Mogollon Snowbirds:
Where did they go for the winter?
A research paper for ASB 567
by Paul Wren
Introduction
The transition
from seasonal round to a sedentary life way in the Mogollon area (as well as in
other regions of the Southwest) is a subject of great interest. The archaeological record provides
significant evidence of sites which represent year-round occupation, but the
seasonally-occupied sites are not so well represented, nor as easily
recognized. This is likely a
natural consequence of both the amount of time that has passed since the
transition to sedentism, and the less permanent nature of seasonal sites.
Locating and
studying seasonally-occupied sites in the Mogollon region is essential to
accurately describing and interpreting the transition to sedentism. It is likely that Mogollon pithouses
from (or contemporaneous to) the Pine Lawn phase (Martin 1943) which do not
feature interior hearths represent summer occupation only, given the cold
winters in the area (Glen Rice, personal conversation).
If these are
indeed summer residences, where are the winter sites? Since they will likely have interior hearths, pithouses
occupied in the winter are difficult to distinguish from year-round highland
sites. Even without such an obvious
indicator as the presence or absence of interior hearths, I believe criteria
exist which can, when combined, be used to reveal seasonal sites occupied
during the winter.
I am concerned
with sites which fall into the "early pithouse" period, which spans
roughly A.D. 200-500 and is known by different
phases in different areas (see Table 1).
Region Early
Phase
Forestdale
Valley Hilltop
Reserve
Area Pine
Lawn
San
Simon Pen‡sco
Table 1: Early Pithouse Cultural Phases
(Haury 1985, Martin
1943, Sayles 1945)
My plan is to
survey existing site reports of early Mogollon pithouse villages, and apply the
following criteria in an attempt to identify winter seasonal sites. I have identified the following
pithouse village features and attributes as potential indicators of seasonal
occupation:
Storage Pits
A seasonal site
should contain more storage pits than a year-round site, specifically to cache
tools which are difficult to carry or prove useful in exploiting only the local
environment.
Pithouse
Construction Quality
One would
expect that houses in a permanent settlement exhibit greater effort and care in
terms of construction techniques and materials. Alternatively, one might expect that houses occupied
seasonally, and then typically for only a few seasons, would be built with less
effort and represent a lower quality of construction. Even though a number of criteria could be used to judge
overall construction quality, the site data available in the literature
restricted my sampling to the number of postholes per pithouse, the regularity
of pithouse shape, and the variance in house diameter within a single
site. I have combined this
information into an index which (hopefully) measures house construction
quality. This index is described
in a later section.
Burials
If people are
living in a particular spot only half of the year, they will only bury half of
their dead there. While this may
be obvious, it is a difficult indicator to measure. It is not enough to compare the number of burials per
village. It is not even sufficient
to compare the number of burials per household-- because the length of
occupation directly affects the number of burials at a site. To be meaningful, the desired
measurement is the number of burials per capita, per time period. Given the nature of existing site
reports (while interior burials are reported, cemeteries are seldom sought out)
and the difficulty in precisely determining the period of site occupation, this
promises to be the weakest criterion.
I will attempt to record the total number of burials at a site, as well
as the total number of pithouses (this can be roughly converted to the number
of residents), and the estimated duration of occupation.
Interior
Hearths
While the
presence of interior hearths is not an indicator of winter sites, their absence
can certainly be used to eliminate candidates from consideration.
Research Approach
After an
extensive search of available publications, I identified the following
archaeological site reports as potential sources of useful data:
Duncan
Site (Lightfoot 1984)
Bluff
Site (Haury 1985)
Bear
Ruin (Haury 1985)
Harris
Village (Haury 1986)
Mogollon
Village (Haury 1986)
SU
Site (Martin 1940-43)
Turkey
Foot Ridge Site (Martin and Rinaldo 1950)
Flattop
Ruin (Wendorf 1953)
Crooked
Ridge Village (Wheat 1954)
Nantack
Village (Breternitz 1959)
Starkweather
Ruin (Nesbitt 1938)
Promontory
(Martin, Rinaldo, and Antevs 1949)
Swarts
Ruin (Cosgrove 1932)
Cameron
Creek Village (Bradfield 1931)
San
Simon Village (Sayles 1945)
Cave
Creek (Sayles 1945)
For this study,
I created a data recording form to use when extracting information from each
site report (facsimiles of the completed forms can be found in Appendix
A). This form included the following
fields:
Site
Information
1.
Elevation
2.
Estimated period of occupation
3.
Number of burials
4.
Number of early pithouses
5.
Number of exterior pits and their maximum diameter, maximum
depth
Pithouse
Information
1.
Greatest diameter
2.
Number of interior hearths
3.
Number of interior storage pits
4.
Number of primary postholes
5.
Shape = Circular ("C"), Rectangular ("R"), or Irregular
("I")
6.
Any other interesting architectural features
7.
Assigned Phase or Dates
Data Overview
Upon closer
examination of all site reports, I determined that some of the sites I had
chosen were not appropriate for this study. I excluded the Swarts Ruin (Cosgrove 1932), Nantack Village
(Breternitz 1959), and San Simon Village (Sayles 1945) as they represented
later period pithouse villages. I
originally recorded data from Nesbitt's 1938 report on Starkweather Ruin (the summary report
for this site can still be found in Appendix A), but excluded it from
statistical analysis because it does not seem to fit well with the other
sites-- the lack of interior hearths, high count of interior pits and high
variance in diameter are at odds with the regularity of house shape and later
phase assignment by the excavator.
Figure 1 is a
map which shows the location of each site included in my analysis.
For several
sites, I recorded information for certain pithouses which I later excluded from
the analysis for various reasons (on the summary reports in Appendix A, these
are the pithouses appearing at the bottom of the table and separated from the
houses by a blank line). For
statistical analysis, my goal was to include for each site only residential
structures from the primary cultural phase which the site represented.
For the Bluff
Site (Haury 1985), I excluded House 5 as Haury believed it to be a ceremonial
structure, and I excluded Houses 11, 14, 19, and 19a since they are assigned to
the Cottonwood and Corduroy phases.
Figure 1: Early Mogollon Pithouse Sites
I excluded
Pithouse A and Pithouse V at the SU Site (Martin 1940, Martin and Rinaldo 1947)
from analysis since they appear to be ceremonial.
At Promontory,
another site excavated by Martin and others (1949), I excluded Pithouse B from
analysis as it also appears to be ceremonial.
Room 2 at the
Winn Canyon Site (Fitting 1973) is distinguished from the
other pithouses
at the site by its larger size and complex interior
features.
Fitting interpreted Room 2 as a ceremonial room, and for that reason I have
excluded it from analysis.
House 34 at
Harris Village (Haury 1986) is excluded from the analysis data since almost
none of the desired data points were available in the site report.
At the Turkey
Foot Ridge site (Martin and Rinaldo 1950), I excluded Pit House K as a
ceremonial room on the basis of its larger size, as well as the unique
rectangular area in the center which is defined by four partially buried logs.
Finally, I am
excluding houses 3 and 5A at the Mogollon Village, since Haury (1986)
identified them as possible ceremonial structures.
Table 2
provides a summary of the data collected for all sites. I have ordered the sites roughly in
chronological order.
The reader
should note that I have not included information on the external storage pits
(i.e., those outside of pithouses) at each site. It is my opinion that the sparse information which I did
collect does not accurately reflect the true nature of external storage pits at
the various sites. Many site
reports did not mention external pits, and their excavation techniques
obviously were not designed with the intent of locating such pits. Some reports mentioned storage pits,
but failed to quantify them or describe them in any way.
I would also
like to point out that the dates presented in the "Occupied" column
are not meant to imply that the sites were occupied throughout the specified
range, but rather that the occupation is estimated to have fallen somewhere
within that range.
Figure 2
provides a graphical view of the total burials per site, divided by the number
of pithouses (this translates into a number which should be proportional to
burials per capita). I had hoped
that burials might give some clue as to how long these people lived at a
certain site, but the actual data appears to be completely useless. There is no discernable trend in
the data, and I suspect that the number of burials discovered at each
particular site has more to do with excavation approaches and research goals
than with the actual number of inhumations.
The average
number of postholes per house at each site (Figure 3) do not present obvious
trends. If, however, one does not
pay too much attention to the SU site or the Duncan site, it becomes apparent
that the early period pithouses have fewer postholes per house. I should point out that the average
house at the SU site is significantly larger in area than those at other
sites. When the average house is
nearly 20 feet in diameter, one should expect more roof supports.
Figure 4 shows
the general trend toward smaller houses moving out of the early pithouse
period, and Figure 5 supports the presumption that more interior storage pits
occur in the early pithouse period.
With the
notable (although not singular) exception of the Cave Creek site (see Figure
6), early pithouse villages show few interior hearths. The Winn Canyon site, representing the
latest "early" pithouse village in my survey, is another exception
(four out of five residences exhibited hearths), but this difference is not
visually obvious due to its position on the graph.
Figure 2: Burials per House, by Site
Figure 3: Average Number of Postholes per House,
by Site
Figure 4: Average House Diameter by Site
Figure 5: Average Number of Interior Pits per
House, by Site
Figure 6: Percent of Pithouses with Hearths by
Site
Construction Quality Metrics
To determine a
measure of the construction quality represented at each site, I have combined
the variance in pit house diameters and the average number of primary postholes
with a number I call the Shape Regularity Index.
This index is higher for a site which contains a larger percentage of
pithouses whose shapes are "Irregular."
The Shape
Regularity Index is computed by first assigning a score of zero to each house
which is either circular or rectangular, and a score of five for every house
whose shape is "Irregular."
The average score for each site is used as its Shape Regularity
Index. In other words, a site whose
pithouses are all Circular or Rectangular will have an index of zero, a site
containing only Irregular pithouses will have an index of five, and all other
sites will fall somewhere between.
Table 3 shows
the pithouse diameter variance, the mean number of primary postholes per house
divided by the mean house area (and multiplied by 20 to normalize this
measurement with the posthole mean), and the Shape Regularity Index for each
site, as well as the composite measure of construction quality.
Table 3: Construction Quality Metrics
The formula for
the construction quality measure is as follows:
Where S2 is the variance of the house diameter
sample, Is is the Shape Regularity Index, Xp is the mean number of postholes per
house for the site, and Xmax is the maximum average number of
postholes observed for any site.
To be able to
sum up the three metrics, I needed all of the measurement scales to be oriented
such that a lower value indicated higher construction quality. Since I'm assuming that more postholes
indicates higher construction quality, I have inverted this scale by
subtracting the posthole index from the maximum posthole value. Thus, a large posthole mean versus
house area will result in a lower index.
It was necessary to multiply this posthole index by a constant to
normalize it with the other metrics-- I chose 20 since it resulted in a maximum
value equivalent to the originally recorded maximum posthole mean.
Figure 7: Construction Quality Index (Lower is
better)
The graphical
representation of the Construction Quality Index is certainly a gratifying one
(see Figure 7), showing a definite increase in house construction quality over
time. Even the most obvious
anomaly-- the SU Site-- can be explained by the "inflated" posthole
measurement, which is a direct result of the larger house size at the site (see
previous discussion of Figure 3 in the "Data Overview" section).
Interpretations
In general, the
information I collected supports the use of most of my previously-discussed
criteria as a method for discriminating early pithouse sites from later ones,
and hopefully can aid in identifying seasonal sites. For convenience, I will reiterate the measurements I believe
are useful:
- Construction Quality Index (based on posthole,
house diameter, and house shape data)
- Interior Pits per House
- Percentage of Houses with Interior Hearths
I expect to
correlate poor construction quality and a higher incidence of interior pits
with seasonal sites, and that the presence of interior hearths can be used as
an indicator of winter occupation.
Based on these
criteria, are there any sites (in addition to the SU site) which seem to be
seasonal? Four sites in my survey
exhibit indications which seem to identify them as seasonal.
The Promontory
site (Martin, Rinaldo and Antevs 1949) not only has a relatively higher number
of interior pits, few interior hearths, and a poor construction quality index,
it also happens to be contemporaneous to the SU site. Additionally, it is located in the Pine Lawn Valley, a short
distance from the SU site. If the
same criteria which indicate the SU site is seasonal are applied to Promontory,
it must also be interpreted as a summer seasonal site.
A close look at
the Bluff site data (Haury 1985) reveals a similar profile, and although the
number of interior pits is much lower than those of the SU or Promontory sites,
the near absence of interior hearths (when considered in combination with the
site's high elevation) seems to rule out winter occupation.
Kent Lightfoot
himself (1984) identifies the Duncan site as a seasonal camp occupied in the
summer, and all three of my criteria support this interpretation to some
degree.
The Winn Canyon
site is particularly interesting.
It it similar to the other presumed seasonal sites in terms of
relatively poor construction quality, and Fitting (1973) identifies it as being
contemporary to the SU site.
Although the raw number of interior pits is not terribly high, a closer
look is more compelling-- each house has at least one storage pit, and the pits
are large and regular in shape.
Additionally, many chipped and ground stone implements were discovered
in some of the storage pits.
The Winn Canyon
site distinguishes itself from the other previously-identified seasonal sites
by featuring a hearth in nearly every habitation. It was on the basis of this difference, combined with its
lower elevation (the Winn Canyon site, at 4610 feet, is almost 1800 feet lower
than the SU site), that I originally identified the Winn Canyon Site as a
possible winter seasonal site.
While the temperatures were no doubt still quite chilly at this elevation,
the winter weather would be much milder than that experienced above 6000 feet in the White Mountains.
Given these
indicators of possible winter seasonal occupation, I examined the Winn Canyon
site report (Fitting 1973) in more detail. The lithic assemblage contained a large number of projectile
points (Fitting theorizes that the resident of one of the excavated houses was
a master craftsman), hinting at hunting as a significant subsistence activity--
as one might expect during the winter.
Further
attention to detail revealed striking similarities between the Winn Canyon and
SU sites. Their geographical
setting and community layout are quite similar-- residences occupy twin mesas
separated by a saddle, and both sites have a possible Kiva at the tip of one
mesa (Fitting 1973). In addition
to this large-scale similarity, there are also many common architectural
features.
In general, the
early pithouses at both SU and Winn Canyon are circular, exhibit large,
interior storage pits in the floors, and have generally East-facing lateral
entries. All of the residential
houses excavated at Winn Canyon can be described as a single type, and are
quite similar to pit houses B, H, J, M, and T at the SU Site (Martin 1940,
1943; Martin and Rinaldo 1947).
Specifically, in addition to the above-mentioned traits, these houses
are similar in size, and have the same posthole pattern, i.e., one large
central post.
I believe this
data supports the interpretation of Winn Canyon as a seasonal site occupied
during the winter. And even though
there is not enough evidence to prove it, there is no reason to rule out the
possibility that the Winn Canyon site may have been the winter home of the very
people who occupied the SU site.
The distance between the two sites (roughly 60 miles) is acceptable for
a community practicing a seasonal round, the two villages are morphologically
similar, and ceramics indicate they are contemporaneous and of the same
culture.
Summary
1. Several published site reports from
Mogollon pithouse villages allowed me to define a set of identifying
characteristics which distinguish early pithouse sites from later ones.
2. I have
developed a set of criteria which I believe can be used to identify early
pithouse villages which are occupied for only part of the year (i.e., seasonal
sites).
3. Using the
aboved-mentioned set of criteria, I have proposed that the following sites were
occupied for only the warmer part of the year: the SU site, the Promontory
site, the Bluff site, and the Duncan site.
4. I have proposed that the Winn Canyon
Site was seasonally occupied during the winter, and that there is no evidence
to rule out interpreting the Winn Canyon site as the winter home of those
people residing at the SU site in the summer.
References Cited
Fitting, James
E.
1973 An early Mogollon
community: a preliminary report on
the Winn Canyon site. The
Artifact vol. 11, no. 1-2. El Paso, Texas: El
Paso Archaeological Society.
Haury, Emil W.
1985 Mogollon Culture in
the Forestdale Valley, East-Central Arizona. Tucson,
Arizona: University of Arizona Press.
1986 Emil Haury's
Prehistory of the American Southwest. Tucson, Arizona:
University of Arizona Press.
Lightfoot, Kent
G.
1984 The Duncan Project: a study of the occupation duration and
settlement pattern of an early Mogollon pithouse village. Anthropological Field Studies no. 6.
Tempe, Arizona: OCRM, Arizona State
University.
Martin, Paul S.
1940 The SU Site Excavations
at a Mogollon village, western New Mexico, 1939. Field Museum of Natural History Anthropological Series
vol. 32, no. 1. Chicago: Field Museum.
1943 The SU Site Excavations
at a Mogollon village, western New Mexico, second season, 1941. Field Museum of Natural History
Anthropological Series vol. 32, no. 2. Chicago: Field
Museum.
Martin, Paul S.
and John B. Rinaldo
1947 The SU Site Excavations
at a Mogollon village, western New Mexico, third season, 1946. Field Museum of Natural History
Anthropological Series vol. 32, no. 3. Chicago: Field
Museum.
1950 Turkey Foot Ridge Site:
A Mogollon village, Pine Lawn Valley, western New Mexico. Fieldiana: Anthropology, vol. 38, no. 2. Chicago: Field Museum.
Martin, Paul
S., John B. Rinaldo, and Ernst Antevs
1949 Cochise and Mogollon
sites, Pine Lawn Valley, western
New Mexico. Fieldiana:
Anthropology, vol. 38, no. 1. Chicago: Field
Museum.
Nesbitt, Paul
H.
1938 Starkweather Ruin: A
Mogollon-Pueblo site in the upper Gila area of New Mexico. Logan Museum Bulletin 6. Beloit, Wisconsin: Logan Museum.
Sayles, Edwin
B.
1945 The San Simon Branch:
Excavations at Cave Creek and in the San Simon Valley. Medallion Papers no. 34.
Globe, Arizona: Gila Pueblo.
