|
|
 |
 |
| |
 Printable version |
Improved cropping practices with tannia,
Xanthosoma sagittifolium (L.) Schott in the Bamileke region of western Cameroon |
Cahiers Agricultures. Volume 8, Number 1, 9-20, Janvier-Février 1999, Étude originale
|
 Résumé
Article gratuit
|
Author(s) : Jean-Luc Schafer |
Summary : In Cameroon, tannia, Xanthosoma sagittifolium (L.) Schott, called macabo or cocoyam, and taro, Colocasia esculenta (L.) Schott, are cultivated from the western highlands to the southern forest region. Araceae crops occupy the second position in terms of national tuber and arrowroot production, with 232,000 t produced in 1988. Cocoyam cropping is twofold that of taro due to its higher productivity, rusticity and nutritional value (Table 1). A fully-grown cocoyam plant has six to seven large leaves, and at maturity the underground part (Figure 1) is composed of a corme (mammy) and five to ten cormels which are eaten, sold or used as seed. The cropping cycle (Figure 2) lasts 9 to 10 months. The crops require 1,500 to 2,000 mm of rainfall, fertile soil and moderate insolation.
An experiment was conducted from 1989 to 1992 on the granitic plateau of Bafou (Figure 3), one of the most important chiefdoms of the Bamileke region. The agricultural system is based on five main mixed intercropping systems, with arabica coffee plantations and intensively-grown food crops. Cocoyam productivity under such conditions is decreasing markedly (1.2 to 3.5t.ha-1), even though this product is a staple food for people in this region.
Since 1989, on-farm research has been carried out to propose solutions for improving tannia yields (and household incomes), while preserving soil fertility. Three methods were evaluated: improving cultivation practices, selecting productive varieties and testing adapted mineral fertilization. In 1989-1990, farming-system diagnosis and exploratory on-farm trials at eight sites identified and prioritized some of the main constraints to tannia production. Improved cultivation practices were developed in collaboration with farmers: choice of fields based on previous crops, careful ridging, seed selection based on type, weight and phytosanitary status, fixed density of 15,100 plants.ha-1, regular weeding and destruction of sprouts.
In 1990-1991, a randomized complete block design (RCBD) with four treatments was used for tests carried out on nine farms, which were selected to represent the wide range of environments in the area. Two varieties of white macabo were planted: the local Bafou variety and a variety from Fongo-Ndeng. Two fertilization levels were used: none for the control treatment (i.e. the local practice), and 80 kg N + 200 kg P2O5, + 200 kg K2O per ha, a level that was determined through research work at the local research station on N and K fertilization and soil analyses. Four treatments were set up: Bafou variety without (T1) or with fertilization (T2), and Fongo-Ndeng variety without (T3) or with fertilization (T4). The farmers used improved cultivation practices on 20 m2 traditional ridging plots and fertilizers were spread three times on the plots (P on plantation under seed, N 45 days after planting and K 90 days after planting). An original and significant (r2 \= 0.993) non-destructive method highlighted changes in the leaf area index (LAI) every 50 days for each plot in a randomized sample. The crop cycle lasted from April 1990 to February 1991, and 1,800 mm of rain fell over this period\; no diseases were observed.
In 1991-1992 a validation trial was carried out on 20 mall farms representative of the wide range of different cultivation practices implemented and environments in this granitic plateau area. Two treatments were conducted: local Bafou variety without fertilization or with 80 kg N + 200 kg P2O5 + 200 kg K2O per ha. Improved cultivation practices were used by farmers on each 72 m2 plot.
Data were analysed with the Stat-Itcf 7 software program, based on ANOVA and regression analysis. The modified stabilility analysis (MSA) technique of Hidelbrand was used to determine recommendation domains according to an environmental index (IM). Analysis of MSA regression lines assessed the technological risks in the different environments. The partial budgeting method based on gross margin was used for economic evaluations for each treatment.
The results of the 1990-1991 trial in 8 blocks (Table 2) indicated that the effect of NPK fertilization was very highly significant on the tannia Bafou variety (22.6 t.ha-1 versus 12.7 t.ha-1) and on the Fongo-Ndeng variety (18.9 t.ha-1 versus 10.1 t.ha-1). The Bafou variety was characterized by higher (p <\; 0.02) mean yield (17.8t.ha-1 versus 74.5 t.ha-1) and higher fertilization response (9.9 t.ha-1 versus 8.8 t.ha-1), as compared to F. Ndeng. The two main yield components were also significantly increased by fertilization: the mean number of tubers per plant increased from 6.6 to 8.9 (p <\; 0.001) and mean weight of a tuber from 116 g to 150 g (p <\; 0.00001). The modified stability analysis technique of Hidelbrand applied to tuber yield indicated that the best treatment involved the fertilized Bafou variety, regardless of the environmental index (mean yield at each of the eight locations). Each r2 of the fourth linear regression was significant (Table 3), indicating that the model was valid considering the wide range of IM (8 to 32 t.ha-1). Figure 4 illustrates the time course of LAI changes according to treatments and mean sunlight levels. Regression curves were determined on the basis of the logistic function, with a goodness-of-fit between 0.987 and 0.999. We noted that fertilization increased the maximum mean LAI from 1.27 to 2.14. The earliest soil cover (LAI \= 1) seemed to enhance photosynthesis activity and soil protection against rainfall erosion during this rainy (200 mm) and low sunshine period (70 h). A significant curvilinear regression (Figure 5) between maximum LAI and yields in each treatment (r2 \= 0.714 × 0.118, Jackknife method) was demonstrated, in agreement with the model of synchronous development of tubers and leaves.
The validation trial (1991-1992) confirmed the RCBD results (mean 11.37 t.ha-1 for the NPK treatment versus 3.43 t.ha-1 without fertilizer for 17 analysed plots). Nevertheless, the mean yield of fertilized plots was twofold lower than the RCBD trial (11.37t.ha-1 versus 22.6 t.ha-1), suggesting that other factors, mainly the previous crop, should be taken into account for fertilization planning (Table 4, Figure 6).
Cost-effectiveness was calculated for the validation trial with 1997 prices and costs (Table 5), 3 years after the CFA franc devaluation. Since 1992, fertilizer costs increased threefold, while macabo retail prices increased 1.7-fold. The mean gross margin was 1,082,000 CFA francs for the 12 fertilized plots without Araceae as previous crops, compared to 288,000 CFA francs for non-fertilized plots and 1,060,000 CFA francs for fertilized potato crop. The five plots with Araceae as previous crop had a negative gross margin. The investment for a smallholder cultivating 500 m2 of pure tannia was about 19,500 CFA francs, which could lead to cash problems. The mean fertilizer ratio in the validation trial (kg of extra crop: kg of fertilizers) was 7.8 and the economic ratio (extra receipts: expenditure on fertilizers) was 3.3. Tannia could thus be both a subsistence and a cash crop, and this innovation was still attractive in 1997, but only on condition that macabo was not preceded by Araceae.
We conclude that well balanced mineral fertilization, based on NPK requirements of tannia and the soil, including improved cropping practices, significantly increased yields and gross margins and had a positive effect on the cooking quality of tubers. Coffee fertilizer (20-10-10), which is currently used by farmers, is therefore unsuitable for fertilizing tannia crops.
We recommend carrying out on-farm trials (i) to determine, over several years, optimum levels of N, P and K of mineral fertilization in typical cropping situations, (ii) to test the effects of the prior history of a field on subsequent crops in the cropping system, (iii) to test the effect of organic fertilization with or without mineral fertilization on yield and on soil organic matter.
The highly fluctuating economic environment and agricultural liberalisation also have marked effects on Bamileke intensive cropping systems and on farmers’ strategies and behaviour. In the light of this new agricultural situation, further research should be carried out to come up with better adapted cropping systems and provide farmers with effective decision-support tools.
|
Keywords : |
|
