger study of ancient Puebloan assemblages from the
northern San Juan basin (Driver 1999*1). Rather than repeat
previously published material, this report offers a short summary of the
zooarchaeological database and
a brief consideration of bone preservation and intrasite spatial variation at Castle Rock Pueblo.
Information about each specimen was entered on a separate line of a
computerized database. The database, along with the database codes and analytical methods, will eventually be available on-line at the Crow Canyon Archaeological Center.
Specimens were considered identifiable if the skeletal element could be
identified. Specimens for which the element could not be determined were
recorded as "unidentified." Taxonomic designations for the identified
specimens ranged from general (for example, small mammal) to species name.
The following fields were used to describe each specimen: provenience,
taxon, element, part of element, state of fusion, breakage, modification,
length, and thickness (long bones only). These fields are discussed more
fully elsewhere (Driver
et al. 1999*1).
Altogether, 2,504 specimens were identified, and another 2,981 were considered unidentified (Table 1). All of the taxa
identified have been found at other archaeological sites in the Northern San Juan region, and most have been identified at
other sites in the Sand Canyon locality (Driver et al.
1999*1). Amphibians and reptiles were not identified precisely and
are probably intrusive.
As is the case with most Pueblo III sites in the Sand Canyon locality, the bird fauna seems impoverished when compared
with that of other Southwestern sites. Other than turkeys, which were
probably domestic (Munro 1994*1), birds were dominated by
raptors, including a large hawk, a large falcon, and a kestrel. Four of
the five owl bones are similar to those of Aegolius sp. and were
collected from Structure 305. Specific identifications, however, could not
be made, because a complete comparative collection of owls was unavailable.
It is notable that a large part of an immature Buteo (hawk) skeleton
was found on the floor of Structure 107. Twenty-two Buteo specimens
were from this skeleton, and another 52 specimens (mainly vertebrae and
ribs) identified only as "medium bird" are certainly from the same
skeleton. This was probably a deliberately deposited bird, and it suggests
that young hawks might have been kept at the site. It should be noted
that most of the specimens identified only as "large bird" are probably from turkeys.
Mammal remains were dominated by lagomorphs (jackrabbits and cottontails). Very small rodents are probably
underrepresented because most small rodent bones are lost during screening. Larger rodents include members of the
squirrel family (chipmunks, ground squirrels, and prairie dogs) as well as
pack rats (Neotoma sp.). The problem of distinguishing intrusive
rodents and lagomorphs from those that were used by people has been
discussed elsewhere for Sand Canyon locality sites (Driver et
al. 1999*1). Carnivores are poorly represented, and, as at many other
Pueblo III sites in this area, bones of domestic dogs are scarce.
Artiodactyls are rare; most of those identified to the genus level are
The relative frequencies of different parts of the skeleton can provide information about cultural practices and preservation.
Element frequencies for any taxon can be derived from the database in conjunction with the coding system. The most
common species at Castle Rock have most parts of the skeleton represented. For cottontails, for example, Table 2 shows
that the cranium, mandible, forelimbs, and hind limbs are well represented. The underrepresentation of some elements can
be attributed to small size (for example, the relatively small numbers of phalanges or the predominance of larger lumbar
vertebrae over smaller thoracic and cervical vertebrae), to fragility (for example, the underrepresentation of vertebrae in
general), or to problems in positively identifying the genus of certain skeletal parts (for example, rib fragments). For
turkeys and large birds (Table 2) there is better representation of different body areas because even the smaller skeletal
elements are larger than many cottontail bones. All of the more common species seem to have been brought to the site as
complete carcasses. Even the artiodactyls display quite a wide range of body parts (Table 2), especially considering the
small number of specimens in the assemblage.
One way in which preservation of fauna can be measured is to compare the relative frequencies of skeletal areas that
preserve well with those that tend to preserve poorly. The simplest way to do this is to compare the frequencies of different
parts of the same skeletal element. For example, in most mammals the proximal end of the humerus is much more
susceptible to damage than the distal end. Using data from the cottontail assemblage, we can count the numbers of different
ends of long bones for which susceptibility to damage is likely to vary. We can expect the proximal humerus and proximal
tibia in cottontails to preserve poorly compared with the distal ends of those same elements, unless preservation conditions
are good. For the radius, it is expected that the proximal end will preserve better. As Table 3 shows, these expectations are
borne out for cottontails, suggesting that the taphonomic pathways have resulted in considerable bone loss. The same
patterns can be seen for the much smaller sample of jackrabbit bones. In the bird skeleton there are less obvious differences
in bone density, but the distal tibiotarsus is denser and more robust than the proximal tibiotarsus. This difference shows up
well in the relative frequencies of turkey and large bird tibiotarsus ends in the Castle Rock assemblage (Table 3).
The faunal assemblage from Castle Rock Pueblo was collected from different types of contexts both within and outside
various types of structures. Although it would be preferable to divide the assemblage in a way that would reflect the
contextual details, this cannot be done, because the resulting small assemblages could not be compared meaningfully. It is
necessary, therefore, to group contexts in various ways. A further problem is the diversity of taxa. In order to reduce this
diversity, eight taxonomic groups were defined and are used throughout this section. The eight groups are as follows:
- Herpetofauna. All amphibians and
- Raptors, etc. All hawks,
falcons, owls, and ravens
- Turkeys. All specimens
identified as turkey, plus all specimens identified as "large bird"
- Lagomorphs. All lagomorphs
(includes jackrabbits, cottontails, and unidentified lagomorphs)
- Sciurids. All members of the
squirrel family (includes chipmunks, ground squirrels, prairie dogs, and
- Other rodents. All non-sciurid rodents)
- Carnivores. All carnivores
- Artiodactyls. All artiodactyls
Using classifications of context types supplied by the excavators, we can consider from which types of contexts the faunal
remains were collected. Table 4 shows that fauna was collected most commonly from secondary refuse, from deposits associated
with collapsed structures, from mixed deposits, from postabandonment
sediments, and from de facto refuse. This means that most faunal
specimens were probably from refuse deposits. Such deposits include
middens as well as refuse that had been reused for construction purposes
(for example, as roofing material) or that was redeposited in rooms after
abandonment. The small quantities of de facto and primary refuse suggest that even
faunal remains found in the fill of structures probably do not reflect the
activities that took place in the structure but, rather, the refuse that
was deposited in
In order to compare the contents of different types of structures in different types of locations, Table 5 organizes structures
into three typeskivas, rooms, and towersby location within the site.
Within-site locations were divided into four zones. Using the site grid as
a reference, "south" structures are on the eastern half of the south side
of the butte and in an area of large boulders on the southeast margin of
the site. "South-central" structures are on the western end of the south
side of the butte. "North-central" structures are on the western end of
the north side of the butte. "North" structures are on the northeast
periphery of the site. In order to compare the contents of structures with
assemblages from open areas, midden deposits
("Nonstructure 1") are also summarized in Table 5.
As I have noted previously (Driver et al. 1999*1), the
main difference between middens and structure fills is that middens
contain relatively more turkey and large bird bones (44 percent of all specimens in Nonstructure 1 as opposed to 22 percent
of specimens in all the structures reported in Table 5). It is difficult to find a ready explanation for this, although the pattern
has been observed at other sites in the locality. It is possible that collapsed masonry structures provide a more attractive
habitat for postabandonment colonization by small mammals, and the lower frequency of turkey remains is the result of the
intrusive, postabandonment addition of noncultural specimens to the fill of structures. Rodent bones are slightly more
common in structures (where they make up 22 percent of all faunal elements) than in middens (where they make up 15
percent). Cottontails (which burrow) occur in higher frequencies in structures (42 percent) than in middens (30 percent),
whereas jackrabbits (which do not burrow) occur in about the same proportion in both. If burrowing cottontails were
responsible for lowering the relative frequency of turkey bones in structures, then we would expect to see the cottontail-to-jackrabbit ratio change between middens and structures. Indeed, cottontails account for 86 percent of the cottontail-plus-jackrabbit assemblage in middens but 92 percent in structures, suggesting that some cottontail specimens might be intrusive
in structures. For all rodents (Rodentia plus Sciurdae in Table 5), it can be seen that midden assemblages contain 15 percent rodents,
whereas the total assemblage from all structures contains about 22 percent rodents. This, too, supports the hypothesis that
rooms attract more burrowing species.
As at other Pueblo III sites in the Sand Canyon locality, the Castle Rock faunal assemblage is dominated by lagomorphs and
turkeys/large birds. Meat was probably procured by local trapping of small game, combined with the production of
domestic turkeys. It is difficult to assess the role of rodents in the economy, but if their presence is due to human activity,
then it reinforces the picture of local trapping. Hunting of larger food animals was relatively uncommon, as was the hunting
of carnivores. Most skeletal parts are represented in the more common taxa, suggesting that complete carcasses were
brought to the site. Birds other than turkeys were probably hunted for purposes other than subsistence. The largely complete
immature hawk skeleton from a kiva floor is the best evidence for the ceremonial or ritual use of birds.
References cited | To borrow, cite, or request permission