Cover Image
close this bookAgro-forestry in the African Humid Tropics (UNU, 1982, 162 pages)
close this folderCurrent agro-forestry activities
View the documentThe integration of livestock production in agro-forestry
View the documentIntercropping of terminalia superba with cocoa and banana in mayombé, people's republic of the Congo
View the documentAn example of agro-forestry for tropical mountain areas
View the documentIntercropping tree and field crops
View the documentPromising trees for agro-forestry in southern Nigeria
View the documentFood crop yield under teak and cassia siamea in south-western Nigeria
View the documentAgro-forestry possibilities in oil palm plantations in the Ivory coast
View the documentEffect of food crops on tree growth in Tanzania
View the documentSelection of leguminous trees for agro-forestry in Cameroon
View the documentForestry aspects of agro-forestry practice in Nigeria
View the documentSummary of discussion: Current agro-forestry activities

Effect of food crops on tree growth in Tanzania

J.F. Redhead and J.A. Maghembe
Division of Forestry, University of Dar es Salaam, Morogoro, Tanzania


Village afforestation must meet the needs of farmers for quick tangible returns. The growth of food crops among trees appears to be a promising incentive. Research to ascertain appropriate species and mixtures, and appropriate espacements and rotations to maximize production, are described.

Good food crop yields are obtained in the first year of intercropping but fall off in the second and subsequent years with trees at close espacement. Wider tree espacement is likely to prolong useful food cropping. Trees intercropped with food crops show growth as good, or nearly as good, as trees grown under clean-weeded conditions and much superior growth to trees raised under spot-weeded or unweeded conditions

It is recommended that similar trials be established for various tree and food crops over a range of ecological sites. These will serve also as valuable demonstration plots.


King and Chandler (1978) emphasize that the term agro-forestry may be applied to a wide spectrum of land-use combinations. These range from taungya (or agrisilviculture) to the selective use of trees as shelterbelts in a predominantly agricultural system.

Many traditional farming systems already include the permanent cultivation of a balanced mixture of tree species and agricultural crops. The trees often provide an immediate economic return in terms of direct products in addition to the indirect benefits of improving the ecological conditions for the food crops and animals.

Examples in high rainfall areas are the integrated farming systems developed in Latin America, where Cordia alliodora and Cedrela odorata are grown over lemon, bananas, and plantains, which are in turn underplanted with coffee, or the Kandy spice gardens of Sri Lanka, where tree crops such as nutmeg, cloves, and jackfruit are underplanted with coffee, bananas, pepper, and vegetables (cf. the paper by Watson in this volume, pp. 6-12). In Africa a good example is the land use on the slopes of Mount Kilimanjaro, where coffee, bananas, and annual food crops are intermixed with timber trees such as Grevillea robusta.

Trees have also been combined with a more limited agricultural cropping programme. For example, the taungya system was designed primarily for the establishment of permanent timber crops. It was most successful in areas where there was a shortage of good agricultural land. King (1968) expresses concern that the peasant labour involved is exploited because the forest services benefit from their efforts and pay no remuneration. However, it is noteworthy that despite dilution with the tree crop, per hectare yields to the farmers are higher than on land available to them outside the reserves. In fact, Hofstad (1978) has calculated that in north-eastern Tanzania the value of the food crops grown under the taungya system may far outweigh the discounted value of the subsequent tree crops.

Peasant farmers in all countries are naturally conservative: change usually involves risk, and they are unwilling to undertake change without convincing evidence of guaranteed personal financial or material advantages. They have very limited capital reserves to sustain them; thus, change should provide immediate, or short-term, returns. Furthermore, peasant farmers belong to closely integrated communities that are economically interdependent. Suggested changes should benefit the whole community and cause minimal disruption to existing social structures and customary practices.

The "felt needs" of a rural community are for a sustainable system of land management that will produce acceptable food for people and animals, maintain fertility and water balance, and provide a renewable source of firewood and building poles within walking distance of the village. A successful agro-forestry system for a farming community must meet specific needs in these areas, must provide obvious short-term personal advantages for the farmer, and, as far as possible, should be consistent with local traditional working habits and patterns of family labour,

Intercropping of selected tree species and agricultural crops may be competitive or collaborative. Traditional systems have developed over time to select effective mixtures; development of new systems will require research into many factors, including appropriate species and mixtures, and appropriate spacings and rotations to maximize production.

The Division of Forestry at Morogoro, Tanzania, is currently initiating research to assess the effects and productivity of intercropping selected tree species with staple food crops over an extended period.

Agro-forestry Research at Morogoro, Tanzania

The rainfall at Morogoro is about 800 mm a year, falling mainly between March and May, but sometimes with a substantial amount in November. Unfortunately for the farmer it is very unpredictable in its regularity. Although Morogoro does not lie within the humid tropics, the research being carried out is of wider application. The investigations are simple in design, the field work is easy to carry out and applicable over a wide range of ecological conditions. Very important also, this type of trial gives quick interim results that can be easily seen and understood.

Trials of Eucalyptus melliodora with Different Crops

The layout of the trial and early results have been previously described by Maghembe and Redhead (1981). Containergrown Eucalyptus melliodora were planted in February 1978 in farmland that had been ploughed and harrowed. The tree seedlings were planted at a spacing of 2.5 x 2.5 m and intercropped with maize, sorghum, and beans, which were planted at 90 x 30 cm, 60 x 15 cm, and 40 x 20 cm, respectively. The same three crops were planted again in 1979 and 1980. Yields of maize and beans were recorded; sorghum was not harvested as the crop was eaten by birds. Both weed-free and unweeded plots of E melliodora were also grown as controls. No fertilizers were used, and the experiment was arranged as a latin square. Each plot contained 5 x 5 trees, and the central core of 3 x 3 trees was periodically measured. The area was clear-felled in March 1981, and the leaf litter and standing biomass assessed. The same food crops were sown again so that yields along with coppice could be ascertained.

The yield of maize was 1,280 kg/ha in the first year and 100 kg/ha in the second year. In the third year it did not flower. Sorghum followed a similar pattern, although yields were not recorded. At 2.5 x 2.5 m spacings, the trees were too close to permit sufficient light for maize and sorghum to grow after the first year. The yields of beans for the three years were poor; in the first year, when good yields were expected, the crop was badly attacked by an unidentified fungus. In the second year the beans appeared healthy but the yield was only 150 kg/ha. In the third year the beans were etiolated and the yield was insignificant. At the time of the first harvest, the mean height of the Eucalyptus in unweeded plots was significantly lower than Eucalyptus in intercropped or weed-free treatments, in which heights did not differ significantly. The trees in the beans and clean-weeded plots were noticeably more robust and more heavily branched than the trees among maize and sorghum, which were spindly due to competition for light (figs. 1-2). Survival of the trees has been approximately 90 per cent, except in the unweeded plots, where half died during the first two years.

FIG. 1. Mean Height of Eucalyptus melliodora Inter" cropped with Maize, Sorghum, and Beans, Compared with Clean-Weeded and Unweeded Trees, Morogoro, Tanzania

FIG. 2. Volume of Eucalyptus melliodora Intercropped with Maize, Sorghum, and Beans, Compared with Clean-Weeded and Unweeded Trees, Morogoro, Tanzania

At three years of age, the stand has yielded more than 1,000 poles/ha suitable for house building and more than 6 m³ /ha of fuelwood. The mean size of trees grown among beans is not significantly different from that of those grown in weed-free plots, and the trees intercropped with maize and sorghum are two-thirds the size of clean-weeded trees. The maize and sorghum yield would clearly make up in value for the reduced volume of the tree crop.

It was apparent from this trial that normal yields of agricultural crops could be expected in the first year with little effect on the tree crop in the case of Eucalyptus melliodora, but for food production in subsequent years wider spacing of the trees is necessary.

Trials of Trees at Different Spacings with Crops

It was primarily to determine how far apart the trees could be and still produce an acceptable volume of firewood or appropriately shaped poles that a series of more elaborate long-term investigations was laid down in 1980. Another objective was to test tree species that would provide both fuel and fodder. Four trials were established: (1) Eucalyptus camaldulensis (planted for fuel and pole production) with maize and beans; (2) Acacia albida (planted for fodder and fuel production) with maize and beans; (3) Leucaena leucocephala (planted for fuel production) with maize and beans; (4) Leucaena leucocephala (planted for fodder production) with maize and beans.

The layout of these investigations has been described by Maghembe and Redhead (1981). Briefly, tree seedlings were container-grown and planted in ploughed and harrowed land. Each investigation has a split-plot layout with food crops (maize or beans) and weeding treatments (spot-weeded or clean-weeded) forming the main plots, and tree spacings forming the sub-plots. The maize was planted at 75 cm x 30 cm, leaving a circle of 50 cm radius around each tree; fertilizer was applied at a rate of 400 kg/ha ammonium sulphate and 200 kg/ha triple superphosphate, in two applications. The beans were planted at 40 cm x 20 cm, leaving a circle of 20 cm radius around each tree seedling; fertilizer was applied at a rate of 200 kg/ha ammonium sulphate and 200 kg/ha triple ammonium phosphate, with the former applied when the beans were well established. Clean weeding was done by harrow, supplemented by hoeing; no fertilizer was applied in these plots. Spotweeding was done by hoe in a circle of 50 cm radius around each tree, as in normal Tanzanian forestry practice; no fertilizer was applied. This main-plot treatment was omitted from trials 2 and 4.

With regard to the spacing, trials 1 and 3 had sub-plots with trees at 3 x 3 m, 4 x 4 m, and 5 x 5 m, respectively. In trial 2, the 3 x 3 m spacing was replaced by a 6 x 6 m spacing in order to maximize fodder production. Similarly, in trial 4, the Leucaena leucocephala was planted in rows 3, 4, 5, and 6 m apart to maximize fodder production. Trees within rows were only one metre apart. In all four trials sub-plots without trees were planted as controls.

Long term soil studies are in progress so that soil under Eucalyptus, Acacia, and Leucaena can be compared with soil under fallow and under pure food crops. The nitrogenfixing activity of the Acacia and Leucaena is being measured with a portable gas-liquid chromatograph.

The trees in all investigations were too young to have had a marked effect on the maize and bean yields during the first cropping season apart from the space they occupied. Differences are expected to show up from the second year onwards, when the canopy starts to close. Unfortunately, the yield of maize and beans was normal in only one set of plots combining Leucaena. In this case, the mean yield of maize was 1,645 kg/ha, which compares favourably with yields obtained on the university farm and is over twice the national average in Tanzania of 670 kg/ha (Acland 1971). The mean yield of beans was 401 kg/ha, an average yield by peasant standards (Acland 1971).

In the other investigations the planting time and tasselling time of maize coincided with a severe drought, and growth yields were very poor and uneven. The drought caused great variation in the interaction between the food and tree crops and masked the significance of the results.

In contrast with the little effect the trees had on the food crops, the maize in particular has had a marked effect on the trees because it is a tall crop casting considerable shade (figs. 3-6). In all investigations, at the time of maize harvest, the trees among maize were taller than those in other treatments-24 per cent taller in both stands of Leucaena and 20 per cent taller in the Eucalyptus The height differences were statistically highly significant except in the Acacia plots, where the maize-grown trees were only 13 per cent taller. After harvest, the clean-weeded trees grew best, and in all cases their height surpassed that of trees in other treatments by the end of the year. The spot-weeded treatment has proved much inferior to the other treatments, as the trees are only two-thirds of the height of clean-weeded trees and approximately 20 per cent less in height than trees intercropped with either maize or beans. These growth trends are emphasized even more by the differences in root-collar diameter (fig. 7).

FIG. 3. Mean Height of One Year-Old Eucalyptus camaldulensis Intercropped with Maize and Beans Compared with E. camaldulensis Grown CleanWeeded and Spot-Weeded, Morogoro, Tanzania. The vertical bars indicate differences significant at P = 0.05; N.S. indicates differences not significant at P = 0.05.

FIG. 4 Mean Height of One-Year-Old Acacia albida Intercropped with Maize and Beans Compared with Clean-Weeded A. albida. The vertical bar indicates a difference significant at P = 0.05; N.S. indicates differences not significant.

FIG. 5. Mean Height of One-Year-Old Leucaena leucocephala Grown for Fuel and Intercropped with Maize and Beans Compared with That of L. Ieucocephala Grown Clean-Weeded and SpotWeeded. The vertical bars indicate differences significant at P = 0.05.

FIG. 6. Mean Height of One-Year-Old Leucaena leucocephala Grown for Fodder and Intercropped with Maize and Beans Compared with That of L. Ieucocephala Grown Clean-Weeded. The vertical bars indicate differences significant at P = 0.05; N.S. indicates differences not significant.

FIG. 7. Mean Root-Collar Area of One-Year-Old Trees Intercropped with Maize and Beans Compared with That of Clean-Weeded and Spot-Weeded Trees, Morogoro, Tanzania. The vertical bars indicate differences significant at P = 0.05; N.S. indicates differences not significant.


The results demonstrate that good food yields can be obtained, at least during the first year of intercropping with trees. The trees benefit more from the weeding associated with growing food crops than they do from the Tanzanian standard practice of spot weeding. Moreover, spot weeding allows a dense growth of grass to grow in the intervening spaces, and this is a serious fire hazard. Clean weeding gives the best growth, but it is not realistic to expect this in community afforestation projects.

It will be interesting to monitor the subsequent yields of food crops at the wider tree spacings. If a valuable pole and fuel crop is nearing harvest by the time food yields drop seriously, agro-forestry in this form could become a standard practice in community afforestation projects. Should tree intercropping prove beneficial in improving soil fertility, this will have great significance for areas where a shortage of land precludes a long natural fallow.

It is suggested that trials of similar design be established for a range of tree and food crops over varied ecological sites in tropical Africa. These will serve as valuable demonstrations and at the same time yield useful data for afforestation.