Of Maize and Men:

A Critique of the Maritime Hypothesis

of State Origins on the Coast of Peru

DAVID
J. WILSON
University of Michigan

The adaptation of preindustrial band and tn'bal groups to enwronments subject to

apenodic downturns of long duration can often be understood by reference to limiting
factors affecting the productivity of the subsistence system. Thai paper examines the t w o
alternative hypotheses of the manlime and agncultural foundations of state society on the
coast of Peru, and by reference to limiting factors affecting the Peru Coastal Current,
argues that coastal mantime groups never advanced beyond an egalitanan tn'bal stage.
Support f o r the acceptance of the alternative agncultural hypothesis of state ongim is pro-
vided by a n ecological analysis of relevant features of the marine and terrestnal biomes,
ethnographic data on manne subsistence wlkzges, and archaeological data f o r the Late
Archaic and Formative periods. [Central Andes, state origins, coastal adaptations, eco-
logical anthropology, limiting factors]

During much of the year, mzny monkeys can obtain enough food in only three or four hours of
gathering each day, and under normal conditions baboons have plenty of time to build the T a j
Mahal. The restriction on population, however, is the lean season or the atypical year [Washburn
and Lancaster 1968:SOS].

INTRODUCTION

WITHTHE DEVELOPMENT OF THE ECOLOGICAL APPROACH in anthropology, it is becoming

increasingly clear that understanding cultural adaptation and evolution in certain en-

DAVID J. WILSON is a Ph.D candidate currently writing his diasenation in anthroplog.

ical archamlogy at the University of Michigan; he recently completed 14 months of re.
search on pre-Hispanic settlement patterns. defensive syatema. roada. and canah in the
Santa Valley, north coast of Peru. with the help of granu from the National Science Foun-
dation. the WcnncrGren Foundation, HEW Fulbright.Hays. and the Center for Field
Rescarch/Earthwatch. After receiving hi B.A. in political science in 1965 at Washington.
he spcnt two years in the Peace C o p in Bolivia. working with the Aymara. Returning to
the U S..he received his M.A. in Spanish in 1971 at San Diego State and his M.A. in an.
thropology in 1974 at Michigan. He has taught Spanish at San Diego State and anthro.
pology at Michigan.Rint and Michigan-Dcarborn. His archamlogical fieldwork includes
proycu in California. MeNco, and Peru

Copvright 1981 bv t h e American A n t h r o p o l o p c a l Association

0002 7294,81 010095 288.3 SO 1

95
94 AMERICAN ANTHROPOLOGIST [89, 1981

vironmental contexts involves understanding the lean season or the abnormal year rather
than the normally prevailing conditions. Bartholomew and Birdsell (1962:27), for exam-
ple, in their study of the Papago of the Lower Colorado region of the Southwest United
States, argued that “anthropologists studying the simpler cultures characterized by few
storage techniques would do well to search for those critical limiting factors which do
determine density.” This paper is concerned with critical limiting factors related to socio-
cultural evolution and the rise of complex society on the coast of Peru during the Late
Archaic and Formative periods (ca. 2000 to 100 B.C.). Analysis and discussion are
directed toward understanding the nature of subsistence adaptations to the two major
biomes-marine and terrestrial-as they relate to the rise of civilization in this area. The
fundamental issue to be examined is which of the two environmental contexts provided
the critical setting for the cultural developments which are most directly antecedent to
state formation on the coast.
If, as argued here, it is the development of a complex of plant domesticates-including
maize- that provided the foundation of pristine state society on the coast, then there ap-
pears to be no exception to the general rule that an agricultural base with high produc-
tive potential preceded the rise of the first states in Peru, Mexico, Egypt, Mesopotamia,
the Indus Valley, and China. If, however, as argued recently by Moseley (1975) and
Feldman (1978), the coastal Peruvian maritime environment is so highly and continuous-
ly productive that it provided the primary subsistence base for incipient stratified society
as early as 1000 B.C., then we have on the coast of Peru an important exception to this
rule.
Several distinct and more or less independent sets of data are marshalled to assess the
relative merits of the maritime and agricultural hypotheses of state formation on the
coast. Listed in the order they will be presented here, these include (1) ecological studies
of aspects of the marine and terrestrial environments; (2) ethnographic studies of tradi-
tional marine subsistence fishing villages: and (9) archaeological studies of coastal sites
dating to the Late Archaic and Formative periods. Taken together, these data support
the primary argument that early coastal societies which focused principally on marine
resources probably did not advance beyond an egalitarian tribal stage. This conclusion,
combined with clear indications of agriculture at all coastal sites whose levrl of socio-
cultural complexity can safely be inferred to be higher than a tribal stage, provides the
basis for acceptance of the alternative agricultural hypothesis.
The environmental setting of the Peruvian littoral is one of the most interesting in the
world. The narrow strip of land sandwiched between the colder waters of the Peru
Coastal Current and the high Andes mountains is located at the appropriate latitudes for
high humidity and dense tropical vegetation. Instead, the terrestrial sector benefits only
slightly from the moisture contained within the tropical air masses sweeping across the
Current. Striking the cold waters of the Current, the air condenses and loses much of its
moisture-carrying capacity, only to heat up again as it moves across the land. Because of
this, little or no rain is ever recorded on the coast, and the land presents one of the
bleakest desert scenes anywhere in the world. Pre-Hispanic and modem agriculturalists
have been essentially confined to the 40-odd narrow river valley oases that break across
the generally arid monotony of the desert.
In sharp contrast to the general biological sterility of the land, the Peru Coastal Cur-
rent became for a time in the 1960s the worlds most productive fishery, with an annual
harvest of ten million metric tons of anchovies alone coming from the ocean off Peru
(Odum 1971; Paulik 1972). Anchovies are only one of a large number of marine specks
that thrive in the rich pelagic waters of the Current. The high productivity of the marine
environment is brought about by continual upwelling which occurs generally along the
coast and brings a large supply of nutrients to the pelagic zone. Feeding upon these
Wilson] OF MAIZE A N D MEN 95

nutrients are phytoplankton and zooplankton, which in turn provide subsistence for an-
chovies and higher levels in the food chain (Gunther 19.36).
The Peruvian fishing industry is currently at only one-fifth the productivity of the
1960s, however, due to the combined effects of heavy overfishing and a recent
devastating intrusion in 1975 of the warm equatorial countercurrent known as EL Nifio
(Cowles, Barber, and GuillCn 1977). Indeed, in the future the industry plans to stabilize
its harvest of coastal waters at one-tenth the highest levels achieved in the 1960s. Thus in
order to understand the nature of a long-term fishing adaptation to the Current, it is ob-
viously not enough to assess its productivity only in terms of the highest short-term levels.
One must also take into account the effects of El Niiio and average yields that can be sus-
tained over the long term. As will be shown in the analysis that follows, this aspect of the
Peru Coastal Current cannot only be applied to achieve understanding of modern
fisheries, but can also be effectively applied in an attempt to understand constraints upon
preindustrial fishing groups during the Late Archaic and Formative periods.
The assertion that complex pre-Hispanic society (i.e., chiefdoms and states) on the
coast arose on an agricultural base has considerable historical precedent. Max Uhle, the
German archaeologist who conducted the first stratigraphic excavations on the central
Peruvian coast in the early 20th century, argued in 1925 that:
. . . fishing is the most natural way of life for primitive man, and fishermen cluster in small
settlements at certain places on the coast and leave us evidence of their earlier presence in the
shellmounds. . . . Shellmounds have another unusual interest because they developed slowly
through different periods and reveal to us in their stratification the way in which culture slowly
developed. . . . The visible remains of cultivated fields . . . form another subject for study. . , .
The practice of agriculture is as old as the civilizations themselves in this continent . . . [ Rowe
1954:86].
Clearly, then, Uhle asserts that agriculture was the primary foundation of stratified socie-
ty on the coast of Peru. Since Uhle’s pioneering work, the agricultural hypothesis has
been supported by a number of investigators, most of whom were themselves engaged in
expanding the archaeological data base that supported it (see Bennett and Bird 1960;
Collier 1961; Kidder 1964; Kosok 1942; Mason 1968). However, no one proposing the
hypothesis has heretofore attempted to provide broad qualitative and quantitative sup-
port for it.
For example, one of the principal assumptions underlying the agricultural argument is
that the population size of maritime groups was rather severely limited relative to that of
agriculturalists (e.g., see Bennett and Bird 1960:68). Paradoxically, while arguing for
these relative limitations, the same investigators have sometimes argued that the inshore
waters of the Peru Coastal Current are extraordinarily rich in marine resources (e.g., see
Kidder 1964:454). As mentioned above, Moseley and Feldman are the principal sup-
porters of the alternative maritime hypothesis, which not only assumes a consistently high
productivity of the Current but also relatively little limitation on the size of prehistoric
fishing groups. An earlier expression of the hypothesis may be found in Lanning’s
(1965:74, 76) article on early humans in Peru where he argues that in the change from a
seminomadic huntinggathering economy to a sedentary maritime economy, the “ex-
tremely abundant” sea life of the littoral fostered such a population explosion that Late
Archaic groups were able to undertake extensive public works. He does not specify the
sociocultural level of these groups.‘
In his recent book on the maritime foundations of Andean civilization, Moseley
(1975:3, 60) has argued in essence that if the Japanese Current along the Pacific North-
west Coast could support societies at a chiefdom level, then there is every reason to sup-
pose that coastal Peruvians living along a “far richer shoreline” would have achieved
civilized society by as early as 1000 B.C. To provide support for his argument, Moseley
96 AMERICAN ANTHROPOLOGIST [83. 1981

discounts entirely any negative effects of El Niiio phenomenon, while pointing to the
clear association between apparently complex public architecture and extensive marine
middens at such coastal sites as Las Haldas, Aspero, Rlo Seco, and El Paralso. Since at
least two of these sites are relatively large in comparison to most other Late Archaic-
Formative sites on the coast (Las Haldas covers 200 ha. and El Paraiso covers 58 ha.), any
attempt to refute Moseley’s hypothesis must deal squarely with their size and internal
complexity, as well as with what is known of their subsistence base.
As will be shown later in this paper, none of these sites presents insurmountable prob-
lems that negate the agricultural hypothesis. Indeed, all of them are clearly part of an
early appearance of a mixed agricultural and maritime subsistence system in the north
central and central areas of the coast.

HUMAN ADAPTATION T O THE PERUVIAN MARITIME ENVIRONMENT

For purposes of the following presentation of data, it is assumed that the operation of
major features of the marine and terrestrial biomes was essentially the same in the Late
Archaic and Formative periods as it is now. Craig and Psuty’s (1968) work in the area of
Paracas Peninsula suggests, for example, that there has been a steady state in the marine
regime for the past several thousand years. This regime is generally characteristic of the
area between 3’ and 18O S., and south into the Atacama Desert of Chile, for the period
in question. Throughout this area the remarkable preservation of organic materials in
sites dating to the Archaic Period (5000 to 1800 B.C.) supports the assertion that the
cold, upwelling waters of the Coastal Current have been present and have maintained
conditions of aridity on the land since the beginning of that period.
Physical and Biological Characteristics of the Current
A primary reason for the exceptionally stable conditions normally characteristic of the
Current is the constancy of the prevailing southerly trade winds along the coast (Coker
1918). The velocity of the Current is less than one mile/hour, and the annual surface
temperature ranges narrowly between about 14O and 18O C. Interestingly, surface
temperature varies only slightly with latitude. Thus, between Paita (5O S . ) and Mollendo
(17O S.) minor variations in temperature are due more to local conditions than to change
in latitude (Murphy 1923). Upwelling of colder water and nutrient salts from the ocean
depths occurs generally between 3 O and 3 3 O S . , although it is strongest off Antofogasta
(Chile), San Juan (Pisco), and Punta Aguja (north coast).
The almost incredibly rich biological productivity of the Current has been mentioned
above. Figures for secondary productivity of the Current, as measured in fish production
during the 1960s, indicate that the annual harvest is 1,680 kilogramdhectare (kg./ha.)
and 355 kilocalories/square meter (kcal./m.2). In comparison, the annual harvest of the
world marine fishery is 1.7 kg/ha and 0.3 kcal./m.2 (Odum 1971:71). Marine life in the
Current is also delicately adapted to the normally stable environmental conditions and
most species, including plankton and anchovies, are stenothemal and stenohaline. To
the west of the Peru Current and to the north beyond about 5O S . , “the richness of sea life
ceases suddenly on a very sharp demarcation line” (Schott 1932:94). This observation has
critical importance, since figures for the relatively low biological productivity in warmer
tropical waters to the north and west of the Current provide a reasonable approximation
of productivity during El Niiio conditions.
Availability of Selected Marine Animals
Several species of fish probably were an important part of the marine food chain upon
Wilson] OF MAIZE A N D MEN 97

which early coastal human populations relied. These species may be divided into two
trophic levels: (1) smaller fish (under 20 cm. long), such as anchovies, at the herbivore
level; and (2) larger fish (over 50 cm. long), such as bonito, r6bal0, and corvina, at the
primary carnivore level. In the calculations of marine carrying capacity (below), it is
assumed that human maritime groups were at the secondary carnivore level (i.e,, eating
larger fish) rather than at the primary carnivore level (i.e., eating smaller fish). This has
critical implications, of course, for the amount of energy these groups had available to
them.
The archaeological record supports the assertion of a secondary carnivore level for the
early maritime groups. For example, at the Early Formative site of Huaca Negra in Vird
Valley, the primary emphasis was on sea lions and larger fish (Strong and Evans 1952).
Clifford Evans (personal communication), of the Smithsonian Institution, corroborates
this on the basis of his recollection of the work carried out in 1946. He reports seeing an-
chovy remains around modem coastal villages, but does not recollect seeing any of the
smaller anchovy-type bones in the excavated materials. Junius Bird (personal communi-
cation), of the American Museum of Natural History, also found no evidence of the tak-
ing of anchovies in his excavations at the Late Archaic site of Huaca Prieta (Chicama
Valley) or at the already-mentioned Huaca Negra site, nor were anchovy remains identi-
fied in the human fecal material collected. In his report on excavations at the Central
Coast site of Asia, Unit I, Engel (1963:77) states that “Small fish such as
anchovetas . . . so frequent in later Chavinoid or Paracas deposits, were not found. . . .”
Thus, although anchovies were used as fertilizer for agricultural crops by later coastal
peoples, they do not appear to have been a popular food in either pre-Hispanic or
modern times (Fiedler 1944).
Further support for the argument that early maritime groups were primarily at the sec-
ondary carnivore level comes from an examination of the behavioral characteristics of the
smaller and larger species of marine fish. Anchovy schools are generally found within 80
km. of the coastline, at about 80-100 m. below the ocean surface UordAn S. 1971). They
appear on or near the surface daily at about 4:OO P.M., following the phytoplankton
which is their principal food source. The anchovy come to the surface a second time
about 1O:OO P.M. (Ministerio de Agricultura 1952). Besides this daily vertical migration,
anchovy schools migrate horizontally on a daily basis. During the day the schools are
located well offshore, and at night come closer in toward shore. It would thus appear that
anchovies were not readily available to early coastal fishermen, especially if fishing were
done from the beaches or in boats just out beyond the surf.
Totora-reed rafts, called caballitos (“Sea Horses”), have been used for at least the past
two thousand years for fishing in the ocean along the Peruvian shoreline. In places such
as Huanchaco (Moche Valley), where they are still used, these one-man boats are
generally anchored out beyond the surf with the bow toward the sea, within 1 km.of the
shore (Fiedler 1944). Caballitos are depicted on pottery vessels as early as the Early Inter-
mediate Period (ca. B.C./A.D. to 700 A.D.) in the Moche culture, but these small
fishing vessels are not found archaeologically since when they become waterlogged they
are simply left on the beach to rot. If one assumes they appeared as early as Late Archaic
times, then coastal maritime groups could have fished out to about 1 km.from the shore.
Since anchovies generally are located much farther offshore than this, it seems unlikely
they were regularly fished. Among the fish that prey on the anchovy are a number of
species that prefer the sandy ocean bottom near the coastal beaches. These include, as
mentioned above, bonito, r6bal0, and corvina (Ministerio de Agricultura 1952). As
shown by the archaeological record, these are the fish that are usually found in Late
Archaic-Formative midden debris.
It is appropriate at this point to refer to the location of major offshore fish concentra-
98 AMERICAN ANTHROPOLOGIST [83, 1981

tions that are exploited by the Peruvian commercial fleet. Of the 37 major zones where
fish tend to concentrate, 19 are located along the coastline to the south of Lima and 18
are located to the north (Schweigger 1943). Except for small gaps in the Jequetepeque
and Ica areas, the zones appear to be spaced more or less evenly along the coast (Figure
1). All of them, of course, are well out to sea from the narrow inshore area utilized by
modern and ancient subsistence fishermen. One of the supporters of the maritime
hypothesis (Moseley 1975) has argued that the early and precocious development of com-
plex society on the Central Coast can be explained by a pronounced differential distribu-
tion of marine biomass along the Peruvian coast. As Figure 1 shows, there is little sup-
port for this argument. Indeed, three of five major commercial fishing centers on the
coast are located well to the north of the Central Coast area. One must therefore look to
other, more compelling reasons for the developments at such sites as Las Haldas, Aspero.
and El Parabo.

Bottleneck Penods in Marine Productivity

As we have seen, the normally high, stable productivity of the Current appears to pro-
vide an exceptionally rich and benign environment for coastal maritime populations. It is
easy to focus on just this aspect of the Current and ignore the implications for human
adaptation of rather frequent, and sometimes lengthy, downturns in productivity. Ben-
son (1972:71), for example, asserts that “The Humboldt Current is one of the world’s
richest fishing grounds today, as it undoubtedly was in Mochica times. Fish would have
been the important protein food, abundant enough to provide for all.”
There are, however, two major phenomena that disrupt the marine environment and
introduce perturbations of variable severity and duration to the coastal ecosystem. The
principal of the two is El Nifio, or the “Christ Child,” since it usually begins at the end of
December just after Christmas. El Niiio is a warm countercurrent sweeping down from
the north which raises water temperature, reduces or stops upwelling, and lowers the salt
content of the Current. It therefore causes severe disruption in the life cycles of marine
species. The phenomenon may last anywhere from 2-12 months, and occurs at intervals
of 6 to as many as 54 years. The second phenomenon is known as uguuje, or the “Callao
Painter,” and is a red tide of dinoflagellates. Like El Niiio its occurrence is unpredict-
able, but it generally occurs between April and July. Accompanied by high water
temperatures, agunje may bring about widespread killing of plankton and fish, migra-
tion of guano birds, and the production of noxious H2S gas from the decomposition of
great masses of the smaller organisms affected (Lavalle y GarcIa 1917; Sverdrup,
Johnson, and Fleming 1942). Aguaje is a short-lived phenomenon in comparison to El
Niiio, however, and it seems unlikely that it would pose a severe adaptive problem for
maritime subsistence groups.
El Niiio is undoubtedly one of the most studied of all oceanographic and meteorolog-
ical phenomena in the world. The two principal bibliographic sources are Gilmartin’s
(1964) Compihci6n Bibliogr6Jcn and Stevenson and Wicks’s (1975) monumental Bibliog-
raphy of El Niiio. Gilmartin lists about 40 entries, while Stevenson and Wicks include well
over 500 entries on El Niiio and related meteorological and oceanographic studies. As I
will attempt to show, El Niiio is exceedingly relevant to understanding the nature of early
coastal maritime adaptation, yet no serious discussion or consideration of its features has
yet been attempted in the anthropological literature. Indeed, it has been at best briefly
mentioned (e.g., see Mason 1968) or summarily dismissed (e.g.. see Moseley 1975).
A number of possible causes of the phenomenon have been advanced, including that it
is brought about by changes in atmospheric pressure gradients, deviation of the Current
away from the coast, fluctuating strength of the trade winds, and oceanic macrovaria-
Wilson] OF MAIZE A N D MEN 99

Fig. 1 . Location of major fish concentrations and coastal fishing communities on the Peruvian
littoral.
100 AMERICAN ANTHROPOLOGIST [83. 1981

tion. Although the causes of El Niiio are still poorly understood, it has recently been
argued (Wyrtki, Stroup, Patzert, Williams, and Quinn 1976) that its occurrence has at
least something to do with the winds, since it is preceded by strong trades and coincides
with their relaxation. Other features of El Niiio are better understood and in the follow-
ing discussion a number of these will be briefly examined, including effects, geographical
extent, environmental signals, frequency and prediction, duration, and biological pro-
ductivity during the lean period.
As it sweeps down from the northern equatorial region, the warm countercurrent
overlies the Current to an average depth of 25 m., bringing water temperatures 5” to 8 O
C. higher than normal. The upwelling of colder water and nutrients ceases, and marine
organisms are forced to migrate or die. Coastal birds, deprived of their normal source of
food and subject to a fatal pulmonary disease during El Niiio conditions, must also
migrate or die (Sverdrup et al. 1942; Naylor 1965). In addition, El Niiio sometimes
brings changed meteorological conditions to the littoral, and some occurrences have been
accompanied by heavy rains with resultant extensive damage to agriculture and commu-
nications systems. It is important to note, however, that severe El Niiio conditions are not
always accompanied by rain and the very visible damage it causes to the normally dry
land. Thus the warm water intrusion is not a necessary or sufficient condition to bring
about rainfall (Prohaska 1973). Severe and sustained downturns in marine productivity
have occurred (e.g., 1953) without any visible climatic effects on the adjacent landmass.
Archaeologists have focused on the visibly destructive aspects of the event - implying, one
supposes, that flood damage is the major and only human adaptive problem - and they
therefore overlook the implications of the frequency, severity, and duration of all El Niiio
occurrences (e.g., see Engel 1963; Lanning 1967; Willey 1953).
Modem human coastal populations are also affected adversely by higher air tempera.
tures, the spread of disease organisms from the masses of dead birds along the shoreline,
an invasion of spiders and snakes from tropical areas to the north, and a substantial in-
crease in the incidence of malaria. In 1925, for example, the death rate from malaria
rose from 2/1000/yr. to 140/1000/yr., as compared to the rate in tropical Panama at
that time of 19/1OOO/yr. (Murphy 1926). Malaria is unlikely to have been a mechanism
of pre-Hispanic population control, however, since it is thought to have been introduced
during the Conquest from the Old World (Dubos 1974). The modern fishing industry is
also severely affected by El Niiio conditions, as mentioned earlier, primarily due to the
disappearance of the anchovy schools upon which it depends (Idyll 1973).
The area of the Peruvian coast between Tumbes (3” S.) and the Paracas Peninsula
(14O S.) is most subject to El Niiio conditions (Murphy 192s; Schott 1932), although the
effects of the 1925 occurrence were reported as far south as Valparaiso, Chile ( S S O S . ) .
The sign of an impending El Niiio during the rare occurrence that brings rain is the for-
mation of enormous black clouds against the western slopes of the Andean cordillera at
the onset of the phenomenon (Eguiguren 1894). During the years when no rain falls, the
sign that the phenomenon is beginning to occur is the migration of thousands of birds
toward the south along the coast. Coastal residents are therefore provided essentially no
advance warning that El Niiio is about to occur.
Data are available on El Niiio occurrences between 1726 and the present (Craig and
Psuty 1968; Eguiguren 1894; Quinn, Zopf, Short, and Yang 1978). There have been 19
occurrences lasting six or more months in the 254 years since 1726. For purposes of this
paper, these are designated as “very abnormal” events. The El Niiio of 1972 was the most
recent of these (Quinn et al. 1978). Other recent very abnormal occurrences were in 1941
(Lobell 1942), 1925 (Murphy 1926), and 1891 (Eguiguren 1894). In addition, there have
been at least 24 other years since 1726 when conditions returned to normal after about
four or five months. These are designated as “abnormal” events. With a total of 43 occur-
Wilson] OF MAIZE AND M E N 101

rences in 254 years (Figure 2), the average length of the period of normalcy between El
Niiio events has been roughly six years. For a time during the late 19th century it was
thought the phenomenon held to a predictable seven-year cycle (e.g., see Sears 1895).
However, in the 254 years for which good records are available, the data show that the in-
terval length may vary from as few as 6 years to over 20 years, and there is no regular
periodicity or predictability (Paulik 1972; Posner 1954; Rudolph 1953).
Interestingly, the length of warm water conditions for each El Niiio appears to vary in-
versely and rather uniformly as a function of increasing latitude. The maximum duration
of the downturn in productivity is in the North, North Central, and Central Coast areas,
and the minimum duration is in the South Coast area. Nevertheless, for all of these areas
one must consider the time it takes for marine biota to reestablish themselves at normal
population levels following the return of upwelling. Very abnormal El Niiio occurrences
may therefore have a maximum actual duration of 12 months or more, while abnormal
occurrences may last 6 to 8 months. Table I shows the duration in days of warm water
conditions during the El Niiio of 1925 for various areas along the Peruvian coast (see
Sverdrup et al. 1942), and the estimated actual duration of the downturn in productivity
for these areas. Support for the estimates of actual duration comes from Moseley
(1975:ll) who states that, following the retreat of warmer water, “the restoration of the
lost biomass is a slower process. Nature can take several years to hone the blade of her
great double-edged sword and bring sea and shore life back to sharpened balance.” For
the North, North Central, and Central Coast areas, the estimated actual duration of the
bottleneck appears to be so long that pre-Hispanic maritime groups would have had to
adapt by limiting their numbers in accordance with lean period productivity or by some
other regulatory mechanism.
Ryther (1963, 1969) has compared the primary production of the open tropical Pacific
“desert” with that of the upwelling area of the Peru Coastal Current in terms of grams of
carbon fixed by photosynthesis per square meter per year (g. C/m.Z/yr). Primary pro-
ductivity of the open tropical Pacific is 55 g. C/m.2/yr. or one-sixth of the 300
g. C/m.2/yr. in the Current. For purposes of the carrying capacity calculations, it is as-
sumed that a reasonable approximation of lean period productivity during El Niiio is
one-sixth that of normal.

Early Coastal Agriculture

Since general accounts of coastal agricultural systems are available in the literature
(Collier 1961; Moseley 1977; Smith 1960), this section will be confined to a brief con-
sideration of data on early coastal Peruvian maize as a basis for later calculations of the
carrying capacity of Late Archaic-Formative agriculture. It is realized that coastal
agriculture of this period consisted of more than just maize. The assumption is, however,
that it did constitute the greatest single percentage of dietary intake compared to any
other cultigen. This is supported by an examination of modem coastal valleys that still
remain more oriented toward traditional subsistence agriculture than toward
agribusiness. For example, in the Santa Valley maize constitutes about 30 percent of all
agricultural crops (ONERN 1972). Sanders and Price (1968) provide support for the pro-
cedure to be followed here in calculating agricultural carrying capacity on the basis of
maize productivity. They argue that, while cereals do not by themselves provide an ade-
quate source of protein:

In Mesoamerica beans, combined with maize, plus a minimal amount of animal proteins, made
up a nutritional complex that resolved this problem. Added to these staples . . . was a great
variety of minor crops that provided the necessary minerals and vitamins to round out the nutri-
102

-
r
\
AMERICAN A N T H R 0POL OGIST

"5

=z

.
'I
[83, 1981

r "
n
0,

L________
Wilson] OF M A I Z E AND M E N 103
I481.E I DURATION OF WARM WATER CONDITIONS AND ESTIMATED ACTL'AL L.ENC'1-H OF I HF l . t : I >
PERIOD BROUGHT A B O L T BY EL N I N O (BASED O N lYP5 O C C C R R E S C E ) a

Estimated actual
duration of lean
Duration of warm period i n months
water conditions before rerurn to
Loca I i t y Latitude Dates in days normal productivity
Lobitos 4OZO'S. 1/20-4/6 76 ~ +
1 0 12
Puerto Chicama 7'40's. 1/30-4/2 63 8-10
Callao 12020's. 3/12-3/27 15 6-8
Pisco 19O 4 O ' S . 3116-3/24 8 2-4

aSource: Sverdrup et al. 1942

tional system (chile peppers were particularly important). Once this balance is achieved . . .
then dense populations could and did result. For this reason in Mesoamen'ca the gauge o f p o t e n -
tial population demity may be taken f r o m the index of productivr'ly of maize alone [pp. 92-93,
emphasis added].
The earliest known maize at sites on the coast of Peru is from an area less than 70 km.
in length, including Las Haldas, Culebras 1, and Huarmey North 1 (Engel 1970; Kelley
and Bonavia B. 196s; Lanning 1967; Moseley 1975). Radiocarbon dates place this maize
primarily in the so-called Cotton Preceramic part of the Late Archaic, or somewhat prior
to 2000 B.C. From this restricted area of the North Central Coast, maize appears next in
the Early Formative site of Pampa Gramalote in the Moche Valley. By the Middle For-
mative, maize is present in the Chicama Valley on the North Coast, and in the Ancbn-
Chill6n area of the Central Coast.
Among the earliest complete maize cobs are those of a variety called Confite Iqueiio
from Los Cerrillos site in the Ica Valley (Grobman, Salhuana, and Sevilla with Mangels-
dorf 1961). From a sample of 126 cobs, the average length was 6.09 cm. A sample of 49
cobs from the earlier, Late Archaic site of Aspero had an average cob length of 5.8 cm.
Kirkby (1979: 126) provides a graph showing the relationship between average cob length
and average yield of dried seed in metric tons per hectare. Using the above cob lengths of
5.8 and 6.09 -or roughly 6.0 cm. -and Kirkby's graph, the average annual single-crop
yield for Late Archaic-Formative maize on the Peruvian coast would be 0.2 t . (metric
tons), or 200 kg., per hectare.

CARRYING CAPACITIES OF
LATE ARCHAIC-FORMATIVE SUBSISTENCE SYSTEMS

Analytical Procedures and Assumptions

In this section an ecological analysis of critical cultural and environmental variables is

carried out. Specifically, I will analyze the carrying capacity of fishing, shellfish beds,
and early maize agriculture. It is assumed that this provides the adaptive context within
which to understand Late Archaic-Formative maritime and agricultural groups on the
Peruvian littoral. Since this paper is a critique of the maritime hypothesis of state forma-
tion on the coast and argues that the agricultural hypothesis is an acceptable alternative,
to compensate for any bias the procedure will be to establish a "best case" maritime adap-
tation and a "worst case" agricultural adaptation, and, finally, compare the two.
104 AMERICAN A N T H R 0 POLOGIS T [as, 1981
For the maritime adaptation, the analysis is oriented toward the question of what is the
highest productivity that could have been maintained by early maritime groups given the
nature of the marine biome and fishing technology. It will be assumed for purposes of the
analysis that Late Archaic-Formative populations had caballito rafts and could fish out
to 1 km. from shore. The important point to establish here is that at best 1 km.2 adjacent
to the shoreline would have been exploitable. It can then be shown that at best the
maritime adaptation supported only eight persons per kilometer of shoreline. This has
important implications for the nature of early coastal levels of sociocultural integration
based on a marine subsistence system.
For the agricultural adaptation, the procedure will be to establish a “worst case” pro-
ductivity estimate. In other words, the analysis will argue for the lowest possible produc-
tivity from the early agricultural system. To be as conservative as possible, Kirkby’s sug-
gested yield of ZOO kg./ha. for an average cob length of 6.0 cm. will be arbitrarily re-
duced by half, to 100 kg./ha. (this does seem conservative, considering that modern
coastal maize yields an average 1,400kg./ha.; see Robinson 1971:304). It will then be
shown that at worst the productivity of early forms of maize on the coast was in a
magnitude of up to six times the productivity of marine fishing.
Another assumption is that marine carrying capacity can be derived from modern pro-
ductivity figures for the Current at the anchovy level; again, this is a best case estimate
since it seems safe to assume that preindustrial technology would have been more limited
in its harvest of the Current. The estimated figure of 335 kcal./m.2/yr. is not altogether
arbitrarily used, however, since it is based on an annual average for the heavily ex-
ploitative fishing industry of the 1960s. It therefore probably represents a harvest at
something more than peak carrying capacity of the Current. For purposes of the analysis,
the 335 kcal./m.2/yr. has been taken as the maximum amount of energy available to the
next higher trophic level (larger fish). It follows, for example, that total energy flow from
the primary carnivore level (larger fish) to the secondary carnivore level (humans) would
be 93.5 kcal./m.2/yr. (see Odum 1971 for a discussion of energy transfer efficiency under
equilibrium conditions).
In estimating the Recommended Daily Allowance (RDA) of a coastal population, the
guidelines set down by the Food and Agricultural Organization of the United Nations
have been followed (FA0 1954, 1957a,b).The hypothetical FA0 referent is that of a
“young” population (preindustrial) with high birth and death rates, but neither growing
rapidly nor contracting in numbers. It is assumed this would characterize an aboriginal
coastal population. The mean annual ambient temperature of the FA0 referent (“Coun-
try I”) is 25O C. The weight of healthy active adults in the population of this country is 50
kg. for men and 40 kg. for women. FA0 suggests an average RDA of 1,994kcal. for all
men, women, and children in this preindustrial context. This figure was adapted to
coastal Peru (ambient temperature averages Z O O C.) by following the FA0 guideline of a
3 percent increase in the RDA for every loo C. below the reference temperature.

Carrying Capacity of Manne Fishing

Assumption 1 . The productivity of the Peru Coastal Current under normal conditions
is 335 kcal./m.2/yr. at the trophic level of anchovies.
Assumption 2. Human populations of the littoral exploiting carnivorous marine biota
(larger fish and sea lions) are secondary carnivores.
Calculation 1 . From Assumptions 1 and 2, and applying the rule of 10 percent energy
transfer efficiency (Odum 1971). it follows that maritime groups would be able to
harvest a maximum 33.5 kcal./m. 2/yr.
Wilson] OF MAIZE A N D MEN 105

Assumption 3 . The per capita RDA for a maritime population would be 2,023 kcal./
persodday, or 738,395 kcal./person/yr.
Calculation 2 . From Calculation 1 and Assumption 3, we may now calculate the per
capita areal requirement under normal oceanic conditions during one year’s time:

738,395 kcal./person/yr.
= 22,041 m.2/person (2 ha. of ocean)
3 3.5 kcal. /m. z/yr.

Assumption 4 . Late Archaic-Formative maritime groups were able to fish coastal

waters to a distance of 1 km. out from the shoreline.
Calculation 3. From Calculation 2 and Assumption 4, we may now calculate the
human carrying capacity of 1 kmY2of ocean adjacent to the shoreline:

100 ha. /km.

= 50 persons/km.2 (or 50 persons/linear km.of shoreline)
2 ha./person

Assumption 5 . The North, North Central, and Central Coast areas of the littoral are
subject to aperiodic lean periods in the maritime environment that last at least 6-12 +
months, and coastal fishermen would probably adapt by limiting their numbers in
accordance with lean period carrying capacity.
Assumption 6. The productivity of the Current during El Niiio conditions is one-sixth
that of normal, or equal to the productivity of the warmer waters of the open tropical
Pacific.
Calculation 4 . From Calculation 3 and Assumptions 5 and 6, we may now calculate the
maritime human carrying capacity of the North, North Central, and Central Coast
areas:

50 persondkm. X 1/6 = about 8 persons/km.

Dljcwsion. In order to throw light on the implications of the above analysis, let us
briefly examine Moseley’s (1975) demographic data for the Central Coast between Anc6n
and Chill611 (including Tank Site, Punta Grande, and El Paralso), a 50 km.-long area
that he argues best reflects his theory that a predominantly maritime subsistence base
supported civilized society by 1000 B.C. Using the figure derived from the above analysis,
the marine subsistence carrying capacity of the Anc6n-Chill6n area is 400 persons (50
km. x 8 persons/km.). This figure does not accord well with Moseley’s minimum popu-
lation estimate of 2,400-5.450 people in the “fishing” settlements of this area. In order
for even the lower end of Moseley’s estimate to be valid for a primarily maritime-oriented
population, 300 lun. of coastline would be needed (300 km. x 8 persons/km. = 2,400
persons). This clearly seems to be too much coastline, since Anc6n-Chillbn would have
had to control an area from about Caiiete to Supe, in addition to keeping anyone else
from living in the valleys of Omas, Mala, Chilca, Lurfn, Rfmac, Chancay, and Huaura.
If the calculations carried out in this analysis are reasonably correct, then 2,000 of the
2.400 persons suggested as a minimum population estimate for Anc6n-Chill6n would re-
quire the presence of a major subsistence alternative to the maritime system. Using
Moseley’s estimated population range, fully 83 percent to 93 percent of the Anc6n-
Chillen population would have had to rely on agriculture for its subsistence. Whether
these people had achieved statehood or not by 1000 B.C., it could not have been based
primarily on marine subsistence.
106 AMERICAN ANTHROPOLOGIST [83, 1981

Carrying Capacity of Shellfish Be&

Assumption 1 . The productivity of Rangea cuneata clams in a Texas estuary (probably

at least as productive as the Peru upwelling area, if not more so; see Odum 1971:360) of
580 kcal./m.2/yr. may be used to estimate the carrying capacity of shellfish beds along
the Peruvian littoral.
Assumption 2 . A given settlement (e.g., Las Haldas) had access to 10 ha. of shellfish
beds. (Although this is a guess, it would seem to be a liberal estimate.)
Calculation 1. The carrying capacity of 10 ha. of shellfish beds, assuming an adapta-
tion reflecting lean period productivity, is:

580 kcal./me2/yr. X 10 ha. X 10,000 me2/ha.

x 1 = about 13 persons
738,395 kcal. /person/yr.
F

Dzscursion. From the above analysis it does not appear that substantial population
numbers could be supported by even a large area of shellfish beds (10 ha. are equivalent,
for example, to an area of beds 50 m. wide by 2,000 m. long). Data from Fiedler (1944)
show that the 1940 shellfish harvest for six ports along the Peruvian coastline was 245,454
kg., or an average of 40,909 kg. per port. Assuming the caloric value per kg. to be 2,000
kcal. (Odum 1971), even this tremendous catch can be shown to feed only 110 persons
per year (assuming anyone wanted to eat only shellfish for one year’s time). If we assume
the adaptation is a preindustrial one reflecting the need to adjust to frequent, aperiodic
lean periods, then the number of persons supportable by one port’s catch is roughly
18 -not far from the figure of 13 suggested by the above calculations. Another important
point to make here is that, even though 40,909 kg. of shellfish meat would not feed very
many people, it would still result in huge piles of shells. Considering this went on for
50-100 or more years, it would look as if great numbers of people had lived on a site
(perhaps even leading archaeologists to argue a “state“ had existed on the spot), when in
actuality only 18 people lived there during this long period.

Carrying Capacity of Early Maize Agriculture

Assumption 1 . Using the 6.0 cm. average cob length for early coastal maize, and
Kirkby’s graph showing the relationship between cob length and yield, the average yield
€or Late Archaic-Formative maize would be 200 kg./ha., reduced to 100 kg./ha. for the
“worst case” adaptation.
Assum tion 2 . Using data for the lowest yielding maize in the world (India: 350
g
kcal./m. /yr., 1,000 kg./ha.; see Odum 1971:54), we may conservatively quantify early
coastal yields in terms of caloric content.
Calculation I . From Assumptions 1 and 2:

1,000 kg./ha. 100 kg./ha.

-
350 kcal./m.2 X kcal./m.2, and solving for X = 35 kcal./m.Z

Calculation 2 . We may now compute the areal requirements for one person over one
year’s time, assuming one crop per year:

738,595 kcal. /person/yr.

- = 21,097 (2 ha. of land)
35 kcal. /m.2
Wilson] OF MAIZE A N D M E N 107

Discussion. The carrying capacity of early maize agriculture is .50 persodha., or 53

persons/km.’ Using Moseley’s population estimates for Anc6n-Chill6n again, the 2 ,0110
persons not supported by the maritime subsistence system would have needed 40 km.2 of
cultivated land to sustain them. This figure represents 33 percent of the land presently in
cultivation in the Chill6n Valley (Robinson 1971).

Comparison of the Maritime and Agn’cultural Systems

The results of the above calculations are shown in Table 11. Comparing the two sub-
sistence systems, it is obvious that at worst the carrying capacity of early agriculture is six
times greater than that of marine fishing. Of equal importance is the probability that,
given the technological limitations of the Late Archaic-Formative maritime adaptation,
the agricultural system was capable of much more cohesive expansion. Uncultivated land
adjacent to that already being farmed could be added to the system by expanding flood-
water irrigation or by developing canals with more “head,” but all within a reasonably
more compact area. The maritime adaptation could have been expanded by developing
oceangoing craft larger than caballitos, but this was apparently not done until the Late
Intermediate-Late Horizon time period (ca. 1100 to 1532 A.D.), as Lothrop (1992) sug-
gests. A maritime group could also have expanded its fishing grounds by extending them
farther up and down the coastline. It seems probable, though, that other groups from ad-
jacent areas or valleys would have been encountered, or that the energy costs in getting to
and from more distant fishing grounds would have been greater than the caloric value of
the items brought back to the settlement.
The results of the above analysis also imply that maritime groups on the South Coast
would have been subject to a much shorter lean period in the marine biome than that in
the areas to the north. Therefore, the marine carrying capacity figure for this area shown
in Table I1 has not been cut by one-sixth, but has been left at normal productivity, A
comparison of maritime and agricultural carrying capacities of the South Coast shows
they are competitive, or of equal productivity (although if Kirkby’s figure of 200 kg./ha.
is not cut in half, then the agricultural system for this area is twice as productive as the
maritime).

ETHNOGRAPHIC AND ARCHAEOLOGICAL SUPPORT FOR THE MODEL

Ethnographic Data

In this section it will be shown that ethnographic data from modem subsistence-
oriented fishing villages on the coast do not conflict with the model of a very limited
maritime adaptation compared to the early agricultural adaptation. Indeed, these data

TABLE 11. COMPARlSON OF CARRYING CAPACITIES OF THE MARINE AND

AGRICULTURAL SUBSISTENCE SYSTEMS DURING THE LATE ARCHAIC
AND FORMATIVE PERIODS ON THE COAST OF PERU.

Marine Fishing Agriculture

Location Penodha. b . 2 Persons/ha. /h.2
North, North Central,
and Central Coast .08 8.0 .50 50.0
South Coast .50 50.0 .50 50.0
108 AMERICAN A N T H R OPOLOCIST [83, 1981

strongly support the arguments that have been made. Two studies have been done of
marine fishing-gathering settlements that offer insights into the nature of both tradi-
tional subsistence and modem commercial maritime adaptations. The first study
resulted from investigations carried out by Gillin (1947) in the community of Moche. The
second involved a survey undertaken by Hammel and Haase (1962) of 47 fishing settle-
ments on the Peruvian coast between Ica and the Ecuadorian border (Figure 1). These
communities ranged in size at the time of the survey from about 10 to as many as 5,000
inhabitants.
Hammel and Haase show that a clear dichotomy of size and function exists between
communities engaged in commercial fishing and those relying upon coastal waters for
subsistence needs. All of the 95 communities involved to a greater or lesser extent in com-
mercial fishing for canneries or fish meal production had populations of over 100 people.
All of the 12 villages engaged in subsistence fishing had populations of fewer than 100
people, with community size ranging from 12-50 people. These villages are located all
along the Peruvian coast and do not appear to vary in size as a function of their proximity
to major offshore fish concentrations (Figure 1). Fishing is done principally from the
shore, although caballito rafts are still used in a few places to fish just out beyond the
surf.
Some insight into the use of caballitos for fishing can be gained from Spanish docu-
mentary sources of the 16th and 17th centuries. For example, Bernabe Cob0 (1964
[1653], II:269; translation mine) states that fishermen of the 17th century:
, . , went out to fish in the ocean together, sometimes with many in the group, but always in

pairs, each man astride his little balsa [caballito]. Once the net is set out, the two men gather it
up and come closer together as they do this, until the balsas meet.
Today, hand lines and small nets are generally used in conjunction with the caballitos.
Still, as Holmberg (1950) points out for Vird fishermen, the subsistence input from
caballito fishing is usually quite limited, especially in comparison to the modern ocean-
going boats used in the commercial communities. Thus, even though the ocean off Virfi
abounds in fish, techniques involving the use of nets, hook-and-line,and reed rafts do not
afford a living for more than a few families. Hammel and Haase (1962:212) also point
out that subsistence "fishing villages tend to be more conservative and poorer than the
average coastal agricultural settlement." The limitations of subsistence fishing technol-
ogy may also help explain why the South Coast area does not exhibit the higher popula-
tion densities during the Late Archaic-Formative time period as predicted by the model.
The ethnographic data alone would not, of course, necessarily indicate that the Late
Archaic-Formative maritime adaptation was limited. But, taken in conjunction with the
indications of the ecological analysis carried out above, it seems clear that the maritime
hypothesis is untenable. Nevertheless, we must still examine the archaeological data to
understand why such apparently complex sites as Las Haldas, Aspero, and El Paraiso
should appear so early on the coast.

Archaeological Data

The analysis of maritime carrying capacity shows that Late Archaic and Formative
Period fishing groups would probably have had to adapt to frequent and long bottleneck
periods by limiting their numbers in some way. The figure of eight persons/km. of shore-
line thus has implications for settlement size and spacing along the Peruvian coast for this
period. If we assume the average maritime settlement size was 100 persons, then these set-
tlements would have been spaced (other things, such as water supply, being equal) at
about 12-km. intervals. The length of the Peruvian coastline is 2,300 km., so the total
Wilson] OF MAIZE A N D MEN 109

number of contemporaneous sites focused on a maritime adaptation could have been as

high as 190. But, considering that this reflects a “best case” maritime adaptation, as well
as the presence of long waterless stretches of desert between valleys, this number is prob-
ably too high. In any case, the total maritime population-assuming only minor inputs
from horticulture-at any one time on the coast would have been less than 20,000 peo-
ple. Sanders and Price (1968) suggest minimum numbers of 25,000 people for a state
society and about 1.000 people for a chiefdom society. It is difficult to imagine how either
level of complex society could have been maintained with a subsistence base that was
primarily maritime.
T o date, slightly over 80 Late Archaic sites- all presumed to be contemporaneous -
have been found on the coast (Moseley 1975). These sites (Figure 3) fall clearly into two
major types: (1) a majority which fit the model as outlined so far, since they are small and
noncomplex; examples include Huaca Prieta, Huaca Negra, Curayacu, Asia, and
Otuma; and (2) a minority of sites that do not fit the model of a limited maritime adapta-
tion, since they are too large and complex: these sites include Las Haldas, Culebras 1,
Aspero, Rlo Seco, Tank Site, and El Paraiso. Before the latter, anomalous type of site is
discussed, it is useful to define briefly the nature of the smaller maritime sites.
Huaca Prieta in Chicama Valley is typical of later Preceramic-Early Formative sites all
along the Peruvian coastline (Bird 1948a, b). It has a very small area of .62 ha. although
its 15 m. height and complex stratification indicate that it was occupied over a long
period of time. The subsistence economy appears to have been based primarily on marine
animals, with some wild plant gathering and gardening of squash, chili peppers, four
varieties of beans, achira tubers, bottle gourds, and cotton. Maize was not found in the
midden debris, nor were any remains of land mammals noted.
The problem of assessing the nature of the larger, complex sites such as Las Haldas,
Aspero, and El Paralso essentially revolves around three issues: (1) whether their size and
internal complexity warrant designating them as “state centers”; (2) the identification
and nature of the other sites in the “state system” of which they are presumed to have
been a part: and (3) whether the subsistence systems of these sites are primarily marine or
agricultural. In discussing the data related to these issues, I will attempt to show that: (1)
all of the sites are too small and internally noncomplex to warrant calling them “state
centers”: (2) the proponents of the maritime hypothesis have failed to identify the lower-
level sites in the control hierarchy of these presumed “states”: and (3) the hypothesis of a
primarily maritime subsistence base can be brought seriously into question.
To provide a framework for understanding the level of sociocultural complexity of the
sites we are dealing with, it is useful to outline briefly major features of nonegalitarian
chiefdoms and state societies, as well as some of their archaeological analogs (see Flan-
nery 1972; Sanders and Price 1968; Service 1975; Wright and Johnson 1975). Chiefdoms
may be defined as ranked societies with a two-level hierarchical system that is manifested
in various of their component features. Social status and group membership, for exam-
ple, are defined in terms of two partially kinship-based classes consisting of a smaller
“chiefly” elite class and a larger “commoner” class. Since these statuses are ascribed from
birth, one of the ways of archaeologically identifying a chiefdom may be the finding of
juvenile burials with lavish sumptuary offerings indicating their inherited higher-class
status. The person who fills the office of “chief’ may serve in both a religious capacity as
priest and an economic capacity as a pivotal redistributor of food and sumptuary goods.
Thus, one may find an “on the ground” association between “chiefly” residences, temple-
platform structures, and large storage facilities. Settlement patterns in chiefdoms may
also be two-level, consisting of a primary “chiefly” level center and a secondary “com-
moner” level of hamlets or villages.
Compared to the urban centers of states, a chiefdom center is usually relatively small
110 AMERICAN A N T H R 0POL OG IS T [8Y. 1981

AChovin d* Hu6ntor

10'
A Wotorh

Fig. 3. Coastal sites of the Late Archaic-Formative Period mentioned in text and major arrhaeologi-
cal regions of the coast.
Wilson] OF MAIZE A N D MEN 111

(about 50 ha., or less) and internally noncomplex. While the size of the center may run to
several thousands of people, most of the population is located in the surrounding hinter-
land. The total population size of a chiefdom usually ranges between 5,000 and 25,000
people, covering a total area of several hundreds of square kilometers. Socioeconomic
pursuits may be characterized as essentially “one-level,” since everyone-except for a few
elite persons -engages primarily in food production. But compared to tribal societies,
the subsistence system of a chiefdom is usually more complex and diverse. Certain
geographical sectors may not only specialize in a narrow range of locally available food
and craft products, but are also “locked into” a system requiring effective channeling of
these products to the chiefdom center for wider redistribution throughout the society.
Probable examples of chiefdom systems in the New World are those centered on Cahokia
(Illinois), La Venta (Mexico), and Chavin de Huantar (Peru).
In contrast to two-level chiefdom societies, states may be defined as societies with a
minimum three-level hierarchy that is observable across a broad range of component
features. Social status and group membership, for example, are organized into at least
three socioeconomic classes including ruler-nobles-military leaders, professionals-over-
seers-administrators, and farmers-commoners. The person who fills the role of ruler may
have total (and often arbitrary) control over the populace, imluding the “right” to have a
sumptuous palace, a massive funerary temple, and the sacrifice of scores of retainers to
accompany him/her in death. Compared to chiefdom sites, state centers are charac-
terized by more monumental structures of larger size, indicating the greater control the
leadership has over corporate labor groups. The size of the centers is usually at least
several hundred hectares and often as many as 10-20 km.2 in area, with residence based
on occupational specialization, and a population numbering at least several tens of
thousands. Below the level of the urban capital in the state hierarchy, one may find at
least a secondary level of urban administrative centers and a tertiary level of villages or
hamlets.
The total population size of a state usually ranges from a minimum of 20,000 up to
several million people, covering a total area of at least several thousands of square
kilometers. Socioeconomic pursuits are divided into at least three hierarchically orga-
nized levels, with a much larger group of nonagriculturalist, full-time specialists than in
chiefdoms. The subsistence system is not only diverse, but places primary reliance on in-
tensive hydraulic agriculture and massive land reclamation projects. While local farmers
in a state may be as self-sufficient as their tribal ancestors, they may also contribute to
and benefit from complex institutionalized redistribution systems centered around
markets or palaces. Examples of New World state systems are Teotihuach (Mexico), and
Moche, Wari, Chimfi, and the Inca (Peru).
The above definitions of chiefdoms and states are obviously not exhaustively complete,
but they do provide an adequate theoretical context for assessing the nature of Las
Haldas, Culebras 1 , Aspero, Rfo Seco, Tank Site, and El ParaIso. Table I11 lists a few “on
the ground” features for each of these sites. In terms of overall size and internal complexi.
ty alone, none of the sites-except Las Haldas and perhaps El ParaIso-meets the
minimum criteria for a state center. However, on closer examination Las Haldas does not
appear to qualify either. The site is characterized generally by very thin and scattered oc-
cupational debris, and an estimated maximum population size of only 2,250 people
(Moseley 1975). Indeed, for none of these sites does Moseley suggest a population size
higher than 3,000 people. None of them has monumental architecture even approaching
that found at the Moche site, a probable state center of the Early Intermediate Period.
The Huaca del Sol alone at Moche covers a huge area of 228 x 136 m., rises over 40 m.
high, and is composed of over 143 million sunbaked adobe bricks. In comparison, a tem-
112 AMERICAN A NTHR 0POL OGIS T [83, 1981
TABLE 111. COMPARISON OF MAJOR FEATURE5 OF COMPLEX LATE ARCHAIC-FORMATIVE SITES OF THE
NORTH CENTRAL AND CENTRAL AREAS OF THE PERUVIAN COAST.

Approx. Nature and Number of Chronology Agricul.

size in depth of known public of public land Maize
Site hectares deposit buildings buildings nearby? present? References
generally one complex
thin and of structures Early Rowe (1963)
Las Haldas 200 patchy over 6.75 ha. Horizon No Yes Grieder (1975)
not Late
Culebras 1 10 reported none (?) Precerarnic Yes Yes Laming (1967)
six structures
over about Late Moseley and
Aspero 13 2 m. 1 ha. Precerarnic Yes Yes Willey (1973)
two small Late Moseley (1975)
Rlo Seco 12 1.5 rn. structures Precerarnic Yes Yes Matos M. (1966)
Late
Tank Site 100) deep(?) none (?) Preceramic No ? Moseley 1975)
one temple
complex with Late Engel (1970)
El Parafso 58 shallow nine buildings Preceramic Yes ? Moseley (1975)

ple at El Parafso-which supposedly provides definite proof that it is a “state

center’’-measures only 50 m. on a side and 8 m. high.
A more important implication of the above theoretical definitions is the need to iden-
tify a state as a system, composed not only of a center but also lower-levelsites in a com-
plex three-level hierarchy. Unfortunately, the proponents of the maritime hypothesis do
not specify any of these lower-levelsites, so there are no data to indicate either what sites
these “state centers” were controlling or what the functional interrelationships among
sites in a system of such presumed complexity would be. With regard to the subsistence
systems of these sites, both the ecological analysis of the marine and agricultural adapta-
tions as well as the ethnographic data would lead us to predict that the explanation for
their “complexity” lies in an agricultural subsistence base. Indeed, maize has been found
at all of the sites in question, except Tank Site and El Parafso. But the latter site, at least,
is located near 90 ha. of land suitable for floodwater irrigation and another 150 ha. of
land suitable for canal irrigation (Moseley 1975). Of all the sites mentioned, only Las
Haldas (on the arid shoreline 30 km.south of Casma Valley) and Tank Site (at Anc6n, 20
km. north of Chill6n Valley) are clearly located at substantial remove from agricultural
land. Yet, at least in the case of Las Haldas, it seems likely that the maize appearing there
is the result of interaction with groups at such contemporaneous sites as Sechln Alto in
the Casma Valley. The latter area would have been dominant in this probable symbio-
economic relationship, if a comparison of the size of the monumental architecture at the
two sites is any indication. The entire area covered by Early Horizon public architecture
at Las Haldas is 6.75 ha., while just one of the several huge pyramid platforms at Sechln
Alto covers 7.5 ha. and stands 35 m. high.
We are left with the question of what level of sociocultural complexity the larger, com-
plex sites represent. It seems likely that sites such as Las Haldas and El Paralso were chief-
dom centers, but the explanation for this early appearance of complex, nonegalitarian
societies on the coast of Peru clearly lies not on a marine subsistence base but primarily
on an agricultural one.
Wilson] OF MAIZE A N D MEN 115

DISCUSSION

If one assumes, as has been done in this analysis, that the maritime adaptation was
subject to aperiodic lean periods of long duration, then it is appropriate to consider both
the theoretically possible range of adaptive responses as well as the archaeological indica-
tions of which of these responses seem to have been present in the Late Archaic-
Formative time period. The discussion of possible responses will provide the basis for then
arguing that population pressure may provide the relevant adaptive context within which
to understand the shift from the relatively limited marine subsistence system to one based
primarily on agriculture.
There are a number of possible adaptive responses for maritime fishing-gathering
groups subject to the lean period brought about by El Niiio. Among these are: (1) storage
of dried or smoked marine products; (2) a temporary shift to alternative local sources of
food; (3) migration away from the coast; (4) linear or areal expansion of fishing grounds;
(5) population limiting mechanisms; and (6) development of a mixed marine-
agricultural subsistence economy. All of these can be considered relevant adaptive
responses if we assume that coastal maritime groups were at high enough population
levels to be nutritionally stressed during the lean period. In quantitative terms, this im-
plies that the population density of such groups would be higher than eight persons/km.
of shoreline, so that less food was available than was needed to maintain the suggested
RDA of 2.023 kcal./person.
For a maritime group continually facing this sort of adaptive predicament, a most
“logical” response would be the development of fish drying and storage techniques.
Nevertheless, the archaeological record indicates a distinct lack of storage facilities in
coastal sites until after the introduction of maize and other cultigens (Lanning 1967).
With regard to this problem, Moseley (1975:46) has suggested that:
The preservability of marine products under preceramic exploitation was highly limited. . . .
.
The lack of preservable commodities . . underscore[s] the possible consequences of macrocyclic
variation [i.e.,El Niiio] because it would be impossible to stockpile against a sudden . . . drop in
the availability of marine resources.
Moseley is correct in implying that it would have been difficult to store against the lean
period, since, as we have seen, El Niiio occurs at unpredictable intervals of 6 to over 20
years.
It seems unlikely that alternative local sources of food would have been readily
available as a temporary dietary supplement. We know that shellfish are just as narrowly
adapted as fish to normal conditions in the Current, and must migrate or die when El
Nifio occurs. Birds, as well, must migrate with their food sources or face starvation and
disease. Both types of food may be ruled out as significant lean period sources of nutri-
tion. The complex of cultigens available to Late Archaic-Formative groups was limited to
mostly “industrial” crops (i.e., as sources of raw material for making fishing and other
gear), and it seems unlikely that they provided an adequate alternative source of
food. In the same vein, it also seems improbable that groups normally used to fishing
year-in-year-outfor their food could suddenly “pick up and migrate” to the zone of an-
nual rainfall in the Andean cordillera. In order to make a case for such a migratory
response, one would have to assume both that suitable unoccupied zones were always
there waiting for maritime groups and that they could effectively adapt to radically new
food-gathering techniques. In any case, it seems unlikely that anyone other than people
in small, highly mobile groups could adapt in such a way.
Other possible adaptive responses would be to expand fishing grounds by extending the
area of shoreline used (to six times the normal area), or by developing oceangoing craft
 114 AMERICAN ANTHROPOLOGIST [83, 1981
capable of getting out to the major fish concentrations. As argued earlier, the presence of
adjacent fishing groups and cost-yield considerations make the first response unlikely,
and the archaeological record shows that the second response did not occur until much
later.
Of the adaptive responses listed above, the most probable ones seem to be population
limiting mechanisms combined with the development of a mixed subsistence economy
based primarily on agriculture. The agricultural system, of course, was also subject to
limiting factors such as total supply and seasonal availability of water, amount of
cultivable land, salinization, and technology. But it was clearly much less limited than
the maritime system. For example, the agricultural population of even a small valley such
as Virfi-with 98 km.2 of cultivable land-was approaching 25,000people by late Early
Intermediate times (Willey 1953). Maritime groups could have maintained themselves at
densities more or less reflecting the lean period carrying capacity through such regulatory
mechanisms as infanticide, senilicide, delay in age of marriage, prolonged lactation, and
other controls used by tribal groups (Kunstadter 1972). While such control mechanisms
unfortunately can only be seen indirectly in the archaeological record, they can be in-
ferred by finding low population numbers and small settlements which are widely dis-
persed. As shown in this paper, this is precisely the situation indicated on the Peruvian
coast during the Late Archaic-Formative time period for maritime settlements. None of
the over 70 primarily maritime sites appears to have a size of more than several hundred
persons, and average distance between sites along the 2,300-km.coastline of Peru is
about SO km.
Thus we have indirect evidence for an adaptation to marine lean periods in the form of
low overall population numbers (fewer than 20,000 people along the entire Peruvian
coast), and small, dispersed maritime settlements. The archaeological sequences of the
better-known valleys (e.g., Virii, Nepeiia, and Moche) suggest that after the Early For-
mative, increasingly wider use was made of their agricultural potential after such
cultigens as maize and beans had been introduced. The adoption of a mixed subsistence
economy based primarily on agriculture, with secondary reliance on marine products,
may well provide the context for the development of redistributive chiefdom economies
in such valleys.
It does not seem unlikely that pressure on the maritime subsistence base was part of the
reason for the shift to a mixed economy. Indeed, it is hard to imagine that such a shift
would have taken place in the absence of such pressure. As we have seen, early maritime
groups probably experienced a continual occurrence of lean periods brought about by El
Niiio. To the extent that a maritime group had more persons than could adequately be
supported during the lean period, nutritional stress would have been felt. Figure 4 shows
how this process could have occurred continually over a long period of time. Potential
stress situations are indicated where the solid line C/luctuafingabsolute carrying capacity
at the human trophic level) dips below the dashed line (actual numbers of persons/km. of
shoreline),
Archaeologists working on the coast of Peru have argued for some time that population
pressure may provide the relevant context for understanding sociocultural evolution in
the area (Cohen 1977;Lanning 1967;Moseley 1975; Patterson 1971). Their assertion has
not been supported, however, by an attempt to deal quantitatively with the archaeolog-
ical and environmental data, nor do their arguments square with their view of a con-
tinuously and highly productive marine environment. Certainly, if the Current remained
always at a high level of productivity, we might expect to see ranked, if not stratified,
society appearing very early on the Peruvian coast, and well before the introduction of
various cultigens. This constancy, or high continuous productivity, is clearly not the case
for the Peru Coastal Current, as has been shown. A model of pressured maritimc
 Nilson] OF MAIZE AND M E N 115

Fig 4 . A model of human maritime adaptation to El Nino bottleneck periods during the Late Ar
chaic-Formative Period on the coast of Peru.

resources at a simple tribal level- as indicated by the archaeological, ethnographic, and

environmental data- does work well to explain both the nature of the coastal maritime
adaptation and the eventual shift to agriculture.
Maize, among other crops, provided the base upon which early coastal states ultimate-
ly developed (e.g., the Moche state at ca. 300 A.D.). But agriculture did not serve deus ex
machina to bring about a miraculous and instantaneous shift to more complex societies.
This point has been understood by some archaeologists for a number of years. For exam-
ple, Collier (1961) points out that established irrigation fanning took hundreds of years
to develop after the introduction of maize and other cultigens to the coast. Although
these crops had no immediate apparent effect on early coastal subsistence-settlement
systems, they provided an initially insignificant "kick' to one sector of the maritime
system. With the gradual development of agricultural techniques and improved varieties
of maize, in time this small initial deviation amplified to a point where agriculture
became a highly productive and primary focus of coastal subsistence, and the foundation
of state society on the Central Andean littoral.

NOTES

Acknowledgments. Research for this paper was carried out over several years at libraries of the
University of Michigan, the Scripps Institute of Oceanography, and the University of California at
116 AMERICAN A N T H R OPO LOClST [83, 1981

Berkeley. During this time I made four separate trips to the Peruvian coast to visit various sites,
some of which are discussed here. I wish to express my deep appreciation to a number of people who
have helped in various ways to improve the logic and rhetoric of the arguments presented. Kent V .
Flannery. chairman of my doctoral committee, has provided much direct encouragement and,
more importantly, has contributed significantly to my intellectual development through his insight -
ful and far-reaching ideas about sociocultural evolution and the rise of complex society. Richard I .
Ford, Jeffrey R. Parsons, and Henry T. Wright have patiently read several earlier versions and have
kindly taken considerable time to comment constructively and suggest improvements. My wife.
Diana Finch Wilson, has been a constant source of encouragement at every stage of the research
and has carefully read the various manuscripts, contributing to the clarity of presentation. Finally.
to Michael E. Moseley and Robert A. Feldman-amicable "adversaries" and proponents of the op-
posing viewpoint about coastal state origins- I wish to express my thanks for the help and advice
they have given me, including orienting me in preparation for one of my trips to Peru. In par-
ticular, I hope Moseley will recognize my debt to him for his intensive research and publications in
coastal Peruvian archaeology over a number of years from which I have profited.
' In The M a d i m e Foundations ofAndean Cidization, Moseley (1975:3) argues that "For pur-
poses of this study civilization was not extant at 3000 B.C.. but it had emerged and was fully estab-
lished some two millenia later." This argument has been continued by Feldman (197827) who,
based on his research at Aspero site, argues that a coastal state focused primarily on a maritime sub-
sistence system arose before the end of the Preceramic Period, that is, prior to ca. 1800 B.C. A
popular account of Moseley and Feldman's theory is found in recent editions of The South Amerr-
can Handbook. In the most recent 1980 edition (Brooks and Candy:605). for example, the editors
state that:
Two Harvard archaeologists. Michael Moseley and Robert Feldman, have come to the conclu-
sion, based on excavations at coastal sites at Anc6n. Supe . . . that an early civilization sprang up
on the coast on the basis of a fishing economy and was well developed by 2100 B.C. (This is in
complete contradiction of one of archaeology's basic tenets that development of agriculture must
invariably precede civilization.)
This statement has appeared in various earlier editions of the handbook, as mentioned, and it
therefore seems fair to conclude that Moseley and Feldman do not disagree with the editors' charac-
terization of their theory, or they would have requested its modification or deletion. My point is
that, considering their formal presentation of the maritime theory as well as its popularization in a
widely read and respected travel handbook, it becomes doubly important that its basic tenets be
critically examined.
It is fair to point out that, while Lanning has argued for rather high continuous marine produc-
tivity and a resultant population explosion among early coastal fishing groups, in his book Peru
Before the Incas (1967:67) he appears to back down somewhat. For example, in discussing the
presence of maize in the uppermost preceramic levels at Culebras 1 and Las Haldas. Lanning states
that "It is tempting to think that regardless of how much sea food was being eaten, the population
surge was mainly a result of the harvesting of such a productive, storable crop."

REFERENCES CITED

Bartholomew, George A.. and J. B. Birdsell

1962 Ecology and the Protohominids. In Culture and the Evolution of Man. Ashley Montagu.
ed. pp. 20-37. New York: Oxford University Press.
Bennett, Wendell C., and Junius B. Bird
1960 Andean Culture History. Garden City, N.Y.: Natural History Press.
Benson. Elizabeth P.
1972 The Mochica: a Culture of Peru. New York: Praeger Publishers.
Bird, Junius B.
1948a America's Oldest Farmers. Natural History 57:296-303, 334-335.
1948b Preceramic Cultures in Chicama and Virfi. In A Reappraisal of Peruvian Archaeology
Society for American Archaeology, Memoir 4:21-28.
Wilson] OF MAIZE A N D MEN 117

Brooks, John, and Joyce Candy, eds.

1980 T h e 1980 South American Handbook. 56th ed. Bath, England: Trade and Travel Publi-
cat ions.
Cobo, Bernabt
1964 [ 16531 Historia del Nuevo Mundo. 2 vols. Madrid: Ediciones Atlas.
Cohen, Mark Nathan
1977 T h e Food Crisis in Prehistory: Overpopulation and the Origins of Agriculture. New H a -
ven: Yale University Press.
Coker. R. E.
1918 Ocean Temperatures off the Coast of Peru. Geographical Review 5:127-135.
Collier, Donald
1961 Agriculture and Civilization on the Coast of Peru. In T h e Evolution of Horticultural
Systems in Native South America. Johannes Wilbert. ed. pp. 101-109. AntropolBgica. Sup-
plement Publication, No. 2. Caracas: Editorial Sucre.
Cowles. Timothy J . , Richard T. Barber, and Oscar Guillkn
1977 Biological Consequences of the 1975 El Niiio. Science N.S. 195:285-287.
Craig, Alan K., and Norbert P. Psuty
1968 T h e Paracas Papers: Studies in Marine Desert Ecology. Reconnaissance Report Vol. 1.
No. 2' (Occasional Publication of the Departmewnt of Geography. No. 1). Boca Raton, Florida
Atlantic University.
Dubos, Rent
1974 Man Adapting. New Haven: Yale University Press.
Eguiguren. D. Victor
1894 Las Lluvias en Piura. Boletfn de la Sociedad Ceogrifica de Lima 4: 241-258.
Engel. Frederic
1963 A Preceramic Settlement on the Central Coast of Peru: Asia, Unit 1. Transactions of the
American Philosophical Society, Vol. 53, Part 3.
1970 Las Lomas de Iguanil y el Complejo de Haldas. Lima: Universidad Nacional Agraria.
F A 0 (Food and Apicultural Organization of the United Nations)
1954 Food Composition Tables- Minerals and Vitamins. F A 0 Nutritional Studies, No 11.
Rome.
1957a Calorie Requirements. Report of the Second Committee on Calorie Requirements. F A 0
Nutritional Studies. No. 15. Rome.
1957b Protein Requirements. F A 0 Nutritional Studies, No. 16. Rome.
Feldrnan. Robert A.
1978 Informe Preliminar Sobre Excavaciones en Aspero, Peru y sus lmplicaciones Te6ricas.
Investigaci6n Arqueol6gica. No. 2 (Marzo):20-27. Trujillo. Peru.
Ficdlcr. Reginald H .
1944 The Peruvian Fisheries. Geographical Review 34:97- 119.
Flanncry. Kent V .
1972 The Cultural Evolution of Civilizations. Annual Review of Ecology and Systematics 3:399
426.
C7illin,John
1947 Moche, a Peruvian Coastal Community. Smithsonian Institution, Institute of Social An-
thropology Publication. No. 3.
(;ilmartin. Malvern
1964 CompilaciBn Bibliogrifica sobre la Oceanografia de las Aguas Litorales de Colombia,
Ecuador y Peru. con Especial Referencia a1 Fen6meno "El Nifio." Instituto de Pesca del
Ecuador, Boletin Cientifico y Ttcnico 1: 1- 15. Guayaquil.
Grieder. Terence
1975 A Dated Sequence of Building and Pottery at Las Haldas. mawpa Pacha 13:99-112.
Grobman. Alexander. Wilfred0 Salhuana. and Ricardo Sevilla with Paul C. Mangelsdorf
1961 Races of Maize in Peru: Their Origins, Evolution and Classification. National Academy of
Sciences. Washington, National Research Council Publication, No. 915.
Gunther. E. R.
1936 Variations in the Behavior of the Peru Coastal Current. Geographical Journal 88:37-65.
118 AMERICAN ANTHROPOLOGIST [83, 1981

Hammel. E. A.. and Ynes D. Haase

1962 A Survey of Peruvian Fishing Communities. Anthropological Records 21:211-229.
Holmberg, Allan R.
1950 Virh: Remnant of an Exalted People. In Patterns for Modern Living. pp. 367-416. Chi-
cago: The Delphian Society.
Idyll, C. P.
1973 The Anchovy Crisis. Scientific American 228:22-29.
Jordan S . , Rbmulo
1971 Distribution of Anchoveta (Engraulis ringens J.) in Relation to the Environment. Investi-
gacibn Pesquera 35: 113-126. Barcelona.
Kelley, David H . , and Duccio Bonavia B.
1963 New Evidence for Pre-ceramic Maize on the Coast of Peru. Nawpa Pacha 1:39-41.
Kidder, Alfred, I1
1964 South American High Cultures. In Prehistoric Man in the New World. Jesse D. Jennings
and Edward Norbeck, eds. pp. 451-486. Chicago: University of Chicago Press.
Kirkby. Anne V . T.
1973 The Use of Land and Water Resources in the Past and Present Valley of Oaxaca. Mexico.
University of Michigan Museum of Anthropology, Memoirs, No. 5. Ann Arbor.
Kosok, Paul
1942 The Role of Irrigation in Ancient Peru. Proceedings of the Eighth American Scientific
Congress. Anthropological Sciences 2: 169-178.
Kunstadter. Peter
1972 Demography, Ecology, Social Structures, and Settlement Patterns. In The Structure of
Human Populations. G . A. Harrison and A. J . Boyce, eds. pp. 313-351. Oxford: Clarendon
Press.
Lanning, Edward P.
1965 Early Man in Peru. Scientific American 213:68-76.
1967 Peru Before the Incas. Englewood Cliffs, N.J.: Prentice-Hall.
Lavalle y Garcia. J . A. de
1917 La Contracorriente Ecuatorial como Causa Determinante del Fenbmeno Marino Cono-
cido con el Nombre de "Aguaje." Boletin de la Sociedad Geogrifica de Lima 33:313-330.
Lobell. Milton J.
1942 Some Observations on the Peruvian Coastal Current. National Academy of Sciences,
American Geophysical Transactions of 1942, pp. 332-336. Washington.
Lothrop, S. K.
1932 Aboriginal Navigation off the West Coast of South America. Journal of the Royal Anthro-
pological Institute 62:229-256. London.
Mason, J. Alden
1968 The Ancient Civilizations of Peru. Baltimore: Penguin Books.
Matos M.. Ramiro
1966 El Period0 Cerimico Inicial en la Costa Central del Perfi. XXXVI Congreso Internacional
de Americanistas, Actas y Memorias, Vol. 1, pp. 509-518. Seville.
Ministerio de Agricultura
1952 Peces Comunes de la Costa Peruana. Pesca y Caza 4:49-71. Lima.
Moseley, Michael Edward
1975 The Maritime Foundations of Andean Civilization. Menlo Park, Calif.: Cummings Pub-
lishing Company.
1977 Waterways of Ancient Peru. Field Museum of Natural History, Bulletin 48:10-15.
Moseley, M. E., and G. R. Willey
1973 Aspero, Peru: A Reexamination of the Site and its Implications. American Antiquity 38:
452-468.
Murphy, Robert Cushman
1923 The Oceanography of the Peruvian Littoral with Reference to the Abundance and Dis-
tribution of Marine Life. Geographical Review 13:64-85.
1926 Oceanic and Climatic Phenomena Along the West Coast of South America During 1925.
Geographical Review 1526-54.
Wilson] OF MAIZE A N D M E N 119
Naylor, E.
1965 Effects of Heated Effluents Upon Marine and Estuarine Organisms. Advances in Marine
Biology 3:63-103.
Odum, Eugene P.
1971 Fundamentals of Ecology, 3rd ed. Philadelphia: W. B. Saunders Company.
ONERN (Oficina Nacional de Evaluaci6n de Recursos Naturales)
1972 Inventario, Evaluacibn y Us0 Racional de 10s Recursos Naturales de la Costa: Cuencas de
10s Rlos Santa, Lacramarca y Nepeiia. 3 vols. Lima: ONERN.
Patterson, Thomas C.
1971 The Emergence of Food Production in Central Peru. In Prehistoric Agriculturr Stuart
Struever, ed. pp. 181-207. Garden City. N.Y.: Natural History Press.
Paulik, Gerald J.
1972 Anchovies, Birds and Fishermen in the Peru Current. In Environment: Resources, Pollu-
tion and Society. William W. Murdoch, ed. pp. 156-185. Stamford, Conn.: Sinauer Asso-
ciates.
Posner. Gerald S.
1954 The Peru Current. Scientific American 190:66-71.
Prohaska. Federico J.
1973 New Evidence on Climatic Controls Along the Peruvian Coast. In Coastal Deserts: Their
Natural and Human Environments. D. H. K. Amiran and A. W. Wilson, eds. pp. 91 107.
Tucson: University of Arizona Press.
Quinn, William H., David 0. Zopf. Kent S. Short, and Richard T. W. K u o Yang
1978 Historical Trends and Statistics of the Southern Oscillation, El Niiio, and Indonesian
Droughts. Fishery Bulletin 76(3):663-678.
Robinson, David A.
1971 Peru in 4 Dimensions. Detroit: Blaine-Ethridge.
Rowe. John H.
1954 Max Uhle, 1856-1944: A Memoir of the Father of American Archaeology. (University of
California Publications in American Archaeology and Ethnology 46: 1- 134.) Berkeley: Univer-
sity of California Press.
1963 Urban Settlements in Ancient Peru. mawpa Pacha 1:l-27.
Rudolph, William E.
1953 Weather Cycles on the South American Coast. Geographical Review 43:565-566.
Ryther. John H.
1963 Geographic Variations in Productivity. In The Sea. Vol. 2. M. N. Hill, ed. pp. 347 -380.
New York: Interscience Publishers.
1969 Photosynthesis and Fish Production in the Sea. Science 166:72-76.
Sanders, William T., and Barbara J. Price
1968 Mesoamerica: the Evolution of a Civilization. New York: Random House.
Schott, Gerhard
1932 The Humboldt Current in Relation to Land and Sea Conditions on the Peruvian Coast.
Geography 17:87-98.
Schweigger, Erwin
1943 Pesqueria y Oceanograffa del Perfi y Proposiciones para su Desarrollo Futuro. Lima:
Compafiia Administradora del Guano.
Sears, Alfred F.
1895 The Coast Desert of Peru. Journal of the American Geographical Society 27:256-271.
Service, Elman R.
1975 Origins of the State and Civilization: the Process of Cultural Evolution. New York:
W. W. Norton.
Smith, C. T.
1960 Aspects of Agriculture and Settlement in Peru. Geographical Journal 126:397-412.
Stevenson, Memtt R.,and Helen R. Wicks
1975 Bibliography of El Niiio and Associated Publications. Interamencan Tropical Tuna Com-
mission, Bulletin 16:451-501.
120 AMERICAN ANTHROPOLOGIST 183. 1981

Strong, William Duncan, and Clifford Evans, Jr.

1952 Cultural Stratigraphy in the Vird Valley, Northern Peru: the Formative and Florescent
Epochs. New York: Columbia University Press.
Sverdrup, H. U . , Martin W. Johnson, and Richard H. Fleming
1942 The Oceans: Their Physics, Chemistry, and General Biology. New York: Prentice.Hal1
Washburn, Sherwood L., and C. S. Lancaster
1968 The Evolution of Hunting. In Man the Hunter. Richard B. Lee and Irven DeVore. eds.
pp. 293-303. Chicago: Aldine.
Willey, Gordon R.
1953 Prehistoric Settlement Patterns in the Vird Valley, Peru. Bureau of American Ethnology
Bulletin, No. 155. Washington.
Wright, Henry, and Gregory Johnson
1975 Population, Exchange and Early State Formation in Southwestern Iran. American A n -
thropologist 77:267-289.
Wynki, Klaus, Edward Stroup. William Patzert, Robert Williams, and William Quinn
1976 Predicting and Observing El Nifio. Science 191:343-346.

Submitted 1 J January 1979

Accepted 30 January 1980
Final revisions received 30 October 1980