The invisible Indians


Material Information

The invisible Indians a history and analysis of the relations of the Cocamilla Indians of Loreto, Peru, to the state
Cocamilla Indians of Loreto, Peru, to the state, A history and analysis of the relations of the
Physical Description:
xxii, 366 leaves : ill. ; 28 cm.
Stocks, Anthony Wayne, 1939-
Publication Date:


Subjects / Keywords:
Indians of South America -- Peru -- Loreto   ( lcsh )
Acculturation -- Case studies   ( lcsh )
Indians of South America -- Social conditions -- Peru   ( lcsh )
Indios de Perú -- Relaciones con el gobierno
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis--University of Florida.
Includes bibliographical references (leaves 353-365).
Statement of Responsibility:
by Anthony Wayne Stocks.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000081981
notis - AAJ7307
oclc - 05112670
sobekcm - AA00004914_00001
System ID:

Full Text







This work is dedicated to the people of
Achual Tipishca who give so freely and
who have received so little.



'v. I ^.^v-^^

------------- SC A


: sh wing ode? dcay
3: distribution of tupitan D
,ndi;rs in pEru O
SSCALE: x aPPrx 1:2 mol'l 2
\ ^ \ I

\ |H -^ ^ --J


The title of this work was chosen in order to drama-

tize the rather unique position of large numbers of native

American Indians in eastern Peru. While a considerable

amount of attention has been given to the problems of

"tribal" Indians, there exists a large class of "detriba-

lized" Indians who are generally ignored. It is usually

thought in Peru that such Indians have integrated into the

white-mestizo culture and that they have disappeared

biologically and as ethnic groups. Were this so, we would

be presented with a selvatic society in which most rural

members would be equivalent to the Brazilian caboclo. In

fact, the term ribereio is generally used in Peru in exactly

the same sense that caboclo is used in Brazil, to describe

the rural frontiersmen of the tropical forest region who

bear a culture which is much closer to European patterns

than to autochthonous native patterns. Furthermore, the

use of the term ribereno in Peru implies that most rural

populations which are not tribal Indians do in fact fit

into this category and that they form a homogeneous group

culturally and socially.

Unfortunately for this view, social reality in eastern

Peru does not match the Brazilian model. While there is

certainly a ribereno stratum which is equivalent to the

Brazilian caboclo (I would call them white-mestizo rural

frontiersmen), there is also a stratum called the cholada

which does not, from all accounts, exist in Brazil. The

cholada is composed of the detribalized, acculturated, but

unassimilated Indians who have made a wide range of adjust-

ments to Peruvian society.

The Cocamilla are members of the cholada. They live in

communities which might be described as native enclaves in

the midst of a world dominated by white-mestizo culture.

They are acculturated to many of the superficial aspects of

white-mestizo existence, appearances which are deceptive.

They retain many of the social forms and practices of

their historical past. They retain them not through some

vague sense of tradition but because they find them useful.

They have special needs and problems which derive from

their position at the bottom of the Peruvian social system

as it is encountered in the tropical forest regions. The

only social status lower than the cholo in eastern Peru

is that of the tribal Indian, and it cannot truly be said

that the tribal Indian is a member of Peruvian society.

The needs of the Cocamilla and the rest of the cholada

are not recognized. It is much more convenient to pretend

that they have disappeared, integrated, assimilated,

"mestizoized." This pretense is facilitated by the degree

of acculturation of the cholada to white-mestizo patterns

which renders them so similar in outward appearance to the

white-mestizo riberenos that they are effectively "invisible."

Hence, the title of the present work. By examining the

history and relations with the state society of what is now

Peru of one representative group of the selvatic cholada,

it is hoped that sensitivity toward the plight of other

members of this social class will be heightened. It is not

a small class in numbers. The Cocamilla and their closely

related cousins, the Cocama, have a population which ap-

proaches 25,000 (Stocks, 1977). If all of the cholada

were counted their population would easily reach 100,000,

a significant percentage (7%) of Lcreto's population of

close to one and a half million persons.

Since the subject will not arise again until the last

chapter of this dissertation, I would like to say something

about the theoretical orientation which is implicit in

much of the organization and understanding of the material

contained in this work. I have been for some time and


continue to be influenced by the perspective in American

anthropology called "cultural materialism" by some (cf.

Harris, 1968, for a thorough treatment of this strategy

for organizing our thoughts about cultural matters).

The debt of cultural materialism to the "dialectical ma-

terialism" of Marx has been generally recognized. I

consider myself to be an economic and ecological anthropolo-

gist. To me, the subjects are entwined to a degree which

makes their separation, even for heuristic purposes,

impractical. Cultural materialism as a research strategy

suggests, among other things, that the determinants of

cultural patterns are to be found in the objective rela-

tions which a given society or culture has with its environ-

ment, both physical and socio-political. Thus, in this

work I have consistently searched out the meaning and the

implications of the relations between the Cocamilla and

the state in the economic relations between them. It is my

belief that these relations are the major determinants

for relations in the socio-political sphere and in the

kinds of ideas and opinions expressed by the Cocamilla and

the members of the Peruvian society with whom they have

related in the past and continue to be related. When the

Cocamilla are considered alone, it is their productive sys-

tem which receives major emphasis in this work.


I am not unmindful of the debt which cultural

materialism owes to earlier thinkers in the social and

natural sciences. Darwinian evolutionary theory is basic.

The work of Leslie White, especially in The Science of

Culture (1949), in stating the basic framework of the

study of culture has influenced me greatly. His "layercake"

model of society devised from earlier materialists including

Marx, in which the technological and economic base sup-

ports "layers" of social organization and ideology is

useful and is implicit in this work. The work of Robert

Murphy in The Dialectics of Social Life (1971) has greatly

clarified for me the relations between the ideological

"layer" and the "layer" of social interaction. My use of

the term "praxis" in the sense of the social and economic

interactions of humans prior to their conceptions of those

interactions follows Murphy. Marshall Sahlins' work,

Stone Age Economics (1972) has helped me greatly in

organizing my thoughts about the domestic mode of produc-

tion, an economic organization which is fundamental to

many groups of South American tropical forest Indians,

the Cocamilla among them. Since I use the terms "mode of

production" and "relations of production" in this work

several times, it is just that the debt to Marxian social


science be made more explicit in this sense. Marx's

conception of mode of production includes the fundamental

relations between people engendered by a determined way

of producing the necessities of life, and considers the

whole implied by the complex of the productive system and

the relations and cultural patterns necessary to carrying

it on. The mode of production, in short, is a socio-

cultural and economic system and interacts with other

systems of the same conceptual order (cf. Godelier, 1977:

15-69). Thus, one may speak of the conflicts between the

Cocamilla mode of production and a capitalist mode of

production and be understood to mean that the conflicts

are on several levels, economic, social, and cultural or

ideological. Marx's work Pre-Capitalist Economic Forma-

tions (1975) has been influential in making me think about

such matters with regard to native Indian societies.

I would like to thank Dr. Charles Wagley for the

material aid, intellectual guidance, and moral encourage-

ment which made it possible to complete this work. It is

difficult to imagine having completed it without him. The

warm support he provides for his students is appreciated

more than he knows. Special thanks also go to the

members of my academic committee, Drs. Paul Doughty,

Maxine Margolis, Anthony Oliver-Smith, and Glaucio Soares

who read and criticized earlier versions of the disserta-

tion. Needless to say, final responsibility for content

and interpretation lies with me. Among my Peruvian friends

I would like to thank Alejandro Camino for his great aid

in securing institutional affiliation in Peru and all of

the staff of the Centro Amazonico de Antropologia y

Aplicacion Practica (CAAAP) in Lima and Iquitos. Alberto

Chirif of SINAMOS first guided me toward the Cocama and

Cocamilla. Norma Faust of the Summer Institute of

Linguistics deserves special thanks for her great help in

making possible my first orientation among the Cocama of

the Ucayali River. My friends in Lagunas know who they

are and they also know that I value their help and their

friendship. P. Fr. Julian Heras O.F.M. was of great help

in guiding my reading in the library at Santa Rosa de

Ocopa. A special debt is owed to Juan de la Cruz Murayari

who shared his knowledge of Cocamilla history unselfishly.

My wife, Kathleen Butkus Stocks, worked as hard during the

field portion of the study as I did. Her contribution is

beyond calculation. Finally, although they never quite

understood precisely why "Don Antonio" was among them, the

people of Achual Tipishca accepted me and my family,

supported us, fed us, entertained us, socialized my

infant daughter Gabriela, and made us one of them in a

way which is slightly incredible to me yet. To them I

owe the most, and it is to them that this work is dedicated.

Research for the dissertation was carried out through

financial help from several sources. The Tropical South

American Research Program under Dr. Charles Wagley at the

University of Florida paid for a field trip of four months

in 1975 during which time I selected a community and

began to learn to deal with the tropical forest environment.

The major portion of the field work, a stay of 19 months

in 1976-1977, was supported by the Social Science Research

Council in the form of a grant for doctoral dissertation

research, by the Department of Health, Education, and

Welfare in the form of a Fulbright-Hayes grant for

doctoral dissertation research #13.441AH60020, and by the

National Science Foundation in the form of a grant for

improvement of doctoral dissertation research #BNS76-09554.

The Social Science Research Council also provided funds

for a six month extension of fieldwork and a six month

write-up period after the field work had ended, a grant

which has saved me much worry and fatigue. My thanks

to all of these institutions and their dedicated staffs.










S. . iv

S. . xvi

* . xviii

. . xix

. . x


Notes. . .


Toward a New Model of Lowlands
Native Societies. .
The Problem of Ethnicity .

COCAMILLA .. .. ..

Proteins and Populations .. ..
Land and Water Features of
the River Valley .
Land and Water Features of
the Upland Plain. .
The Larger Region. .


S. 1

. 16

. 17

S. 34
S. 43




. 70
. 73

. 77




The Aboriginal Subsistence Economy
Social Organization. .
Adornment. . .
Ideology and Custom. .
Notes. . .

1640-1680 . .

S. 82
. 85
. 92
. 93
. 98

Summary . .
Notes . .

LIFE 1680-1820 . .

The Departure of the Jesuits .
Military Rule in Mainas. .
Summary. . .


. 142

. 143



The Native Response to Independence. 171
Competition for Indian Labor .. 176
Haciendas Appear on the Lower
Huallaga . .. 180
The Rubber Boom. .. . 187
The Barbasco Years .. 195
Recent Times . ... 201
Notes. .. . 207


Agriculture and Land Use .
The Tipishca Economy in
Qualitative Terms .
The Position of the Agrarian Bank.
The Agrarian Band and Rural
Impoverishment. . .


. 218
. 225

S. 234






Land and Water Rights .
Fishing Rights or Starvation .
Present Nutritional Levels .
Summary. . .
Notes . .




The Nature of the Egalitarian
Community--Cocamilla Social
Organization . .. 253
The Community and the State. ... 264
The School as a Model of the Social
Environment of Achual Tipishca. .. 270
Summary. . .. 276


Identity as Peruvians .
The Class of the Apellido Humilde.
The Cocamilla as Cocamilla .
Community Identification .
Shamanism and Cocamilla Identity .
The Fiesta System .
Other Boundary-Maintaining Customs
Summary . .


Today's Cocamilla .
The Future of the Cocamilla .



. 300

S. 305
. 308

. 315

. 331

. 334





BIBLIOGRAPHY. ......... ....... 353



















Climatological Data from Yurimaguas
Experimental Station--1976. .

Cocamilla Population Decline 1638-1681.

Cocama Population Decline 1559-1681

Major Epidemics in Mainas Missions
1638-1681 . .

Population of Lagunas from 1670 to 1768

Lagunas Political Organization 1790

Populations in the Lower Huallaga
Region 1790-1864. . .

Exports of Some Lowland Products
from Loreto 1862-1870 . .

Tipishca Populations as Shown in
Various Censuses. . .

Land Use in Achual Tipishca in 1976
and 1977 . .

Average Time Per Day Spent on Various
Subsistence Activities by Age and Sex

Sample Planting Schedule for Floodplain
Agriculture . .

Expenses of Growing and Marketing Rice
in Achual Tipishca. . .

Time Inputs Per Adult Male Equivalent
(AME) Per Year Per Hectare on Manioc,
Plantains, and Jute . .


. 57

. 138

S 139



. 163


S. 185

S. 200

. 212

S. 216


. 229

. 231

LIST OF TABLES (Continued)

Table Page

8.6 Expenses of Growing and Marketing
Jute in Achual Tipishca. ... 232

8.7 Bank Debt in Achual Tipishca Over a
Three Year Period. . 235

8.8 Degree of Community Involvement with
Agrarian Bank vs. Value of Material
Prestige Items in Huallaga Communities. 236

8.9 Results of Three Dietary Studies in
Achual Tipishca at Varying Points in
the Flood Cycle . .. 244

8.10 Proportions of Food by Weight Supplied
by Various Categories of Activity 245

10.1 Partial Reconstruction of the Annual
Calendar. . ... 293













. 58





S 258




Graph of Twenty-One Year Average
Rainfall in Yurimaguas . .

Subsistence Activities of Adult Males
on a Daily Basis . .

Plantain Prices Per Stalk in Lagunas
1976-1977 in Soles . .

Groups of Patrilineal Kinsmen in
Achual Tipishca. . .

Cocamilla Use of Spanish Kinship Terms

Cocamilla Use of Primo-Tio-Sobrino-
Abuelo Terms . .

Typical Marriage Exchange Over Two
Generations. . .

Typical Cross-Cousin Inter-Generational
Marriage . .


Showing Modern Day Distribution of
Tupian Indians in Peru. . .

Showing General Region of Study with
Locations of Some Native Groups at
Contact . . .

Showing the Limits of the Mainas
Missions . .

Showing the Early Mainas Missions
Discussed in this Chapter .

Showing the Location of Prominent Towns
in the Lower Huallaga Region in the 19th
and 20th Centuries. . .

Showing Part of the Lower Huallaga River
Valley and the Location of the Study
Community, Achual Tipishca. .














Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Anthony Wayne Stocks

December, 1978

Chairman: Charles Wagley
Major Department: Anthropology

This dissertation addresses itself to the problems of

the Cocamilla Indians of Loreto, Peru,and other similar

groups of tropical forest Indians of Peru who are acculturated

to white-mestizo patterns, but who are not assimilated or

well-integrated into Peruvian society. The Cocamilla form,

in effect, is a native "enclave" which is part of the rural

class structure of eastern Peru. Cocamilla ethnohistory pro-

vides the background for understanding how this situation

came about as a result of historical processes. It is

asserted with supporting evidence that the Cocamilla and

other Peruvian native Indians in similar structural posi-

tions actually form a "native" social class which is dis-

tinguishable both from the less-acculturated tribal Indians

and from the poor rural white-mestizos who form the other

component of Loreto rural society. This social class has

special problems because of the discordance produced by

the thorough insertion of native ethnic groups such as the

Cocamilla into the regional variants of Peruvian national

class structure, an insertion which demands of them patterns

of behavior which are alien to them. The Cocamilla, a group

of about 5,000 7,000 people, have survived nearly 340

years of contact with western society by working out an

arrangement whereby they supplied the developing Peruvian

society with labor as canoemen and the products of the lakes

and rivers of the Peruvian varzea, using techniques which

have only recently been modified. Since their own lands,

until the late 19th century, were essentially valueless to

the Peruvian society in terms of agriculture and lumber,

they were allowed to fill this economic niche without being

forced to make drastic modifications in the life-style

generated by their domestic subsistence-oriented mode of

production. This situation has changed in the 20th century

and Cocamilla institutions are undergoing rapid and radical

changes today.

Based on an analysis of the conflicts between the

Cocamilla native communities (defined by the qualities of

being egalitarian, non-intrusive, and essentially kinship-

based communities) and the bureaucratic, hierarchical, and


highly formal institutions of the state with which they

are in constant contact, recommendations are made for

recognizing the Cocamilla's right under Peruvian law as

native Indians. These rights must include rights to both

land and water resources in order to permit them to adjust

to the developing Loreto society on terms which include

the possibility of cultural and economic survival as an

ethnic group.




The Cocamilla are a group of about 5,000 native Ameri-

can Indians living in three major and a dozen minor com-

munities who speak a language called Cocama, a member of

the Tupi linguistic family. They are generally considered

to be a sub-group of the Cocama. The Cocama number about

20,000 people and inhabit the lower Ucayali River flood-

plain in eastern Peru from about latitude 610' South to

the river's mouth. The Cocamilla inhabit the lower

Huallaga floodplain from about latitude 5028' South to

the mouth. These were the historical limits of their

territory, but both groups have expanded considerably in

recent times. Tupian Indians may now be found along parts

of the floodplain of the Marahon, Upper Amazon, Nanay,

Pastaza, and lower Napo rivers, to name but a few.

Neither group has a long history in Peru. Archaeological

evidence indicates that the Peruvian Tupians, including

the Cocamilla and Cocama, probably arrived as part of a

historical migration of Tupian Indians not more than two

or three hundred years before the conquest. They were part

of an expanding population and were reported to have been

extremely bellicose in many of the early historical sources.

I first became interested in the Cocamilla Indians as

I become interested in most things, through slight ex-

perience with them. In 1973 I was traveling down the

Ucayali River in eastern Peru on a decrepit wooden passen-

ger boat. I had been studying in the Andes and wanted to

see more of the country. Part of my plan was to see some

of the eastern lowlands.

My first view of the eastern flanks of the Andes where

the tropical forest begins as a cloud forest above Tingo

Maria was breathtaking, even at 4:00 A.M. in the rain from

a crowded bus. The dry Andean air was suddenly charged

with heat and showerbath humidity. The moon lit up the

thick vegetation on the right of the bus and the impossibly

sheer and deep canyon on the left. We soon became stuck

in the mud while waiting for a landslide to be cleared.

At Pucallpa a couple of days later, my wife and I

found a boat traveling to Iquitos, 1200 kilometers

downstream. The trip was revealing in terms of the river

life and commerce. I was confused at the terms used by

the Cantonese owner and his river pilot to refer to the

occasional Indians (mostly Shipibo, Conibo, and, farther

down, Cocama) we saw living along the shores and

paddling canoes. They were all called chamas or cholos1

by the boatsmen without distinction as to their ethnic

group. Since I had been interested in the so-called

cholos in southern Peru, and since the people called

cholos on the river hardly seemed to be in the same posi-

tions as their highlands counterpart, I became interested

in them.

In 1975 I returned to Peru with my wife, having by

then determined to work in the tropical lowlands as my

professional area specialty in anthropology. In the

course of making contacts in the capital, Lima, we visited

the offices of SINAMOS (National Social Mobilization

Support System), the political-organizational arm of the

1968 military junta. There we talked to Alberto Chirif,

the head of the support office for native communities.

His office was in charge of organizing Amazon native

Indians into political and juridical entities called

native2 communities" so that they might receive land

titles, learn to deal with bank loans, receive identity

papers, and in general defend themselves against white-

mestizo traders and patrons. All of this was being done,

at least on paper, under law #20653, which, since 1973,

has defined the rights of Amazon ethnic minorities, and

has regulated the use of lands in eastern Peru under the

700 meter contour, which was the lower limit of the

Agrarian Reform.

Mr. Chirif indicated an interest in the Cocama Indians.

He pointed out to me that on his maps of the locations of

240 native villages of tropical forest Indians there was

not a single Cocama or Cocamilla village shown. He said

they had been able to organize no Cocama native communities,

and in fact his office did not know which villages along

the Ucayali, Amazon, Maranon, and Huallaga rivers might be

Cocama or Cocamilla. This was true despite the fact that

the Cocama/Cocamilla ethnic groups, with approximately

20,000 population comprise one of the largest native

Indian groups in eastern Peru, and lie along the flood-

plains of the most traveled rivers in the country.

The puzzle, of course, had a simple solution. The

Cocama and Cocamilla were believed to have lost their

"tribal" identity and to have largely merged with and

integrated into that large sector of riverbank frontiers-

men frequently called riberenos in Peru and caboclos in

Brazil. Although no one I talked to in Lima during the


month I stayed there actually said it, there seemed to be

an unstated assumption that if lowland tropical forest

Indians were no longer "tribal," then they were no longer

Indians at all.

This unvoiced assumption actually became voiced later

in the summer in the heart of historic Cocama country,

Requena, where the local head of SINAMOS, the man who,

along with his other duties, was in charge of organizing

the native communities, told me that they had not organized

any Cocama under law #20653 because he did not consider

them to be natives. "They are now mostly crossed with

the mestizos," he said. He saw his job with regard to

"real" natives as helping them to "revalidate their cul-

tural patterns." This was taken to mean getting them to

"preserve their dances and folklore." The Cocama and

Cocamilla had no colorful dances and no one had asked them

about their folklore since 1935. Their villages looked

like the frontiersmen's villages, so much so that the

Requena office was not sure where they could find a Cocama

village. The Cocama, they said, "were too mixed with the

rural agriculturalists and urban slum people to identify."

The only reason I failed to become persuaded that the

Cocama and Cocamilla had disappeared as ethnic groups was

that I was traveling in a small boat with a Cocama family

and they unhesitatingly identified certain communities as

places where old people spoke the language, the dialecto.

Furthermore, the white-mestizos on the Ucayali, the

upper Amazon, lower Maraion, and lower Huallaga rivers,

also knew where idiomeros were to be found and were quite

willing to point out their communities.

On the advice of Norma Faust of the Summer Institute

of Linguistics, the Cocama language specialist, I journeyed

to Lagunas on the lower Huallaga River to see the Cocamilla.

They were supposed to be more conservative of language and

custom than the Cocama. It seemed to be true. Three large

and several small communities of Cocamilla were identified.

Most of the children understood the language, even if they

did not speak it.

I saw no reason why the communities of Cocamilla on

the Huallaga River should not have titles for their land

and protection under the law as native Indians. They

were stable communities in a geographic sense. They

expressed an interest in securing titles to communal land.

The only drawback was that they had never been visited

by a government organizer, and they had no idea of the

law or of their rights. I decided to study the question

of what sort of lands might be desirable for them to

secure title to, a study which implied a study of their

subsistence patterns and needs.

I returned to Peru in July of 1976 for the field por-

tion of the study. By then I was familiar enough with the

Cocamilla so that I naively assumed that they were not

socially "invisible" to the authorities, even if the

more dispersed Cocama seemed to be. I spent the next

nineteen months disabusing myself of that idea as I tried

to convince the officials in the closest SINAMOS office

in Yurimaguas to come and at least talk to the Cocamilla

about their rights as native Indians. The office responded

that the Cocamilla were "campesinos" (peasants), and thus

could only be organized on an equal basis with the white-

mestizo rural agriculturalists. SINAMOS promoters in

Yurimaguas were busy organizing Chayahuita Indians, among

other groups, into native communities. The Chayahuita at

least looked like the popular conception of Indians, the

men with pudding-bowl haircuts, and the women in crude

black homespun skirts and short blouses with bracelets

and beads. The last Cocamilla who wore such clothes had

died two years before. Yet in phenotypical appearance

the Cocamilla were as Indian as the Chayahuita.

As I began to understand something about Cocamilla

social and economic life from living in one of their

villages, I began to see that they were discriminated

against by most, if not all, of their white-mestizo

neighbors. Their history as mission Indians, and later

as peons, had ill-fitted them to be thrust into the

capitalist-extractive or capitalist-agricultural economy

of Loreto as individuals or as communities without some

sort of "head-start." Their historical heritage as

Christian Indians seemed to be strong enough to keep them

distinct as an ethnic group, at least up until the pre-

sent, but they lacked the access to political and

economic power which the ribereno white-mestizo frontiers-

men had.

As I lived among the Cocamilla I began to see that

they still formed a distinct society in which up to 90%

of them found marriage partners with other Cocamilla. I

watched their customs deprecated by the white-mestizo

teachers, their lakes violated by commercial fishermen,

and their daughters go away to near bondage as servants

in urban centers. In such circumstances others have

spoken of "ethnocide." I wish itwere that simple. But

at times there almost seemed to be an element of malevolant

conspiracy in the way the entire Peruvian society in con-

tact with them persistently tried to get them to act

like anything but Cocamilla.

The Cocamilla had long-since developed cultural means

of dealing with such pressures from white-mestizo society.

They had become "closed" in the sense that they did not

speak easily to outsiders and the men tried to behave

like the white-mestizos when they were around them. When

their language was mentioned,they shuffled their feet

and laughed nervously.

But in the past 30 years direct domination by mestizos

who formerly lived in their communities and appropriated

their labor, and direct contact with the public school

system had eroded their old cultural models, and had

destroyed the old system of authority developed during

the Jesuit period 300 years before. Cash cropping had

brought them into close relations with the Peruvian econ-

omy, not as peons any longer, but as independent small

producers. A "generation gap" had appeared as the

younger members of the community tried to deal with new

developments in their life-plans. The Cocamilla were

no longer "tribal" Indians in any sense of the term, but

neither were they white-mestizos, either as a society or

as a culture.

As my hopes faded for seeing them secure titles

for their communal land and their fishing rights protected

in the immediate future, my plans for the kind of work I

would write about them changed. I have always thought that

anthropologists should observe some sort of large-scale

reciprocity with the people who provide the data for their

work. That we who study reciprocity in finely-shaded

detail should ignore such a fundamental seems to me to be

the most gross form of ingratitude. I speak here not of the

difference between so-called "applied" or "action"

anthropology and "theoretical" anthropology, but rather

to the simple dictates of elementary human relations.

The present work thus grew out of experience or praxis,

if you will, with the real conditions of life for the

Cocamilla. It will try to lay bare the historical reasons

for what might be called the Cocamilla "particularity,"

and to show that their specialness has broad ramifications

in their relations with Peruvian national society. It

will argue that the Cocamilla and other native minorities

in a similar socio-economic position deserve special con-

sideration. It will be a very different work, in short,

than it started out to be. The ecological and historical

data are made to serve an argument which is politically

biased in favor of the Cocamilla. I think it must be that

way if I am to write it at all.

A word about method. There are three axes of data

collection integrated in this work. The first is his-

torical, the second ecological, and the third ethnographic.

In terms of history I was interested in the past of

one specific ethnolinguistic group, the Cocamilla. Towards

that end I studied all available documents in the community

in which I lived. In Lagunas I consulted all the docu-

ments in the Catholic mission, and in the meagre public

archives available. In Yurimaguas, the provincial capital,

I examined church archives, consulted the public library,

and examined unpublished documents in the files of the

sub-prefecture. I consulted published sources insofar as

they were available to me on interlibrary loans. I col-

lected data on the people's own version of their history,

and conducted interviews to cross-check the information

as much as possible. There is at least one general

academic work in English on the conquest of the Peruvian

montana (Werlich, 1968) with an extensive bibliography for

those interested in the macrohistory of the region. The

history I wish to write deals essentially with the Coca-

milla experience over the course of more than 330 years

of direct domination by the culture-bearers of the Western

World. I wish in this way to arrive at an understanding

of the sociology of eastern Peru through the Cocamilla


The data on the current Cocamilla economy and ecology,

the subsistence system, and cash cropping, were gathered in

a systematic way. For calculating the use of time by age

and sex I relied on five visits each day for one year to

houses selected at random by non-replacement sampling, at

times also selected at random in 15 minute blocks from

6:00 A.M. until 7:30 P.M. The activities of each house-

hold member were recorded and later given a classification

number, the first three digits of which are based on the

Outline of Cultural Materials (Murdock et al., 1971), and

the last two digits based on the activities specific to the

Cocamilla culture. This material was collected specifically

so as to be analyzed by computer. This method was suggested

by Johnson (1974). Over the course of a year over 14,000

observations were accumulated.

More specific data on agriculture, hunting, and fish-

ing were collected by the use of three key informants.

Each of these informants was carefully selected as to

reliability and representativity after five months

experience in the community. Each was given a watch and

asked to provide daily reports. The information requested

involved the time they spent traveling to their work, the

time worked at each daily activity, the people present,

the location of the work, the tools used, and the weight

in kilograms of the product they brought home. All fish

were to be weighed before cleaning, and all animals were

to be weighed before gutting. These reports also included

the subsistence activities of other household members and

the inputs of non-household members into the fields or

fishing activities of the person filling out the form.

Communal work sessions attended, for example, would include

the names of all persons present. These reports were

obtained daily for a full calendar year and presented a

picture of a domestic mode of production which included

the entire family, not simply data on the activities of

adult males. The reports were carefully analyzed daily

and recorded in large account books, the activities of

each person working in the system being separated. The

data were set up to be processed by computer.

Nutritional data were obtained by sitting in selected

houses for periods of five to seven days weighing every-

thing that family members ate. This was as much of a

burden on the Cocamilla as it was difficult for us. Since

the families frequently eat from common dishes, we weighed

what went into the pots and took notes on what was indivi-

dually consumed. The leftovers were weighed before being

thrown to the chickens, as were the fishbones. These data

were collected in February when the water was rising

toward flood stage, in March at "high water" during the

rainy season, and in August at the dry season "low water"

time. Nutritional data are the most difficult data to

collect since the close cooperation and patience of all

family members is required and I owe a large debt of

gratitude to the Cocamilla families who assisted in this


Ethnographic material was provided by the technique

of participant-observation, which is another way of saying

that my family and I lived with the Cocamilla, worked with

them, danced with them, marketed with them, shared ritual

coparenthood with them, were occasionally laughed at by

them, and never really understood them as well as we

wanted to. We collected geneologies in order to compre-

hend their social organization with some sort of time

depth. We collected census data and tried to be

systematic about exploring their social world. We took

extensive notes on what we saw and heard, and classified

the notes along standard ethnographic categories for

purposes of analysis.

I cannot say that we came to think as the Cocamilla

think--I have too much respect for the power of one's own

sociocultural system in forming personality and cognitive

categories to believe that we could ever "think" like the

Cocamilla, but we did come to understand some of the

reasons why the Cocamilla behave as they do, and at

times I felt very close to them indeed.


iThe use of the terms cholo and cholada will prove
troublesome to some. I am aware that the terms are used
in many different senses in Peru. The literature on
so-called cholos and the process sometimes called
cholification has become fairly abundant. I do not
intend to argue that my usage of the term corresponds to
its usage in the rest of Peru. The people called cholos
in the tropical forest region provide yet another example
of the varied use of the term and I use it here in a
regional sense to refer to acculturated native Indians
who are integrated into the Peruvian national class
structure. In this sense the term probably has more
than regional significance but I leave it to others to
draw that conclusion.

2This terminology provides the framework for the use
of the term "native" in this work. By native, I refer to
members of ethnic groups which are American Indians. This
usage corresponds to the terminology used by government
of Peru in referring to Indian societies in the tropical
forest regions. I consider members of the cholada to be
natives when they form native communities as defined by
the qualities of being essentially egalitarian communities
in which social and economic relations are kinship-based.
When members of the selvatic cholada do not live in such
communities, they may be considered to have advanced far
along the path to assimilation into white-mestizo culture
and society. To be native, then, is less a matter of
biology than of cultural patterns including residence and
social organization.



The Peruvian system of stratification,
whether in the colonial, republican, or
contemporary periods, has never made
use of a strictly binary opposition
between Indian and non-Indian. But
neither has it been characterized by
a pluralism of juxtaposed groups--that
is, at once isolated from each other yet
all forming part of the same economic
space and subject to the same legal sys-
tem, as, for example, is the case with
the Jews, Italians, and Puerto Ricans of
New York (Bourricaud, 1975:351).

The above quote illustrates a problem. The problem

arises when statements about the nature of Peruvian society

such as this one are applied to the eastern lowlands of

Peru. The fact is that the history and sociology of the

lowland tropical forest region tend to make attempts at

generalizations about Peru problematic. The long co-

existence of many ethnic groups in the lowlands missions,

groups which shared the same "economic space," and yet were

ethnically and socially "isolated from each other" for

nearly two centuries in many cases, serves as an example.

The lowland tropical areas simply do not fit the model

for highland and coastal Peru. The land tenure system

which heavily influenced the relations between Spanish

and Indians in the highlands of all the Andean countries

did not apply. Probably for the above reasons the lowland

tropical forest region has, until very recently, been

excluded from consideration in discussions of social

stratification in Peru, usually on the basis that it is

"sparsely populated" (Larson and Bergman, 1969:4) if it is

mentioned at all. More frequently it is not mentioned.

Actually, with a population of close to 1,500,000

people the montana accounts for over 10% of Peru's total

population, hardly a negligible number. It is only the

fact that this population is spread over the 60% of Peru

which is or was covered by various sorts of tropical

forest that allows the region to be dismissed so easily.

If the same number of people were gathered in one place,

their social composition would doubtless arouse more

interest and attention.

But if this region is not to be analyzed with the

rest of Peru, how is it to be thought of? The history of

post-conquest geo-politics in the Amazon basin has meant

that the native populations of the tropical forest regions

of the Andean countries and those of Brazil have had very

different histories; the Amazon basin in general cannot

be thought of as a social unit in any sense. Peru, for

example, became a refuge area for at least four large groups

of riverine Tupian Indians between 1549 and 1700 (Bollaert,

1861:2-3; Edmunson, 1922:119-127; Vasquez, 1881). Spanish

military control in what is today lowland Peru determined

considerable cultural divergence in the upper Amazon from

the lower Amazon which was controlled soon after 1500 by

the Portuguese. For the Andean countries of South America

the lowland tropical forest was always a far-off hinterland.

Difficulty of access through the rough topography of the

eastern valleys of the Andes, covered with thick cloud

forest growth, tended to protect the Peruvian native forest

and river Indians from total destruction. On the other

hand, Portuguese slavers were more or less free to wreak

havoc on Brazilian natives by traveling up the broad

Amazon River and its tributaries from Para.

To be sure, the destruction of forest and river Indians

during the years from 1885 to 1912, when world demand for

rubber was dependent on the dispersed wild resources of

the Amazon Basin, took place in both the Peruvian and

Brazilian Amazon. Labor-intensive extraction methods

were frequently accompanied by the slavery and starvation

of captive Indian populations (Valcarcel, 1915; Singleton-

Gates, 1959). But by this time Brazil had few easily

accessible native Indians left and it is symptomatic of

the relatively more numerous Peruvian Indian population

that most of the infamous "correrias" (Indian roundups)

of this period took place in Peru. Today the results of

the very different history of the Peruvian Amazon and

that of Brazil is starkly suggested by simply population

figures. Brazil has approximately 120,000 native Indians

in the entire country (Supysaua, 1974) while Peru has

approximately 212,500 (Uriarte, 1977) in the Amazon

region alone.

It must be emphasized that the situation described

above is not due to pre-conquest population differences.

By all accounts the Brazilian Amazon River valley and its

tributaries were more densely populated than any part

of the Peruvian Amazon (Medina, 1834; Acuna, 1918).

Rather it is due to an accident of geo-politics which gave

a large region of the Amazon basin to Peru, a country

which could not efficiently dominate or exploit it. The

Andean mountains were the barrier, and Lima, the capital

of Peru, was on the coast. In between were up to six

million Quechua and Aymara Indian peasants who provided

at once, labor for the new colony, and a management problem

which has occupied and plagued the dominant but minority

white-mestizo society ever since the conquest.

Yet the Peruvian Amazon in its recent social history

is very much a part of the Peruvian nation, even though it

forms a distinct region within national boundaries. How-

ever, when Mariategui took up an analysis of Peruvian

society in 1928 he dismissed the montana completely as


The montana still lacks any significance
sociologically or economically. It may
be said that the montaWa or better put,
the floresta, is a colonial dominion of
the Peruvian state. (Mariategui, 1928:151,
translation mine)

At the time the lowland tropical forest was the sixth

largest lumber exporter in the world. Rubber export had

reached enormous proportions a few years before. The

department of Loreto had had two attempts within twenty

years at revolution. The scandals of the Putumayo River

in which the rubber gatherers enslaved and destroyed as

many as 40,000 native Indians were well-known history

(Singleton-Gates, 1959).

In the early part of the 20th century when people

turned their gaze to the lowlands in eastern Peru, it was

through strangely tinted lenses. Disagreements about

the nature of eastern Peruvian society usually revolved

about racial categories. Wilheim Sievers' well-known

Peruvian and Bolivian geography (Sievers, 1931) divided

Peru, as do most geographers, into three regions, the

sierra, the coast, and the montana. His estimate of

the population of Loreto (then in dispute with Ecuador)

was 150,000 people. Wiesse (1921) had computed earlier

that roughly one-half of these were Indians while 45.8%

were white or mixed. Sievers disagreed and estimated

that 60-70% were mixed "bloods." What all of this meant

in socio-cultural terms is difficult to fathom. Saenz had

this to say:

The population of /-eastern Peru_7 apart
from the small groups of mestizos and
mestizo-blancos is formed by the Jivaros,
the Chunchos, and other tribes. They
speak a great variety of languages from
the primitive inje-inje to others more
developed such as aguaruna-campa. Most
of these people live from hunting and
fishing and lead a nomadic life, although
some have initiated the agricultural
step, for they are dedicated to the
cultivation of small fields on the
riverbanks. (Saenz, 1933:13-14, trans-
lation mine)

That views such as these were extremely unsophisticated

hardly needs to be pointed out. In fact, earlier accounts

were much more sophisticated. In 1791, a hundred and

forty years before, Fr. Manual Sobreviela, the dynamic

Fransciscan guardian of Santa Rosa de Ocopa, had visited

the lower Huallaga River region (Aristio, 1861:29-50).

His map clearly recognized the distinction between the

"faithful" Christian Indians (fieles or almas) and the

non-Christian Indians (infieles or savages) in the area

bounded by the Andes and the Ucayali River. In addition

he noted a stratum of white-mestizo governors and

priests. The distinction between fieles and infieles

was more than a label of religious convenience. It also

implied very different life styles between Indians

gathered in white-mestizo dominated communities who were

given Christian instruction and whose socio-political

organization had quasi-military hierarchies imposed on it,

and autonomous groups of Indians who avoided contact

with the whites and whose socio-political organization

was essentially egalitarian.

Travelers in the 19th century such as Herndon (1853)

were even more sophisticated about the montana society,

perhaps because the society itself was becoming more

complex than a simple Indian/White distinction would

indicate. In Herndon's accounts of the lower Huallaga, he

distinguished between free forest Indians who were

captured and sold as slaves to Tarapoto residents, peas-

ant farmers without ethnic label (apparently the com-

pletely assimilated remains of former tribes such as the

Yurimaguas and Aysuares in the town of Yurimaguas, for

example), Indians in towns under secular and religious

authority, such as the Aguanos in Santa Cruz and the

Cocamillas in Lagunas, and finally the white-mestizo

stratum, including priests, military governors, and


Such elaborations as the above should have provided

the basis for some coherent theory of social structure in

the montana, but by the 20th century it appears that the

White/Indian distinction had only become more polarized;

in the current sociological imagination the Indians had all

become reconverted into savages who spoke primitive

languages such as Saenz's above mentioned "inje-inje."

Part of the problem was that the only social scientists

who wrote about eastern Peru were anthropologists, and

they were totally uninterested in the current sociology

of the tropical forest regions. Farabee's 1907 expedition

to the eastern Peruvian lowlands, for example, makes no

comment on the condition of the Indians from whom he

collected data, beyond the bare mention that most of them

were "working for" rubber patrons (Farabee, 1922, 1, 77, 81,

96, 136, 152, 163). Even the terrible scandals of the

Putumayo (cf. Congress, 1913, 160; Singleton-Gates, 1959;

Valcarcel, 1915) seemed to stimulate no interest among

anthropologists to do anything except to hasten to "salvage"

what cultural data they could before aboriginal cultures were

completely destroyed. Tessman's (1930) treatise on eastern

Peruvian Indians gives but few hints about the real condi-

tions of their existence and lays great stress on "pure"

scientific data. Studies in the acculturation genre were

not done in the region until recently. Anthropologists,

in short, worried a great deal more about the provenience

of "culture traits" than they did about slavery.

The net effect of such writing by ethnographers be-

fore 1945 was to give the impression, quite falsely, of

an upper Amazon Basin full of wild Indian tribes who were

just coming into contact with traders and scientists.

Even Espinosa's monograph on the Tupian Indians (including

the Cocamilla, the classic "fieles," Indians who had been

living in or closely connected with missions for nearly

300 years at the time), mentioned merely that they were

dominated by patrons and described their ideal culture

patterns without further reference to their socio-political

environment (Espinca, 1935).

One may argue that the ethnographers of the time were

merely fulfilling the demands of their profession. It is

not, after all, the job of an icthyologist to be con-

cerned with birds, nor of the medical doctor to analyze

social structure. While this may be true, it cannot be

seriously maintained that contemporary socio-political

environments have so little to do with ethnography, even

if the goal of the ethnographer is only historical recon-

struction of a culture. It is currently acceptable theory

that the environment in which social data is gathered

has a great deal to do with its interpretation and its


Another factor which influenced the re-conversion

of the popular image of tropical forest social structure

from the fairly complex schemes along the great rivers

hinted at by Herndon and other travelers to the simple

White/Indian distinctions of the 20th century may have

been the growing pressures by a mercantile-extractive

economy for "civilized" workers. San Roman (1975) has

noted a change in the 18th century consideration of the

Indian as "pagan" who must be "christianized" to the

Indian as "savage" who must be "civilized." He lays the

burden for explaining the change in emphasis to the

expulsion of the Jesuits from the region in 1767 and the

increased demands on the Indians by a growing and secular

trading, agricultural, and extractive economy during the

19th and 20th centuries.

It is true that the role of workers in extractive

and large-scale agricultural enterprises often demands

that they understand and participate to some degree in the

money economy as they become pawns in a supposedly

"rational" system. It is probably not necessary for them

to understand the system completely to do so. One may

note in this connection that the term for the so-called

"half-breeds" who helped the rubber gatherers round up

Indians as workers was "racionales" which was equivalent

to calling them "civilized" (Singleton-Gates, 1959:230).

But neither the 19th century nor our own era has been

free of missionizing efforts, and the Jesuit era was

hardly free of encomiendas, extraction, trading, and

slaving, sometimes aided and abetted by the priests them-

selves. The expulsion of the Jesuits and their replace-

ment by a series of secular and Franciscan priests in

the late 18th century Mainas missions merely reflected

a relative shift in the power structure of the tropical

forest regions. It tipped the balance in favor of more

direct secular domination of the Indians by the white-

mestizo minority of traders and land-owners. Thus, the

image of a halcyon Jesuit period during which the Indians

were merely protected and instructed in religious matters

must give way to a more realistic assessment of the long-

term economic importance of Indian labor to all montana


By the first half of the 20th century the distinction

between Christian and non-Christian Indians in the tropical

forest regions had become blurred, and Indian society

began again to be considered as one pole of a racial-

cultural dichotomy which had as its counterbalance the

white-mestizo sector. It may be that the image of two

somewhat disconnected sectors, a "modern" or "progressive"

one, and a "backward" or "traditional" one was easier to

hold in the low tropical forest regions because of the

writings of ethnographers mentioned above, but such

works as Roger Casement's report to the British govern-

ment on the rubber camps (Singleton-Gates, 1959), the

congressional report on slavery in Peru (Congress, 1913),

and Harry Hoy's (1946) description of the Peruvian logging

industry were pregnant with implications of the links

between the allegedly separate sectors and complex

stratifications of local society based on access to


The 1940 Peruvian census reflects the lack of

detailed knowledge of the society of the region. The

census reported that at least 321,341 people lived in

Loreto but only 52% of them were actually counted (Censo,

1944:3). The other 48% were only estimated. The social/

racial classifications were White-Mestizo, Indian, Black,

and Yellow. Loreto was considered to be 65% Indian (Censo,

1944:46-47). The black and yellow populations were a

scant .16% of the total population.

The picture is not uniformly dark for this period,

however. Avencio Villarejo, in 1948, published Asi es la

Selva and in 1959 published La Selva y el Hombre which

located Indian groups with an excellent map and classified

them into such categories as "civilized" and "semi-civilized"

with much information on the current status of Indian

societies at the time (Villarejo, 1948, 1949).

In 1972 Sefano Varese published a study for the

International Work Group for Indigenous Affairs (IWGIA)

in which he tried to make sense of the society of the

tropical forest region. His work is the first serious

attempt to understand the contemporary social reality of

the Peruvian tropical forest Indians. A major theoretical

base for his analysis was the premise that the analysis of

the political and economic position of "tribal minorities"

was inseparable from the analysis of the entire surround-

ing society. This was true because the empirical con-

ditions for all aspects of their existence is imposed by

the dominant society. Nor can an analysis of their position

be made without taking into consideration the international

economic and political environment in which the Peruvian

national society exists (Varese, 1972:3-4).

All Peruvian tropical forest Indians are now and have

for some time been directly affected by the international

demand for a series of forest products, and the national

demands for certain cash crops. All tropical forest

Indians, with no exceptions, supply labor or products to

a national or international market, and all, with no

exceptions, have at least some material needs which cannot

be satisfied without participation in those markets. The

factor of national and international economic demands which

leads to concrete social, economic, political and cultural

relations between tropical forest Indians and the dominant

white-mestizo society is the most basic element of the

social reality of the tropical forest region.

Varese's schema, admittedly preliminary, for the

sociology of the lowland tropical forest region involved

setting up a polarity between the native societies and

the "white" society which I have called white-mestizo.

Native societies are defined by Vanese as "ethnolinguis-

tic minorities" rather than "tribes" to avoid giving the

impression that any given group is characterized either

by complete internal cultural homogeneity or by possessing

contiguous territory. The term seems specifically designed

to include the cases of the Campa and the Aguaruna, both

of which peoples are dispersed, and both of which have

rather distinct sub-cultural variations within the same

language group.

Having polarized the Indian/Non-Indian populations on

the basis of ethnic, cultural, and linguistic differences,

and especially differences in access to the means of produc-

tion and power, Vanese presents an analysis of selvatic

society based on occupation and relative power. White

society is divided into three categories. Category one

is an urban based commercial sector and includes large

land owners or administrators. Category two is a rural

mixed group of small farmers and commercial traders and

outfitters. Varese calls these groups intermediaries

from the native viewpoint. Category three includes the

non-commercial network of government officials, military

police, and missionaries whom Varese calls "the communica-

tors of white society." This group is especially charac-

terized being ethnocentric and racist, and it forms the

channel by which national economic and political power is

transmitted to the tropical forest region.

The native population is grouped together into one

category, but with the admission that generalizations are

difficult within this sector, since occupation is variable,

and some incipient local stratification is visible with

native bilingual teachers and native boat-owners. In

terms of access to economic and political power, however,

the native population is always at the bottom of an

"asymmetric pyramid" vis a vis the white-mestizo sectors.

The native populations along the navigable rivers tend to

be "disintegrated and atomized" by the demand for labor on

the part of the extractive economy, while tne highland

forest communities tend to cohere and consolidate around

their territory" (Varese, 1972:19).

The units of Varese's analysis of the native sector

are (1) ethnolinguistic groups divided into, (a) highland

forest and, (b) lowland forest regions, on the basis that

the two regions have been characterized by two distinct

sorts of frontiers, demographic (colonists who occupy land)

in the case of the highland forest groups, and economic

extractivee industries for the most part) in the case of

the lowland forest groups. Ethnolinguistic groups are

further divided into "native communities" which are

operationally defined as follows:

.-The Native Community is_7 the stable
socio-economic unit, bound to a specific
territory, with a type of settlement which
can be either nuclear or dispersed, which
recognized itself as a community, and which
is distinguished from other neighboring
socio-economic communities, native or not

Furthermore, Varese indicates that the two sorts of

frontiers have had differential effects on the highland

forest and lowland forest peoples. The choices of the

highlanders have been to retreat into marginal areas, to

be absorbed as agricultural laborers through a process of

being detribalizedd" with resultant loss of land, or to

retain small and inadequate pockets of land in settled

regions. The choices of the lowland forest and river

groups are more subtle: (1) "De-tribalization. .through

a process of proletarization and ethnic disintegration

or dissolution" (Ibid., 10), but in most cases with the

retention of land for those who remain in the rural areas;

(2) Retreat to remote areas to avoid contact.

Toward a New Model of Lowlands
Native Societies

To arrive at an understanding of the current sociology

of native groups in the lowland forest, it is necessary

to revise some of Varese's concepts. This should be done

first by examining his units of analysis. The term

"ethnolinguistic group" is troublesome. It makes an

assumption which is difficult to support. That assumption

is that language is somehow necessarily related to

ethnicity. As Barth (1969) has shown, ethnicity may be

defined (for some purposes) as the maintenance of boun-

daries between social groups over time by means of cul-

tural differences. While language is usually one of

those cultural differences, there is no a priori way of

determining that it must be. The focus on language leads

to discussions of a native group which may vary widely

internally in social organization, responses to local

pressures, and degree of self-identity as natives as if

it were a homogeneous social unit. In other words, non-

linguistic factors may be much more important than lin-

guistic factors in determining the nature of a given group's

adjustment to white-mestizo society. A focus on linguistic

factors also diverts attention from native groups which

maintain ethnic boundaries (or, as in the case of the

Cocamilla and most other members of the cholada, have

them imposed) in spite of partial or total loss of the

original native language. It is not difficult through

empirical research in the lowland forest to demonstrate

this point. In the highlands it is adequately demonstrated

by the peasant populations of the Cajamarca region whose

cultural differences are ignored because they speak

Spanish rather than Quechua.

Politically, the consequences of the emphasis on

language has led to the denial of the "native" status of

large sectors of autochthonous populations in the major

river valleys who may be losing at least the public use

of their languages and who are now under strong pressure

from a new wave of white-mestizo and highland Indian

agricultural frontiersmen. Populations such as the

Cocamilla are biologically self-sustaining and clear

ethnic barriers separate them from white-mestizo settlers

and wage workers. Nevertheless, the focus on language

has meant that they have no protection for their communal

lands even today, and their fish-protein base is being

constantly diminished by white-mestizo commercial fisher-

men without any attempt by the government to give them the

guarantees which "natives" are supposed to have under the

law. Most of the Cocamilla speak Spanish, while generally

adults under 30 understand but do not speak Cocama.

If linguistic criteria are removed from "ethnolinguistic

groups" we are left with "ethnic groups." Two concepts

are implied which require definition if we are to arrive

at a new model for comprehending the social reality of

the lowland forest. These concepts are "ethnic" (with the

implication in this case of "native ethnic") and the con-

cept of group. Groups are easier to deal with. Native

society in the upper Amazon has its locus in the unit of

the extended family (broadly defined) both today and in

the past. Numbers of extended families may form nucleated

settlements of individual houses, nucleated settlements

of one large house, or more dispersed "neighborhoods."

The nucleated settlements or neighborhoods tend to interact

socially around cultural institutions which usually take

the form of one or more integrative rituals throughout the


The field of regular inter-personal interaction over

the yearly cycle is conceived by the native as constituting

a social unit which may be conceived in many cases simply

as a kinship unit. This unit is distinguished from other

social units of the same order. In the great majority of

cases the unit has a name. It is this unit which largely

corresponds to Varese's "native community" and his defini-

tion can be accepted with the provision that the term

"stable socio-economic unit" does not necessarily imply

geographic stability except within very large areas, nor

does it imply that personnel do not freely flow across

community boundaries without destroying the conceptual,

territorial, or economic unit. The phrase, "bound to a

territory" must not be taken to mean that the territory

does not change as ecological conditions change. In

addition, it should be stressed that Varese's definition

of the "native community" says nothing about its socio-

political composition, a feature of prime importance in

the dealings of the native community with the state as

Chapter IX will show.

Groups of communities form larger aggregates which

may be objectively and subjectively identified by cul-

tural criteria which may or may not include language.

They may or may not be contiguous in space. Such aggre-

gates have little or no importance today in economic

or political terms in the lowland forest. They may, how-

ever, be conceptually significant in the broader social

organization of the group. Native ethnic group may be

used to refer to this aggregate or communities in the

abstract, or they may be given conventional "linguistic"

names such as Cocama, or Campa. Such terms, as pointed

out above, have more historical value than current socio-

logical value. Given the broad ties of social organiza-

tion and ethnic boundaries usually marked by linguistic

features, this ethnic and linguistic group corresponds to

what is frequently called a "tribe" in the literature.

Native communities in the lowland forest are articu-

lated to the national society in a variety of ways. We

may conceive of a continuum ranging from more to less

domination in their dealings with white-mestizo sectors.

The few communities with relatively little contact tend

to be completely dominated by a few patrons whenever they

are in contact, while communities in permanent and con-

tinuous contact along the great rivers choose the middle-

men with whom they deal with relatively greater autonomy,

including working directly as individual community members

with agricultural development banks, or working as wage

laborers in extractive industries.

One might conceive of the above continuum as con-

sisting of the poles "colony" to "class." Colonized

communities have access to upward social mobility only in

the very limited local context. The range of contacts

with the dominant society is narrow; perhaps a few traders,

lumber patrons, or missionaries may be the only members

of the dominant society which will have dealings with

them in a year's time. Their relatively isolated position

means that the prices they pay for material goods from

the dominant society is much more dependent on the quality

of the social relationship they have with the traders than

the actual market prices of the items.

The members of communities fully in the class system

may move up through that social system if they are able

to shed their ethnic identities through geographic

migration and subsequent education. This ability depends

on a knowledge of white-mestizo values, economics, and

politics which is quite outside the range of the colonized

(see Delgato, 1968, for a more general discussion of

social mobility in the Peruvian class system). In terms

of middlemen, the native community member in the class

system has relative autonomy in choosing which middlemen

he will deal with and the prices for his labor and

products are much more dependent on market factors than

are prices in the colony. In the "colony" native com-

munity prices are strongly influenced by the dominant-

subordinate character of the white-mestizo/native relation-

ship. Prices also vary according to the relative lack

of market information available to the native. To sum

up, the communities tending toward incorporation in the

regional class system have a much greater number of social

and economic connections with the rest of society than does

the colony, and those connections fall along a much wider

qualitative range. A major point made by this thesis is

that the communities tending toward incorporation in

the class system are the communities historically iden-

tified as mission Indian or Christian Indian.

Such people are called the cholada by the dominant

white-mestizos of the Peruvian lowlands to distinguish

them from the tribal (colony) Indians. The term cholada

refers to the members of these communities in the context

in which it is normally used in the lowland tropical

forest area of Peru. The socioeconomic position of the

Cocamilla cholo is more similar to that of the rural

agricultural village Indians of the highlands in that the

term in the tropical forest region does not imply the

upward mobility of the persons to which it refers as

is usually the case in the highlands. One semantic

connection with the term as it is used in the highlands

is that the cholo of the tropical forest region stands

between persons called Indian and persons called mestizo

just as the cholo of the highlands does in many instances.

Thus, it refers somewhat imprecisely to obviously

acculturated native Americans. It should be added that

even though, objectively, upward mobility is possible

for only a few of the selvatic cholada, their insertion

into the national class system means that they are con-

scious of the possibilities of such mobility. Tribal

Indians simply are not sufficiently involved in the

national structure to care about individual social

mobility within it.

The term "class" is used above advisedly. The broad

front of contacts which native communities in the regional

stratification system have with white-mestizo sectors,

and their similar lack of access to the means of produc-

tion, may not be enough per se to identify them as

belonging to a social class. However, among the Cocamilla

at least, there is a definite and growing sense of class

identification frequently expressed in terms of native

versus Spanish surnames (apellidos bajos vs. apellidos

altos) which transcends the ethnic group. The identifi-

cation as cholo is closely tied to the identification

as apellido bajo and is, in fact, a class identification

based upon cultural criteria and difference in access to

political and economic power.

To complete the methodological/conceptual inventory

required for a new model of lowland tropical forest

native ethnic groups now requires dealing with the ex-

tremely problematic terms "native" and "ethnic." As many

studies in the Peruvian highlands show, distinguishing

the "Indians" from the non-Indians is not as simple as

it might seem (Adams, 1953; Arguedas, 1952; Borricaud,

1954, 1975; Hammel, 1961; Mangin, 1955, 1965; Metraux,

1959; Mishkin, 1946; Nunez del Prado, 1951; Tschopik,

1952). This is true in part because Indian identity is

socially stigmatized all over the Andean countries.

It is popularly thought that in the tropical forest

areas the problem is somehow simplified; the "Native

Community Law" (20653 ),for example, does not even approach

the problem of defining who the natives are, and Varese's

article which is discussed above conspicuously avoids

it. It seems to be assumed that the natives will step

forward and identify themselves, or that there will be

criteria, clear to all, on which to base an application

of the law. The laws and their application clearly imply

that if the natives do not so identify themselves, and

if they are in the economic position of being subsistence

mixed farmers who sell small surpluses to the regional

markets, then they will be treated like the Cocamilla as

campesinos (peasants) along with white-mestizo settlers

in a similar structural position. They are thus classed

campesinos regardless of other factors in their mode of

production such as discrimination in access to credit,

organization of labor in communal forms, and lack of

familiarity with white-mestizo bureaucracy which put

them at a disadvantage with regard to the white-mestizo


The Problem of Ethnicity

When Metraux (1959:227) rejected the list of ob-

jective cultural traits which supposedly defined the

Indian in the Peruvian highlands (i.e., coca chewing,

birthplace, craft characteristics, etc.) as being too

imprecise, he proposed the criteria of ascription and

self-ascription in order to settle the issue. He pre-

ceded Barth by ten years in this formulation (Barth,

1969). Barth, however, took the issue farther, and

probed the logic behind ascription and self-ascription.

He concluded that ethnicity depended on the maintenance

of boundaries between social units over time by means of

cultural differences. Following Barth, however, one

cannot predict what "cultural differences" will provide

such boundaries (i.e., ritual, food habits, dress and

ornaments, etc.). Thus, Barth's definition is not a

scientific definition, if one defines science as the art

of generating statements which have predictive power.

Barth's "definition" is, in fact, merely an empirical

description (Cf. Gomes,1977:38-40 for an interesting

discussion on this point). Another problem is that if

we follow Barth and define ethnic groups as socio-cultural

systems enclosed by cultural boundaries, a certain amount

of tautological reasoning automatically follows. State-

ments such as "such and such ethnic group disappeared

because it failed to maintain its boundaries" become

common. If the question is changed to a query about why

a given ethnic group failed to maintain its boundaries,

the answer usually becomes, "because it was detribalizedd'

or weakened as an ethnic group." If an ethnic group is only

defined by its boundaries, then the two terms are synonymous.

In any case, such a definition of an ethnic group is of

little practical help in determining if a given group is or

is not an ethnic group.

The fact is that most studies of ethnicity and ethnic

groups begin with the assumption of the existence of such

groups and proceed with an examination of their boundary-

maintaining mechanisms. But what calls such groups into

being and what is the cause of their persistence, or alterna-

tively their demise, are two separate questions. The cause

of their persistence and the means of their persistence are

also distinct questions.

In the usage applied here it is necessary to separate

the term "ethnic group" from a usage common to the social

sciences. Wagley and Harris (1958:244-253) point out that

the term has become synonymous in some sectors with

"minority group." As such, ethnic groups can only be

considered to have existed since the rise of state

societies which included them, thereby making them

minorities. But this usage misconstrues the nature of

ethnicity. In the sense used here, ethnicity would be

the sense of belonging to a cultural group regardless

of whether it is a "tribe" or an "ethnic minority."

Ethnic group, then, refers to the group of primary cultural

identification regardless of the group's objective status

with regard to the state society. If an ethnic group is

included in a state society, as all modern ethnic groups

are, it becomes an ethnic minority.

In the above sense, ethnic groups appeared on the scene

as social organizations long before the rise of the state.

Generations of anthropologists have indicated that in a

pre-state condition, ethnic groups might be defined as

territorial groups organized by theories of consanguinity

(or simply geneology) and affinity and residence rules.

Boundary mechanisms vary, but language differences, or at

least certain linguistic features, are common. Such terri-

torial units must be distinguished from most groups claiming

"ethnicity" in the modern world, and the term is used in an

extremely variable way today. The features common to both

are that the theme of material interest runs as a motive

through the course of history, and all ethnic groups like

all social groups must solve problems of continuity, com-

munication, authority, and ideology, as well as boundary

problems in order to survive.

There is a further problem with Metraux's and Barth's

formulations of the nature of ethnicity which is a purely

practical problem in operationalizing a definition which

can be used in the lowland tropical forest region of Peru.

Eidham (1969) has shown that in a situation in which ethnic

identity is stigmatized, the aspect of self-ascription as

a member of an ethnic group tends to be acted out only

in private situations. The ethnic boundaries are operating,

but they are not easily visible in public life where the

two sectors interact. -In the case of eastern Peru this is

precisely the case because of the situation of the Cocamilla

and other native ethnic groups as conquered and dominated

peoples. This makes it highly unlikely that many Cocamilla

or any other Indians in their position are going to step

forward and identify themselves as such unless they see some

very concrete economic or social advantage in doing so.

Since the Cocamilla and others in their position are rarely

offered information as to possible economic advantages of

the public assumption of Cocamilla identity, they do not

proclaim publically their ethnic membership.

All this leads to a dilemma: What, in fact, is the

minimal definition of an ethnic group which can be

operationalized and has objective significance? The answer

is clearly that such a definition is as impossible as

defining a concept of "word" which will be valid in all

languages. Pursuing the linguistic analogy, in the same

way that all non-gestural words may be assumed to have a

carrier, the voice, all ethnic groups have an essence; like

all distinct societies they may be assumed to have provided

for biological and social continuity. Putting the methodolo-

gical cart before the horse, this means that such groups must

have a theory of consanguinity (geneology) and affinity

which is expressed in behavior, which in turn implies rules

of inclusion and exclusion, and the authority to enforce

them (Wagley and Harris, 1958:1-14, 237-295).

Fortunately, just as the concept of "word" may be

defined structurally for a given language, a concept such

as native ethnic group may be formulated in terms of

regional social behavior. In the upper Amazon region of

Peru, a native ethnic group might be defined as follows:

A native ethnic group is a cultural group,
the members of which are inhibited from
marrying members of the larger society of
Peruvians. It derives its composition from
the resultant endogamy and from the appli-
cation of its own exogamic principles as
distinct from Catholic canonic law to persons
outside the primary kin group.

The above definition depends for its validity on two

social facts. The first is that persons with "native"

surnames infrequently marry persons with Spanish surnames

in the region. The second is that all native ethnic groups

have theories of consanguinity which are different from

white-mestizo sectors. The Cocamilla, for example, pro-

hibit marriage to any person bearing one's own paternal

surname under the theory that they are one "blood." Thus,

exogamic principles are extended to persons outside the

group which a white-mestizo would consider the "primary"

kin group (i.e., Ist, 2nd, and 3rd cousins).

Since theories of consanguinity loom as important in

native social organization, it may be predicted that life

in native communities will be organized along kinship lines

whether or not the community is "tribal" or detribalizedd"

and part of the class system. It may also be predicted that

because of the relations between social organization and

the rituals and symbols which make social life coherent,

any native ethnic group will have at least one integrative

ritual at regular intervals which will tend to be exclusive

and which, while it may be formally similar to that of a

neighboring ethnic group, will have a content which expresses

the individuality of the group practicing it. It is important

to stress again that the detribalizationn" of native com-

munities is not synonymous with their disappearance.

The question of what a native ethnic group is and how

one may determine whether a given ethnic group is or is not a

native ethnic group is now largely solved. The remaining

question, as far as this dissertation is concerned, is one

of determining what is the essential relation between the

native community and the state. The native extended family

mode of production and the importance of kinship units in

native social organization provide the basis for a theory

of this relation. These two features of native life mean

that native communities will tend to be politically acephalous

and organized around related groups of families rather than

central political figures with overall authority as in the

case of white-mestizo culture. Given this organizational

base, a theory of the relation between the native community

and the state would necessarily be one of their mutual and

natural antagonism, for the state is a bureaucratic organi-

zation which depends on the adherence to formally agreed-

upon rules. The state cannot tolerate the acephalous com-

munity. Thus, we may predict that the native community will

only be drawn into the orbit of the state by deception and/or

violence, and it will be maintained in the state system only

by duress. Whenever the state tries to impose itself upon

the native community directly on a day to day basis, one

may predict that conflicts and clashes will occur which can

be traced to the opposed nature of the state and the native

community social and economic organization. When these

clashes are overcome,it will only be at the expense of

converting the native community into a "non-native" com-

munity, that is by changing native cultural behavior into

acceptable white-mestizo behavior.

The chapteswhich follow are arranged in two parts.

After a brief description of their ecological setting and

some basic information about their life-style pre-

historically, Cocamilla history is explored in order to

examine the changes they have undergone as a result of their

contacts with the Western world through time, and, I hasten

to add, as a result of the internal logic of their own social

system as it is supported by their mode of production.

Special emphasis will be given to their position in the

developing social stratification of the region. Two related

problems will also be explored in the course of this work.

The first problem was outlined in the first part of this

chapter. It is a problem of conceptualizing the lowland

tropical forest region as something other than the two-class

or two-race region it is commonly considered to be. The

Cocamilla and others like them form a lower class which is

neither Indian in its own conception, or mestizo in either

their own or in the mestizo view. It is a class

ideologically based on their historical separation from other

Indians as Christian Indians, and created by the praxis

which that status implies. Though called cholos their

position in the regional economic structure and their his-

torical trajectory has little to do with the term cholo as

it is used in the highlands or on the coast (cf. Fuenzalida,

1970). The second problem concerns the units of sociological

analysis. There are the household, the native community,

and the native ethnic group.

The historical depth seems essential to sound sociologi-

cal analysis, but it would be included in any case, if for

no other reason then for the very practical reason that the

Cocamilla are now in a rapidly changing and developing world,

and, as Levy (1972) has observed, the most cruel and stupid

mistakes in modernization result from ignorance about the

history of the people involved.

The second section will continue the theme of the

relations between the Cocamilla and the state society in

which they are included, this time synchronically. The

interface of the relationship will be examined at the level

of the economic system, the level of social organization,

and the level of ideology and identity. A final chapter


will draw conclusions as to the reasons for their survival

thus far from the perspective of cultural materialism,

and attempt to project a trajectory for them based on

possible actions by the Peruvian government.



The eastern lowlands of Peru, mostly covered with

trees and shrubs except where human use has destroyed it,

presents two orders of problem to the Peruvian nation.

The region presents formidable problems in settlement,

strictly from a tactical and ecological standpoint. An even

more serious problem is that few if any people have ever

understood the region thoroughly in both its social and

ecological aspects. To make a sweeping statement even more

sweeping, I will include most of the people responsible for

planning and implementation of planning in the tropical

lowlands in the ranks of the ignorant. To calm outraged

friends and colleagues, the writer must be included among

their ranks as well, for the complexity of the problem is

overwhelming, and the division of labor among academic

disciplines mitigates against global views.

To further the point about how little the region is

known in general, even after 330 years of occupation and

the recent intense search for oil in the Peruvian lowlands,

there are no topographical maps available for any of the

lowland forest. The impression is gotten from the literature

which points out how little slope there is in the 4,167

kilometers from the Atlantic Ocean (Elevation O meters) to

Iquitos (Elevation 300 meters) along the Amazon River, that

the entire Amazon basin is as flat as a table. This im-

pression pervades even the thinking of normally careful

planners. The Swiss project of cattle-raising at Genaro

Herrera on the Ucayali River, for example, had serious

problems with cattle when the planners cleared the forest for

pasture. Rolling hills, holes, and deep gullies severely

prejudiced their effort. The terrain throughout the lowland

forest region is very accidental within limits of fairly

low relief, and this has many consequences for native agri-

cultural practices. The old people among the Cocamilla

used to say that when the earth changed from water to land,

the waves got frozen into place.

Different people split up the totality of the forested

lands east of the Andes in different ways. The usual ceja/

selva distinction based on altitude above sea level was

echoed in Pulgar Vidal's Geografia del Peru (1968) when he

divided eastern Peru into Selva Alta and Selva Baja with

the former occupying the altitude of 400 to 1,000 meters,

and the latter occupying the elevations under 400 meters.

Ecologists today are able to improve on the scheme with the

concept of "life zones" (Holdridge, 1967) based on three

factors: average annual bio-temperatures (the annual average

of temperatures between 0C. and 300C.), average annual

rainfall in millimeters, and the potential evapotranspira-

tion ratio(mean annual temp. in C. X 58.93), which expresses
( mean annual rainfall in mm. )

the relation between rainfall and evaporation. By Hol-

dridge's system much of the lowland forest in the Cocamilla

area would be classified as "premontane tropical moist

forest." Table 3.1 shows the relevant climatological data

from 1976 (Bandy and Benites, 1977), and for comparison,

the average rainfall from a 21 year sample. These data were

taken in Yurimaguas, Loreto Department, and are very similar

to figures for Iquitos, indicating that the totality of the

area immediately along the lower Huallaga, lower Maranon,

and lower Ucayali rivers may have a similar climate.

One should be aware, however, that ecologists currently

working on the problem now believe that the entire lowland

forest under 400 meters is probably a mozaic of premontane

tropical moist, tropical wet, and tropical dry forest,

depending on purely local convection cells which seem to

remain persistently in some regions (Ewell, personal

mm mm mdm l

n r n o ri

co co co co co 00r r r o co m
o c m r -

.H a) >
-4 >1 Z r-
a M ::
<4 : n '

0 u o0
0< n 0 a

c \

d a





0- 0
I- ',

Hn '

n LO 0 in

UwLL UI Ile~j'eJ lNluouLj






^ e

G< r-

communication). This variability may go a long way toward

explaining some of the cultural diversity noted in the region,

even among non-riverine Indians.

The life zone classification does not tell the whole

story, however. Various micro-environments within a life

zone have profound consequences for human occupation. The

conventional distinction between riverine Indians and

inland Indians (or forest Indians) reflects great differ-

ences in adaptation to the two major poles of subsistence

resources in the lowland tropics.

Proteins and Populations

Lowie observed that the major part of the human popula-

tions of the lowland forests were more adapted to the rivers

than to the areas between the rivers. While he recognized

that some people were canoeless, he considered the possession

of "effective river craft" a diagnostic feature of the cul-

tures of the lowland tropical forest (1963:1).

Meggers (1970) pointed out certain similarities in

density of settlement and ease of subsistence among

populations living in the limits defined by the margins

of the river valleys. She contrasted these populations

with the more scattered and migratory populations in the

areas between the rivers which she supposed to be an adapta-

tion to poor soils and low protein availability, to name

the two most important variables.

Lathrap (1970) agreed with the basic division of

cultural types (while disagreeing with Meggar's assessment

on general population movements) and took the point far

enough to insist that Panoan Indians who lived inland from

the Ucayali River were "de-cultured" refugees from river

populations, driven inland at some previous time by fierce

competition for the fertile lands of the river valleys.

They had been unable, he said, to sustain the relatively

rich symbolic life of the riverine Panoans, primarily

because the poor soils and protein resources of the inland

areas did not permit large and enduring settlements.

Gross (1975) took both Meggars and Lathrap to task

for emphasizing relative soil quality rather than relative

richness in animal proteins between river floodplain and

inland areas. He brought a great deal of data to bear

on the subject and emphasized the very low ratio of animal

mass to total biomass in the inland areas. The distribution

of proteins, he maintained, was sufficient in and of itself

to explain population size and stability. His Peruvian

data were taken from Bergman (1974) who studied the

subsistence patterns of a group of Ucayali Shipibo as an

example of riverine subsistence. These Shipibo had no

access to high ground and their entire agricultural cycle

was based on the rise and fall of the river, for in the

rainy season the river flooded all their lands.

While conceding the essential validity of the riverine/

inland distinction, it seems wise to hedge a bit. Recent

studies have shown rather large settlements of Matses

(Mayoruna) between the Ucayali and Yavari rivers, hardly

the scattered fugitives predicted by Lathrap's model

(Romanoff, personal communication). Campos (1977) has

studied Shipibo on the Aguatia River far from the major

river valleys who seem to exploit both river and inland

areas about equally. Previous studies of the animal mass/

total biomass have not included stream fish, and no studies

of this kind have been conducted in eastern Peru. In all

of the area treated by this study, small and large streams

(quebradas) wind their way inland for long distances, often

carrying surprising quantities of fish species unique to

the stream environment, and easily exploited with fish

poisons. Even the floodplain Indians such as the Cocamilla

did a considerable amount of hunting and gathering in the

inland forests during flood times. Thus, there seem to

be groups who exploit both riverine and inland environments

and no simple separation of the two kinds of groups is likely

to be accurate.

Land and Water Features of
the River Valley

The area of interest to this study is mainly the lower

Huallaga River valley between Yurimaguas and the confluence

of the Huallaga and the Maranon. It will have reference,

however, to the Maranon as far upstream as Borja and as

far downstream as the mouth of the Napo River, the Pastaza

River as far upstream as Lake Rimachi, and the Ucayali from

its mouth to the area around the current site of Tierra

Blanca, some 12 days upstream by canoe. For this reason,

these areas are included in Map 1, but the following

description is of the lower Huallaga region, the most signi-

ficant area for the Cocamilla. The river valley dynamics,

of course, have wide application.

Thirty miles downstream from Yurimaguas, near the town

of Santa Maria, the Huallaga River leaves the confining low

hills through which it has flowed rapidly, being unable to

easily widen its channel, and slows down abruptly, dumping

an enormous load of sand just below Santa Maria. From that

point on, it becomes a wide threaded river, moving leisurely

at about 4-5 kilometers per hour, and meandering back and

forth across its valley boundaries. Between Santa Maria

and Yurimaguas, entrenched meanders indicate that the uplift

which provides the inland high ground between the Huallaga

River and the current location of Xeberos is more recent

than some of the Andean uplift, and in fact earthquakes are

frequently felt all over the lower Huallaga region.

Below Santa Maria the width of the river floodplain

averages close to 5-8 kilometers. As the river migrates

back and forth over this valley,it has a predictable dynamic

cycle. The cycle is born in the fact that while the water

tends to run in the shortest straight line from A to B,

following gravitational laws, the unequal pulls of river

channel and sand bars tend to make the river currents

shuttle back and forth, forming curves (vueltas). Distance

between riverbank communities is often expressed in terms

of the number of vueltas between them.

The curves lengthen into long and elaborate U forms

as the faster-moving currents on the outside of the curve

wear away the bank and the slower currents deposit sand

on the inside of the curves in longer and longer beaches.

These beaches are the major resource area of the river

channel itself, or were in pre-conquest times, for fish

could be speared along their margins, and turtles, turtle

eggs, and iguana eggs could be gathered in July and August

on the beaches themselves. The main part of the river

channel was effectively sterile for native use, for they

lacked the means to exploit the resource.

Eventually gravity wins the struggle of the vuelta

and the river shortens its channel by cutting off the U and

isolating it into an oxbow lake (cocha or tipishca) connected

to the main river by deep canals (canos) which serve as

channels to fill and drain the lakes as the river goes up

and down from day to day and seasonally in response to local

and regional rains. In high water time, from February until

May (and a less drastic flood in November and early December)

other channels to the lakes open up (sacaritas) and in all

seasons the lake and river system forms one connected water


The river is never stable. It is always either rising

or falling and the lakes connected to it rise and fall in

rhythm. Tremendous surges of water go in and out through

the canos and they may change the direction of their

current quite abruptly as the lakes change from fill to

drain. The rise and fall of the lakes and rivers is an

important feature nargizing the floodplain ecosystem for

it allows rapid incorporation of plant material cycling

from an enormous area of forest into the lakes and into the

total water system. The drastic rises and falls (as much

as two meters in 24 hours) of the lakes create an unstable

ecosystem for the fish populations. The instability favors

the propagation of "R" selected species, selected for the

ability to grow and reproduce rapidly, and this feature of

the ecosystem dynamics helps explain why it is that the lakes

and rivers of the lowland tropical forest are such an extra-

ordinary resource for native populations. The lakes,

especially, are rich in fish and reptile populations.

Figure 3.1 shows the rise and fall of one such lake on a

weekly basis for 1977.

Eventually, the oxbow lake silts in and its canos close

except in high water. The Cocamilla vividly express what

happens then when they say the lake "dies." Aquatic plants

take over and the lake is gradually made uninhabitable for

fish, although some "dead" lakes have large populations of

paiche (arapaima gigas), the fish known as pirarucu in Brazil

and the largest freshwater fish in the world, for a long

time before they are extinct. Eventually the lake becomes

a muddy bog colonized by a few tree species, waiting for

the next migration of the river channel to destroy it.

3 66
C3 <


wI 0

o >


C" <

7 .
I0 0
o -


o -
7A1 1 1
f 0 I^- '-n cn

-13A1 V3S 3A08GV S8313H 3iVINIX08ddV

The total length of time from birth of a vuelta to

the death of the resultant lake may vary widely. One

vuelta observed over the 19 month field portion of this

study lengthened the arms of its U over 1,000 meters after

being relatively stable for 50 years, putting a community

on the cutting bank into full flight inland. Once a lake

is formed it seems unlikely, from the historical evidence

on riverine communities gathered in this study, that the

lake can exist for more than 200 years.

One other feature of the lakes should be noted. During

the flood season from February until May many fish move into

the lakes and into the surrounding forest which is flooded.

In the forests they can feed more directly on plant

material including the direct utilization of fruits, but

they become very hard to catch by native techniques. This

creates a protein crisis at least once a year. A related

feature of the flood is that when the lake goes back into

its normal boundaries the fish populations become more and

more dense as the size of the water body is reduced until,

at a critical point sometime in June or July, they begin

to migrate out of the lake in phenomenal numbers. The

local term for such migrations is the "mijano" and at

such time fishing in the canos, through which the fish

must pass, is very productive. Natives today sometimes try

to block the channel temporarily and net as many fish as

possible. Leaving the lakes, the fish migrate to the

rivers and move upstream, forming pockets of local rich-

ness for the communities which happen to be where the fish

are at any given time.

The land features of the floodplain take their forms

from the dynamics of the river. Every time the river spills

its banks it deposits the greater part of its sediment load

within 30 meters of the channel, creating rather uneven

natural levees which tend to be higher in elevation than

the lands immediately in back of them. The zone of the

levee (the banda) is the zone normally used today for river-

bank agriculture. The crops most frequently grown are

plantains, corn, sweet manioc, taro, peanuts, sweet

potatoes, and recently jute (malva urena).

Old levees along old river channels and old oxbow lakes

may be high enough so that only exceptionally high floods

in the rainy season can cover them. These areas are called

restingas and are an.important land resource for the

natives, not only for agriculture, but also for hunting.

When the flood waters rise in the floodplain, terrestrial

animals tend to concentrate on the restingas and hunting

them becomes easier. Many Cocamilla today only hunt such

animals when the restingas become islands, and a special

verb, restinguiar, describes the activity.

The lower lands of the floodplain are of two kinds.

Lands lying so low that even moderate rises in the water

level will flood them, and which tend to act as rain catch-

ment basins, are called tahuampas and are not used for any

purpose except for gathering and occasional hunting of the

tapir and the capybara which inhabit them. Lands which

are not high restingas but which have adequate drainage

are called bajiales (also used as a generic term for any

lowlying area), and are farmed and hunted. These lands are

of exceptional significance today for cash-cropping and

they flood annually, but with a much lower rate of sedi-

mentation than the banda. Today they are mainly used for

jute, corn, watermelon, and various herbs.

The last significant land forms of the river valley

are the beaches along the river which are exposed from late

May until November. These are of two kinds. Sandy beaches,

over which water flows too rapidly to deposit silt and

which are not particularly good for any crop, are called

playas. As mentioned earlier, these beaches are signifi-

cant fishing and gathering resources. Beaches over which

water runs slowly have an annual layer of silt deposited

on them and are used to plant beans, peanuts, and (recently)

rice. These beaches are called barreales. Appropriate

crops can be grown on them at a comparatively light cost

in labor energy.

Land and Water Features of
the Upland Plain

The drainage of the uplands is far from complete. All

over the upland areas are rain catchment basins which tend

to be swampy during most of the year. These are called

bajiales, the same as the annually flooded farmlands of

the river basin. The ground between the bajiales is called

altura, a term which refers to any land which never floods

from the rivers. Nevertheless, some of the altura does

flood, not from the rivers but from local rainfall rushing

down the larger quebradas which cut their way toward the

river from the divide between the Huallaga and Ucayali

rivers. These quebradas can be quite good sources of fish,

for many of them are born not in the leached sediments of

the ancient ocean and lake beds of the Amazon basin, but

rather in sedimentary bedrock which is comparatively rich

in nutrients. The native names for these quebradas fre-

quently express a distinction between "white water"

(yurac yacu) and "black water" (yana yacu), a distinction

which some ecologists think is significant in terms of the

environment the two types of streams provide for fish


Many quebradas have their own small floodplains which,

in cases of heavy rainfall flood for a period ranging from

two hours to two days, creating a permanent micro-environment

with more friable soils than the usual upland soils. These

areas are good for hunting and for farming. The crops

planted on them include most of the repertory of the banda

and the bajiales (i.e., plantains, corn, manioc, sweet

potatoes, taro, herbs,and sometimes (recently) rice and

jute). These lands are recognized in the native classifi-

cation system for soils as yarinales, meaning that the

yarina palm (phytelephas macrocarpa) is a dominant under-

story plant. The palm has cultural significance as the

source for the gable covering on native houses. A worm

which grows in the palm is consumed, and the fruit of the

palm became valuable for a time as button material before

the widespread use of plastic buttons.

The sandier upland soils are recognized as irapayales.

The irapaya, an unidentified plant with a palmate arrange-

ment of long laurel shaped leaves, is used for roofing

material and the soils on which it grows are recognized

as being good for cultivating manioc and a root used for

fish poison barbascoo), but not for plantains or corn.

Land along quebradas which do not flood, and which tend

to have heavier clay soils than those that do flood, are

also valued, possibly because of the drainage offered by

the quebrada. These soils are used for plantains and

(recently) occasionally rice.

On the west side of the Huallaga River, the floodplain

extends much farther inland, and relatively little of the

land forms are presently used for farming. The settle-

ments tend to be either on the river bank or on the shores

of oxbow lakes.

The impression is widespread that the adaptation of

the aboriginal populations in all of the Amazon valley was

mainly to the rivers. The generic term in eastern Peru,

for example, for rural residents is ribere'nos. A close

reading of the Cocamilla adaptation today, however, reveals

that the floodplain lakes are much more important than the

rivers as fish protein resources. This, as pointed out

before, is because some of the lakes are closer than the

rivers to the sources of primary production emanating from

the forests, and they are, therefore, much richer in fish

populations. They are also easier to exploit. The

lakeshores and valley margins are closer to alturas, and

the present Cocamilla farming system involves exploiting

both alturas and bajiales (used here in the generic sense

to include restingas; in other words, all floodplain land

which is farmable except for the beaches). Even popula-

tions along the west bank of the river who do not have

easy access to inland alturas, normally have one or two

fields in the east bank alturas, sometimes a day's travel

or more from their communities. These fields are a sort

of insurance against extremely high or long-lasting floods,

guaranteeing that some food will be available even in

cases of emergency. In all probability the Cocamilla in

pre-conquest times did not use the alturas east of the

river because of conflicts with the Aguano-Chamicuro people.

The Larger Region

The valley of the lower Huallaga River is only one of

the valleys which concern the Cocamilla historically,

although by far the most important. A small part of it

was their homeland at the time of contact, and they have

subsequently occupied most of the lower 100 kilometers of

it. They have recently occupied lands along the lower

Maranon Valley, and have begun to occupy some of the rivers

and quebradas north of the Maranon River. The displaced

and the dissatisfied swell the migrant barrios of Yurimaguas,

Lagunas, and Iquitos. Their familiarity with the region is

long-standing as will be seen. At least from the mid-19th

century on and possibly before, they used lake Rimachi,

near the Pastaza River as a major resource for Manatee and

Paiche fish, traveling up to 11 days in canoes to get

there, and staying until the Muratos drove them out. Before

the conquest they and the Cocama in confederation were the

scourge of the entire lower Huallaga River and the Maranon

from Borja to the Ucayali River. According to some accounts,

they prevented other Indians from occupying the shores of

the river or the accessible floodplain. After they were

reduced to a few starving and sick Indians by the 1680

epidemics they became river explorers and travelers for

whites. The theatre in which the history of the next

chapter takes place is this larger area of operations with

special emphasis on mission sites. They will be traced

from contact to the present and at all times will be seen

in opposition to the dominant Spanish and later white-

mestizos. Their history is a history of action and

reaction, not the blending of peoples. It reflects the


fixed intent of peoples to remain separate under conditions

of direct domination. First, however, a few pages will

be devoted to reconstructing what is known about their

culture before contact, and their placement with regard to

other ethnic groups. Map 3.1 shows the region pertinent

to the Cocamilla and Cocama.


ul o~
S- 4
*^~~ru *X ^? U 3


,I -- -_.,
s...i iJ" ^ i-.
y ,- ?

A $~-. 4.. <, : ..

^ ) ,.^ ......
-J-- _
^J ; ru'
i \\



The prehistory of the Cocama and Cocamilla in the

natural realm described in the previous chapter is some-

what obscure. If Lathrap (1970:145) is correct in identify-

ing the central Ucayali River Caimito complex as belonging

to the ancestors of the Cocama, the probability is that

the Cocama fissioned from the Omagua not too long before

the 14th century A.D., and pushed on upstream and into the

Ucayali. They must have established themselves as far south

as the Tamaya River at that time. Lathrap's model of Tupian

expansion implies that they had been pushing the Panoans

upstream on the Ucayali River and that the process is

ongoing today. It is difficult to reconcile this view with

the historical data for the Cocama. At the time they were

discovered in 1557 by the expedition of Juan de Salinas

Loyola (Jimenez de Espada, 1897:LXXIII), they were isolated

between a group called the "Benorinas" and the Panoans to

the South. It seems extremely improbable, using Salinas

Loyola's information on distance (and factoring it for a

comparison between his estimates in leagues and actual

kilometers on modern maps) that the Cocama were distributed

at that time any farther south than Tierra Blanca, and prob-

ably not even that far. This would indicate that the

Panoan Indians had been steadily pushing back the Cocama

for 200 years, and not the reverse. By the 17th century,

the Cocama were indeed at war with the Panoans to the

South, and the Cocama were forced to have their largest

concentration of population on the southern frontier.

The Cocama, by 1619, had separated into two groups,

the Cocama proper and the Cocamilla. The direction of

fissioning was not upstream on the Ucayali, but rather

across the isthmus which separates the Ucayali and Huallaga

drainages in their lower reaches. On the Lower Huallaga

River the Cocamilla (frequently called the Cocama of the

Huallaga or simply Guallagas in the old literature) had

formed at least one village somewhere in the vicinity of

the mouth of the Shishinahua River, probably on the western

side of the river valley of the Huallaga (Figueroa, 1904:

78-84; Jimenez de Espada, 1897:CLX). The Cocamilla were

at peace with the Xeberos Indians, an inland group between

the Huallaga and Maranon rivers, and had assimilated to