• TABLE OF CONTENTS
HIDE
 Front Cover
 Title Page
 Table of Contents
 Abstract
 List of tables and figures
 Acknowledgement
 Atlantic Honduras
 People, land, and occupations
 Livelihood strategies
 Comparisons between livelihood...
 Conclusions
 Appendix A. TWINSPAN method
 References
 Back Cover
 Copyright






Group Title: Group paper - CIMMYT Natural Resources Group ; 95-01
Title: Land and livelihoods
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00080058/00001
 Material Information
Title: Land and livelihoods patterns of rural development in Atlantic Honduras
Series Title: CIMMYT Natural Resources Group paper
Physical Description: vi, 28 p. : ; 28 cm.
Language: English
Creator: Buckles, Daniel, 1955-
Saín, Gustavo
Publisher: CIMMYT
Place of Publication: Mexico D.F
Publication Date: 1995
 Subjects
Subject: Land tenure -- Honduras   ( lcsh )
Rural development -- Honduras   ( lcsh )
Land use -- Honduras   ( lcsh )
Labor supply -- Honduras   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Spatial Coverage: Honduras
 Notes
Bibliography: Includes bibliographical references (p. 27-28).
Statement of Responsibility: Daniel Buckles and Gustavo Sain.
 Record Information
Bibliographic ID: UF00080058
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 35769373
issn - 0258-8587 ;

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
    Abstract
        Page iv
    List of tables and figures
        Page v
    Acknowledgement
        Page vi
    Atlantic Honduras
        Page 1
        Page 2
        Page 3
        Page 4
    People, land, and occupations
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
    Livelihood strategies
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
    Comparisons between livelihood strategies
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
    Conclusions
        Page 22
        Page 23
        Page 24
    Appendix A. TWINSPAN method
        Page 25
        Page 26
    References
        Page 27
        Page 28
    Back Cover
        Back Cover
    Copyright
        Copyright
Full Text
,-- 0706


CIMMYT


Land and Livelihoods:
Patterns of Rural Development
in Atlantic Honduras
Daniel Buckles and
Gustavo Sain














Natural Resources Group
Paper 95-01











CIMMYT
Natural Resources Group
Paper 95-01


Land and Livelihoods:

Patterns of Rural Development

in Atlantic Honduras

Daniel Buckles and
Gustavo Sain*


Daniel Buckles is an anthropologist and was with the International Maize and Wheat Improvement Center
(CIMMYT), Mexico, when the work described in this paper was done. He currently works at the International
Development Research Centre (IDRC), Ottawa, Canada. Gustavo Sain is a CIMMYT Regional Economist for
Central America and the Caribbean. The views and interpretations herein are those of the authors and should
not be attributed to CIMMYT.







































CIMMYT is an internationally funded, nonprofit scientific research and training organization.
Headquartered in Mexico, the Center is engaged in a research program for maize, wheat, and triticale,
with emphasis on improving the productivity of agricultural resources in developing countries. It is one
of several nonprofit international agricultural research and training centers supported by the
Consultative Group on International Agricultural Research (CGIAR), which is sponsored by the Food and
Agriculture Organization (FAO) of the United Nations, the International Bank for Reconstruction and
Development (World Bank), and the United Nations Development Programme (UNDP). The CGIAR
consists of some 40 donor countries, international and regional organizations, and private foundations.

CIMMYT receives core support through the CGIAR from a number of sources, including the international
aid agencies of Australia, Austria, Belgium, Brazil, Canada, China, Denmark, Finland, France, India,
Germany, Italy, Japan, Mexico, the Netherlands, Norway, the Philippines, Spain, Switzerland, the United
Kingdom, and the USA, and from the European Union, Ford Foundation, Inter-American Development
Bank, OPEC Fund for International Development, UNDP, and World Bank. CIMMYT also receives non-
CGIAR extra-core support from the International Development Research Centre (IDRC) of Canada, the
Rockefeller Foundation, and many of the core donors listed above.

Responsibility for this publication rests solely with CIMMYT.

Printed in Mexico.

Correct citation: Buckles, D., and G. Sain. 1995. Land and Livelihoods: Patterns of Rural Development in
Atlantic Honduras. NRG Paper 95-01. Mexico, D.F.: CIMMYT.

AGROVOC descriptors: Honduras; farmers; living standards; production structure; land productivity;
land management; rural development; sloping land; farm surveys
AGRIS category codes: E50, E16
Dewey decimal classification: 333.76









Contents


Page

iv Abstract
v Tables
v Figures
vi Acknowledgments

1 Atlantic Honduras

5 People, Land and Occupations

10 Livelihood Strategies

15 Comparisons between Household Strategies

22 Conclusions

25 Appendix A. TWINSPAN Method

27 References









Abstract


This paper examines the impact of land distribution on the livelihoods of farm households in
Atlantic Honduras. Using survey data and multivariate analysis, we link the distribution of land
among households to the development of distinct livelihood strategies involving the production of
particular crops, wage labor, and diverse forms of non-farm employment. The analysis indicates
that the distribution of land has a strong influence on the strategic allocation of household land
and labor resources to various uses and occupations, which in turn creates structural limitations
on opportunities to improve livelihoods and manage land in a sustainable fashion. Concluding
remarks point to research and policy issues raised by the analysis.










Tables


Page
3 Table 1.

6 Table 2.

7 Table3.


9 Table 4.

11 Table 5.

14 Table 6.


15 Table 7.

15 Table 8.

16 Table 9.

17 Table 10.


18 Table 11.


19 Table 12.


20 Table 13.


Table 14.

Table 15.


Land uses in the Department of Atlantida, Honduras

Farmers' reasons for leaving their place of origin, Atlantic Honduras

Distribution of land in the hillside zone and in the Department of Atlantida,
Honduras

Ranches in the hillside zone, Atlantic Honduras

Variables used in the TWINSPAN multivariate analysis

Group indicator variables at three division levels created by TWINSPAN using
1992 survey data

Household groups in the hillside zone, Atlantic Honduras

Average land holdings and land uses of household groups, Atlantic Honduras

Percentage of farm owned and rented, by household group, Atlantic Honduras

Percentage of households cultivating various crops and average cropped area
(excluding 0 values) for each household group, Atlantic Honduras

Average years of continuous cultivation and average fallow periods, by
household group, for land under first-season maize, Atlantic Honduras

Percentage of households that sold half or more of the harvest in 1991-92,
Atlantic Honduras

Family labor resources and labor hired-in during the 1992 second season, by
household group, Atlantic Honduras

Livestock ownership, by household group, Atlantic Honduras

Percentage of households engaged in various occupations, by household group,
Atlantic Honduras


Figures


Page
2 Figure 1. Atlantic Honduras

3 Figure 2. Average weekly rainfall, Finca Buena Vista, Atlintida, 1989-91

13 Figure 3. Farm household classification generated by TWINSPAN









Acknowledgments


The authors wish to thank Hugo Perales Rivera for suggesting the multivariate analysis
technique used in this paper, and for his direct contributions to development of the analysis.
The insights of Larry Harrington, Paul Heisey, and Sally Humphries, as well as several
anonymous reviewers, helped us improve the clarity and focus of the analysis. Kelly Cassaday
and Mike Listman reviewed the paper for style and clarity, and Jos6 Crossa provided
significant input into the technical annex. The authors accept responsibility for all remaining
errors and for specific interpretations of the data.










Land and Livelihoods: Patterns of Rural
Development in Atlantic Honduras

Daniel Buckles and Gustavo Sain


Concentration of landownership is a central
and persistent feature of the Honduran
countryside. Recent data on land distribution
indicate that some 53% of the arable land in the
nation is owned by 3.7% of the farming
population (SECPLAN, 1994). The social and
environmental consequences of land
concentration in Honduras and elsewhere in
Central America have been reported widely in
the literature (Kaimowitz 1995; Humphries
1994; DeWalt, Stonich, and Hamilton 1993;
Walker 1993; Ruben 1991; Thorpe 1991;
Brockett 1990; Galvez et al. 1990; Leonard 1987;
Ruhl 1984). Problems of rural poverty, rural-
urban migration, political unrest, deforestation,
and land degradation have been traced to the
inequality of land distribution in the region.
The limited employment opportunities created
by extensive cattle ranching and other
dominant forms of rural land use are related
causes of these problems (de Janvry and Garcia
1992).

This paper examines the impact of land
distribution on the livelihoods of farm
households in Atlantic Honduras. The paper
also attempts to extend the analysis beyond the
specific influence of access to land by taking
into account an array of household resources
(including land) and occupational decisions. In
Atlantic Honduras, as in many parts of the
developing world, farm households pursue
various means of maintaining themselves
including the production of particular crops,
wage labor and diverse forms of non-farm
employment. Using survey data and
multivariate analysis, we link the distribution
of land among households to the development


of distinctive livelihood strategies. The analysis
indicates that the distribution of land has a
strong influence on the strategic allocation of
household land and labor resources to various
uses and occupations, which in turn creates
structural limitations on opportunities to
improve livelihoods and manage land in a
sustainable fashion.

The paper begins with a description of the
study area, highlighting the regional
agroecological and settlement history. A
subsequent account of population, migration,
land rights and occupations focuses on the
hillsides of Atlantic Honduras, a territory
recently inhabited by households displaced
from other parts of Honduras. The livelihoods
of the hillside population are then described
through a multivariate analysis of survey data.
Comparisons between livelihoods follow,
showing the implications of strategic decisions
made by households. Concluding remarks
point to research and policy issues raised by
the analysis.

Atlantic Honduras

Honduras comprises 18 major administrative
units, called departments. The Department of
Atlantida is located on the northern coast of
Honduras, bordering the Caribbean, at
approximately 16 N latitude (Figure 1).
AtlAntida encompasses some 4,251 km2 in a
long strip running east to west along the
coastal plain and northern side of the Nombre
de Dios mountain range. The Department of
Col6n borders Atlantida on the southeast and
the Department of Cortes on the northwest.









These three departments make up what is
known as the Northern Coast of Honduras.

The Department of Atlantida has a population
of approximately 243,000, 60% of whom live in
urban areas. This relatively high concentration
of urbanites (compared to a 45% urban
population nation-wide, reported by the World
Bank, 1993) is due primarily to the importance
of La Ceiba, the third largest city in Honduras.
La Ceiba figures highly in the agricultural
history of Honduras transnational
corporations based on banana production were
established there and in Tela during the early
part of the century but has been eclipsed in
recent decades by the industrial center of San
Pedro Sula and the Port of Cortds, both in the
neighboring Department of Cortes.

The regional population is mainly mestizo, the
descendants of Spanish and Indian peoples. A
small garifuno population brought from Africa


during the 18th century inhabits fishing
villages all along the coast, but virtually no
indigenous population remains. Although
Tolupan Indians originally lived along the
coast and in the mountains of northern
Honduras, they have been greatly reduced in
numbers and are currently limited to a few
communities in a neighboring Department,
Yoro (Spahni 1982, p. 70). Moskito Indians live
further east in the Department Gracias a Dios.

The Nombre de Dios mountain range
dominates the regional landscape. Old
sedimentary materials from the ocean floor
were pushed up during the Tertiary era to form
peaks some 2,400 meters above sea level (masl),
running parallel to the coastline. The high
mountains drop abruptly to a coastal plain only
30 km wide, where the region's best
agricultural land is concentrated. Pastures and
plantation crops are the main regional land
uses but maize, beans, and rice are also










important crops to some sectors of the regional
population and in some areas within the
Department (Table 1). Forests still cover nearly
one-fifth of the Department.

The sudden rise of the Nombre de Dios range
interrupts moisture-laden prevailing winds
from the Caribbean, generating high annual
rainfall throughout the Department of
Atlantida in a bimodal distribution. Average
annual precipitation at Buena Vista, Atlantida,


Table 1. Land uses in the Department of
Atlantida, Honduras

Area
Land use (%) (ha)

Annual crops 11 17,812
Permanent crops 15 23,915
Cultivated pastures 33 54,363
Natural pastures 17 27,111
Fallow land 4 7,054
Forests 18 29,249
Urban/non-arable 2 3,250
Total 100 162,754

Source: SECPLAN, 1994.


is approximately 3,063 mm, with some rain
during most of the year (Figure 2).1 The first
rains usually begin in June, establishing the
primera or "first season." Rains are light at this
time and subject to considerable variability
from year to year, creating a production risk for
farmers planting first-season crops. The
heaviest and most consistent rainfall on the
Atlantic Coast coincides with the last trimester
of the year (October-December), initiating a
second major cropping season known as the
postrera or "second season." Rainfall during the
last part of the second season is erratic,
however, and crops run the risk of inadequate
moisture. The rains drop off sharply in April,
interrupting most agricultural activities. May is
the driest month of the year. This short, sharp
dry period is known as verano or "summer."

Temperatures near the coast reach their peak in
May, averaging 28.80 C in this month; the
coolest month is January, which has an average
temperature of 23.90 C. The mean annual
temperature in La Ceiba, located at sea level, is
26.30 C. These climatic conditions are classified


150 Jan Feb March, April May June July Aug Sept



50 y
Oct Nov Dec
1 5 9 13 17 21 25 29 33 37 41 45 49 52
Weeks
Figure 2. Average weekly rainfall, Finca Buena Vista, Atlantida, 1989-1991.
Source: Finca Buena Vista Experiment Station, unpublished data.

1 Rainfall data from several other sites, including La Ceiba, San Francisco del Saco, Tela, and Jutiapa, follow similar
patterns.


First season









as humid tropical (Zufiiga Andrade 1990;
Pineda Portillo 1984).

The sharp differences in the distribution of
rainfall between the Atlantic Coast of
Honduras and the rest of the country (Zufiiga
Andrade 1990) have important implications for
the cropping patterns used by farmers in
different parts of the country. The interior and
southern Honduras experience prolonged
annual dry periods, limiting maize production
in most of the country to one season per year
and producing sharp seasonal increases in
maize prices. In contrast, because some rain
falls during most of the year in Atlantic
Honduras, farmers can produce two maize
crops every year. Maize prices are highest
during the second season (postrera), providing
farmers in Atlantic Honduras with an
economic advantage over farmers in other
regions limited to first-season maize
production.

The altitudinal gradient of the mountain range
creates three major agroecological zones within
the Department: coastal plain, hillsides, and
upper slopes and scarps (Irias and Szaraz 1991;
Labelle, Mendieta, and Sanchez 1990; PDBL
1991). As noted earlier, the flat terrain and
rolling hills of the coastal plain are the best
agricultural lands in the region. Slopes are
typically less than 10% and do not exceed 20%
throughout the zone; topsoil is commonly
deeper than 60 cm. The soils, created from the
continental shelf and recent marine
depositions, are relatively fertile Entisols, with
moderate pH values. Humid tropical forest was
the primary vegetation in this area (SECPLAN
1989; Holdridge 1982), although most of it was
removed for ranching and agriculture prior to
the 1940s (Yuncker 1939, cited in Ludeke 1987).
The main land uses are pastures for dual-
purpose cattle; banana, pineapple, and African


palm plantations; and some rice production. In
recent years, pasture production has been
spurred by the development of a regional
capacity to process milk and milk products for
national and international markets (Humphries
1994, p. 198).

The hillside zone, the primary focus of this
study, is less suitable for agriculture than the
coastal plain, yet it is in the hillsides where
basic grain production is concentrated. The
soils, broadly classified as Alfisols, Inceptisols,
and Entisols, are derived mainly from hard
metamorphic rock originating in the Paleozoic
era (PDBL 1991; Simons 1969). Soil textures
throughout the hillside zone include silty clay
and silty clay loam, with pH values from 6.5-
7.0. The topsoil is moderately fertile, but tends
to be poor in phosphorous and relatively thin
(40-60 cm) (PDBL 1991, p. 12). The landscape is
rough. Three-quarters of the buffer zone
surrounding the Pico Bonito national park, an
area not untypical of the hillsides of the region,
is characterized by slopes in excess of 30%, and
slopes on one-quarter of the area exceed 75%
(Rodriguez Torres 1992, p. 159). As a result, the
topsoil is highly susceptible to erosion,
especially when the soil surface is unprotected
during periods of crop cultivation or as a result
of overgrazing.

Very humid subtropical forests comprise the
primary vegetation of the hillside zone, much
of which has been deflected to crops, natural
pastures, and secondary forest. Maize, beans,
and upland rice are the most important annual
crops in the hillside zone, but pastures
dominate all other land uses, as on the coastal
plain. Approximately half the land is under
pasture, according to estimates using a
geographic information system for the buffer
zone surrounding the Pico Bonito national park
(Rodriguez Torres 1992, p. 110).









The mountain zone is generally unsuitable for
agriculture, consisting of very steep slopes and
undeveloped, thin soils (Labelle, Mendieta, and
Sanchez 1990). Hard igneous rock thrust to the
surface during the Tertiary era has evolved into
Ultisols whose top horizon is typically less than
40 cm thick. Slopes through much of this zone
exceed 50% (some exceed 100%), creating a high
risk of erosion after the forest cover is cleared.
Humid subtropical lower montane forest is the
primary vegetation from 800 to 1,800 masl, with
cloud forest at higher elevations (Holdridge
1982; PDBL 1991). Both mountain forest types
are under increasing pressure from loggers,
ranchers, and shifting cultivators migrating into
the region. The destruction of broadleaf forests
in Honduras is estimated at 46,000 ha/yr, a rate
that will quickly claim all remaining forests
(Silviagro 1994).

The Atlantic Coast has a long history of
settlement, but the population remained low
throughout most of this century. Extensive land
concessions were made by the Honduran
government to the United Fruit Company and
predecessors of the Standard Fruit Company in
exchange for construction of railroads in the
region (Ellis 1983, p. 64). The companies
established large banana plantations and
extracted much precious wood from the region,
but completed only one of the several railroad
lines they were to construct. Parts of these
holdings were later sold to Honduran nationals
and military families turned ranchers and
plantation owners (Euraque 1993). The demand
for labor in plantation agriculture and cattle
ranching was and remains low, limiting
employment opportunities for landless
workers. Malaria also impeded large-scale
settlement in Atlantic Honduras until effective
treatment became available in the 1950s.

In recent decades, the region has become one of
the main destinations for rural populations


displaced from southern and western
Honduras. The development of agro-exports in
these areas during the 1950s and 1960s (cattle,
cotton, shrimp) displaced many peasants from
the fertile valleys and coastal plains of
Honduras and caused numerous agrarian
conflicts (Stonich 1992; Thorpe 1991; Ponce
Cambar 1990; Posas 1980; White 1977; del Cid
1976). Relatively little employment was created
by these activities (Stonich 1992; Howard-
Borjas 1989), and slow urban-based industrial
growth could not absorb the landless
population (DeWalt, Stonich, and Hamilton
1993; Brockett 1990). However, modest land
reform programs during the 1970s created
peasant collectives from unused and remote
lands held by the state and transnationals,
helping reduce social unrest in Honduras
relative to other countries in the region
characterized by unequal land distribution
(Thorpe 1991).

Land reform programs and the apparent
abundance of land drew landless households
into Atlantic Honduras, and the population
grew at an annual rate of 4.2% between 1970
and 1990, compared to 3.4% yearly during the
same period for the nation as a whole.
Population density in Atlantic Honduras rose
from 35 to 57 inhabitants/km2. Much of the
growth was concentrated in settlements on the
sloping lands of the Nombre de Dios mountain
range, a population to which we now turn.

People, Land, and Occupations

An estimated 6,761 households, with an
average of 6.4 persons per household, live in
approximately 110 small towns and hamlets
scattered throughout the hillsides of Atlhntida.
The data presented in this paper were derived
from a 1992 survey of 126 families in 16 hillside
villages in the municipalities of Jutiapa and









Tela, AtlAntida. Villages were selected at
random from a complete list of hillside villages
in the region, stratified by municipality, with
the chance of selection in rough proportion to
the population of the village. The list of hillside
villages was developed in consultation with
extension agents for the Atlantic Coast District
of the Secretariat of Natural Resources (SRN).
A 1988 Secretariat of Health village-level
census provided the names of heads of
households and data on family size. In each
village, households were randomly selected
from the census for inclusion in the survey. No
female-headed households were selected;
possibly a result of the small number of such
households in the hillside zone and the
consequent low probability of selection. It is
also possible, however, that the census did not
systematically account for female-headed
households, a household type often overlooked
by outsiders. An additional requirement for
selecting a household was that its members had
to have planted maize during the 1992 winter
cycle or the previous summer cycle. This
requirement was in keeping with the survey's
focus on maize production practices (described
in detail in Buckles et al. 1992). It may also have
been a source of bias against the selection of
female-headed households engaged mainly in
non-agricultural activities. In each village,
heads of households were randomly selected
from the census for inclusion in the survey. The
survey questionnaire was tested and revised,
enumerators were trained during a three-day
workshop, and each questionnaire was
reviewed by the coordinators of the survey at
the end of each day. The survey was completed
over three weeks.

The hillsides of Atlantic Honduras were settled
mainly during the 1970s and early 1980s and
now have a relatively stable population. The
survey data indicate that more than three-
quarters of the hillside families interviewed in


1992 migrated to the Department from other
parts of the country. Most, however, reported
that they had been living in the same hillside
village for more than 15 years. By contrast, the
upper slope of the mountain range, generally
unsuitable for agriculture, is an active frontier.
Almost half the families interviewed by
Humphries (1994, p. 171) in three mountain
villages had arrived there within the previous
five years.

Farmers' reasons for leaving their home
communities are overwhelmingly related to a
lack of access to land and to land degradation
(Table 2). Almost half the farmers interviewed
in 1992 reported leaving their home
communities because they had no or
insufficient land of their own, while one-
quarter cited the declining quality of their land
as the reason for migrating. Humphries (1994,
p..172) also found that land degradation a
problem perceived by farmers as declining
yields and increasing aridity was a
frequently cited reason for migration. In her
study, as well as our own, the Honduran war
with El Salvador in 1969 was cited by some
farmers as the reason they fled from western
border communities.

While the scarcity of land forced households to
migrate, its apparent abundance drew them to
Atlantic Honduras. Until the late 1960s, the
sloping lands of the region were virtually


Table 2. Farmers' reasons for leaving their
place of origin, Atlantic Honduras

Households
Reason (%) (number)

No access to land 45 47
Land degraded 24 25
Parents migrated 14 14
Personal conflicts 9 9
War with El Salvador 8 8
Total 100 103









unoccupied, and land could be claimed simply
by clearing the forest cover and registering the
claim with municipal authorities. Most forested
land in Honduras is state property, subject to
rights of usufruct (dominio util) or squatters'
rights.2 Some 80% of all landowners in
Honduras do not have legal title to the land
they occupy and use (SEDA 1993, p. 79).

Squatters' rights, while less flexible than titled
forms of property, do provide the landless with
the opportunity to transform labor into
property rights on state land. In Atlantic
Honduras, land taxes on the order of US$ 2/ha,
paid annually to municipalities, provide
squatters with the right to sell or bequeath land
to those they choose. The transfer of property
rights requires only that the buyer assume the
obligation to pay the land taxes to the
municipality. Squatters cannot, however, use
these land rights to finance investments in
production; banks and other lending
institutions do not recognize squatters' rights as
guarantees against farm loans. This does not
represent an important limitation for most
farmers, however, as farm credit is extremely
limited throughout the region.


The 1992 Agricultural Modernization Law
facilitates titling of national land held in
usufruct (Ruben et al. 1993; Pino and Thorpe
1992), but the cost of this legal process
(approximately US$ 30/ha in Atlantic
Honduras, plus annual land taxes of US$ 6/ha)
is prohibitive for poor farmers.3 It is likely that
only commercially oriented farmers and
ranchers have the wherewithal to take
advantage of titling opportunities, a
development that may accelerate dispossession
and further land concentration (Walker 1993;
Pino and Thorpe 1992). On the coastal plain,
transnationals and national estate owners are
making use of this legislation to purchase prime
coastal territories from disaffected cooperatives,
adding new impetus to land concentration in
favored areas (Ruben et al. 1993, pp. 71-73).

While settlement of the hillsides is relatively
recent, the distribution of land is already highly
concentrated (Table 3). Survey data from 1992
indicate that 58% of the land in the hillside zone
was owned by only 17% of the land holders, in
holdings of more than 20 ha. For most of these
households, ranching was the central
component of their production strategy (see


Table 3. Distribution of land in the hillside zone and in the Department of Atlntida,
Honduras
Survey, 1992, National Agricultural Census, 1993,
hillsides Department of Atlantida
Size of land Percentage of Percentage of Percentage of Percentage of
holding (ha) households land owned households land owned
0.1-2 17.5 1.3 41.0 2.7
2.1-5 28.9 8.6 20.4 5.0
5.1-10 17.5 10.4 11.8 6.5
10.1-20 18.6 22.2 10.1 11.0
> 20 17.5 57.6 16.7 74.8


2 Recent changes in legislation allow for private ownership of forested land, a reform that is gradually transforming
the forestry sector in Honduras.
3 Walker (1993) makes a similar point for farmers in the Rio Aguan region of Honduras. Note that taxes on titled land
are higher than on land held under squatters'rights.









below). By contrast, 46% of the land holders
owned less than 10% of the total land area, in
holdings of 5 ha or less. Agricultural census
data for the Department of AtlAntida reveal an
even higher degree of land concentration for
the region as a whole; almost three-quarters of
the land is owned by less than 20% of the land
holders. This may reflect a higher degree of
land concentration in the prime lowland area
compared to the hillside zone, an issue we will
revisit later.

The concentration of landownership in the
hillside zone is an important but not absolute
limitation on access to land for farming. Some
21% of the farm families interviewed in the
hillside zone did not own farmland yet were
engaged in farming. These households
depended upon the land of others. Humphries
(1994, p. 179) argues that a social organization
favoring relations of dependency between
older landowners needing support in their old
age and younger, landless sons and sons-in-law
standing to inherit land has created a form of
extended family landownership. Farmers gain
access to farmland from other family members
in exchange for cash, labor, or a share of the
harvest.

A more formal land rental market is also well
developed in Atlantic Honduras, a market
partly created by pasture management
practices. Large-scale landowners rent out
fallow land to small-scale and landless farmers
who clear the land for annual crops. After a
few cycles, this land is transformed by the
farmers or the landowners into pastures for
grazing cattle, a process documented
throughout Central America (Kaimowitz 1995;


Stonich 1992; Leonard 1987; DeWalt and
DeWalt 1984; Brockett 1990). In Atlantic
Honduras, the conversion of farmland into
pastures may not be permanent, however. The
maintenance of permanent pastures under
conditions of extremely high rainfall, extensive
grazing, and seasonal overgrazing typical of
the hillsides of Atlantic Honduras is very
costly, prompting many ranchers to allow their
pastures to revert gradually to fallow land,
which is once again loaned out to farmers who
reestablish the pastures for the ranchers. Thus,
through land rental markets there is a relatively
constant movement of land from fallow to
crops to pasture and back to fallow again. The
land-rich benefit from the low costs of pasture
establishment through these arrangements,
while the land-poor gain access to some
farmland. Three-quarters of all households
surveyed in 1992 rented-in some land -
typically a hectare or so for maize and other
annual crops.4

While the concentration of landownership is
ameliorated by land rental markets and
extended family ownership, farmers with
limited or no land of their own typically cannot
devote all of their productive time to farming.
About 69% of all hillside households surveyed
relied partly on the off-farm earnings of at least
one family member. Almost one-half of the
total male labor force was engaged part-time or
full-time as day-workers, petty traders, loggers,
artisans, and seasonal construction workers.

Day-workers are employed by other farm
households to assist in land preparation and
planting, weeding, and harvesting annual
crops. Ranchers employ day-workers to


4 Few households reported renting-in pastures during the survey, even though informal interviews indicated that the
practice was common among larger-scale ranchers. As a result, the importance of land rental markets as a means of
gaining access to pastures among ranchers is probably greatly underestimated. This weakness in the survey data also
biases estimates of total farm size among ranchers. We maintain, however, that more complete data would not
change the general patterns identified.









establish and manage pastures. Other forms of
off-farm wage employment reported by hillside
farm households include seasonal highway
maintenance, work in small factories, and
public sector employment (mainly primary
school teachers).

Hillside farm households also engage in
various forms of self-employment. Petty
trading (fruit, bread) and craft work (mainly
baskets) provide limited cash earnings to meet
the household subsistence needs of some
families. Others extract logs from the high
forests of the Sierra Nombre de Dios and dress
them by hand for sale as timber. Logging is
profitable but subject to various problems, such
as the insecurity of tree tenure and inconsistent
regulations (Humphries 1994; Rodriguez
Torres 1992; PDBL 1991). For these reasons and
for the sheer physical strain of the work,
logging is usually a complementary activity for
relatively young households also engaged in
agriculture.

Landownership allows independence from off-
farm employment and the diversification of
farm enterprises. Some landed households
have opted to cultivate commercial crops such
as chile, cacao, and coffee. Others have
specialized in livestock production, including
cattle and pigs. About one-quarter of the farm
families surveyed in the hillside zone owned
cattle, in herds ranging from a few to 125 head
(Table 4). Most of these ranches produce milk


Table 4. Ranches in the hillside zone,
Atlantic Honduras
Households
Size of ranch (ha) (%) (number)
0 72 91
1-5 17 21
6-10 6 7
>10 6 8
Total 100 127


for local cheese manufacture or for sale to a
regional milk processing facility with
collection centers located at various points in
the Department. About one-quarter of these
ranches, typically the more isolated ones, raise
beef cattle for sale in regional markets. Many
hillside farmers aspire to establish ranches,
since cattle are less risky and more profitable
than other regional agricultural alternatives.
Humphries (1994, p. 200) calculates that a
rancher with only three milk-producing cows
can realize profits as high as the average
producer of basic grains, but with
considerably less effort and risk.

Pigs, owned in small numbers by 37% of the
surveyed population, are a poor household's
form of livestock production. While pigs are
mainly a means of savings, pig production can
be used to convert excess grain and the by-
products of cheese-making (whey) into cash,
thereby providing a stepping-stone into cattle
ranching. Slightly more than half the
households surveyed owned one or two
horses used to transport grain from the field to
the home. In a few cases, farmers rented out
their horses or their services as muleteers.

Land and livestock also provide the basis for
investment in small businesses. Cattle brokers
with trucks of their own buy cattle in the
hillside communities and resell them to
slaughterhouses in regional urban centers.
These merchants may also use their vehicles to
transport milk produced by other ranchers to
regional collection centers. A popular form of
investment among ranchers is to establish
small stores stocked with dry goods
(machetes, rope) and food items (rice, salt,
canned milk).









Livelihood Strategies


The availability of land in Atlantic Honduras
drew families to the hillside zone and initially
allowed settlers to establish farms. However,
continuous migration into the zone has
produced a land tenure system characterized
by inequalities in the distribution of land
resources. This section uses multivariate
analysis of the 1992 survey data to help
determine the impact of access to land on the
development of livelihood strategies. The
analysis uses a hierarchical classification of
households, combining various criteria such as
landownership, occupational profiles, and land
uses, which normally are considered separately
in the characterization of farming populations.
The classification applies not to individuals as
such but rather to households, an approach in
keeping with the domestic character of rural
livelihood. When compared to analyses based
on individual profiles only (e.g., head of
household) or key variables such as farm size
(cf. Galvez et al. 1990; CEPAL 1982),
hierarchical classification of household data
offers a picture of strategic relationships among
an assembly of household resources and
occupational decisions. As will be shown in
this and the subsequent section, the livelihoods
or means of maintaining households developed
by the land-rich and the land-poor differ in
ways that strongly influence land use patterns
and other features of hillside agricultural
systems.

Households are analysed here as a single unit
of production. This is not to say, however, that
the intra-household distribution of resources
and gender division of labor have little impact
on livelihood strategies and agricultural
production. Women in Atlantic Honduras
engage in a wide range of productive activities
not easily identified and quantified using the


categories and methods of surveys. For
example, although women were often present
during the interviews and asked to report their
economic activities, in many cases men
provided the responses to these questions, even
when women researchers posed the questions.
This important weakness in the survey data
and analysis could have been mitigated by an
explicit strategy to design a survey instrument
and process capable of examining gender
differences (Poats, 1991; Thomas-Slayter, Esser
and Shields, 1993).

Survey data describing household land, labor
and capital resources, occupational profiles and
crop production activities were arranged in an
ordered two-way table using the computer
program TWINSPAN (Appendix A).
TWINSPAN stands for "Two Way Indicator
Species Analysis," in reference to its
development for the classification of ecological
communities and the identification of species
indicative of these communities (Gauch 1982;
Jongman, der Braak, and van Tongeren 1987;
Hill 1979; Gauch and Whittaker 1980). The
method differs from clustering techniques in
that it is based on a two-way (case-by-variable)
ordination. This allows for the aggregation of
cases using some variables for one level of
classification and other variables for
subsequent levels, an advantage over
clustering techniques which depend upon
single measures of similarity or dissimilarity.
The ordination is done by reciprocal averaging,
an extension of the weighted averages
algorithm (Jongman, der Braak, and van
Tongeren 1987). While not free of problems
common to multivariate analyses, such as the
order in which cases and variables are entered
or large displacement distance of the sample
points (van Groenewoud 1992), the method is
less affected than clustering techniques by the
particular array of variables used.










The classification presented below is based on
22 variables from the survey data pertaining to
household resources and activities (Table 5).
The availability and allocation of land was
analyzed through its relationship to variables
for the land area in hectares under different
crops (MAIZE, RICE, BEANS, OTHER),
pastures (PASTURE), and fallow (FALLOW).
Measures of the slope of the main maize field
cultivated by the household (SLOPE) and an
index of yield potential of this same field
(YIELD) provided data on the quality of
household land resources. Variables for the
number of cattle (CATTLE) and pigs (PIGS)


provided data on specialization in livestock
production. Variables indicating the market
orientation of maize and other crop production
(SALESM and SALESO) provided information
on farm sources of income, while household
dependence upon off-farm sources of income
was analyzed through its relationship to the
total number of weeks per year households
engage in various forms of off-farm
employment (PEON, WLOFF, SELF, LOGGER).
Business activities (BUSINESS) were measured
with a binary variable. Variables for family size
(FAMILY), the use of family labor on the farm
(LABOR), and the age of the male head of


Table 5. Variables used in the TWINSPAN multivariate analysis
Variable and
category Unit Description
I. Land resources


MAIZE
RICE
BEANS
OTHER
PASTURE
FALLOW
SLOPE
YIELD


II. Livestock
CATTLE
PIGS
III. Market orientation
SALES
SALES
IV. Off-farm employ
PEON
WLOFF
SELF
LOGGER
BUSINESS
V. Family labor resort
LABOR
FAMILY
AGE
VI. Financial resource
CREDIT
HIRED


(ha) First- and second-season maize area
(ha) Rice area
(ha) Bean area
(ha) Other cropped area (mainly commercial crops such as cacao, coffee, and chiles)
(ha) Pasture area
(ha) Fallow area
(%) Percentage slope of main maize field
(%) Actual first-season yield as a percentage of reported maximum first-season yield

(no.) Number of cattle
1 for 1-4 pigs, 2 for 5-8 pigs, 3 for >8 pigs, and 0 otherwise

1 for sale of half or more of annual maize production, 0 otherwise
1 for sale of half or more of annual crop production (excluding maize), 0 otherwise
ent
(wks) Weeks per year in day-work (all family members)
(wks) Weeks per year in wage work (all family members)
(wks) Weeks per year in self-employment (all family members)
(wks) Weeks per year in logging (all family members)
1 for ownership of a business (store, brokerage), 0 otherwise
urces
(no.) Adult male equivalents of family members working on the farm
(no.) Adult male equivalents of family size
(yrs) Age of male head of household
es
1 for use of credit for maize in 1992, 0 otherwise
1 for employment of day-workers during the second season for 1-10 days, 2 for
more than 10 days, and 0 otherwise










household (AGE) were included to provide
data on the availability of family labor
resources and the stage of development in the
household cycle. Finally, the availability of
financial resources was analyzed through its
relationship to access to credit for maize
production (CREDIT)5 and the employment of
day-workers on the farm (HIRED). This
assortment of variables provided the greatest
and most pertinent information available from
the survey on household resources and
production profiles.

The distinctions between households indicated
by the multivariate analysis are presented in
Figure 3 showing the data matrix and Table 6
presenting the indicator variables at three
division levels. This analysis takes into account
the eigenvalues resulting from reciprocal
averaging. As described below, distinctions
indicated by the analysis were examined and
subjective criteria were used to evaluate their
relative importance.

The first classification level produced two
groups, one comprising approximately two-
thirds of the sample. The main difference
between the two is that Group A contains
households with a large crop and livestock
component in their production strategy,
whereas Group B contains households with
considerable involvement in various forms of
off-farm employment. At this level, the
program identifies five particularly indicative
variables: PASTURE, CATTLE, HIRED,
GUAMIL, and PEON (Table 6). Area under
pasture and fallow (guamil), number of cattle,
and employment of hired hands are inversely
related to off-farm employment as day-
workers. This relationship shows the relative
independence of ranchers and other land


holders from off-farm employment, especially
poorly paid day-work.

At the second classification level, four groups
are created with 11 indicator variables
(Table 6). In Groups C and D, number of cattle
(CATTLE), sale of annual crops other than
maize (SALESO), the production of rice (RICE)
and crops other than basic grains (OTHER),
access to credit (CREDIT), and land area under
pasture (PASTURE) are inversely related to the
ownership of pigs (PIGS). These relationships
are indicative of a distinction between ranchers
on the one hand and smaller, diversified
farmers and small livestock producers on the
other.

In Groups E and F, logging activities
(LOGGING), area under fallow (GUAMIL),
and pasture resources (PASTURE) are inversely
related to off-farm employment as day-
workers. Once again, the relative independence
from poorly paid wage-work as a peon attained
by households with land resources is apparent.
Independence from all forms of off-farm
employment is not indicated, however.
Group E includes households engaged in
logging and other forms of self-employment,
including petty trade (SELF) and small stores
(BUSINESS). The data suggest that although
households with moderate land resources can
avoid poorly paid day-work, they cannot rely
on farming alone, but must also secure some
off-farm employment.

The variables for fallow land resources
(GUAMIL) and area under crops other than
basic grains (OTHER) are indicators of
differences at classification level three between
two relatively large groups, M and N. The
number of pigs, and production of rice and


5 Credit for maize, while very limited, is more common


than credit for other crops.








1--
Farm households (samples)

11 1 1 11 11 11 1 1 111 11 1 1 1 1 1 1111 1
5679 492561668924517834008901235656772 102 2512251211201315589299111 3466911 37 4222480689 03347134592 11447788803381102267 70
Variables 516293089632365002272840304275930440318416140327853163656717792782492177961662259569597513750711052610489684545883430894188394


PAST URE
CATT LE
CRED IT
PIGS
HIRE D
OTHER R
RICE
GUAM IL
SALE SO
BUSI NESS
WAGE
BEAN S
MAIZ E
SALE SM
FAWO RK
YIEL D
SLOP E
AGE
FAMI LY
SELF
LOGG ING
PEON


88888878787767668677484554575445566688------------------565526434337--------------------------------------4-------------------
-86655546644444-435444-4--2-43-24234542---------------------------43----------------------2-----------------------------------
2--2222---22222---2---------- 2------------------------------------------------------------------------------------------------
----434-3-222-222-224244-444----434322-4-442-------2-2--------322-------2-2---4-2--5-2---3--4-6----------------2-------2--2--
33-2333-32333233333?23333-2-223222322232--2-2--33--2--3---2------------------- 2----------3233-33-----2---22-222-3-----2-----23
8-5-5-54-4564-56-5344----434---4-----32344--3---------5-4-5-2--4-5---2------645---32424-434-4444-------------------------2---
666544433-4443-4454--4444-----43411-----1-----3--4--44-2-442-----------------444445-5---4-----3-4433-3-------------5-5---
88688655558566-56676-67-66-85686686568-445-66-485-667854865868-4645865865545667645-------653----------------------------------
222222-222222-22222----222-2222222----22-2---------222222-222----2----------2--2-22--22---2222222-22-2--2---------------2-2---
2-2---------------2--------------------2-------------------------------------------------2-------------------2 -------------
-------------------566757-----------------------------------4---------------------------------------------------------446778
43-4-2-23-6443-44551443344455445344--344-4-3---3433356-5414435--331-3-34-4-443-544442-5-5546565-3-321-14354-444------2--43--
867866747568676666856766566666666666666666555446655766646666666566555556564566646565766756558686556355565665466665554544566555
22222-222-2222222-2-222-2-222-2222-2--2---2--2-222222-2--2--2--222-22--222--22--- 22---2---22222---2-22----2222-2--2--2-2-22
343434232223222444223432332332222244322224322232322342323222222323422233222344322443322434224422222223333322322432222222233222
37546566566574567464666844766466666564746744666664666647556644564566347765616655665546663556566566666558735S765785545466635646
66777677767776667788768767687777778767765867668878866775668S877578866757867776775766667445776767765678785685688876677788767766
766677676766666666777776766766666667776667676666666677766666776666766676666677766676766777667666666666776666666866666666677666
344434333434334344234443343443334244433334323233443344344333334334422334333444444444424444334323334233444424433443343334344342
---------------------------------4--4----------------46--4-4-------------------------45744-------------------------665--
-------------------------------------56-666656-666---676----------------------4--------------------------------------------
-----2--------------------------63----------------------44---45-4446544446653546466435363---634665666-64536565-566866-563---


111
111
111
110
110
10
10
10
10
011
011
0101
0101
0101
01001
01000
01000
01000
01000
00
00
00


00000000000000000000000000000000000000111111111111111111 111111111111111111il1111 1111111111111111111111111111111111
000000000 111111111looooo000000000000000llillill1111111111111110lli01 1lli01 111111
0000111111111111111000000111111111111100000000000001111100000000011111111111100000000001111111q10000000000000000011111 1 ill1111 11
000000111111111001111000000000111100000000000111110000001111110011100000000000000000000000011111111100000q0000000000000000000000001111111
000111000000001 000111111 0000111111111000010011110000110000000000 1111000000000000011000 0Olliililooolli
00000111 000001 000001111 000000001110001111100011 000000111111111110000011

GROUPS


Classification
levels used


B


.....I .C^:::l bII:DI :;I...E II ..........III.. F I


Figure 3. Farm household classification generated by TWINSPAN.


I


:: :; :: :: N :::: .........
. . . . . . . ..!! !









beans, are more common in Group M, whereas
wage employment is more common in Group
N. It would appear from these indicators that
farmers in Group M have developed relatively
diversified crop production strategies
compared to Group N. In both groups, off-farm
wage employment is consistently present in the


livelihood strategy, only more so for Group N.
This distinction between small-scale farmers
(M) and more subsistence-oriented workers (N)
is the only distinction made at classification
level three that is maintained in the final
classification (Table 7).6


Table 6. Group indicator variables at three division levels created by TWINSPAN using
1992 survey data (N=128)
Indicator
variables First-level division
Group A (N=38) Group B (N=88)
PASTURE High Very low (when present null CATTLE)
CATTLE High Very low (when present null HIRED and high PEON)
HIRED High Very low (when present null PASTURE and CATTLE, and high PEON)
GUAMIL Relatively higher Relatively lower
PEON Null (when present high High
PASTURE and CATTLE)
Indicator
variables Second-level division
Group C Group D Group E
Diversified Diversified Medium Group F
ranchers (N=19) farmers (N=19) farmers (N=30) (N=58)
PASTURE Higher Lower
CATTLE Higher Lower
RICE Higher Lower
OTHER Higher Lower
PIGS Lower Higher
CREDIT Present Absent
SALESO More present More absent
PEON Absent Present
LOGGING Present1 Absent
GUAMIL More common Less common
PASTURE Some present Absent
SELF -Present Absent
BUSINESS Some present Absent
Indicator
variables Third-level division
Group M Group N
Groups Small farmers Subsistence workers
G, H, I, J Group K Group L (N=28) (N=30)
GUAMIL Present Absent
OTHER Present Absent
1 Not present in all cases, but when present PASTURE absent; LOGGING completely absent in other group.










Comparisons Between
Livelihood Strategies

The classification of sample households using
multivariate analysis represents divisions
among households involving the unequal
distribution of land, labor, and capital.
Differences in land uses and occupational
profiles are also considered. This section
examines the impact these divisions have on
the strategic allocation of household resources
and opportunities to improve livelihoods and


manage agricultural land. In our discussions,
we will refer to the five household groups that
emerged from our final classification based on
the analysis: ranchers (Group C), diversified
farmers (Group D), medium-scale farmers
(Group E), small-scale farmers (Group M), and
subsistence workers (Group N).

Differences in farm size and land use among
groups are presented in Table 8. Ranchers
control more land both in production and in
fallow than other groups. They have roughly


Table 7. Household groups in the hillside zone, Atlantic Honduras
Diversified Medium-scale Small-scale Subsistence
TWINSPAN Ranchers farmers farmers farmers workers Total
classification (Group C) (Group D) (Group E) (Group M) (Group N) sample
Percentage of
households 15.1 15.1 23.8 22.2 23.8 100.0
Number of
households 19 19 30 28 30 126


Table 8. Average land holdings and land uses of household groups, Atlantic Honduras

Land holdings Diversified Medium-scale Small-scale Subsistence
and land use Ranchers farmers farmers farmers workers P(F)
Farm size (ha) 32.0 12.3 7.7 5.1 2.0 *
Cropland (ha) 7.3 3.5 2.8 3.2 1.8
Fallow land (ha) 8.5 4.2 4.3 1.7 0
Pasture (ha) 15.1 4.5 0.5 0 0.2*
Permanent tree crops (ha) 1.1 0.1 0.1 0.2 0 n.s.
Note: ** = significant at 0.1% probability; n.s. = non-significant.

6 Several other distinctions are noteworthy, however. Differences between Groups I and J (not shown in Table 7) are
indicated by the importance of off-farm wage employment (mainly semi-skilled construction work) to a few
households comprising part of a larger Group (D) of diversified farmers. This may represent a strategy on the part of
households with a moderate amount of land and livestock to capitalize or expand their farming operations by
bringing in additional cash through off-farm employment. While reasonable, this distinction is not used in
subsequent analyses because the resulting groups are very small (6 cases for Group I). Differences between Groups K
and L (not shown in Table 7) are indicated by only one variable, the area under pasture. The output of TWINSPAN
does indicate, however, other variables important to the division, identified as tending either positively or negatively
to one group or the other in the classification. The variables for day-work (PEON) and area under fallow (GUAMIL)
are both positively preferential at this division level, whereas logging (LOGGING) and hired help (HIRED) are
negatively preferential. A distinction between loggers on the one hand and small-scale farmers engaged in pasture
production and off-farm employment on the other is suggested, but the differences between these two groups were
subjectively judged to be insufficiently unique to warrant the recognition of distinct household production strategies.
For these reasons, the aggregation of households (Group E) at division level two was maintained.









half of their total farm area in pasture, but
cropland controlled by this group is substantial
as well, reflecting a strategy of diversification
rather than specialization in cattle ranching
alone. The dual strategy of diversified farmers
engaged in some crop and livestock production
is also brought to light by data showing the
distribution of land resources among crops,
fallow, and pastures for this group. Medium-
and small-scale farmers dedicate as much land
to crops as diversified farmers, but they do not
manage pastures. Their practice of some
fallowing and use of permanent tree crops
distinguishes them from highly specialized
subsistence workers.

The close relationship between farm size and
land rights is illustrated by data presented in
Table 9 showing the proportion of farmed area
owned and rented for each household group.
Groups with larger farms (ranchers; diversified
and medium-scale farmers) are much less
dependent upon land rental markets for access
to land. By contrast, subsistence workers
depend almost entirely upon rented land. Land
rental markets also play a significant role in the
farming systems of small-scale farmers. These
observations underline the importance of land
rental markets to large sectors of the hillside
population and point to relations of exchange
and interdependency between the landed and
the landless. Field observations suggest that
hillside ranchers, as well as larger, urban-based
landowners not included in the survey sample,


play the role of land brokers to land-poor
households. This system allows ranchers to use
their capital in fallow land to establish pastures
at little or no direct cost, while providing
landless workers with access to farmland.

While the existence of a land rental market
provides the land-poor with the opportunity to
rent-in some farmland, the potential of these
households to increase or diversify crop
production is not fundamentally altered
(Table 10). The data indicate that subsistence
workers tend to specialize in maize production
alone, on average cultivating 1.3 ha of maize in
either season. Only half of the households in
this group cultivate beans, and less than one-
third cultivate rice, typically very small fields.
In keeping with the limited land resources of
this group, other annual crops and commercial-
scale tree crops are rarely grown.

Small- and medium-scale farmers have more
substantial farms. Households in both groups
typically cultivate maize, on average between
1.5 and 2.0 ha during each season, as well as
small fields of beans. Rice is cultivated by
fewer than one-third of the households in these
groups. A sizable proportion of small-scale
farmers tend tree crops, a strategy for
diversification uncommon among medium-
scale farmers. Medium-scale farmers more
typically diversify into producing pastures for
rent or producing their own animals (Table 9).


Table 9. Percentage of farm owned and rented, by household group, Atlantic Honduras
Farm Diversified Medium-scale Small-scale Subsistence
property rights Ranchers farmers farmers farmers workers P(F)
Percentage of farm:
Owned 96.3 94.0 83.2 58.2 10.8 *
Rented 3.7 6.0 16.8 41.8 89.2
Total 100 100 100 100 100
Note: *** = significant at 0.1% probability.









The crop production profile of diversified
farmers reflects the tendency for households in
this group to produce all of the basic grains
grown in the region. Most of these farmers
produce beans and nearly half engage in rice
production as well. The proportion of
households producing other annual crops does
not differ significantly from other farmer
groups. The average maize area for members of
this group is somewhat higher, however,
surpassing 2 ha during the second season.

Ranchers present the most diversified crop
production strategy. The average maize area is
quite large, 3.8 ha during both seasons, but the
bean area does not differ from that of other
groups. Rice, a difficult crop requiring
relatively good land and careful weeding, is
grown by almost 90% of the households in this
group, whereas other annual crops (mainly
chiles) and tree crops are also much more
common than for other groups. This crop
production profile reflects the capacity of
ranchers to muster the land, labor, and
financial resources needed to cultivate a wide
range of crops. It also underlines the non-
specialized nature of livelihood strategies
among households of this group; these
ranchers have not abandoned agriculture


altogether but rather continue to rely on a
range of activities and land uses for their
livelihood. This finding does not conform to
the narrowly defined logic of enterprise
development whereby larger farms specialize
and smaller farms maintain diversified
production strategies. One possible explanation
for this is that the management and
supervision costs associated with diversified
strategies are still not great for larger farms in
Atlantic Honduras.

Maize and other annual crops traditionally
have been grown using techniques of shifting
cultivation characteristic of the humid tropics
of Mesoamerica. Trees and other fallow
vegetation are slashed and burned to prepare
the land for a short period of cultivation with
annual crops followed by an extended period
of fallow. The machete, ax, hoe, and digging
stick are the main farm implements, the plow
being used only rarely.

The key to sustainable shifting cultivation
systems in Atlantic Honduras and elsewhere is
the length of the cropping and fallow periods.
Continuous cultivation of tropical land leads to
a decline in yields and an increase in weeds,
the general reasons for field shifting (Weischet


Table 10. Percentage of households cultivating various crops and average cropped area
(excluding 0 values), for each household group, Atlantic Honduras

Diversified Medium-scale Small-scale Subsistence
Ranchers farmers farmers farmers workers
Crop % (ha) % (ha) % (ha) % (ha) % (ha) P(F)
Maize
1st season 100 (3.8) 100 (1.7) 93.3 (1.6) 96.4 (1.5) 90 (1.3)
2nd season 94.7 (3.8) 94.7 (2.4) 86.7 (1.8) 96.4 (2.0) 90 (1.3)
Beans 73.7 (0.7) 89.5 (0.6) 70 (0.6) 75 (0.8) 50 (0.4) n.s.
Rice 89.5 (1.0) 47.4 (0.4) 30 (0.4) 28.6 (0.8) 30 (0.5)
Other annual crops 42.1 (0.6) 21.1 (0.3) 33.3 (0.5) 25 (0.2) 0 (0.0) n.s.
Tree crops 57.9 (1.9) 26.3 (0.4) 6.7 (0.7) 42.9 (0.6) 3.3 (0.2) n.s.
Note: ** = significant at 0.1% probability; ** = significant at 0.5% probability; n.s. = non-significant. Values in
parentheses correspond to average values of the cropped area in hectares.









Honduras report that after only one or two
cycles of cropping, maize yields decline to less
than 800 kg/ha and the time dedicated to
weeding a particular field doubles. The yield
decline appears to be the most important cause
for these farmers to abandon a field, a finding
consistent with studies in other regions with
similar land use patterns. Increased weeding
costs, while relevant, are considered
manageable so long as yields reach acceptable
levels.

In the absence of external inputs, lengthy
fallows are needed to restore agricultural
potential exhausted by cultivation. Farmers in
Atlantic Honduras distinguish two important
stages of fallowing, only one of which is
considered suitable for cultivation. A field
abandoned to natural regrowth is called a
guatal for the first three years. During this
stage, the vegetation is composed of grasses
(zacates) and tree species (monte) in roughly
equal proportions. A guatal is relatively easy to
clear but little "nourishment" (abono) is
produced by burning the vegetation.
Furthermore, "bad weeds" (mala hierba) abound
and grow quickly in the cleared field,
increasing weeding costs. Cultivation of
guatales is avoided. If left uncultivated, a guatal
will become a guamil after about five years, a
fallow composed mostly of woody tree species.
Clearing a guamil is more time consuming
because of the abundance of trees, but the field
is "well rested" (descansado) and consequently
better suited to cultivation. The ash from a


guamil will fertilize crops for a cycle or two and
initially the field will be relatively free of
grassy weeds. Older guamiles of up to 10 years
present no particular advantages, and if left
longer require considerably more labor to clear.

While cropping periods of two-to-three cycles
and fallow periods of five-to-10 years are
preferred, cropping patterns in Atlantic
Honduras vary considerably around this norm.
The survey data indicate that cropping periods
for first-season maize ranged from as little as
one cycle to as many as seven cycles around an
average of 1.9. Cropping periods of only one
cycle were common. Fallow periods prior to
clearing land for first-season maize ranged
from one to 15 years, with an average slightly
below the minimum period needed to establish
a guamil (4.2 years). These measures, while
highly variable, suggest that bush-fallow
rotations in Atlantic Honduras are on average
as intensive as they can be within the
parameters of shifting cultivation. Further
intensification is likely to result in yield
declines and possibly land degradation.

The average number of years of continuous
cultivation of land under first-season maize
and average fallow periods prior to first-season
maize cultivation differ only slightly among
groups (Table 11). Medium-scale farmers are
the only group with fallow periods long
enough to establish a guamil, although this
average does not differ significantly from that
of other groups. Ranchers and diversified


Table 11. Average years of continuous cultivation and average fallow periods, by
household group, for land under first-season maize, Atlantic Honduras
Diversified Medium-scale Small-scale Subsistence Chi-
Ranchers farmers farmers farmers workers square
Continuous cultivation 2.2 2.2 1.7 1.5 1.9 n.s.
Fallow period 4.5 3.0 5.9 3.8 2.7 n.s.
Note: n.s. = non-significant.









of other groups. Ranchers and diversified
farmers on average cultivate their land for
slightly longer periods than other groups. The
absence of significant differences among
groups, however, suggests similar problems of
yield decline and land degradation due to
frequent cultivation.

Household groups may share similar crop
production problems but their ability to
respond to constraints is not the same. Whereas
two-thirds of all ranchers and almost three-
quarters of all diversified ranchers applied
fertilizer to their maize, only one-quarter of all
medium-scale farmers and a third of all small-
scale farmers used fertilizer on maize. Slightly
more than half the subsistence workers used
fertilizer on maize, possibly reflecting their
greater access to cash compared with other
groups dependent upon farming alone. Only
10% of the surveyed population most of
whom were ranchers received credit for
maize production in 1992. Given these
widespread cash and credit constraints,
research focused on maize yield declines
resulting from frequent cultivation should
probably include options other than fertilizers,
if the problems of all household groups are to
be addressed. For example, current fallowing
periods allow considerable scope for the use of
green manure cover crops to intensify shifting



Table 12. Percentage of households that sold
Atlantic Honduras


cultivation, a practice requiring no purchased
external inputs (cf. Buckles and Barreto 1995).

Data on the percentage of households within
each group that sold half or more of the 1991-92
harvest of various crops (Table 12) shed light
on the relative importance of various market
transactions to each group. While most hillside
farmers (71%) sell little or none of their first-
season maize, diversified farmers are more
likely than other groups to put their harvest on
the market. This tendency reflects the greater
dependence of diversified farmers on income
from crop production, compared with ranchers.

Second-season maize is more commonly sold
on the market by all groups, in response to
much higher maize prices during this season.
Sales of second-season maize are very common
among ranchers, in keeping with the much
larger maize area cropped by members of this
group. Diversified farmers and medium-scale
farmers also tend more than other groups to
sell second-season maize. As noted above, the
maize area for these households also tends to
be greater during the second season, a
production strategy not common among
subsistence-oriented farmers and workers.

The proportion of households selling half or
more of their bean harvest is low for all groups.



half or more of the harvest in 1991-92,


Diversified Medium-scale Small-scale Subsistence Chi-
Crop sold Ranchers farmers farmers farmers workers square
1st-season maize 26.3 42.1 20.0 21.4 33.3 n.s.
2nd-season maize 73.7 57.9 50.0 32.1 40.0
Beans 21.1 31.6 30.0 21.4 3.3
Rice 52.6 15.8 10.0 10.7 20.0
Other crops 31.6 5.3 13.3 10.7 0 **
Notes: *" = significant at 0.1% probability; ** = significant at 0.5% probability; = significant at 1.0% probability;
n.s. = non-significant.










Beans are commonly cultivated on the hillsides
of Atlantic Honduras, but most of the
production is destined for household
consumption. Nevertheless, small amounts of
beans sold by numerous farmers throughout
the region account for almost half the beans
consumed in La Ceiba, the country's third
largest city (Matute Ortiz 1992, p. 36).

The sale of production varies among groups
much more for rice than for beans. Ranchers
are clearly the most important rice producers,
both in terms of average area cultivated and
tendency to market the harvest. This group is
also the only one to market sizable proportions
of other crops such as chile, fruit, coffee, and
cacao. Overall, however, the level of home
consumption of annual crops is high among all
groups, including ranchers and diversified
farmers, in keeping with the subsistence
orientation of most agricultural activities in the
region.


No significant differences in family size or age
of the head of household were found among
groups surveyed that might point to the role of
the family development cycle in the rise of
livelihood strategies (Table 13). The availability
of family labor resources, while undoubtedly
important to individual households, has no
group profile. However, the capacity of
households to employ non-family laborers does
vary from group to group. Data on the use of
non-family labor by household groups
highlights the advantaged position of ranchers
and, to a lesser degree, diversified farmers.
Ranchers hired workers for an average 28
person-days to assist in clearing, planting, and
weeding second-season maize.

The group distribution of livestock is also
extremely skewed (Table 14). All households in
the rancher group own cattle, with average
herds of some 19 animals. Diversified farmers
have many fewer head of cattle but a larger
number of pigs, reflecting a strategy of gradual


Table 13. Family labor resources and labor hired-in during the 1992 second-season, by
household group, Atlantic Honduras
Diversified Medium-scale Small-scale Subsistence
Ranchers farmers farmers farmers workers P(F)
Average family size 6.2 6.3 6.3 7.1 6.2 n.s.
Average age of male
head of household 43.6 46.8 43.5 44.3 38.1 n.s.
Average days of labor
hired-in 27.8 12.5 7.9 0 3.7
Note: *** = significant at 0.1% probability; n.s. = non-significant.


Table 14. Livestock ownership, by household group, Atlantic Honduras
Diversified Medium-scale Small-scale Subsistence
Livestock Ranchers farmers farmers farmers workers P(F)
Average no. of cattle 19.2 3.2 0.2 0 0
Average no. of pigs 2.1 4.2 .7 1.8 0.1
Note: *** = significant at 0.1% probability.









accumulation of capital in livestock through
pig production. Complementarities between
very small-scale dairy and pig production may
also account for the development of this
strategy (Humphries, pers. comm.). Ranchers
earn their income through the direct sale of
milk, but the smaller herds of diversified
farmers oblige them to transform milk into
cheese for profit. Whey, a by-product of cheese-
making, can be used to fatten pigs, increasing
the profitability of this activity as well.

Medium- and small-scale farmers may own a
pig or two as a means of savings. By contrast,
subsistence workers cannot support livestock,
further evidence of structural limitations on
their livelihood strategy.

The classification of livelihood strategies can be
extended to an analysis of multiple occupations
among hillside families. As noted previously,
some 69% of all households surveyed depend
to some degree on the off-farm employment of
a family member. Specific forms of off-farm
employment are more common, however,
among some groups than others (Table 15).

Day-work on the farms and ranches of other
households is primarily the domain of small-
scale farmers and subsistence workers with
little land of their own on which to employ
family labor resources. These two groups,


representing 46% of the households surveyed,
account for 93% of the time dedicated to day-
work. Small-scale farmers and subsistence
wage workers averaged 16 and 26 weeks per
household per year, respectively, as day-
workers. These averages highlight the greater
commitment of subsistence workers to day-
work compared to more independent small-
scale farmers. By contrast, other groups rarely
engage in day-work. This form of employment
is physically demanding yet very poorly paid;
most day-workers earn approximately US$ 1.25
per day slashing brush to clear fields for
cultivation or planting, weeding, and
harvesting.

Only 10% of the households surveyed engaged
in non-farm wage employment, typically of
two types. Members of diversified farm
households who were employed off-farm had
relatively stable and better paying jobs as
school teachers, workers in small factories, and
skilled journeymen. These workers were all
literate, a skill that makes it possible for them
to acquire better jobs and through employment
accumulate land and livestock. By contrast,
subsistence workers employed in the non-farm
sector typically worked seasonally or
temporarily in highway construction. They had
a much lower level of literacy; only one-third of
the wage workers in this group could read.
Family labor cannot be productively employed


Table 15. Percentage of households engaged in various occupations, by household group,
Atlantic Honduras
Occupation Diversified Medium-scale Small-scale Subsistence Chi-
(wks/yr) Ranchers farmers farmers farmers workers square
Day labor 5.3 10.5 20.0 89.3 80.0
Off-farm wage work 0.0 31.6 3.3 0.0 20.0
Self-employment 0.0 10.5 13.3 17.9 10.0 n.s.
Logging 0.0 0.0 43.3 3.6 0.0
Business 15.8 0.0 6.7 3.6 3.3 n.s.
Note: *** = significant at 0.1% probability; n.s. = non-significant.









on the limited farms managed by subsistence
workers, nor can their labor skills fetch wages
above the subsistence level.

Logging is important to only 11% of the
households surveyed; virtually all loggers were
classified as medium-scale farmers. All but one
of these households belonged to a logging
cooperative which provided them with access
to community forests and assistance in
marketing finished lumber. As noted above,
logging is physically demanding and
constrained by uncertain access to suitable
forest resources and the risk of having finished
lumber confiscated by government officials
(Humphries 1994, p. 206). Nevertheless, the
financial rewards of logging can be
considerable, as mahogany and Spanish cedar
fetch a good price on regional markets.
Logging families reported dedicating an
average of 30 weeks per year per household to
extracting precious woods, a level of
employment generating an estimated
US$ 1,000-2,000 per family, depending on the
type and quality of wood cut.7

Self-employment in the hillside zone is
important to 11% of the households surveyed;
family members engaged in petty trading
(fruit, bread), craft work (mainly baskets) or the
management of small food stands from their
homes. Unlike day-work, non-farm wage
employment, and logging, small-scale self-
employment has no particular group
orientation; a small proportion of households
in all groups except the ranchers reported the
employment of some family members, mainly
women, in petty trade, crafts, or food


preparation. By contrast, most small businesses
have been established by ranchers reinvesting
income in small stores or trucks for brokering
animals and milk, an opportunity closed to
other household groups.

Conclusions

Survey data and multivariate analysis made it
possible to establish the interplay of variable
land resources, land uses, labor resources, crop
production strategies, and livelihood strategies.
The analysis reveals a high degree of social
differentiation and the development of distinct
means of maintaining households, based on the
distribution of land and other resources. These
livelihood strategies reflect structural
limitations on the opportunities of land-poor
households and impose constraints on their
land management practices.

Ranchers represent 15% of the households
surveyed. They own cattle and pigs, pastures,
and fallow land. Crop production is typically
diversified, including above-average areas of
maize, beans, and rice. The harvest of annual
crops is frequently sold on the market. The
financial resources of this group allow them to
establish small businesses, such as stores or
livestock brokerages, and to avoid low-paying
off-farm employment. This group probably
also has the internal resources needed to
undertake land conserving and productivity
enhancing investments in agriculture. The
current concentration of maize credit among
households in this group is probably
unwarranted.


7 The estimate is based on the average number of weeks per year of logging activities reported by households (30
weeks) and on Humphries' (1994) calculations of logging income (US$ 143-285 per month). Humphries notes that
logging is limited for most individuals to three to four months per year. While our estimated average is much higher
(7.5 months per year), it reflects the level of employment in logging activities reported by the entire household, not
individual loggers interviewed by Humphries.









Diversified farmers, also representing some
15% of the surveyed households, have fewer
cattle on average than ranchers but more pigs,
a less land-intensive form of livestock
production. Nevertheless, these households
own enough land to grow a wide variety of
annual crops. They also control some pasture
and fallow land resources. Diversified farmers
do not need to rely on day-work for their
livelihood, but they may engage in relatively
stable and better-paying forms of off-farm
employment, such as factory work and
teaching. Investment of family labor in off-farm
employment seems to provide these
households with additional opportunities to
accumulate land and livestock, an association
that points to the role which improvements in
wages and rural employment opportunities
might play in rural development.

Medium-scale farmers represent 24% of the
households surveyed. They own land,
including sizable fallow areas, on which they
grow small quantities of maize, rice, and beans.
Some land may be under permanent tree crops.
Their land resources are too limited, however,
for livestock production. Farming activities
allow these households to avoid low-paying
day-work, but most opt to complement
farming with logging, facilitated by
membership in logging cooperatives. The
decision by medium-scale farmers not to invest
labor resources in commercial crop production,
despite adequate land resources, raises
questions about the profitability of this form of
production. High marketing costs for crops and
economies of scale may be restricting the
commercial production of food by these
households.

Small-scale farmers, representing some 22% of
the surveyed households, own some cropland
but very little of it is under pasture or in fallow.


Access to land through rental markets is very
important to members of this group. Crop
production by small-scale farmers is usually
limited to maize and beans, typically for
subsistence. The absence of livestock from the
livelihood strategy of small- and medium-scale
farmers suggests that there may be
opportunities for the promotion of soil-
conserving, productivity-enhancing practices
involving the conservation of crop residues as
mulch among households in these groups. As it
is, small-scale farmers are forced by their
limited capital in land to employ family labor
resources in day-work and some forms of self-
employment (crafts and petty trade) to make
ends meet. This dependence on off-farm
employment among small-scale farmers
implies in turn that little labor would be
available within the household to invest in land
management practices.

Subsistence workers are a large group,
accounting for 24% of the households
surveyed. These households frequently engage
in low-paying off-farm employment as day-
workers, their primary source of income. Crop
production among subsistence workers is
focused exclusively on maize, the main
subsistence crop in the region. As with small-
scale farmers, subsistence workers' land
resources are too limited to permit crop
diversification, and livestock are beyond their
means. These households are highly dependent
upon rental markets for access to land.

The picture of class relations which emerges
from the analysis, while revealing concrete and
meaningful differences between households, is
far less polarized than one might expect. The
concentration of land in the hillside zone is
tempered by extensive land use patterns
compatible with the development of active
land rental markets and exchanges between the










relatively land-rich and households with
nothing but labor resources. Furthermore, the
average size of land holdings among ranchers
is modest (32 ha), suggesting that land
concentration by hillside ranchers does not give
rise to basic problems of landlessness in the
hillside zone.

More fundamental inequalities, especially with
respect to property rights, hold between the
hillside population as a whole and ranchers on
the coastal plain. Land suitable for intensive
agriculture on the coastal plain was occupied
early on by large landowners. Rapid migration
into the region from the 1970s to the present
quickly saturated the limited employment
potential of a labor-shunning cattle industry
and mechanized plantation agriculture on the
plain, forcing new arrivals onto the fragile
hillsides and, more recently, mountain slopes
of the Nombre de Dios range.

Qualitative evidence suggests that larger-scale
ranchers residing in lowland communities not
included in this study are acquiring the more
accessible and better quality land in the hillside
zone for seasonal grazing of herds (D. Buckles'
field notes; Humphries 1994). This
development is likely to have a significant
impact on the agricultural systems of the
hillside zone. Cropping frequency is already as
intensive as it can be within the parameters of
shifting cultivation- Further reductions in the


availability of land in the hillside zone
prompted by land concentration among
outside ranchers is likely to result in increased
cropping frequency and possibly land
degradation.

The viability of the livelihood strategies of a
large proportion of the hillside population is
also threatened. The establishment of
permanent pastures on the best hillside land by
well-financed ranching operations would
change the land use system that currently
supports an active land rental market, a
development that would in turn undermine the
precarious situation of land-poor households
dependent upon these markets for access to
land. The willingness of small farmers to invest
in land management practices depends in large
part on their perceptions regarding security of
access to land, a perception increasingly
influenced by the influx of large-scale coastal
ranchers. Paradoxically, technical changes in
pasture management practices facilitating the
intensification of pasture management in the
hillside zone may also have negative social
effects under conditions of increasing land
concentration, insofar as the improved practice
undermines land rental markets. If this
association is true, then research on technical
change in the cattle sector cannot be separated
from the broader context of land and
livelihoods prevailing in the Atlantic region as
a whole.









Appendix A
TWINSPAN Method*


Survey data describing the household
production activities and resources of 126
families were arranged in an ordered two-way
table using the computer program
TWINSPAN. TWINSPAN stands for "Two
Way Indicator Species Analysis," in reference
to its development for the classification of
ecological communities and the identification
of species indicative of these communities
(Gauch 1982; Jongman, der Braak, and van
Tongeren 1987; Hill 1979; Gauch and Whittaker
1980). The program constructs a classification
of cases which is used in successive iterations
to obtain a classification of variables according
to their case preferences. The two
classifications are then used together to obtain
an ordered two-way table that expresses the
variables' relations as succinctly as possible.

Two-way Matrix Structure

A two-way data matrix has a two-way
structure with cases on one side and attributes
on the other; most commonly, cases
(individuals) are rows and variables
(attributes) are columns. In this study, the cases
are samples of farm households and the
attributes are related to land resources,
livestock, market orientation, off-farm
employment, family labor resources, and
financial resources. The data used in the
analysis are a mixture of qualitative and
quantitative attributes analyzed using the
TWINSPAN method.


Reciprocal Averaging

Reciprocal Averaging (RA) is an ordination
method that consists of an eigenanalysis
(finding eigenvalues and eigenvectors) of a
two-way data matrix. Computationally it is
similar to principal component analysis, a
common ordination method, but it uses an
iteration algorithm for simultaneously
estimating the RA scores for cases and
attributes.

In the first iteration, ordination scores are
assigned to arbitrary rows and weighted
averages are used to obtain column scores from
these row scores. The second iteration produces
new row scores by weighted averages of the
column scores, and then new column scores are
produced by weighted averages of the row
scores. Iterations continue until the scores
stabilize (that is, the scores converge on a
unique solution, and the solution is not affected
by the initial choice of arbitrary row scores).

The first RA axis has the property of
maximizing the correlation between rows and
columns (cases and attributes, respectively).
When RA first-axis ordination scores of cases
and attributes are arranged in rank order, the
resulting matrix will have its largest value
concentrated along its diagonal (this is a
fundamental property of RA), so that similar
cases are located together and dissimilar cases
are far apart; likewise, similar attributes are
together and dissimilar ones are located far
apart.


SThis appendix was prepared with the assistance of Jose Crossa, CIMMYT, and Hugo Perales Rivera, Colegio de
Postgraduados, Montecillo, Mexico.









Reciprocal averaging has two weaknesses. The
first one is arch distortion of the second axis
(when plotted against the first axis) and the
compression of the first axis ends relative to the
middle. However, RA is less affected by the
particular array of variables used than
clustering techniques.

TWINSPAN

TWINSPAN uses RA to classify cases and
attributes. The classification technique is a
divisive strategy (polythetic divisive) that
begins with all the members together in a
single cluster and then successively divides the
members into hierarchies of smaller clusters.
The data are first ordinated by reciprocal
averaging, then those cases with extreme RA
axes are emphasized such that the cases are
divided in two clusters by partitioning the RA
axis near its middle. The dichotomous splitting
procedure is then repeated on the two-case
clusters to give four clusters, and so on until
each cluster has a minimum number of
members. Attributes are classified using the
same procedure and the case and attribute
hierarchical classifications are used together to
produce an arranged data matrix. The resulting
case and attribute hierarchies may be depicted
in a dendrogram. The strategy followed by
TWINSPAN to make the dichotomous splitting
can be summarized as follows:

* A primary ordination of the cases is made
using Reciprocal Averaging;
The cases are divided into two clusters by
making a dichotomous split near the middle
of the RA axis;
Cases that are preferential to one cluster or
the other are identified;


* A refined ordination using the differential
cases as a basis is made, and;
* The refined ordination is divided to obtain
the desired dichotomous split.

The program also allows for the use of
nominal, ordinal, and continuous variables in
the same data matrix. This is done by
establishing a range of values (1, 2, 3, 10, 20, 50,
80 in our case) which subjectively express an
optimal measure of variability for all variables.
Thus, in our case, the values 1, 2, and 3 mainly
represented the nominal variables while the
rest expressed the continuous variables. The
resulting "pseudovariables" all have equal
weight in the construction of the classification,
a distinct advantage over clustering techniques,
which use a secondary dissimilarity matrix
instead of the original data matrix.

In contrast to most standard clustering
techniques that classify only cases (or
attributes), the TWINSPAN method classifies
both simultaneously. The TWINSPAN method
has been used in ecology and community
studies where observation of the abundance of
a number of species in a number of samples is
recorded. However, the method can be applied
to any two-way data matrix, such as those
obtained in plant breeding and agronomy trials
in which genotypes are evaluated across
different environmental conditions, or in
genetic resource conservation systems where
data are collected for several attributes of
germplasm accessions and across different
years and sites. The method can also be applied
to the study of social differentiation and
stratification, as illustrated in this study.











References


Brockett, C. 1990. Land, Power, and Poverty: Agrarian
Transformation and Political Conflict in Central America.
Boston, Massachussetts: Unwin Hyman.

Buckles, D., I. Ponce, G. Sain, and G. Medina. 1992. Tierra
cobarde se vuelve valiente: Uso y difusidn delfrijol de
abono (Mucuna deeringiana) en las laderas del Litoral
Atldntico de Honduras. Mexico, D.F.: CIMMYT.

Buckles, D., and H. Barreto. 1995. Intensificaci6n de
sistemas de agriculture tropical mediante
leguminosas de cobertura: Un marco conceptual.
Paper presented at the Annual Meeting of the
PCCMCA, Tegucigalpa, Honduras. Mexico, D.F.:
CIMMYT.

CEPAL. 1982. Economia campesina y agriculture empresarial.
Mexico, D.F.: Comisi6n Econ6mica para America
Latina (CEPAL).

de Janvry, A., and R. Garcia. 1992. Rural Poverty and
Environmental Degradation in Latin America.
International Fund for Agricultural Development,
Staff Working Paper 1. Rome, Italy: IFAD.

del Cid, R. 1976. Reform agraria y capitalism dependiente.
Tegucigalpa, D.C.: Universidad Nacional Aut6noma
de Honduras (UNAH).

DeWalt, B.R., S.C. Stonich, and S.L. Hamilton. 1993.
Honduras: Population, inequality, and resource
destruction. In C.L. Jolly and B.B. Torrey (eds.),
Population and Land Use in Developing Countries.
Washington, D.C.: National Academy Press.
Pp. 106-123.

DeWalt, B.R., and K.M. DeWalt. 1984. Sistemas de cultivo en
Pespire, Sur de Honduras: Un enfoque de
Agroecosistemas. Tegucigalpa, D.C.: Instituto
Hondurefo de Antropologia e Historia and the
International Sorghum and Millet CRSP
(INTSORMIL).

Ellis, F. 1983. Las transnacionales del banano en Centroamerica.
San Jose, Costa Rica: Editorial Universitaria
Centroamericana.

Euraque, D.A. 1993. The Absent Honduran Oligarchy and the
Central American Crisis: A Historical Interpretation.
Trinity College, Department of History.

Galvez, G., M. Colindres, T. M. Gonzalez, and J. C.
Castaldi. 1990. Honduras: Caracterizacidn de los
productores de granos bdsicos. Temas de Seguridad
Alimentaria No. 7. Panama, Panama: CADESCA.

Gauch, H.G. 1982. Multivariate Analysis in Community
Ecology. Cambridge, UK: Cambridge University
Press.


Gauch, H.G., and R.H. Whittaker. 1980. Hierarchical
classification of community data. Journal of Ecology
69: 537-557.

Hill, M.O. 1979. TWINSPAN a FORTRAN Program for
Arranging Multivariate Data in an Ordered Two-way
Table by Classification of Individuals and Attributes.
Ithaca, New York: Cornell University.

Holdrige, L. 1982. Mapa ecol6gico de Honduras. San Jose,
Costa Rica: Centro Agron6mico Tropical de
Investigaci6n y Ensefianza (CATIE).

Howard-Borjas, P. 1989. Perspectives on the Central
American crisis: "Reactionary despotism" or
monopoly capital? Capital and Class 39: 51-81.

Humphries, S. 1994. Land use in humid tropical hillsides
of Central America: A case study of migrant farmers
in the Atlantic Littoral area of northern Honduras.
In CIAT Hillsides Program Annual Report, 1993-1994.
Cali, Colombia: Centro Internacional de Agricultura
Tropical (CIAT). Pp. 156-216.

Irias, D., and G. Szaraz. 1991. El desarrollo del Bosque
Latifoliado de Honduras: Experiencias en dreas de manejo
integrado. Unpublished MS.

Jongman, R.H.G., C.J.F. der Braak, and O.F.R. van
Tongeren. 1987. Data Analysis in Community and
Landscape Ecology. Wageningen, The Netherlands:
Pudoc.

Kaimowitz, D. 1995. Livestock and Deforestation in Central
America in the 1980s and 1990s: A Policy Perspective.
IFPRI Environment and Production Technology
Division Discussion Paper No. 9. Washington, D.C:
International Food Policy Research Institute (IFPRI).

Labelle, L., M. Mendieta, and F. Sanchez. 1990. Sistema de
clasificacidn de tierras por capacidad de uso: Plan de
desarrollo del Bosque Latifoliado. La Ceiba, Honduras:
Proyecto Desarrollo del Bosque Latifoliado (PDBL),
Program Forestal Honduras-Canada.

Leonard, H.J. 1987. Natural Resources and Economic
Development in Central America: A Regional
Environmental Profile. New Brunswick, USA, and
Oxford, UK: Transaction Books.

Ludeke, A.K. 1987. Natural and Cultural Physical
Determinants of Anthropogenic Deforestation in the
Cordillera Nombre de Dios, Honduras. PhD thesis.
College Station, Texas: Texas A&M University,
Recreation and Resources Development.

Matute Ortiz, R. 1992. Mercado de Frijol Comin en la
Ciudad de la Ceiba. BA thesis. La Ceiba, Honduras:
Universidad Nacional Aut6noma de Honduras
(UNAH), Centro Universitario Regional del Litoral
Atlantico.











PDBL (Proyecto de Desarrollo del Bosque Latifoliado).
1991. Estudio de base de component agroforestal. La
Ceiba, Honduras: Programa Forestal Honduras-
Canada.

Pineda Portillo, N. 1984. Geografia de Honduras.
Tegucigalpa, D.C.

Pino, H.N., and A. Thorpe. 1992. Honduras: El ajuste
structural y la Reforma Agraria. Tegucigalpa, D.C.:
Centro de Documentaci6n de Honduras (CEDOH)
and Postgrado Centroamericano en Economia
(POSCAE), Universidad Nacional Aut6noma de
Honduras (UNAH).

Pino, H.N., A. Thorpe, and R. Sandoval Corea. 1992. El
sector agricola y la modernizaci6n en Honduras.
Tegucigalpa, D.C.: Centro de Documentaci6n de
Honduras (CEDOH) and Postgrado
Centroamericano en Economia (POSCAE),
Universidad Nacional Aut6noma de Honduras
(UNAH).

Poats, S. 1991. The Role of Gender in Agricultural
Development. Issues in Agriculture No. 3.
Washington, D.C.: Consultative Group on
International Agricultural Research (CGIAR).

Ponce Cambar, M. 1990. Honduras: Politica agricola y
perspectives. In Honduras: Realidad Nacional y Crisis
Regional. Tegucigalpa: Centro de Documentaci6n de
Honduras (CEDOH). pp. 249-278.

Posas, M. 1980. Honduras at the crossroads. Latin American
Perspectives 7: 45-56.

Rodriguez Torres, G.A. 1992. Diagn6stico Socioambiental
y Estrategia de Manejo para la Zona de
Amortiguamiento del Parque Nacional Pico Bonito,
La Ceiba, Honduras. MSc thesis. Turrialba, Costa
Rica: Centro Agron6mico Tropical de Investigaci6n
y Ensenanza (CATIE).

Ruben, R. 1991. El problema agrario en Honduras.
Tegucigalpa, D.C.: Centro de Documentaci6n de
Honduras (CEDOH).

Ruben, R., F. Fufez, R. Sandoval, M. Ponce, E. Baumeister,
and E. Galvez. 1993. La compraventa de tierras de la
Reforma Agraria. Tegucigalpa, D.C.: Editorial
Guaymuras.

Ruhl, J.M. 1984. Agrarian structure and political stability
in Honduras. Journal of Inter-American Studies 26:
33-68.

SECPLAN. 1994. IV Censo Nacional Agropecuario 1993.
Tegucigalpa, D.C.: Secretaria de Planificaci6n,
Coordinaci6n y Presupuesto, Honduras
(SECPLAN). 8 volumes.

SEDA. 1993. Plan de accidn ambiente y desarrollo.
Tegucigalpa, D.C.: Secretaria del Ambiente,
Honduras (SEDA).


SECPLAN. 1990. Compendio estadistico agropecuario.
Tegucigalpa, D.C.: Secretaria de Planificaci6n,
Coordinaci6n y Presupuesto, Honduras
(SECPLAN).

SECPLAN. 1989. Perfil ambiental de Honduras. Tegucigalpa,
D.C.: Secretaria de Planificaci6n, Coordinaci6n y
Presupuesto, Honduras (SECPLAN).

Silviagro, S. de, R.L. 1994. Analisis del sub-sector forestal
de Honduras. Unpublished MS.

Simons, C. S. 1969. Informe al Gobierno de Honduras sobre los
suelos de Honduras. Rome, Italy: United Nations.

Spahni, J.C. 1982. Los indios de Am'rica Central. Guatemala,
Guatemala: Editorial Piedra Santa.

Stonich, S.C. 1992. Struggling with Honduran poverty:
The environmental consequences of natural
resource-based development and rural
transformations. World Development 20: 385-399.

Thomas-Slayter, B., A. Esser, and D. Shields. 1993. Tools
for Gender Analysis: A Field Guide to Methods for
Bringing Gender into Sustainable Resource
Management. Blacksburg: Virginia Polytechnic
Institute and State University.

Thorpe, A. 1991. America Central no puede tener democracia
con hambre: Las political de la Reforma Agraria en
Honduras antes de 1982. Documento de Trabajo No. 3.
Tegucigalpa, D.C.: Postgrado Centroamericano en
Economia (POSCAE), Universidad Nacional
Aut6noma de Honduras (UNAH).

van Groenwoud, H. 1992. The robustness of
correspondence, detrended correspondence, and
TWINSPAN analysis. Journal of Vegetation Science 3:
239-246.

Walker, I. 1993. El impact de las politics de ajuste
structural sobre el medio ambiente en Honduras.
Tegucigalpa, D.C.: Universidad Nacional
Aut6noma de Honduras (UNAH).

Weischet, W., and C.N. Caviedes. 1993. The Persisting
Ecological Constraints of Tropical Agriculture. New
York, New York: Longman Group.

White, R. 1977. Structural Factors in Rural Development:
The Church and the Peasant in Honduras. PhD
thesis. Ithaca, New York: Cornell University.

World Bank. 1993. World Development Report 1993:
Investing in Health. New York, New York: Oxford
University Press.

Yuncker, T.G. 1939. Notes on a semi-arid region in the
Agudn River Valley, Republic of Honduras. Torreya
39: 133-137.

Zutiga Andrade, E. 1990. Las modalidades de la lluvia en
Honduras. Tegucigalpa, D.C.: Editorial Guaymuras.






























































International Maize and Wheat Improvement Center
Centro Internacional de Mejoramiento de Maiz y Trigo
Lisboa 27, Apartado Postal 6-641, 06600 M6xico, D.F., M6xico











COPYRIGHT NOTICE

CIMMYT International Maize and Wheat Improvement Center

The CIMMYT International Maize and Wheat Improvement
Center holds all rights to the source text and shall be
considered the copyright holder for the text and images of
these publications.

The CIMMYT International Maize and Wheat Improvement
Center has made this publication available to the University of
Florida, for purposes of digitization and Internet distribution.

The CIMMYT International Maize and Wheat Improvement
Center reserves all rights to this publication. All uses,
excluding those made under "fair use" provisions of U.S.
Code, Title 17, Section 107 are restricted.

Contact the CIMMYT International Maize and Wheat
Improvement Center for additional information and
permissions.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs