Professional Documents
Culture Documents
Stark1981. The Rise of Sedentary Life
Stark1981. The Rise of Sedentary Life
^ V O L U M E O N F fr=Ii^r=]r=Jr=Jr=Jr=Jr=Jr=Jr=Jr=Jr=Jr=J
=jp=ir=dr=ip^i—J r= = Jr= Jr= Jr= Jr= Jr= Jr= Jr= Jr= Jf= Jr= iJr= Ji= ir= Jr= Jr= Jr= Jr= = Jr= Jr= Jr= u r= Jr= Jr= Jr= Jr= ir= 3
BARBARA L. STARK
345
BARBARA L. STARK
Therefore, a shift away from the degree of chaic period patterns, and we would have to
hunting that had been characteristic in ear become more concerned with the reasons
lier periods may have engendered some of why plant domestication did not occur until
the preconditions for subsequent changes much later in the prehistoric record. Other
leading to village agriculture. wise, a relatively late peopling of the New
It is the preconditions and selective factors World by specialized big-game hunters
which fostered development of sedentary would suggest that the Post-Pleistocene tim
food production that are of particular interest ing of sedentarization and cultivation is less
in this review. I attempt to reconcile certain problematic.
interpretations of the rise of sedentary Meso- The nature of the settlement patterns and
american farming in a systemic, multicausal subsistence practices of the earliest inhabi
model. Population growth, sedentarization, tants of Mesoamerica remains unclear de
food storage, distance costs, ecological con spite discovery in Mexico of two remarkable
ditions, and cultivation-domestication are sites suggesting occupation as early as 20,000
the processes and factors I incorporate in the B.C. Publication is incomplete in both cases,
model. Finally, I explore briefly some of the rendering my review provisional. Before de
probable consequences of settled farming in scribing these sites, I stress that we must shy
Mesoamerica. Throughout, I separate sum from firm conclusions based on such a small
maries of new archaeological evidence from sample of sites. Until the sites are completely
discussion of processes of change. Earlier published, it is useful to suspend final judg
reviews of some of these topics appear in ment about dating and cultural and faunal as
Aveleyra Arroyo de Anda 1964, MacNeish sociations (Haynes 1969). My purpose is sim
1964b, and Mangelsdorf, MacNeish, and ply to indicate the nature of the evidence as
Willey 1964; Warwick Bray (1976, 1977) pro it has been described and some of its possible
vides more recent surveys. implications if accepted.
Williams and Armenta Camacho began a which lends credibility to the mollusk date
project which located five sites: Hueyatlaco, associated with the flake. However, uranium
Caulapan, El Horno, El Mirador, and Teca- series dates for the sites of Hueyatlaco and El
caxco. Hueyatlaco is the most completely Horno gave unreasonably old dates.
published (Irwin-Williams 1967a), but Caula Early deposits at Tlapacoya in the Valley of
pan is the most adequately dated. Hueya Mexico lend support to the case for human
tlaco proved to be an important site because occupation by 20,000 B.c. Beginning in 1965
it produced a stratified series of small as a series of excavations probed the flanks of
semblages associated with extinct fauna. José the Cerro de Tlapacoya beside Lake Chaleo.
Luis Lorenzo (1969) subsequently challenged At Tlapacoya I, a series of three radiocarbon
the associations for a number of reasons, but dates agree in placing an early lakeshore be
several scientists observed the artifacts and tween about 25,000 and 20,000 b . c . A large,
excavations and found no cause to doubt rock-edged hearth intruded into this shore
their validity, which also has been defended deposit from just above; it is the earliest such
in detail by Irwin-Williams (1967b, 1969). feature known from Mesoamerica, with a
Of particular interest at Hueyatlaco is the charcoal date of 24,000 ± 4,000 radiocarbon
technologically simple material from Deposit years (A-794B). Two additional hearths were
I, the oldest cultural stratum. Six artifacts discovered in subsequent excavations, one of
were recovered—three flakes and three which yielded a date of 21,700 ± 500 ra
edge-retouched pointed pieces, probably diocarbon years (1-4449). Two piles of non
projectile points, which were found near un articulated bones occurred near the hearths
gulate and proboscidean bones. This as (Mirambell 1978:224); black bear, a heavy
semblage was stratigraphically below others species of deer, and white-tailed deer are
with bifacially retouched projectile points. among the fauna.
Three other Valsequillo sites, which are The slightly deeper lakeshore below the
stratigraphically earlier in the geology of the hearths contained a discoidal scraper made
area, produced unifacially edge-retouched on a flake; most of the other artifacts from
lithics also. The oldest is the mastodon kill this stratum, which are reported to number
and butchering site of El Horno; among the over 2,500, have not been completely illus
unifacially edge-retouched tools were imple trated, but are described as flakes, blades,
ments described as scrapers, burins, and and cores made from both local and nonlocal
pointed pieces which may have been projec materials (Mirambell 1973:8; 1974:61-62;
tile points. El Mirador yielded a single edge- 1978). Retouch is infrequent, and debitage is
retouched point made on a blade, found in a plentiful. Pleistocene fauna was abundant in
stratum with horse, mammoth, camel, mas the lakeshore deposit, but no direct associa
todon, and other fauna. Tecacaxco produced tions have been announced between most of
a number of chert flakes, some blades, and a the fauna and artifacts.
few edge-retouched artifacts described as At Tlapacoya II a lacustrine deposit con
scrapers and knives. At Caulapan, located in tained a tree trunk dated to 23,150 ± 950 ra
a nearby barranca, a single edge-retouched diocarbon years (GX 0959); nearby was a
flake appeared in a stratum with fossil mol- well-made obsidian blade, which would fur
lusks which were radiocarbon dated to 21,850 ther support the case for early human oc
± 850 radiocarbon years (W-1895). A pro cupation, except that the fine quality of the
boscidean vertebra from the same deposit blade seems anomalous.2
was dated by uranium series to 20,000 ±
1,500 years and 22,000 ± 2,000 years. An
Processes of Change
other uranium series date and three other ra
diocarbon dates from the geological section The two charcoal hearth dates from Tlapa
at Caulapan are in appropriate relative order, coya have led many scholars to view at least
347
BARBARA L. STARK
the timing of sedentarization, cultivation, ness of prey like mammoth. He views killing
and domestication makes sense only for the big game as an occasional activity that took
Archaic period. Currently, we cannot say advantage of animals who became mired.
that Paleoindian I hunting closely resembled Ruth Gruhn and Alan Lyle Bryan (1977) sug
Archaic patterns, which in many cases in gest that the functional distinctness of kill
volved carefully organized rotation among sites and campsites obscures the fact that
seasonal plant and animal resources. more generalized hunting and collecting
The best summary for this period is a cau were part of subsistence in many areas.
tion. Researchers have only begun to ac In terms of Mesoamerican data them
cumulate data and postulate processes of selves, perhaps the most telling argument
change. The bearing of these controversial against the prevalence of specialized big-
discoveries on the rise of sedentary farming game hunting has come from MacNeish’s
remains uncertain, but conceivably could (1964b) and Kent V. Flannery’s (1967:144)
raise a question like that often posed in the evaluation of paleoenvironmental and »subsis
Old World—why did domestication occur tence information from the Early Ajuereado
when it did? phase in the Tehuacan Valley, Puebla. Prior
to about 7600 R.C., cool, dry climatic condi
tions are indicated by Flannery’s faunal anal
Paleoindian II
ysis (see Schoenwetter 1974:297). In the
Mesoamerican data from the Paleoindian II semiarid Tehuacan Valley, this may imply a
period are widely scattered, although more lack of extensive forage for the larger her
plentiful than those from Paleoindian I (see bivores. Nor is there any indication from the
Aveleyra Arroyo de Anda 1964 for review of cave deposits that people in Early Ajuereado
earlier discoveries). Campsites and kill sites times hunted mammoth or most other large
are among the archaeological contexts, and Pleistocene forms. Rather, subsistence was
mobile hunting is the best-documented sub closer to the mixed Archaic period pattern of
sistence pursuit. Mammoth kill sites have hunting and gathering, with much of the
been discovered on the margins of lakes in prey consisting of small animals. However,
the Valley of Mexico, and there have been some game was taken which later became lo
scattered surface finds of fluted projectile cally or completely extinct, such as antelope,
points, which in the U.S. Southwest and horse, and a large rabbit. Mesquite, grass
Great Plains are found in association with kill seeds, and tree fruits were gathered plant
sites of mammoth (Clovis points) and extinct foods in Ajuereado times. MacNeish (1964a:
bison (Folsom points). Nonetheless, scholars 532) aptly summarizes this subsistence pat
disagree about the role of big-game hunting. tern with the comment: “They probably
One school of thought tends to view the found one mammoth in a lifetime and never
Paleoindian II period as a time when highly got over talking about it.” At issue in the de
successful, specialized big-game hunting de bates about Paleoindian subsistence is the
veloped or was introduced—a pursuit which degree of contrast between the Paleoindian
relied partly on the use of marshy areas and II and early Archaic periods and whether the
other natural aids to facilitate entrapping and timing of cultivation and domestication un
dispatching animals (Martin 1973). However, derstandably awaited Post-Pleistocene ex
many researchers have been skeptical of the tinctions. However, new Paleoindian II evi
prominent role of megafauna in subsistence, dence does not yet resolve the debates.
preferring to view the kill sites as one aspect
of a seasonal or mixed subsistence strategy.
Archaeological Evidence
Lorenzo (1975:32) argues against the idea
that big-game hunting was feasible on a gen In the Basin of Mexico, recently discovered
eral basis because of the size and dangerous sites reveal a stratigraphic association of
349
BARBARA L. STARK
Pleistocene megafauna and cultural remains, Thus, the scant Cueva Blanca remains re
although actual kill sites have not been lo semble the kinds of evidence from Tehua
cated. Angel Garcia Cook (1975) reports the cán—late Pleistocene and early Post
discovery of two pieces of worked or utilized Pleistocene camps possibly used by groups of
bone at Los Reyes La Paz in a lakeside set a few hunters who obtained relatively small
ting. Dates between 10,000 and 6000 B.c. animals. Lack of plant remains at Cueva
are proposed. Associated fauna in the stra Blanca probably is due to poor preservation
tum included mammoth, among a variety of rather than any marked difference from Te
other animals. At Cerro Chimalhuaque in huacán in subsistence.
Chimalhuacan-Atenco, García Cook (1968) New surface finds of fluted points have
discovered an obsidian flake associated with been reported elsewhere in Mexico. From
Pleistocene fauna, mainly mammoth. How the Laguna San Marcos and Laguna Zocoalco
ever, there was no indication of butchering. shores in Jalisco, Lorenzo (1964) examined
The deposit is correlated with the Becerra obsidian fluted points that appeared Clovis-
Formation, which has produced other Termi like. Phil C. Weigand (1970) found a fluted
nal Pleistocene associations of mammoth and point fragment in the talus below a rockshel-
man. To date little has been published about ter near Mezquitic, Jalisco. Clovis points
cultural remains at Tlapacoya XVII, where a have been reported from sites in Sonora (Ro
stratum dated to 7970 ± 220 b . c . (1-6897) bles Ortiz 1974; Braniff 1976:26), and Garcia
yielded a variety of lithic artifacts and a Cook (1973b) found a fragment of one near
human skull (Mirambell 1973:7-8; 1978: San Juan Chaucingo, Tetla, Puebla (cited in
224). Lorenzo 1975:37).
Outside the basin, the Valsequillo site of A campsite was discovered in highland
Hueyatlaco contained in its upper deposits Guatemala at Los Tapiales, but no bone or
artifacts with bifacial retouch, bipointed pro other indications about subsistence were
jectile points, and, in one case, a stemmed preserved (Gruhn and Bryan 1977). Among
projectile point. One of the strata included a the lithics were part of one fluted point and
horse kill and butchering station (Irwin a channel flake. Unifacial points and leaf
Williams 1967a). shaped bifaces, possibly projectile points,
Between 1966 and 1969 four preceramic were also part of the assemblage. Cutting,
sites in Oaxaca were tested (Flannery 1970), scraping, and woodworking tools are de
but only one produced remains from the Pa- scribed by Gruhn and Bryan as well. The site
leoindian II period. At Cueva Blanca, zone F contained four hearths, which produced
contained no artifacts, but it did produce some problematic radiocarbon dates on char
Pleistocene animal bones including fox, rab coal. The hearth dates seem too young to be
bit, small rodents, and a gopher-tortoise. associated with fluted points, but stratigra
Some bones may have been artificially frac phic charcoal dates suggest a more reason
tured and burned. The pollen record in able figure of 9000-8000 b . c . for the occupa
dicated a late Pleistocene date; because of tion. The small size of the site and its
overlying radiocarbon dates from zone E, artifactual remains bear out the notion of
Flannery suggested a date prior to 9050 B.C. small mobile groups during this period. Two
and perhaps as early as 11,000 b .c . km away Gruhn and Bryan (1977:253-254)
Zone E produced an ash layer and at least discovered La Piedra del Coyote, a site
one hearth, implying a brief occupation. The which yielded a few stone artifacts in its
few stone tools recovered were not tem lower levels. Most of the associated radiocar
porally diagnostic, but three radiocarbon bon dates were near 8000 B.C. Gruhn and
dates from the hearth ranged between about Bryan suspect this is a site similar to Los
9000 and 8800 b .c . Deer and cottontail were Tapiales.
among the bones of Post-Pleistocene fauna. During a survey of the highland Quiche
350
THE RISE OF SEDENTARY LIFE
Basin, Kenneth L. Brown (1980) identified indian to Archaic times involved marked
additional Guatemalan Paleoindian sites, but réadaptation from considerable hunting of a
has not specified the number because of the variety of large herb herbivores to hunting of
difficulty of segregating Paleoindian from smaller game, such as deer or rabbits, com
early Archaic period occupations. Two sites bined with plant collecting. This is a debate
had a Clovis-like projectile point, and two in which the middle ground appears more
others had a leaf-shaped or Lerma point, sug tenable than either extreme. Since early
gesting a Paleoindian II date. Other kinds of populations in Mesoamerica were small,
artifacts were judged to be possibly of Paleo groups could gravitate to environments with
indian origin if they were heavily patinated a combination of varied resources for easy
like the diagnostic early projectile points. hunting and collecting. Areas with herds of
Unlike Brown’s Archaic period sites, Paleoin herbivores would presumably be attractive.
dian sites included few or no large “base Therefore, I am inclined to view Paleoindian
camps,” but were small sites which could subsistence as likely to have incorporated
have resulted from limited-activity occupa more hunting than many Archaic period oc
tions or temporary campsites. Brown sug cupations, although one need not conceive of
gests they could be remains of “small, scat this as implying predominantly big-game
tered, and fairly nomadic bands.” hunting.
A quarry and workshop site of Turrialba in For several reasons the Early Ajuereado
eastern Costa Rica is a specialized site cross phase data do not necessarily contradict this
dated to Paleoindian II (Snarskis 1979). idea. The Tehuacan Valley may not have
Turrialba produced Clovis-like and fishtail been particularly attractive to early hunters
points which helped define the antiquity of or hunter-gatherers during the Paleoindian
the component. However, the site does not II period if the climate was cooler and drier.
clarify settlement or subsistence patterns. In fact, occupation of the area may have been
At El Bosque, Nicaragua, geologic depos delayed until fewer and fewer groups could
its yielded Pleistocene fossils and two enig pursue large herbivore hunting in more de
matic features which might be the result of sirable locations. The Early Ajuereado Phase
human activity (Gruhn 1978). Until further is thought to date sometime between 10,000
information is obtained from the site, it can and 7600 B.C. (Johnson and MacNeish
not be evaluated. 1972:18), but its placement is not well fixed
In summary, none of these sites resolves within that span, and it is possible that the
dilemmas about Paleoindian II subsistence, deposits are from the latter part of that pe
although in several there is a juxtaposition of riod. Regardless of the timing of the Early
cultural remains and Pleistocene animals, in Ajuereado phase, the presence of antelope
cluding large game species. In general, for and horse probably made the Tehuacan Val
this period subsistence seems to have in ley more attractive for hunting than during
cluded a variety of prey, not all of which were the Archaic period. Even though these ani
as large as mammoth and several of which mals did not predominate as prey, their meat
were small. The role of plant foods is uncer contribution to the diet in comparison to
tain. The size and content of sites suggest im small animals like jackrabbits would have
permanent camps or activity stations left by been greater than the minimum number of
small bands. individuals would imply. This point generally
has not received sufficient emphasis in the
literature (e.g., Bray 1976:80-81). Loss of
Processes of Change
antelope and horse and their replacement
I have described the schism of opinion about with deer entailed no major dislocation of
Paleoindian II subsistence, which in turn de technology and subsistence, since a mixed
termines whether the transition from Paleo- hunting and gathering strategy was still prac
351
BARBARA L. STARK
ticable during Archaic times. But Archaic pe in the literature. In general, nomadic versus
riod subsistence is thought to have been diffi sedentary settlement patterns are endpoints
cult in the late winter season, a condition on the two separate continua of greater and
which may not have characterized the Early lesser annual mobility and of degree of per
Ajuereado phase to the same degree because manence of sites. There are ethnographic
of the greater amount of winter fauna. Mac- cases of biannual movement between vil
Neish (1967b: 300; 1978:140) infers a higher lages; in these cases sites are relatively per
percentage of meat in the Ajuereado diet manent and the settlement pattern is semi
than in that of succeeding periods. sedentary. But if hunting and gathering
In summary, data from Paleoindian II are groups follow a sufficiently regular annual
so scattered and variable that it is difficult to subsistence-settlement cycle, their campsites
determine what regional shifts of population can be reoccupied year after year—a mobile
and subsistence may have occurred. Admit pattern with relatively permanent sites.
tedly a generally greater reliance on hunting We are most familiar with sedentarism in
is hard to demonstrate during that time, but the context of hamlets, villages, or even
regardless of how much stress we place on larger settlements, although it need not oc
big-game hunting per se, a decline in the cur in aggregated communities (Flannery
role of hunting or of its feasibility in given 1972b: 24). But since sedentarism in aggre
seasons may help explain the establishment gated settlements7eventually became typical
of more broadly based Archaic period subsis in Mesoamerica, I will focus on the develop
tence, which included a considerable amount ment of villages and agriculture.
of seasonally organized plant collection. I will use the term “agriculture” to refer to
Early Archaic period subsistence patterns in a subsistence pattern in which the majority of
turn form part of the ecological preconditions food is produced by plant cultivation—which
for the processes of cultivation and domes ordinarily involves some clearing of land,
tication. sowing, perhaps weeding, and harvesting.
However, early cultivation probably was
quite minimal in effort, perhaps merely the
Archaic Period
weeding or extra watering of natural stands of
Two key Archaic period changes, the devel economic plants. “Cultivation” means aiding
opment of sedentarism and agriculture, ap plant growth through deliberate human ac
pear to have been interrelated in much of tion. “Domestication” involves genetic al
Mesoamerica, although there is reason to teration of plant populations by human selec
think that the link was not universal. By sed tion, whether conscious or not. A fairly
entarism I mean habitation in one location substantial amount of cultivation over a suffi
year-round, although special-purpose trips or ciently long time tends to result in some se
seasonal, short-term activities may take some lection because seeds are saved to sow the
or most inhabitants away for brief periods. following season. This activity can readily
Ordinarily sedentarism also implies con lead to selection of genetically different
tinuity of residence for at least a few years. plants, even if inadvertently. However, lim
Semisedentarism is a somewhat ambiguous ited cultivation need not result in domestica
term which can be applied two ways: to year tion. A common example can be found with
long habitation in a single place without nec economic trees, which may be spared when
essary continuity from year to year, and to land is cleared or may be favored as seed
year-to-year habitation in a single place with lings. The result may be an artificial increase
departure of most of the inhabitants for no in the density of the plants, not a change in
more than one season of the year. The latter their genotype.
pattern is rare in ethnographic records from Cultivation is a key to the development of
Mesoamerica, but it is described elsewhere agriculture, but domestication plays an im
352
THE RISE OF SEDENTARY LIFE
portant contributing role because it generally 311; Patterson 1971; Flannery 1972b; Mose
results in selection for greater productivity of ley 1975; Niederberger 1979).
edible plant parts. Greater productivity per To recognize that the food supply deter
unit labor is important in making food pro mines whether groups can be sedentary is
duction more worthwhile energetically and not to rule out the contribution of other fac
more attractive as a subsistence strategy. tors to sedentarism or its lack. Obviously the
During the Archaic period we might rea adequacy of wild foods depends on the num
sonably expect nomadism or semisedentar- ber of mouths to feed. Despite an adequate
ism to have been characteristic in envi food supply, situations such as warfare or
ronments and time spans in which food necessary exchanges with other groups may
production or localized wild resources did require a more mobile settlement pattern.
not suffice for a year-round diet, or did so in There are other positive factors which make
frequently. Certainly during the centuries sedentarism attractive. It does not penalize
when plant cultivation and domestication people for heavy or bulky possessions, and it
were beginning in the semiarid highland val makes construction of more durable struc
leys of Mesoamerica, we might expect a se tures less costly in long-term effort. How
quence of changes leading to sedentarism ever, without an adequate localized food sup
like that described by MacNeish (1972; Mac- ply, groups cannot remain sedentary, a fact
Neish, Peterson, and Neely 1975): an initial which makes this settlement option an ener
occupation by nomadic bands of hunter-gath getic (and nutritional) issue. Since most high
erers who seasonally harvested wild plants, land localities in Mesoamerica from which we
then a gradual lengthening and growth in have adequate data either were not highly fa
size of summer harvest encampments, a pro vorable for stationary hunting, fishing, and
cess tied to plant cultivation and domestica collecting or had marked seasonality in spa
tion, and, finally, establishment of semised- tially distinct natural foods, sedentarism and
entarism and then sedentarism in small food production tend to be linked. Archae
villages. This path to sedentarism is depen ological data from the Archaic period do
dent on improvements in food production point to a causal interdependence of seden-
and is currently the accepted scenario for the tarization and cultivation-domestication in
rise of settled farming life in Mesoamerica. some regions, but other information suggests
There is a logic behind sedentarism which the possibility of early sedentarism based on
pertains both to its proposed interdepen natural resources.
dence with food production and to another
pathway to sedentarism, one based on exploi
Archaeological Evidence
tation of certain naturally advantageous lo
calities. If nearby resources suffice, sedentar Our archaeological evidence and ideas about
ism is an attractive least-cost solution to the the Archaic period derive principally from a
energy expenditures associated with travel series of concerted efforts to trace the origins
for getting food (Jochim 1976:17-19; G. of agriculture. Several major research proj
Johnson 1977:489; Pyke, Pulliam, and Char- ects investigated semiarid highland valleys
nov 1977:144-145). Therefore, some fa with dry caves offering good preservation of
vored natural localities may be suitable for organic substances, particularly plant re
sedentary life regardless of whether the in mains. Consequently, our knowledge of the
habitants domesticated plants (or animals) history of plant cultivation and domestication
and farmed, and sedentarism may precede is biased toward upland crops. Some of the
plant cultivation or may occur independently resulting information about the Archaic pe
of it. This is now a widely recognized eth riod from Tamaulipas and Chiapas has been
nographic and archaeological fact (M. Coe reviewed previously (MacNeish 1958; Mac
and Flannery 1967:104; MacNeish 1967c: Neish and Peterson 1962; MacNeish 1964b;
353
BARBARA L. STARK
Mangelsdorf, MacNeish, and Willey 1964). 1968a; 1970). In Oaxaca fewer components
In addition, we now have nearly completely were investigated, but a cave and rockshelter
published results from the major research suggest autumn collecting and hunting
program in the Tehuacan Valley (Byers 1967; camps of small bands (Guila Naquitz) and
MacNeish, Nelken-Terner, and Johnson winter deer hunting (Cueva Blanca). The
1967; MacNeish, Peterson, and Flannery open-air site of Gheo-Shih is thought to have
1970; F. Johnson 1972; MacNeish et al. 1975) been a summer camp perhaps occupied
and some information from investigations in jointly by several of the smaller fall-winter
the Valley of Oaxaca (Flannery 1970). bands. At the Texcal cave site in the Valse-
Excavated caves and a few open sites in the quillo area, Puebla, Roberto García Moll
Tehuacan Valley provide the most extensive (1977) excavated levels pertaining to two
and detailed set of data on the Archaic pe Archaic phases, Texcal I (7000-5000 b . c .)
riod. There, nomadic hunting and gathering and Texcal II (5000-2500 b . c .). N o flora was
slowly gave way to settled farming through a preserved, but fauna indicated hunting of a
long Archaic period sequence. Starting in variety of relatively small animals such as tur
the terminal Pleistocene and early Holocene tle and rabbits, along with some deer. The
Ajuereado phase (prior to 6800 b . c .), the evidence suggests temporary camps.
Tehuacan sequence proceeds through the El Despite the greater comprehensiveness of
Riego (6800-5000 b . c .), Coxcatlan (5000 the Tehuacan data, they contain ambiguities
3400 b . c .), Abejas (3400-2300 b . c .), Purron which affect interpretations of settlement and
(2300-1500 B.C.), and Early Ajalpan phases subsistence patterns. Relatively small oc
(1500-1150 B.C.), by which time sedentary cupation floors and activity areas constitute
communities predominate in the settlement the early archaeological contexts. Deter
record. Pottery appears in Purron times, a minations of seasonality and even of distinc
convenient terminus for the Archaic period. tions between some of the activity areas rely
Throughout this sequence, the representa partly on a variety of perishable seasonal in
tion of ancestral wild and then domesticated dicators. Because certain seasonal remains
plants in the diet increased, with cultivation were found infrequently or in small num
apparently established in the El Riego phase. bers, there is a possibility for sampling error
Tehuacan inhabitants probably themselves (see Bray 1976:77). Estimates of band size at
participated in the domestication process particular encampments also pose problems,
and also gradually adopted domesticates from since the estimates depend on whether oc
other areas in Mesoamerica where selection cupation floors are subdivided into separate,
of slightly different plants was underway. seasonally distinct components—a decision
MacNeish (1978:163; MacNeish 1975:498) which also affects our view of the longevity of
has proposed that increasing use of storable the occupation. Spatial segregation of clus
wild plants and cultivars allowed two changes tered remains is potentially a good clue to
in small band encampments: (1) bands were separate occupation episodes. However, spa
able to stay in one place for longer periods, as tial segregation and composition of clusters
witnessed by increased numbers of multisea vary, and some divergence of opinion is pos
son occupations through time, and (2) larger sible both about activity area definitions and
bands were able to locate in the same place, about which areas within a stratum should be
witnessed by an increase in the number and grouped into a single occupation (Brumfiel
size of larger occupation floors. MacNeish 1977:211).
has suggested that there were a few seden For example, an El Riego stratum in Cox
tary pit house hamlets by the Abejas phase. catlan Cave (site Tc 50, stratum XVI) has
In general, this Tehuacan pattern is fairly several loci defined as activity areas by Mel
consistent with the less complete informa vin L. Fowler and MacNeish (1975:257
tion from the Valley of Oaxaca (Flannery 262; MacNeish, Peterson, and Neely 1975:
354
THE RISE OF SEDENTARY LIFE
368). According to seasonal indicators, all remains, the pattern of component sizes and
seasons of the year are represented in the lengths of occupation, and architecture all
stratum.8 Fowler and MacNeish divide the point to declining mobility during the Ar
stratum into two seasonally distinct compo chaic period.
nents, each of which contains more than one Up to this point, little has been said specif
activity area; there is a fall-winter macroband ically about the subsistence patterns which
camp and a summer-fall microband camp. underlay changing settlement patterns. In
Could this level represent a nearly year-long the Tehuacan and Oaxaca valleys, reconstruc
occupation by at least a few of the people? tions of subsistence emphasize an annual
They note that the possibility cannot be round of activities tied to the summer rainy
ruled out. The possibility of year-long en and winter dry season regime (MacNeish
campments by at least one or two families 1967b; Flannery 1967, 1968a, 1976b:114-
could signal an earlier appearance of occa 115; C. E. Smith 1967). Deer hunting and
sional semisedentarism than has been gener rabbit trapping, cactus fruit and leaf collec
ally acknowledged. The subsequent Cox- tion, tree legume and fruit collection, and
catlan phase has three components which wild grass and amaranth harvesting were
include all seasons of the year (site Tc 50, Archaic period subsistence pursuits. Fall
stratum XIII, activity areas A-E; Tc 50, XII, acorn and piñón nut collecting also were part
A-D; and Tc 50, XI, A -F; see Flannery of the Oaxaca pattern (Flannery 1970). Flan
1967:162; C. E. Smith 1967:233; but cf. nery (1968a) outlines this seasonal cycle as an
Fowler and MacNeish 1975:273-285). This equilibrium system of Archaic period hunt
greater number is especially interesting in ing and gathering, which eventually broke
view of the fact that only twelve excavated down as plant cultivation grew in impor
components occur in the Coxcatlan phase, tance. Originally, maize (Zea mays) and other
compared to thirty for El Riego. grain constituted a minor part of the diet.
I point out these possible year-long or The nature and timing of Mesoamerican
nearly year-long components because they plant domestication has been reviewed else
may move the onset of some semiseden where (Mangelsdorf, MacNeish, and Willey
tarism back before the construction of pit 1964; MacNeish 1967b; Flannery 1973; Pick
houses in the Abejas phase. This interpreta ersgill and Heiser 1977). I wish to concen
tion does no great violence to MacNeish’s de trate on some general issues important for
velopmental scenario, since clearly semised understanding plant domestication. Despite
entarism was not the dominant settlement the wealth of relevant Archaic period discov
mode earlier, and a gradual decrease in mo eries in Tehuacan, Oaxaca, and elsewhere,
bility is exactly what he has argued for. An there remain certain problems in chronicling
all-season component in a thin cave stratum domestication (Flannery 1973). Foremost
does not suggest great residential continuity among them are (1) the frequent lack of ade
in caves from year to year. However, neither quate samples of appropriate plant parts and
is the one excavated pit house a completely (2) controversies about the characteristics of
convincing case for either sedentarism or a putative wild forms. Excellent preservation
hamlet in Abejas times (Flannery 1972b: 37). in dry caves has tended to obscure the fact
The pit house contained no evidence about that organic remains there do not necessarily
seasonality, it was not a very substantial con have a one-to-one relationship with diet
struction, and 17 test pits in the vicinity un (Brumfiel 1977:209), and the preserved parts
covered no other evidence of structures and may not constitute an ideal sample for botani
only a few artifacts. Despite the fact that ar cal analysis.
chaeologists clearly can disagree about what Botanical controversies particularly re
constitutes adequate evidence of sedentar volve around the antiquity of domesticated
ism, in the Tehuacan Valley seasonal dietary maize. The earliest cobs recovered from Cox-
355
BARBARA L. STARK
catlan Cave, ca. 5000 B.C., are wild maize ac vation and selection, although particular
cording to Paul C. Mangelsdorfs reconstruc plants do not always appear at identical times
tion of the ancestral form, but are partially in other sequences. Not surprisingly, in sub
domesticated teosinte according to George sequent phases in the Tehuacan Valley, rough
W. Beadle’s opposing view, which has a con estimates of the relative contributions of wild
siderable amount of current evidence to sup and cultivated foods chronicle a slowly grow
port it (Flannery 1973; Beadle 1977). This ing role of cultivated (and domesticated)
debate has not been settled (MacNeish foods in Archaic period subsistence (Mac
1978:148-149). Since teosinte per se does Neish 1967b: 296-300).
not occur in the Tehuacan record, Beadle’s The foregoing settlement and subsistence
view implies that partly domesticated teo data indicate a close chronological relation
sinte was adopted from elsewhere, as was ship between cultivation-domestication and
true of several of the domesticates used in sedentarization, both of which proceeded
early Tehuacan cultivation. In contrast, Man- very slowly. Similar processes may have been
gelsdorf’s and MacNeish’s position implies underway in other semiarid highland valleys
that Tehuacanos were among those who do in Mesoamerica. The introductions of nonlo
mesticated the plant. cal domesticates in Archaic period sequences
A palynological study is instructive about and the natural distributions of the wild rela
the problem of domestication from an ar tives of domesticates make it appear likely
chaeological standpoint. James Schoenwet- that people in multiple localities participated
ter’s (1974) identification of Zea pollen at in early cultivation-domestication. However,
Guila Naquitz Cave in Oaxaca, dated to ca. there is considerable environmental varia
7000 B.C., is accompanied by the argument tion among highland regions, and I presume
that the ancient inhabitants selectively har that they also varied in their involvement in
vested and probably cultivated the plant, re cultivation, domestication, and sedentariza
gardless of whether we choose to view it as tion. Perhaps in some regions cultivation was
wild teosinte or domesticated maize (which only sporadic or even nonexistent through
currently are classified as two species of the most of the Archaic period; a transition to set
same genus). Cultivation and selection be tled farming may have depended on inmigra
haviors, along with subsequent changes in tion of farming people or on conditions which
the processing characteristics or productivity favored adoption of a package of relatively
of domesticates, are the archaeological topics productive domesticates from neighbors.
of greatest interest in the early Archaic pe For example, it is possible that the rela
riod. Any morphological changes in certain tively late evidence for maize in the pollen
plant parts, taken as evidence of domestica record from the cave of Santa Marta, Chia
tion, are a means to this end. pas, is due to a relatively late adoption of do
Because of the difficulty of inferring these mesticates there (macrofossil plant remains
behaviors, scholars have adopted slightly dif from the site have not yet been analyzed;
ferent positions about the timing of cultiva MacNeish and Peterson 1962; García Barcena
tion and domestication of particular plants. et al. 1976). If the Upper Grijalva Basin did
However, the presence of maize, avocado, follow a different pattern from the Tehuacan
chile, beans, amaranth, and cucurbits as and Oaxaca valleys, it may have been due in
probable cultivars or domesticates in the part to greater rainfall there and different re
Coxcatlan phase in the Tehuacan Valley sug source availability. In a similar vein, Brown’s
gests that cultivation and selection occurred (1980) examination of surface survey data
during the preceding El Riego phase and from the Quiche Basin leads him to speculate
some of its temporal equivalents (MacNeish that Archaic period adaptation to that lusher
1967b: 293). Data from other localities in environment differed from that in the Tehua
Mexico do not contradict this timing of culti can and Oaxaca valleys because cultivation
356
THE RISE OF SEDENTARY LIFE
and domestication apparently were absent or or seasonal movements would lead to build
scant until agricultural populations entered ing of dual or multiple residences. A further
the area. complication is the complexity of the Chan
Before examining key processes of change tuto settlement pattern; Voorhies found some
during the Archaic period in the areas that evidence in fauna, soil analysis, and the bed
did witness the development of settled farm ding of strata for seasonal influxes of addi
ing, it is appropriate to consider a different tional people with, by implication, either a
set of evidence which derives from sites in semisedentary or mobile settlement pattern.
terpreted as sedentary or semisedentary oc Plant remains were not preserved in Chan
cupations dependent on wild resources, not tuto deposits, but manos, metates, and ob
cultivars. These sites do not unequivocally sidian chips did occur. The latter artifacts
demonstrate early sedentarism in Meso might have been used in plant food process
america independent of plant cultivation, but ing, but we do not know whether plant
they suggest its possibility. cultivation was involved. Subsistence foods
In Latin America the idea of early forager included estuarine marsh clams, fish, rep
sedentarism and semisedentarism histor tiles, and a few land animals.
ically has been associated with coastal archae S. Jeffrey K. Wilkerson’s (1975) Palo Hueco
ology (Sauer 1952:23-24; M. Coe and Flan phase at Santa Luisa on the coast of northern
nery 1964, 1967:104; MacNeish 1967c:311; Veracruz is also a late Archaic period occupa
Binford 1968; Moseley 1975; but cf. Voorhies tion, dated to about 3000 B.C. Wilkerson in
1978:17-18; see Flannery 1972b and Nie terprets evidence of terrestrial’hunting, mol-
derberger 1979 for reviews of early noncoas lusk gathering, and fishing as indicative of
tal sedentarism elsewhere). River mouths year-round village life, but no architectural
with estuaries closely juxtapose plentiful features were found (but see Wilkerson 1972:
brackish-water fauna with terrestrial plants 844-852 for discussion of alternatives). Palo
and animals and have provoked the greatest Hueco lacked manos and metates, which
amount of discussion of non-agricultural sed makes use of plant foods doubtful (no plant
entarism, although other coastal aquatic en fossils were preserved). The extent of the de
vironments may offer a similar subsistence posits, distributed along 1 km of a river bank,
richness (cf. Patterson 1971:319; Moseley may have suggested a village occupation to
1975). Wilkerson, but there is no demonstration as
The Tlacuachero site during the Chantuto yet that permanent residence occurred. With
phase (3100-2000 B.C.) on the coast of Chia present evidence we cannot classify Palo
pas offers the best evidence to date. Barbara Hueco as sedentary because it may have
Voorhies’ (1976:38-39, 99) excavations un been a revisited camp location. Three or four
covered a clay stratum or floor with two pos other coastal sites in Veracruz may have Ar
sible post holes. A burial was placed through chaic period deposits (MacNeish 1967c: 311;
the floor, which later was a typical village Wilkerson 1975; Medellin Zenil 1975:87
practice. She interprets this Tlacuachero fea 90), but we know too little about them to war
ture as a probable house floor, partly because rant their discussion here. Farther north,
special floors are not constructed for seasonal MacNeish (1958:173-193) identified coastal
or temporary habitations ethnographically. Archaic sites in Tamaulipas during survey
The house interpretation forms the basis for but, again, we do not have much information
her judgment that the Chantuto phase in about them. Only a few were subjected to lim
cluded sedentary habitation. However, lack ited testing. He regards the Archaic period
of perishable organic materials or sufficient components as brief camps.
artifacts in association with the floor blocked The aceramic Matanchen shell midden de
conclusive demonstration of year-round oc posits at Ceboruco (SB-4) in Nayarit are prob
cupation. It is possible that semisedentarism ably late Archaic period (ca. 1760 b . c .), but
357
BARBARA L. STARK
appear to represent seasonal occupation, pri tuarine biotopes, although they do manifest
marily for shellfishing (Mountjoy, Taylor, and seasonal changes, do not show the discrete
Feldman 1972; Mountjoy 1974:114-115). ness and variety of seasonal markers that typ
Only four flakes and three worked cobbles ify the semiarid highlands. Plant remains
definitely pertain to the Matanchen complex. typically are not preserved. Archaeologists
Similarly, at Puerto Marquez in Guerrero the working with coastal components on all time
deepest levels lacked pottery and contained horizons struggle to resolve whether occupa
only a few obsidian and chert artifacts. A ra tions were seasonal or permanent (cf. Stark
diocarbon date of 2940 ± 130 b .C. came from 1977 :214). In preceramic times the suite of
the bottom of the preceramic deposit, termed artifactual and architectural evidence that
the Ostiones phase (Brush 1965, 1969:90). could bolster ambiguous food remains is ex
Charles F. Brush (1969:97) interpreted the tremely limited. We may never conclusively
deposits as a village but, as Voorhies (1976: solve this problem—and certainly, for the
6; 1978:8, 16) notes, his conclusion was sup early Archaic period, we may have great diffi
ported only by the volume of unbedded de culty locating appropriate sites because of
posits, which is insufficient evidence of sed- subsidence and/or rising sea level.
entarism. However, MacNeish (1978:184; In view of this bleak outlook, it would be
personal communication) reports there was a well to keep in mind ethnographic cases of
clay floor (from a house?) in the Ostiones de semisedentary and sedentary non-agricul
posits, which might strengthen the case. tural occupation in favorable coastal loca
Further southeast on the Pacific coast of tions, such as the historic Calusa on the west
Panama at the Archaic period coastal site of coast of southern Florida (Goggin and Sturte-
Cerro Mangote, both aquatic and terrestrial vant 1964). This instance establishes, I think,
resources were used. The balance between the ecological soundness of the concept of
the two is uncertain (Ranere and Hansell non-agricultural coastal sedentarism even
1978:50-51). Probably some plant foods though the example is tied to the availability
were prepared as well, but we do not know if of aquatic and terrestrial fauna and of wild
they were cultivated since the only evidence roots in that part of Florida. Obviously we
is the presence of edge-ground cobbles and cannot indiscriminately generalize to all
milling stones (McGimsey 1956). Recently coastal situations, but since much of the
inland rockshelters have been discovered southern Florida aquatic fauna is found in
which contain remains of both terrestrial and the Gulf of Mexico, the Calusa example is
coastal aquatic animals, but the former pre certainly pertinent.
dominate (Ranere and Hansell 1978). An In a noncoastal setting, recent work at the
thony J. Ranere and Pat Hansell (1978:57) lakeside site of Zohapilco in the Basin of
point out that a transhumant settlement pat Mexico led Christine Niederberger (1976,
tern is a possibility, although permanent hab 1979) to argue for sedentarism based on wild
itation of some sites such as Cerro Mangote resources. At Zohapilco in the Playa phase
cannot be ruled out. (6000-4500 B.c.) food remains and artifacts
Clearly coastal Archaic period remains were recovered in and around hearths (no ar
have not provided conclusive evidence of chitecture was encountered). Foods and pol
early sedentarism independent of agricul len indicate that a supply of diverse resources
ture, although it is a subject of continued ar was available locally year-round. Playa phase
chaeological discussion (see Bray 1976 for an inhabitants hunted land animals such as dem
alternate view). Archaic period components and rabbits, they fished, and they captured
are late enough that domesticates could have summer amphibians and reptiles as well as
been adopted from highland groups. How winter migratory birds. They also gathered
ever, it is crucial that we keep in mind the amaranth, teosinte, cucurbits, and other
limits of our information. Coastal and es summer plants. Lorenzo and Lauro González
358
THE RISE OF SEDENTARY LIFE
Quintero (1970) and Niederberger (1979) during that time: population growth; reduc
speculate that Zohapilcans may even have tion in the diversity of exploited food re
cultivated some of these plants. sources and concomitantly the cultivation
The settlement pattern at Zohapilco is a and domestication of some plants; sedentar-
subject of contention (MacNeish 1978:187), ization; increasing food storage; and, possi
partly because full details on flora and fauna bly, increasing exchange. I will discuss each
level by level have not been published yet. I of these, along with some of the explanations
suspect that even with a complete account that have been offered to account for them.
there will be disagreement, since only three Most of these processes appear to be interre
Playa levels appear to have adequate food lated in the gradual economic transformation
preservation to allow seasonality assign to sedentary agriculture that characterizes
ments. Neiderberger’s argument for seden- the Archaic period. The interrelationships
tarism hinges on two things: (1) a hearth area will be summarized as a systemic model for
in Layer 23 included both summer foods and the development of sedentary agriculture in
winter migratory water fowl (Niederberger the Mesoamerican highlands.
1979:136); and (2) there were ample re POPULATION GROWTH AND SEDENTARIZA-
sources available year-round to support sed- t io n .Population growth cannot be demon
entarism. Other levels at Zohapilco are more strated for all of Mesoamerica. Evidence for
ambiguous. Layer 25, for example, contained it comes from the steady increase in total site
several winter migratory birds, but only teo- area in the Tehuacan Valley. Why, at least in
sinte grains represented summer or rainy some areas, might population have been
season occupation (Niederberger 1976:43). growing? The question of demographic
Could the teosinte be from stored seeds? growth among hunter-gatherers has attracted
Layer 24 produced summer foods and bird considerable recent attention (see reviews by
remains, but whether from migratory or P. Smith 1976:19-25; Harris 1978). David
year-round fowl is not yet clear (Nieder R. Harris (1978:408-409) notes that a num
berger 1976:42-43). Despite these prob ber of factors could disrupt the typically low
lems the notion of occasional nearly year-long population equilibria maintained by hunter-
occupation at a Zohapilco encampment is gatherers. An idea propounded by Lewis R.
plausible to me. The safest interpretation Binford (1968), Robert W. Sussman (1972),
would be that some of the levels derive from and Richard B. Lee (1972b) seems relevant to
semisedentary base camps, since even a level the Mesoamerican Archaic period. Among
with summer and winter indicators could mobile groups relatively long birth spacing
have been produced by intermittent use oc seems characteristic, maintained possibly by
casioned by departures on hunting or gather a variety of mechanisms such as long nursing
ing trips. Like the possible nearly year-long periods, infanticide, or sexual abstinence. Its
components in the Tehuacan Valley, Zoha advantages derive from the fact that infants
pilco does not suggest long year-to-year oc typically must be carried, and more than one
cupation in a single location. The reasons are small child poses a considerable logistic prob
a lack of architecture and relatively fine bed lem. If groups become less mobile, shorter
ding of the strata. Apparently intermittent birth spacing and consequently population
reoccupation of the site occurred. growth might ensue.
Declining mobility (perhaps with occa
sional semisedentarism) characterized the
Processes of Change
Tehuacan Valley, and sedentarism may have
I have briefly characterized settlement and existed exceptionally early in other favored
subsistence evidence in the Mesoamerican locations. Declining mobility and especially
Archaic period. At present I would argue that semisedentarism may have reduced the ad
the following were key processes underway vantages of maintaining longer birth spacing.
359
BARBARA L. STARK
Without selective pressure in its favor, them edible and (2) softening of the glume
longer birth spacing and a low demographic around maize kernels and development of a
equilibrium may have given way to slow nonshattering rachis which both allow easier
growth. Admittedly the evidence for early threshing of maize kernels (provided we as
semisedentarism is equivocal, but evidence sume teosinte is ancestral to maize). Mac
for population growth does exist, and I sus Neish (1975:498) has suggested that the
pect that the cause may have been sedentar- Tehuacan pattern of regular rotation among
ization. As I will explain in the next section, resource zones facilitated cultivation and se
population growth in areas like the Tehuacan lection because bands returned to previously
Valley may have made seasonal food procure occupied locations during the annual cycle.
ment more difficult, establishing a situation Thus, they could return to plots for
in which plant cultivation increasingly be harvesting.
came a more attractive subsistence strategy Both seasonal procurement cycles and ge
(MacNeish 1978:227). Although there has netic mutations among plants are important
been a tendency by some scholars to place ecological preconditions, the former for cul
undue emphasis on the single factor of popu tivation and the latter for selection. But
lation growth in the transformation to village these ecological conditions alone do not ex
farming, this does not mean that it played no plain why plant selection and cultivation con
role in all cases as Bronson notes (1977:30). tinued for millennia in a slow process of
CULTIVATION-DOMESTICATION AND FOOD change away from the successful seasonal
STORAGE. Reduction in the diversity of re pattern of foraging and hunting (Meyers
sources exploited for food went hand in hand 1971:118). This is an especially puzzling
with development of plant cultivation change because increasing cultivation of
and domestication. As domesticated maize, plants generally requires more labor per unit
squash, chile, avocados, beans, and other yield than harvesting them in the wild, and,
plants increased in the diet, wild foods de for a considerable time, plant productivity
clined. This change seems indisputable de was not dramatically altered by the selection
spite the crudeness of the measures used to process. The issue we need to address is the
gauge the relative contributions of different nature and strength of incentives or selective
foods (MacNeish 1967b: 296-300). The rea pressures that would encourage increased
sons for such a transformation have attracted cultivation-domestication as a subsistence
world-wide archaeological attention. I note strategy. I am not persuaded that plant cul
only a selected set of explanations for the ori tivation on a minimal basis requires any radi
gin of agriculture, concentrating on those cal inventions by hunter-gatherers, since it
that seem pertinent to Mesoamerica. seems well within their botanical knowledge
At the outset it is important to recognize and skills. Therefore, it is crucial that we
the role of biological and ecological condi identify strong selective pressures that
tions. It is a truism that wild ancestral forms would encourage continuation of cultivation
must be available and exploited by humans as experiments.
a necessary antecedent condition. Moreover, Gregory A. Johnson (1977) observed that
as various biologists have emphasized, cer cultivation may have developed to increase
tain economically beneficial mutations must the patch size or density of a collected re
occur in wild populations if human selection source or to move resources closer to settle
of characteristics is to succeed (see summary ments in order to reduce travel time to other
in Flannery 1973). Many of the early muta patches. This suggestion is useful, but the
tions selected seem to have affected plant energetic gain in travel saved has to be
processing more than yield. For example, weighed against the costs of clearing, plant
there was (1) softening of the seed coat in ing, and any other activities connected with
beans, which facilitates soaking to make cultivation. It has not been demonstrated
360
THE RISE OF SEDENTARY LIFE
that this alone is a powerful enough incen source, but, in general, plant food diversity
tive, particularly in the Tehuacán Valley, and quality would have been at their lowest
where seasonal movements may have in during late winter foraging. One solution to a
volved only a 10-15 km distance at most. poor resource period is to increase collection
Another suggestion is population pressure of storable foods in an antecedent season (see
(Meyers 1971; MacNeish 1975:498). Binford Flannery 1973:296). Many plants which ul
(1968) first raised the question of Post-Pleis timately were domesticated were storable.
tocene population disequilibria and their MacNeish (1973:85) also notes that con
possible effects on subsistence. This argu centration of cultivated trees near camp
ment, as restated by Flannery (1973), is that would have abetted seed storage by facilitat
optimal zones with sedentary populations ing use of fall tree fruits and delaying the use
will produce excess population. Resource of cached seeds. Another reason that food
stress will then appear in more marginal en storage might have been important is tied to
vironments which absorb this population risk. Rainfall variations create good and bad
growth. It is in marginal areas that groups years in the semiarid highland valleys (Mac
will experience subsistence difficulties and Neish 1975:499; 1978:143; see also Chapter
have reason to experiment with plant cultiva 3, this volume). Although cultivated plots
tion. But in Mesoamerica, as Flannery even in favorable areas might suffer during a
(1973:296) notes, we do not have evidence of bad year, just as would wild resources, any
rapidly expanding population in the areas yields they produced would be exceptionally
where we have our earliest evidence of plant valuable, especially if they mitigated the
cultivation-domestication. However, the ap winter season.
plicability of this model in a modified form is Although the Archaic period record defi
still open to question. There is evidence of nitely shows food storage, evidence for in
population growth in the Tehuacán Valley, creasing food storage is equivocal. For ex
and it is best evaluated not on an absolute ample, in the Tehuacan Valley, the ratio of
scale but on a relative one. Because the win cache pits and storage/refuse pits to exca
ter season is especially poor in wild food re vated components jumps from 17 percent to
sources, even slight increases in population 30 percent between the Ajuereado and El
may have caused subsistence problems not Riego phases, but it remains at about 30-36
strictly in relation to carrying capacity, but in percent through Abejas times.9 Although we
relation to the returns to labor through hunt cannot clearly document an increase in food
ing and gathering, perhaps by requiring storage, I consider it reasonable. The pres
longer or more frequent foraging trips. The ence of cultivars in some of the caves itself
winter season is a pressure point for popula suggests storage (Flannery 1976b: 116). Cul
tion growth. tivation-selection consequently would be a
Aside from its effects in combination with solution to the need for storable resources,
population growth, the annual low period in especially as population grew. Cultivation
the seasonal food cycle may have operated as would have had the additional advantage of
an independent selective factor favoring cul increasing the size and perhaps the density of
tivation. The late winter season in the Tehua stands of economic plants without extra in
cán Valley was the skimpiest in foods. While creases in travel and transport costs, as John
deer hunting continued in the winter it may son has suggested. Population growth, the
have been less successful; stored foods were lean winter season, and annual rainfall fluc
exhausted; and harvestable fruits and seeds tuations are three factors which would select
were unavailable until near the start of the for storage and cultivation.
spring rains (Flannery 1968a, 1973:296). The storability of different resources war
Year-round cactus foods, such as maguey, rants further discussion because it also helps
would have been the major “fail-safe” re explain why agriculture developed so slowly
361
BARBARA L. STARK
in Mesoamerica. Several resources which 1973:296), and the maize plants in a plot will
ripened in the late summer and early fall had tend to ripen together. James Schoenwetter
excellent storage potential. Mesquite pods (personal communication) pointed out that
(mesocarp) and seeds can be dried and ground mesquite pods on different trees and even on
into flour (see recent review in Doelle 1976). the same tree ripen at slightly different
Grass seeds such as Setaria and Zea dry and times, which reduces the “peak load” of
store well, as does amaranth. If my argument human harvesting, but which also exposes
about the role of a poor winter season is cor the pods to more animal predation in compe
rect, I would expect domestication and culti tition with humans.
vation to emphasize storable wet-season A third reason to be cautious about the
products. This expectation is somewhat 1500 b .C. threshold is that comparison of the
borne out because both amaranth and maize weight of shelled corn to the weight of (pre
become important grain crops.10 Several sumably dried) mesquite pods may be mis
other cultivars were also storable, for exam leading, since, first, more than just maize
ple, cucurbit seeds, beans, and chile. may have been planted in fields and, second,
What about mesquite? Flannery (1973: we are not sure which parts of the mesquite
297-299) perceived this storable tree crop to pod generally were used. Ethnographic data
be a major economic “block” to the expan in Northwestern Mexico and the U.S. South
sion of maize cultivation. To move Zea from west suggest that so much labor is required
its natural barranca locations to adequately to process the extremely hard seeds that
watered additional land would require clear often only the pod is used to make flour
ing mesquite groves along waterways in the (Doelle 1976:53-60). William H. Doelle
valley bottom. Flannery suggests that mes (1976:65) estimates that (conservatively) per
quite productivity would inhibit Zea expan haps only 30 percent of the pod weight is edi
sion on the floodplains until grain production ble and used among Pima Indians. Thus,
reached 200-250 kg/ha, thereby making the maize farming may have been expanding
labor of clearing mesquite groves worth onto river floodplains earlier than Flannery
while. Based on Anne V. T. Kirkby’s (1973: suggests, and Abejas phase waterway pit
126) estimates of maize yield from cob houses may reflect an early facet of such a
length, Flannery notes that this threshold slow shift.
of maize production was not reached until Semiarid highland valleys with mobile
about 1500 B.C., close to the time when hunter-gatherers are not the only settings
floodplain villages became characteristic in which have been proposed for agricultural
the archaeological record. I would agree that experimentation. Voorhies (1976:30) and
mesquite, a plentiful storable wild food, Niederberger (1979:141) both suggest that
would slow down the development of food early sedentary groups who made some mar
production based on grain staples. ginal use of weedy grasses or of other potential
Alternate or competing storable resources domesticates, such as root crops, may have
alone may account for the extreme slowness been predisposed to cultivate the plants be
with which grain agriculture developed in cause they were sedentary. This is a notion of
Mesoamerica, but I question whether we preadaptation expressed by Coe and Flan
should view the mesquite effect as a thresh nery (1967:104) in a different context (cf.
old tied to the end of the Initial period. Fowler 1971). A rather similar idea is embod
First, sedentarization as a process antedates ied in Carl O. Sauer s (1952) concept of root
floodplain villages. Second, a more gradual crop domestication by South American, riv
expansion of maize may have occurred be erine-dwelling, tropical hunter-gatherers (see
cause of differences in harvesting patterns also Harris 1972). The argument is that an
between the two foods. Teosinte-maize rip appropriate settlement pattern, sedentarism,
ens relatively late in the season (Flannery allows ready selection and manipulation of lo
362
THE RISE OF SEDENTARY LIFE
cally available and exploited plants. Seden adopt cultivars from other areas in part be
tary groups can more readily care for plots and cause of the same considerations. The key to
be on hand at the right time for harvesting. this explanation is that if foraging travel costs
This explanation makes sense for some are sufficiently high, some cultivation near
plant manipulation, but it does not account favorable localities may not be more labor in
for a steady economic transformation into an tensive than gathering.
agricultural economy. Agriculture generally There are other possible paths toward
requires greater labor investment than col plant cultivation for sedentary groups. Al
lection of wild resources because, minimally, though we might expect that sedentary
clearing and planting are added steps. Thus, groups would experience population growth,
explanation of domestication clearly de Binford (1968) noted that in optimal zones
mands consideration of more than settlement which allow sedentarization, groups can
pattern if we are to understand the origin of readily accommodate to population growth
agriculture. Because we lack evidence about by shedding excess population to more mar
where or how sedentary groups may have be ginal habitats. But there are other possibil
gun to cultivate plants, I simply state some of ities as well. It appears likely that early sed
the possible reasons for the step. entarization, if it promoted plant cultivation,
Bennet Bronson (1977) points out that in domestication, and agriculture, did so proba
some cases locational constraints may have bly for a combination of sociological and eco
been important in addition to local popula nomic reasons.
tion size, since the latter is important not so Population growth in optimal locations also
much in relation to carrying capacity as in re could lead to competition for control of them;
lation to the relative costs of subsistence competition would make larger-sized groups
strategies. One of his examples of how loca advantageous (see Chagnon 1968:39-44).
tional constraints might operate is a situation Once present, such groups might then find
in which a fixed attraction to settlement augmentation of natural resources attractive.
tends to keep people in one locality at the This possibility does not seem particularly
same time that they also use some plant re likely to me for Mesoamerica because none of
sources which are distributed fairly evenly the putative sedentary sites is necessarily
but infrequently over the landscape; an ex very large nor are such sites plentiful around
ample of an attraction to settlement would be resource areas.
a rich protein source like fishing grounds. In Barbara Bender (1978) argued for more at
such a situation, plant supplies near camp tention to the role of social organization in
become quickly exhausted, requiring longer stimulating greater food production and even
and longer foraging trips. As the labor invest sedentarization. Alliance systems and leaders
ment in gathering food increases, the labor who desire to increase their authority or
investment in cultivation may be offset, es status are singled out by her as independent
pecially since the labor investment in early factors which potentially could encourage in
cultivation presumably was small. creased food production through plant culti
This explanation seems apt for tropical root vation. Strong leadership incentives seem
crop domestication. Individuals of a given unlikely in Mesoamerican instances where
species tend to be widely scattered in tropi small mobile bands have been reconstructed
cal forests, while concentrated protein sup for the early Archaic; whether any sedentary
plies may exist only along rivers or coasts groups may have had more complex social or
(Meggers 1971:6-38; Gross 1975). Although ganization is unknown. However, it seems
Mesoamerica is not generally identified as more likely that alliances or leadership po
an area where root crop domestication oc sitions would play a role later as farming
curred, coastal or riverine foragers in the spread or as larger and more settled groups
lowlands may have been predisposed to increased their reliance on it. Rather, a dif
363
BARBARA L. STARK
ferent social organizational factor might have Florida maintained stable villages without
encouraged early cultivation among small agriculture leads me to suggest that many en
sedentary groups: maintenance of necessary vironments naturally rich in food resources
social relations. are unlikely to favor development of cultiva
Ethnographic studies of hunter-gatherers tion and domestication, although sedentary
and, for that matter, many agriculturalists, settlement may facilitate later adoption of ag
show that mating networks, exchange, and riculture.
mutual assistance, among other things, often exchange . I have deliberately left discus
require wider social contacts than does sheer sion of one other process, increasing product
economic self-sufficiency (see Wobst 1974, exchange, until last. Increased product ex
1976 on mating networks). Seasonally mobile change is poorly documented for the Archaic
hunter-gatherers such as the Bushmen (R. period, but plays a more definite role once
Lee 1972a) typically maintain a flexible social sedentary farming is established. I focus on
group and broad-based social ties. Their so product exchange rather than shared styles,
cial contacts are established and maintained which can be interpreted as evidence of com
in the course of seasonal movements and of munication and exchange of ideas. Product
frequent visiting. In contrast, small groups of exchange is subject to fewer disputes about
sedentary hunter-gatherers would have an the existence of the exchange process; it is
enduring economic tie to localized resources, easier to trace because degrees of stylistic re
but at the same time they would need to semblance and the historical origins of re
maintain social contacts with an adequate semblances do not have to be assessed. How
number of other people. These contacts ever, among mobile or seasonally mobile
would not be maintained automatically in the groups, the distinction between product ex
course of subsistence movements. change and direct procurement may be par
To reduce the travel costs of visiting and ticularly troublesome. It seems likely that
its possible disruption of local subsistence, during the Archaic period any exchange was
people might opt for some increase in popu reciprocal, perhaps in “down-the-line” pat
lation density, presumably expressed both in terns (Renfrew 1975:41-45), since we have
growth of “contagious” distribution around evidence only of egalitarian social organiza
favorable resource locations and of larger lo tion and small groups.
cal groups. If declining natural food returns That Archaic period exchange did occur is
were reached (because of locational con indicated by two facts: (1) obsidian and shell
straints) before group size exceeded the occur in sites quite far from sources (Mac-
effectiveness of mechanisms of social control Neish 1958:144; Voorhies 1976:93; Wilker-
to settle disputes, individuals might find the son 1975), and (2) some domesticates appear
labor costs of cultivation acceptable. in local areas after they apparently were de
At present, Mesoamerican data are insuffi veloped elsewhere earlier. Comparisons of
cient to shed light on the possible roles of lo the Tehuacan and Tamaulipas sequences are
cational constraints, competition, or social especially instructive about the spread of cul-
interaction in encouraging cultivation among tivars, for they show diversification of agri
sedentary hunter-gatherers. To substantiate culture continuing up to the Spanish Con
these ideas would require careful study of quest. The transition to agriculture during
relative energetic costs of a number of ac the Archaic period was probably as much de
tivities. At this point, we are still searching pendent on exchange of early cultivars as it
for unequivocal evidence of precocious sed- was on their domestication because each in
entarism independent of plant cultivation creased the flexibility and productivity of ag
(even though instances are known else ricultural strategies.
where). However, the fact that historic groups Obsidian, a key commodity in Mesoameri
in the U.S. Northwest Coast and southern can sourcing studies, clearly was exchanged
364
THE RISE OF SEDENTARY LIFE
by late Archaic times. Use of obsidian ex movement may have made slight population
tends back into the Paleoindian period in increase possible. Any such increase, if it
Mesoamerica (e.g., Stross et al. 1977). In the were felt in areas like the Tehuacan Valley,
Archaic period, for example, it occurs in the could exacerbate seasonal food procurement
Chantuto phase in Chiapas and predominates problems by forcing more frequent or longer
among chipped stone in the Palo Hueco procurement trips. In such areas the low pe
phase at Santa Luisa.11 But no coastal sites riod in the annual diet cycle is itself an inde
provide a long sequence during which we can pendent factor which likely selected for food
monitor exchange. Obsidian is scarce or ab storage by hunter-gatherers; the irregularity
sent until near the end of the Archaic period of annual precipitation presumably also en
in the Tehuacán Valley, where local materials couraged storage. Any increase in the use of
appear to predominate; their ready accessi stored foods would, in turn, make some sea
bility, in contrast to obsidian, may have re sonal movements less frequent and would al
duced the demand for lithic exchange. The low population growth in a “deviation ampli
sequence at Zohapilco, which is nearer to ob fying” process (Flannery 1968a). To increase
sidian sources, is intriguing because it ex storable plant resources, some people may
hibits a slight tendency for the percentage of have begun a small amount of cultivation
obsidian among lithic artifacts to rise from with a minimal investment of effort. If I am
Preceramic through Preclassic times (Nieder correct about the advantages of food storage,
berger 1976:56). Thus, obsidian exchange is cultivation would have been adopted as a
indisputable during the Archaic period, but minor but expanding element of subsistence.
we need additional evidence to establish a As cultivation began to involve the storing
general pattern of increase. However, I feel and planting of seeds, selection and domes
it is a possibility to which we should be alert. tication became more probable outcomes
The reason is that increasing exchange may and presumably began in a more consistent
well prove to be closely connected to and de manner, gradually increasing the attractive
pendent on the other processes I described, ness of cultivation as a strategy. One result of
especially sedentarization. I will explore this these processes was slightly larger and less
relationship in greater detail as I examine the mobile local groups, which could not be
consequences of sedentary life. maintained without continued cultivation.
SUMMARY. Explanations concerning pro This is the only model of settlement-sub
cesses of change in the Archaic period are in sistence changes for which we have much
terrelated. We can most appropriately con evidence: slowly declining mobility, slow
ceptualize the relations among these changes population growth, and slow plant cultiva
in a multicausal systemic framework (Flan tion-domestication in the Tehuacan Valley.
nery 1968a), one which accommodates a slow In other ecological situations, where great
shift in subsistence strategies (see Bray 1976, er sedentarism was possible initially, a dif
1977:238). At the outset, least-cost travel ferent set of factors and processes of change
considerations would lead to human occupa might have occurred. At unusually favorable
tion of localities with concentrated natural areas, such as lacustrine or riverine-estuarine
food resources sufficient for year-long or settings, competition for resource access
nearly year-long occupation. This seems to might have made larger local groups advan
be a reasonable premise in general ecological tageous—a situation which under some con
terms (Pyke, Pulliam, and Charnov 1977: ditions might lead to plant cultivation. Or
144-145). This settlement-subsistence op perhaps the requisite social networks of sed
tion, although not possible in most Meso- entary hunter-gatherers would most easily be
american environments, appears likely to maintained by increasing group size and den
lead to some population growth. Even in sity—which, again, under some conditions,
non-optimal areas, spatially limited seasonal might lead to plant cultivation. Perhaps the
365
BARBARA L. STARK
366
THE RISE OF SEDENTARY LIFE
low it in time. Our understanding of the nities in Mesoamerica, we can consider the
nature and causes of change will not be consequences of this way of life. According to
served by simply assuming that all changes in Flannery (1972b: 47) and Robert D. Drennan
religion, social organization, technology, sub (1976b: 349), one consequence is changes in
sistence, and settlement pattern in the few social institutions. They both comment that
centuries immediately following establish when disputes arising from regular face-to-
ment of sedentary farming are somehow di face interactions cannot be resolved through
rect effects of it. Sedentary life may “permit” the residential mobility of seasonal foraging,
a number of developments to occur, but my new ritual and social mechanisms to increase
purpose is to isolate what seem to be its more solidarity and perhaps to mediate disputes
immediate effects without following their fur are advantageous, and such mechanisms
ther consequences. appear early in the ceramic record in Oa
As a preface to my consideration of the xaca—e.g., possible dance or ritual sodal
effects of sedentary life, I note two descrip ities. Flannery and Drennan reason that
tive observations by Flannery (1972b) about since agriculturalists are closely tied to fixed
the nature of early Mesoamerican villages land and crop resources, greater territorial
(combining hamlets and small villages). Flan control is characteristic, and people are not
nery observed that the layout of early Meso as free economically to move away when dis
american villages indicates nuclear family putes arise. Therefore, new methods of social
units of food production and storage. Not all integration are selected for, particularly the
sedentary communities take this form; for ex types of crosscutting social institutions char
ample, circular hut compounds generally acteristically found in egalitarian “tribal” so
suggest more communal production and stor cieties (Service 1962; Fried 1967).
age. Because production is not immediately Many other processes of change during
pooled into a common store in nuclear famil early sedentary farming times represent a
ial villages, they can more readily allow sur continuation of those that began in the Ar
plus production for prestige-related social ac chaic period. For example, MacNeish (1967b)
tivities. This organizational feature Flannery and Paul Tolstoy et al. (1977:99-100) note
traces back to the Archaic period, when small that plant selection and the degree of reli
bands of seasonally mobile foragers had to di ance on food production, as opposed to hunt
vide into nuclear family groups in the dry sea ing and gathering, continued to grow. These
son because of the nature of food resources processes appear to be related to another,
and procurement. continued population growth, which can be
Flannery also notes that once small farm argued to be an effect of sedentary farming
ing villages are formed, they have a defen (although initially it may also have been a
sive and competitive advantage over isolated cause).
farming families (or less stable small foraging Population growth may have occurred not
bands). Early villages would be unlikely to only because of the relaxation of incentives
be deposed from their land and crops by for long birth spacing. Many agricultural sys
smaller, less coherent social groups. Thus, tems have a peak load of labor during field
we can view early small villages as a settle preparation, planting, and harvesting, which
ment-subsistence development which, once makes the help of older children and close
established, was not likely to disappear in kin desirable; this constitutes a positive in
favor of isolated farming families or mobile centive for offspring. There is some evidence
foragers unless there was a basic change in from surveys in the Basin of Mexico and Oa
ecological conditions. xaca that some slow population growth did
Given these preliminary observations characterize early ceramic periods (J. Par
about the nature of early sedentary cominu- sons 1976; Tolstoy 1975; Feinman and Kowa-
367
BARBARA L. STARK
lewski 1979). Although I would agree that we nities, then some settlements may be in a
cannot simply assume that population growth better relative spatial position than others
always occurs (Cowgill 1975b), there seem to (see Renfrew 1975; Wobst 1976). Central
be some reasonable economic bases for it in positions have received considerable atten
this area, and the archaeological record tends tion for development of social and economic
to verify it. differences, although other possibilities have
Undoubtedly population growth led to been raised (Brunton 1975). Additionally, the
fissioning of social units as group size and historical sequence of village founding and
consequently the incidence of disputes in fissioning may make older communities and
creased. Perhaps also there were some situa lineages social foci for intercommunity ritual,
tions in which immediately accessible local social, and economic relations (see Goldman
food resources were sufficiently limited in re 1963:292).
turns to labor to prompt fissioning, although Obviously, economic or communication
this notion has been called into question by differentials can have social consequences
studies in the Valley of Oaxaca (Flannery among communities only if people in dif
1976b: 111; Feinman and Kowalewski 1979). ferent settlements interact. Mating networks
We do not have to conceive of the economic are one type of interaction. Others can be
effects of population growth solely in terms of readily envisioned. Occasional, localized
“carrying capacity” problems. Relative re poor farming years would make intercom
turns to energetic costs are a more realistic munity social ties and mutual aid advan
framework for human decision-making. In a tageous. Erratic rainfall patterns in many of
more definite example of economic controls the semiarid highland valleys certainly allow
on settlement pattern, “filling in” of the land for this eventuality. Differences in local foods
scape appears to have followed a progression or other resources could lead to exchange
from more desirable to less desirable farming (Service 1962:133-143; 1975:75-76; Fried
habitats (Flannery et al. 1967; Tolstoy 1975; 1967:183 but cf. 115, 129; Sanders 1956).12
J. Parsons 1976). Demand for nonlocal resources such as obsid
It is in such a growth-dispersal situation ian would also prompt contact and exchange
that a further effect of sedentary farming life among sedentary communities. For example,
becomes apparent. Differentials in resource Voorhies (1978 :18) has argued for the early
access and control arise among communities importance of long distance exchange be
spreading through a nonhomogeneous en tween upland and coastal groups. With in
vironment. This may take the form of more creasing populations, exchange volume could
and less advantageous access to the same sort be expected to grow proportionately. Three
of resource, e.g., poorer or better farmland aspects of sedentary life facilitate elaboration
(M. Coe 1974), or to different resources, such of exchange. The increasing numbers of vil
as localized minerological outcrops (Tolstoy lages would reduce, per capita, the cost of
et al. 1977:102). Out of these differentials transport of goods because of economies of
may grow others. For example, some less de scale; permanent dwellings would facilitate
sirable land may, with investment in irriga storage of goods; and fixed local groups would
tion facilities, prove to be highly productive. make communication and the organization of
Michael H. Logan and Sanders (1976) pro exchange easier. During the subsequent Ear
vide a model for highland Mexico which in ly Horizon, studies of exchange in the Valley
corporates the effects of irrigation in this of Oaxaca and at San Lorenzo Tenochtitlan,
fashion. A third kind of differential might ap Veracruz, tend to confirm growth of exchange
ply to both relatively homogeneous and non as well as changes in its organization (Cobean
homogeneous environments: if necessary so et al. 1971; Flannery 1976a: 283-328; Zeitlin
cial networks (e.g., for connubial or material 1978). Exchange in durable exotic products is
exchange) extend beyond individual comrnu- the best-documented form of interaction
368
THE RISE OF SEDENTARY LIFE
Early sedentarism based on naturally avail tra- and intercommunity relations was pro
able foods may have existed in Mesoamerica foundly altered both by new relations to
in some localities. Conceivably under certain more fixed resources and by changed condi
conditions, such a settlement pattern pre tions of resource exchange. The alteration
disposed people toward cultivation-domes undoubtedly became more pronounced as
tication, an idea which is still largely con settled farming spread to more regions.
jectural. In any case, it seems clear that many Social differentiation apparently devel
groups in such favored locations eventually oped relatively quickly in Mesoamerica un
either adopted agriculture or engaged in ex der these conditions, but the development
change for agricultural products. The ques was confined initially to certain key regions.
tion of why sedentary agriculture spread to Presumably the Olmec “heartland” in the
so many parts of Mesoamerica is one which I Gulf Coast lowlands was one of the most im
have not addressed in this article, and to portant. We do not have a record of the tim
which Mesoamericanists have given little at ing and nature of sedentarization, cultivation,
tention. However, to understand the growth and domestication there, and it may have
of sedentary farming as a predominant sub been an area to which agriculture spread or to
sistence pattern in Mesoamerica, the reasons which agriculturalists moved. An important
for its spread will have to be examined in fu topic in future research should be acquisition
ture research. I identify this topic as a dis of a better Archaic and Initial period archae
tinct research question on the grounds that ological record from the Gulf lowlands and
initiation of cultivation-domestication was identification of the conditions which pro
tied to particular ecological circumstances. moted sedentary farming and social differen
Such situations need not be identical with tiation in that area. Currently we understand
the conditions under which agriculture better the consequences of sedentary farm
spread or was adopted as a subsistence strat ing in the highlands than in the lowlands,
egy by other hunter-gatherers. and it is possible that different models will be
In summary, within Mesoamerica there is required to account for each.
no reason to assume that a single combina
tion of conditions selected for cultivation-
domestication or for sedentarization. By the
Acknowledgments
same token, conditions appropriate to under
standing the Mesoamerican case do not nec My colleagues whose good advice helped me
essarily apply to other areas of the world improve this paper have my warmest thanks.
where these processes occurred. However, I James Eder, Gary Feinman, Sylvia Gaines,
do not argue for the uniqueness of every in Richard MacNeish, Fred Plog, Reynold Rup-
stance of such change. Rather, I expect many pé, James Schoenwetter, Barbara Voorhies,
of the same variables to prove to have been Phil Weigand, and David Wilcox deserve
important, even though they may have inter credit for many improvements in it, although
acted in slightly different ways. they have no responsibility for its shortcom
After farming villages became widespread ings. I thank Joaquín García-Barcena, Ruth
in Mesoamerica, certain processes underway Gruhn, Cynthia Irwin-Williams, José Luis
in the Archaic period continued (at least in Lorenzo, Richard MacNeish, Carol Ruppé,
some areas). Population growth and con Jeremy Sabloff, and Paul Tolstoy for their
tinued plant selection and improvement are kind assistance with reference materials, and
examples. Thus, there is reason to see con I appreciate Terri McCarty’s having typed
siderable continuity in processes of change the manuscript on short notice.
rather than a radical break signaling a com
pletely new era in Mesoamerican prehistory.
But despite continuities, the nature of in-
370
THE RISE OF SEDENTARY LIFE
371
BARBARA L. STARK
however, divergent opinions about migrations ex 9. I attempted also to calculate the frequency
ist, depending on scholars assessment of both Pa- through time of dry season components which had
leoindian periods (Gruhn and Bryan 1977). It is milling stones because milling stones are associ
appropriate to note that the early sites from Meso ated predominantly with wet season plant food
america, especially when considered in conjunc processing. Therefore, their occurrence in a dry
tion with the Ayacucho evidence, pose problems season component might indicate their use on
for Paul S. Martin’s (1973) recent explanation of stored plants even if the plants were not pre
Pleistocene extinctions, which he attributes to a served (of course, artifact curation cannot be ruled
relatively late migration into the New World of out). There are some dry season components with
highly skilled hunters who swept through both milling stones, but the marked growth of multi
continents, decimating unwary game. Rather, season components results in too few components
these sites suggest there were earlier occupants from only the dry season to give a meaningful sam
who hunted some of these animals, although this ple.
fact does not eliminate human agency from some 10. Tropical root crops partly fit this argument
role in Pleistocene extinctions. Processes of popu because they can remain in the ground until
lation growth (and possibly decline) are tied to the needed, but a resource-poor season—and hence
Martin hypothesis (MacNeish 1976), but there are several aspects of a highland model—seems ques
too few Mesoamerican and other Paleoindian I tionable for the tropical lowlands.
sites to allow cogent discussion of demographic 11. Voorhies (personal communication) reports
change. that some Chantuto obsidian has been analyzed
7. Archaeologists have not devoted much at and assigned to sources. Tajulmulco may have
tention to possible exceptions or to the manner in been a source from which obsidian was obtained
which sedentarism was integrated with seasonal directly, but El Chaval was more distant and
hunting, harvesting, or collecting trips. Later consequently a potential source of exchanged ob
cave components in the Tehuacan and Oaxaca val sidian.
leys show that such trips and encampments con 12. Timothy K. Earle (1977) questioned the
tinued on a reduced scale in periods which were general role of exchange in foods and other staples
predominantly sedentary (MacNeish, Peterson, in the development of early chiefdoms, but his
and Neely 1975; Flannery 1976b: 111). case example from Hawaii is in an area with rela
8. There are some differences of opinion about tively close spacing of different resource zones
seasonal placement of a few of the diagnostics; (within ca. 12 km), which makes them accessible
MacNeish, Peterson, and Neely (1975:348-351) from a single settlement. This need not be the
treat some tree pods, such as mesquite, and case for all regions. For example, in the Lower Pa-
grasses and amaranth as spring indicators, but late paloapan Basin in Veracruz, zonation is on a larger
summer-fall is a more common interpretation; scale (Stark 1978).
see C. E. Smith 1967:232; Flannery 1967:159;
but cf. Flannery 1976b: 114-115.
372