CARIBBEAN Vol. XX
Sociedad Caribefia de Cultivos Alimenticios
Association Caraibe des Plantes Alimentaires
OF THE 20th ANNUAL MEETING ST. CROIX, U.S. VIRGIN ISLANDS OCTOBER 21-26, 1984
THE EASTERN CARIBBEAN CENTER, COLLEGE OF THE VIRGIN ISLANDS and THE CARIBBEAN FOOD CROPS SOCIETY
Small Farming Systems in
Las Cuevas Watershed, Dominican Republic
Food and Resource Econ. Dept.
University of Florida, B. Glade, FL
Food and Resource Econ. Dept.
University of Florida, Gainesville, FL
State Secretariat of Agriculture
Santo Domingo, Dominican Republic
Farming systems in the watershed are described. Data were
gathered by a multi-disciplinary team during the summer of
1982. Sample size was 182. Farming systems identified were
(1) irrigated with a rice base, (2) rainfed with an annual crop
base and minor permanent crops, and (3) predominantly cof-
fee farms which were not studied in detail. Irrigated farming
systems consisted of sole crops of beans, peanuts and maize in
addition to rice. Rainfed systems consisted of both sole crops
and crop mixtures of beans, maize and pigeon peas. Predomi-
nant crop mixtures are beans-maize, beans-pigeon peas, and
beans-pigeon peas-maize. Peanuts are predominantly sole
cropped. Unique to these systems are: (1) the "convite"
system-a soLial form of labor sharing for harvest; (2) bean
trading between rainfed and irrigated systems to maintain
fresh seed; (3) peanuts to provide cash; available as a loan
from the processing plant. Rainfed systems and increasing
population pressures subject the watershed to high rates of soil
erosion and resulting siltation downstream.
Keywords: farming systems; crop systems; crop associations;
Most islandsin the Caribbean are experiencing ecological prob-
lems of uncommon magnitude. Small territories and population
pressures are pushing farmers to marginally productive steep-
lands. The Dominican Republic is no exception. Antonini et al.
(1981) have summarized the problem in the following manner:
The Dominican Republic today is faced with serious prob-
lems of erosion due to the widespread practice of slash-
and-bum agriculture and the prevalence ofshallow soils on
steep slopes. Substantial amounts of top soil are lost due
to poor management practices and there results greatly
reduced soil fertility and crop productivity . As a conse-
quence of increasing population pressure and a depleting
resource base, the farm-pasture-fallow cycle is rapidly be-
ing shortened, the land's capability for sustained produc-
tivity is diminishing, andincreasingly more marginallands
are being brought under cultivation and extensive use. The
results of this deterioration of the Dominican steeplands is
affecting not only farmers within the high watersheds, but
it s also influencing the government's ability to develop
and maintain water resources for meeting energy,
agricultural and human needs (p.4).
Solving these problems is no easy task. But it seems obvious
that knowing the resources and farming practices prevalent in
these areas provides a good starting point for developing conser-
vation policies. For that reason, this paper describes the farming
systems in Las Cuevas watershed of the Dominican Republic.
The results reported are part of a multi-disciplinary research ef-
fort sponsored by the State Secretariat of Agriculture of the
Dominican Republic, the Association of Caribbean Universities
and Research Institutes, and the University of Florida. The final
objective of this inter-institutional agreement was to develop an
integrated management and protection plan for Las Cuevas
VOL. XX-PROCEEDINGS of the CARIBBEAN FOOD C OPS SOCIETY
The Study Region
Las Cuevas watershed is located on the southwestern flank of
the Cordillera Central and covers approximately 600 km2. An-
tonini et al. (1981, pp. 16-19) have described some of the region's
most important characteristics.
More than 80% of the watershed is mountainous with elevations
ranging between 2,200 and 2,800 m. Plateaus and interior
lowlands account for about 13%, while the rest of the area con-
sists of flood plains and river terraces.
Dry subtropical conditions prevail in the westernmost portion
where the average temperature is 24.5 C and the mean annual
rainfall is about 750 mm. These conditions exist in 12% of the
total area. A wet subtropical life zone covers 15% of the basin,
with more humid conditions but still irregular rainfall distribu-
tion. Some 52% of the area is characterized by low montane wet
forest, with irregular rainfall and periodic frosts. The very wet low
forest accounts for 18% of the area, where mean annual
precipitation is 55% higher than evapotranspiration. Finally, 3%
of the watershed is very wet montane forest in the eastern por-
tion, with average temperatures of 17.7"C and mean annual
precipitation of 1,750 mm. Frosts are more frequent and average
rainfall exceeds evapotranspiration by 60%.
The population of the watershed is 85 % rural and 15 % urban.
Population estimates for 1981 were 39,411 inhabitants, with an
average density of 65 persons per km2. The population is ex-
pected to double by the year 2,000.
About 7,000 families live in the study region. More than half
live in small rural settlements (100 to 2,500 persons) in the hills
or along the terraces and flood plains adjoining the Las Cuevas
river and its tributaries. The greatest number of families live in
isolated hamlets and small agricultural villages. One road con-
nects these villages with Padre Las Casas, the only urban center in
the area, and most interconnecting dirt trails become impassable
during the rainy season.
MATERIALS AND METHODS
The study area was visited by a multi-disciplinary group in
March, 1982. All members were placed on small teams and a type
of "sondeo" following Hildebrand (1981) was conducted. The
results of this visit provided valuable input for developing a ques-
tionnaire which was pre-tested near the end of June. Once the
questionnaire was redesigned, the final interviews were con-
ducted during the month of July.
The relevant population used was 5,609 households existing in
Las Cuevas watershed according to the 1980 population census.
Systematic sampling methods as outlined by Mendenhall et al.
(1971) were used to draw the sample size. The key variable in the
design was the level of household income, and accuracy of the
systematic sample was based on the assumption of random
household income and the advantages of systematic sampling in
the field work (Mendenhall et al., 1971). From the total popula-
tion of 5,609 households, a sample of 286 was selected. Because
the survey was multipurpose, 104 sampled households did not
operate any agricultural enterprise. Thus, the final sample for
this study included 182 agricultural households.
For the purpose of this analysis, relevant classification criteria
were needed. Ruthenberg (1980) has outlined this need as
In the process of adopting cropping patterns and farming
techniques to the natural, economic and socio-politieal
conditions of each location and the aims of the farmers,
more or less distinct farm systems have developed. In fact,
no farm ir organized exactly like any other, but farms pro-
ducing under similar natural, economic and socio-institu-
tional conditions tend to be similarly structured. For the
purpose ofagricultural development, and to devise mean-.
ingful measures in agricultural policy, it is advisable to
group farms with similar structural properties into classes.
It is important in this context that relevant classification
criteria are used and no single criterinb allows the forma-
tion of meaningful classes (p. 14).
The cropping patterns and farming systems found in the water-
shed were analyzed by clustering the most important crops of the
watershed with homogeneous technology and characteristics. The
classification was based on relevant factors such as irrigated and
rainfed land, type of land preparation, type of planting, level of
input usage, yields and crop cycles.
The final classification of the farming systems included:
1. short-cycle sole crop systems on irrigated land (beans, rice,
peanuts and maize) or on rainfed land (beans, maize,
pigeon peas and peanuts);
2. short-cycle crop mixtures on rainfed land (beans-maize,
beans-pigeon peas and beans-pigeon peas-maize);
3. permanent sole crop system (coffee); and
4. permanent crop mixture system (coffee-bananas).
RESULTS AND DISCUSSION
Both large and small farms are found throughout the water-
shed although half of them contain 3 ha or less land. Excluding
the nine largest farms with 64% of the area, average farm size is
about 5 ha. Slightly over half the farms had only one parcel; the
others were divided into two or three parcels (Table 1). Half the
land area is in pasture, forest or bush and only 10% is in annual
(short season) crops (Table 2).
Coffee is the most important crop and is sole cropped on 675
ha (Table 3). It is associated with bananas on another 71 ha and
with a mixture of other crops on 5 ha. Rice and peanuts are only
sole cropped, 78% of the beans are sole cropped, but only 18%
of the maize and 10% of the pigeon peas are grown alone (Table
4). The most important annual crop associations are bean-pigeon
pea, bean-pigeon pea-maize and bean-maize (Table 5).
TABLE 1. Land distribution by numbers of parcels on each farm and percentage of
total area in Las Cuevas watershed, Dominican Republic, 1982.
Farms Total area
parcels Number Percentage Ha. Percentage
1 100 55.0 941 36.1
2 67 36.8 967 37.1
3 15 8.2 697 26.8
Total 182 100.0 2605 100.0
TABLE 2. Land uses found in the 182 surveyed fans in Las Cuevas watershed,
Dominican Republic, 1982.
Classification Area (ha) of total
Short-cycle crops 255 9.8
Permanent crops (excluding
coffee and coffee-banana) 49 1.9
Coffee and coffee-banana 746 28.6
Fallow land 203 7.8
Pasture, forest or bush 1352 51.9
Total 2605 100.0
TABLE 3. Land use by sole crop systems in the 182 surveyed farms in Las Cuevas
watershed, Dominican Republic, 1982.
Percentage of total
Crop Area (ha) Including coffee Excluding coffee
Rice 23 2.4 7.8
Beans 181 18.7 61.6
Peanut 18 1.8 6.1
Maize 7 0.7 2.4
Pigeon pea 5 0.5 1.7
Coffee 675 69.7
Other 60 6.2 20.4
Total 969 100.0 100.0
TABLE 4. Land use by crops grown alone and associated in the 182 surveyed farms
in Ias Cuevas watershed, Domincan Republic, 1982.
Area (ha) Percentage of crop
Crop Sole cropped Associated Sole cropped Associated
Beans 181 51 78.0 22.0
Maize 7 31 18.4 111.6
Pigeon pea 5 43 10.4 89.6
Coffee 675 71 90.5 9.5
PROCEEDINGSof the CARIBBEAN FOOD CROPS SOCIETY-VOL. XX
sloping land. Soil loss is minimal for permanent crop systems
which are prepared and planted only once over many years. The
following sections describe these farming systems.
Short-cycle Sole Crop Systems on Irrigated Land
These systems are located on lowlands. The four crops rotated
include rice, beans, peanut and maize (Table 6).
Rice is planted once a year, maize and peanuts are planted
twice and beans three times. Land preparation is the same for the
four crops: with oxen and a plow. For 0.6 ha of rice, a nursery bed
of 10 m, where 45 kg of seed are broadcast, is prepared.
Although four major types of farming systems emerged from
the classification, this paper is concerned mainly with short-cycle
sole and associated crop systems on irrigated and on rainfed land.
The emphasis placed on these systems is related to soil conserva-
tion concerns. Short-cycle crop systems require land preparation
one to three times a year. This permits soil erosion when farming
TABLE 5. Associations of beans, maize and pigeon pea in the 182 surveyed fans in
Las Cuevas watershed, Dominican Republic, 1982.
Association Area (ha)
Bean-pigeon pea 17
Bean-pigeon pea-maize 16
Other bean associations 8
Other pigeon pea associations 10
Other maize associations 5
Total bean associations 51
Total pigeon pea associations 43
Total maize associations 31
Rice transplanting by hand occurs one month after land
preparation. A seed drill with a mule is used to plant the other
three crops. Seed rates are about 73 kg ha 1 for rice, between 73
and 87 kg ha I for beans and for peanuts and maize 73 and 2Z
kg ha-', respectively.
Fertilization practices are absent in peanuts and maize.
However, from 145 to 363 kg ha I urea are applied twice to rice,
and urea or the formula 15-15-15 are used on beans at a rate of
between 73 and 254 kg ha .
Weeding is done with machetes in all four crops. No insec-
ticides are applied to either rice or maize. Beans receive an ap-
plication of insecticide mixed with liquid N, while powdered in-
secticide is applied once to peanuts.
Harvesting dates vary among the four crops and wide yield
fluctuations are present. In the case of beans, most of the farmers
harvest by "convite." Under this system, the owner of the farm
prepares a large meal for the men, women and children working
in the harvest and no cash payment is involved. The people are
willing to provide their labor because they receive the same help
when they harvest their own crops. This is a system of mutual ex-
change of labor for mutual help among the farmers and is carried
out in a festive mood. The "convite" is also practiced in other
systems where beans are involved; i.e., beans as sole crop and
crop mixtures on rainfed land. Perhaps one contributing factor is
the more complex harvesting and packaging process involved in
bean harvesting. After being pulled by hand, the dry plants are
collected over a canvas and mules are passed over them to thresh
out the grains. Then the grains are cleaned and sacked.
Marketing for beans and maize is through middlemen at the
farm gate on these lowland farms where access to roads is more
general. Rice is sold to local millers. All peanut production is
bought by "La Manicera," which is the only peanut processing
plant in the country. This company finances all peanut produc-
tion and deducts that Mpoey at harvest time. Farmers feel that
peanut production is not profitable but they plantthis crop as a
means of obtaining cash from the loan to subsist during the
period when they do not have another feasible choice.
TABLE 6. Characteristics of short-cyde sole crops systems on irrigated land in Las Cuevas watershed, Dominican Republic. 1982.
Land Preparation Planting Fertilization
Crop Rotation Time Means Time Means (kg ha- ) Time Type (kg ha-') Time Means Pesticides Time (kg ha-) Marketing
Rice (R) B P Feb.- Oxen Apr.- Hand 73 May(l) Urea 145-363 May- Hand None Aug- 1450-3630 Local millers
with May June(l) Jun.(machete) Sep.
Bran (B) R B March Oxen Apr. Seed 73-87 May Urea 73-254 May Hand 1 applic. Jul 72L-1452 Middlemen
Aug. with Sep. orill Oct. or Oct.(machete) of insect. Dec. at farm gate
Dec. plow Jan. with Feb. 15-15-15 Feb. mixed with Marcl,
Peanut(P) R B M March Oxen Apr. Seed 73 None May Hand 1 applic. July 544-1452 Processing
Aug. with Sep. drill Oct.(machete) of powdered Nov plant
plow with insect.
Maze (M) B P March Oxen
Apr. Seed 22
May Hand None
VOL XX-PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY
at farm gate
Short-cyce Sole Crop Systems on Rainfed Land
Table 7 shows that bean, maize, pigeon pea and peanut are the
main short-cycle sole crop systems grown on rainfed land. With
the exception of pigeon peas, which is grown only once a year,
crops can be planted twice each year. Land preparation starts in
March for all crops; the second time for bean, maize and peanut
is in August. It is performed by hand with a machete in the case
of pigeon pea, while an oxen with plow or a machete are used for
the other crops.
Planting occurs within a month after the land has been
prepared. All crops are planted by hand with a machete although
in some cases peanut planting is done with a seed drill and a
mule. Seed rates for bean, maize, pigeon pea and peanut are 73,
43, 14, and 80 kg ha- 1, respectively. Some of the farmers who
harvest beans in July-August provide bean seed to those who
plant in September with the agreement that when these farmers
finish their harvest in December, they will return an extra 50% of.
the seed borrowed. Under this system, the farmers harvesting in
July and August conserve the germination quality of the seed and
obtain a 50% bonus. The farmers planting in September need no
cash for seed purchases.
The use of fertilizers and pesticides is not common. Weeding is
done during the same time of the year for all crops and is per-
formed with a machete. Yields-are lower than those obtained on
irrigated land owing in part to the occurrence of periodic
Marketing under these systems, in which are produced on the
more accessible rainfed lands, is very similar to that for irrigated
systems. Peanut production is sold to "La Manicera" under the
same contract discussed above. The other three crops are sold to
middlemen who come with their trucks to purchase the output at
the farm gate or nearest road. Sometimes, pigeon peas are
transported by mules to be sold in the nearest market.
Short-cycle Crop Mixture Systems on Rainfed Land
Bean-maize, bean-pigeon pea, and bean-pigeon pea-maize
(Table 8) are the three most important crop mixtures found on
rainfed land. They are grown only once a year with the exception
of bean-maize which is produced twice each year. The bean-
pigeon pea-maize crop mixture is mostly found on more remote
farms in the uplands. Farmers argue that under this system, if
one crop fails it is still possible to obtain some production from
the others; that is, this system guarantees them the possibility of
a certain amount of food for their families.
TABLE 7. Characteristics of short-cyde crops systems on rainfed land in Las Cuevas watershed, Dominican Republic, 1982.
Crop Rotation Time Means
Rate 1 and Weeding
Time Means (kg ha ) Pesticide Time Means
Time (kg ha -) Marketing
Ocean March Oxen w/plow or April Hand 73 Rarely Hay Hand July 167-725 Middlemen
Aug. hand (machete) Sept. (machete) used Oct. (machete) Dec. at farm gate
Maize March Oxen w/plow or April Hand 43 Rarely May-June Hand Aug.- 363-1183 Middlemen
Aug. hand (machste) Sep. (machete) used Oct-Nov.(machete) Sept. at farm gate
Pigeon pea March Hand April Hand 14 None May-June Hand Jan. 363-1088 Middlemenat
(machete) (machete) (machete) (for 2-3 farm gate or
months) in mules to
March Oxen w/plow or April Seed drill
Aug. hand (machete) Sep. w/mule or
None May Hand July 239-624
Oct (machete) Dec.
TABLE 8. Characteristics of short-cycle crop mixture systems on rainfed land in Lasl Cuevas watershed, Dominican Republic, 1982.
Crop mixture Rotation Time Means
Planting Fertilization Harvesting
Rate and Weedine Yield
Time Means (kg ha-) Pesticide Time Means Tim (kg ha 1) Marketing
Bean-maize March Oxen w/plow or April Hand Bean: None May Hand Bean: 181-725 Middlemen
hand (machete) (machete) 50-73 (machete) July at farm gate
P.Pea: P. Pea: 145-617
March Oxen w/plow or March Hand Bean: None April Hand Bean: 290-653 Middlemen
Bean-maize Aug. hand (machete) Sep. (machete) 50-73 Oct. (machete) July; at farm gate
15-22 Maize: 218-580
Bean-pigeon pea- March Oxen w/plow or
maize hand (machete)
April Hand Bean:
None Continu- Hand Bean: 109-317 Home
ously (machete) July cnumption
P.Pea: 240-435 co mp
PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY-VOL. XX
Land preparation takes place in March for all crops. A second
crop of bean-maize requires land preparation in August. For all
systems, the land is prepared with oxen and plow or with a
Planting is done with a machete, opening a small hole in the
soil and dropping in the seed. Seeding rates are similar in all
systems except that less bean seed is used in the bean-pigeon pea-
Fertilizers and pesticides are not used in these systems.
Weeding is done by hand with a machete and, in the case of
bean-pigeon pea-maize, is a continuous activity carried out by all
members of the family.
Yields vary among the three crop mixtures and from those ob-
tained on sole crop systems. Output of bean-maize and bean-
pigeon pea is sold to middlemen. The production from the bean-
pigeon pea-maize crop mixture is consumed at home.
This paper has described the small farming systems in Las
Cuevas watershed of the Dominican Republic. Several important
characteristics, some of them unique in this area, were found.
The role of crop associations in these farming systems was a
relevant finding. For example, 82% of the maize and 90% of the
pigeon peas are grown in association with other crops. Further-
more, all output from the bean-pigeon pea-maize association is
devoted to home consumption.
Two systems of mutual help among the farmers were also
found. One consists of the exchange of bean seed between those
harvesting in July-August and those planting in September.
The former conserve the germination quality of the seed and ob-
tain a 50% bonus, while the latter do not need cash for seed pur-
chases. The "convite" system is the means by which farmers
harvest their bean crops without incurring labor expenses.
Farmers are willing to provide their labor because they receive the
same help when they harvest their own crops.
Peanut production, although not profitable, is carried out as a
means of obtaining cash from a loan to subsist during the period
when they do not have another feasible choice.
Although a relatively small amount of all the land in the water-
shed is devoted to short-cycle crops, these systems produce high
rates of soil erosion that are unacceptable. Increasing population
pressures are likely to worsen this problem. The importance of
these systems to the farmers' diets preclude any policy that would
prohibit their future production. Perhaps the problem could be
alleviated by the development of improved technology that
would bring about increased production in the uplands on a
smaller land area.
1. Antonini, G., K. Ewel, R. Fisher,J. Sartain. P. Hildebrand. T. McCoy. H.
Safa and I. Russo. 1981. Integrated training and research program in natural
resources management for the Dominican Republic with special applications for Las
Cuevas watershed. Gainesville, FL: University of Florida Press.
2. Hildebrand, P.E. 1981. Combining disciplines in rapid appraisal: the
sondeo approach. Agr. Adm. 8:423-432.
3. Mendenhall, W., L. Ott and R. Scheaffer. 1971. Elementary survey sampl-
ing. Second Ed. North Scituate, MA: Dusbury Press.
4. Ruthenberg, H. 1980. Farming systems in the tropics. Oxford, New York:
Oxford University Press.
VOL XX-PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY