About Bioline  All Journals  Testimonials  Membership  News  Donations

African Crop Science Journal
African Crop Science Society
ISSN: 1021-9730 EISSN: 2072-6589
Vol. 4, Num. 2, 1996, pp. 167-176
African Crop Science Journal
Vol.5. No.2, pp. 167-176 1997

Soil and foliar phosphorus determination in Rwanda:

procedures and interpretation

P. Drechsel, B. Mutwewingabo^1, F. Hagedorn^2 and C.W. Wortmann^3
Scientific consultant in Rwanda; present address: IBSRAM, P. O. Box 9-109, Bangkhen, Bangkok 10900, Thailand

^1 Faculty of Agronomy, National University of Rwanda, B.P. 117, Butare, Rwanda
^2 Institute of Soil Science, University of Bayreuth; present address: WSL, 8903-Birmensdorf, Switzerland
^3 CIAT, P.O. Box 6247, Kampala, Uganda

(Received 21 August, 1995; accepted 16 June, 1996)

Code Number: CS96054
Sizes of Files:
    Graphics: Line drawings (gif) - 31.1K
              Tables (gif) - 15.3K   

Revised reference values of available soil phosphorus by Bray I below which bean (Phaseolus vulgaris L.) and sorghum (Sorghum bicolor L.) yield response to applied P is expected, were 6 and 9 ppm, for January and July soil sampling dates, respectively. Reference values determined for interpretation of foliar test results of bean differed from those estimated previously. The critical foliar level of P concentration was estimated to be 0.40%. The importance of modifications of the Bray I procedure for soil P analysis was assessed using the modifications of three soil laboratories in Rwanda. Variations in shaking time, soil to extraction solution ratio, and soil sieve size were found to significantly affect available P values using the Bray I procedure. The modifications of one laboratory gave values which were 75% higher than those of the original Bray I procedure. Errors in preparation of standard solutions and in calculations resulted in values ranging from 2.5 to 300% of the correct value.

Key Words: Available phosphorus, DRIS, foliar analysis, Phaseolus vulgaris, Rwanda, soil analysis


Des valeurs de reference revisees pour des teneurs du sol en P assimilable determine par Bray I, en dessous desquelles on s'attend e une reponse positive des cultures (Phaseolus vulgaris L. et Sorghum bicolor L.) e un apport en phosphore, etaient 6 et 9 ppm pour des echantillons preleves en janvier et en juillet respectivement. Des valeurs de reference etablies pour l'analyse foliaire du haricot se distinguent des valeurs publiees avant. Une teneur de 0.40% P dans les feuilles est proposee comme niveau critique pour l'alimentation du haricot. L'importance des modifications de la methode "Bray I" pour l'analyse des teneurs du sol en phosphore assimilable a ete revelee dans trois laboratoires au Rwanda. La methode a ete modifiee de plusieurs faons differentes et les resultats provenus de differents procedes ont ete compares. Les valeurs trouvees pour le phosphore assimilable dependaient significativement de la duree d'agitation, du rapport sol/solution d'extraction et du tamisage (e 0.5 ou 2mm). Les modifications des valeurs depassaient dans un laboratoire de 75% les valeurs estimees par la methode originale Bray I. Des erreurs faites lors de la preparation des solutions etalon et calculs ont fait aboutir e des valeurs entre 2,5 et 300% de la valeur correcte.

Mots Cles: Analyse du sol, diagnostic foliaire, DRIS, Phaseolus vulgaris, phosphore assimilable, Rwanda


Rwanda's prosperity depends on agricultural production which is constrained by low soil fertility, especially low soil P availability (Neel et al., 1976; Vander Zaag, 1982; Pietrowicz, 1985; Mutwewingabo and Rutunga, 1987; Mutwewingabo, 1989; Rutunga, 1991). Vander Zaag (1982) estimated an average requirement of 220 kg ha^-1 of triple superphosphate (44 kg P ha^-1) to increase available soil P adequately in Rwanda. In Gikongoro and Butare areas, Mutwewingabo (1989) estimated that up to 4000 kg ha^-1 of TSP (800 kg P ha^-1) are needed for a short-term amelioration.

Soil P determinations in Rwanda are done by analysis of P-adsorption isotherms (e.g. Mutwewingabo and Rutunga, 1987), or more routinely, using the Bray I procedure according to Bray and Kurtz (1945) of estimating available soil P. The Bray I extractant is appropriate for low-to-medium CEC soils with high amounts of Al-P (Cope and Evans, 1985). The procedure is often modified in soil laboratories in Rwanda, as well as in other countries (Bingham, 1973; Page et al., 1982). In Rwanda, however, the effects of the modifications are not considered in the interpretation of the soil test results. In addition, the commonly used reference values for interpretation of Bray I results appear to lack a sound basis for their use in Rwanda (Agrar- und Hydrotechnik GmbH, 1986, 1993).

Good interpretation of foliar data depends on adequate reference values. Comparing the Diagnosis and Recommendation Integrated System (DRIS) with the critical nutrient level approach, Wortmann et al. (1992) found that DRIS provided a better basis for interpreting foliar test results of bean. While they validated a set of DRIS norms estimated from a broad-based data set, their results indicate that interpretation can be further improved with locally estimated DRIS norms.

This research had three objectives: to estimate adapted soil reference values for Bray I extractable P (Singh, 1979); to determine the effects of the methodical modifications on Bray I results; and to estimate specific DRIS norms for bean grown in Southern Rwanda.


The investigations were carried out in Southern Rwanda. Soil and foliar reference values were estimated using data collected from a long-term fertilizer experiment (1990-1994) conducted at the Rubona research station of Institut des Sciences Agronomique du Rwanda (ISAR-Rubona) (Hagedorn, 1995; Steiner et al., 1995). The trial was arranged in a split-plot design with 12 treatments consisting of two green manure species (Cajanus cajan (L.) Millsp. and Tephrosia vogelii Hook.f.) as sub-plots, and four fertility management treatments as the main plots: control; farmyard manure at 10t ha^-1 yr^-1; lime at 2t ha^-1 yr^-1; and 40 kg ha^-1 yr^-1 each of N, P2O5 and K2O applied as 240 kg NPK (17:17:17). Trials were of six replications. The test crops were sorghum, cv. Ikinyaruka, associated with the green manures in the main rainy season, and bean, cv. RWR 221, in the short rainy season.

Soil samples were taken at 0-20 and 20-40 cm depths several times during the year. Soil P was extracted using the procedure of Bray I and measured photometrically. Samples of the youngest (uppermost), mature leaf blades at early flowering of beans grown at ISAR-Rubona, as well as in various other locations and projects in South Rwanda (n=100), were collected. Foliar samples were analysed for N, P, K, Ca, Mg, Mn, Fe, and Zn using an atomic absorption spectrophotometer, photometer and CNS-Analyser of Universities in Butare and Bayreuth. Statistical analysis and DRIS calculations were carried out with SPSS/PC+ (Drechsel, 1994).

Modifications of the Bray I procedure were observed in each of the well-equipped laboratories of ISAR-Rubona, the Agronomic Faculty of the National University in Butare, and PASI (Project Agricole et Social Interuniversitaire) in Butare. Modifications of Bray I evaluated in this study were: sieve size (0.5 or 2mm); soil to extractant ratio (1:7 or 1:10); time of shaking (one or five minutes); and direct filtration versus decantation after centrifugation at ca. 3000 rpm for 10-15 minutes (increasing total contact time between soil and solution from about 20 to at least 40 minutes). Four soils (Table 1) from the ISAR stations in Rubona, Karama and Gakuta (Fig. 1) were analysed in each laboratory. We analysed three replicates of the samples to compare the modifications. In addition, the samples were analysed in each labaratory by their own staff and according to their modifications to check for errors in preparation of standards and in calculations.


Reference values for the interpretation of soil phosphorus. Bray I results for the 0-20 cm soil layer reliably reflected the soil P available to the crop and were related to crop yield (R2 = 0.62; Fig. 2). Available P levels in the subsoil were generally low and not significantly related to crop yields. The critical Bray I values determined for a Typic Sombriudox (Table 1) were similar for sorghum and bean (Table 2), but were affected by time of sampling. Bean and sorghum yield response to applied P is expected if Bray I results are less than 6 and 9 ppm P, when soil is sampled in January or July, respectively (Fig. 2). Available P levels were relatively high early in the season because of the rapid mineralization of soil organic matter that occurs with the onset of the rains (Pietrowicz, 1983; Hagedorn, 1991), but lower and more stable when sampling was done late in the season (January and June/July). Sampling at harvest is preferred as P levels are relatively stable due to low rates of P uptake by plants and P release by mineralisation, and it allows sufficient time for soil analysis before planting of the next crop. In addition, it is possible at harvest to see the variations in crop growth which may indicate need for additional sampling of the field (Cope and Evans, 1985).

These estimates of critical soil P levels (Table 2) are in the range of those currently used (the range for low P is 1-9 ppm; Agrar- und Hydrotechnik GmbH, 1986), but are more precise and have a better empirical basis for conditions in Rwanda than the currently used values. Care should be taken, however, in applying these results to other crops and in other regions. In different parts of the USA, for example, critical soil Bray I-P levels range between 1 and 11 ppm for soybeans, 3 and 22 ppm for maize, and 5 and 65 ppm for potatoes (Bingham, 1973). A distinct critical value of 15 ppm for field beans is reported from Michigan (Bingham, 1973).

Bray I modifications and systematic errors. Increasing shaking time from 1 to 5 minutes increased extractable P values by 10-20% (significant at P<0.05; Fig. 3) in three of four soils. The effect was not significant for the Gakuta soil where extractable P was very low (<2 ppm). Increases of 10-20% are small relative to the 400% increase in shaking time as, apparently, most of the P release occurs early in the shaking process. Also, the total soil with solution contact time did not differ greatly (21 versus 25 minutes).

Soil P values were significantly higher (18-24%) for the soil to solution ratio of 1:10 as compared to the 1:7 ratio (Fig. 4), and in agreement with results of Randall and Grava (1971).

Sieve size effects were tested using one soil only. Extractable P was 6-7% more with the smaller sieve size (P<0.05). An opposite effect was found during the same study for total carbon. Approximately 40% less C, in form of small pieces of organic material, passed through the smaller mesh. However, these organic particles are not a source of Bray P.

Large differences in the total contact times between soil and solution of more than 40 versus about 20 minutes were caused by the centrifugation of the samples in comparison with the direct filtration. However, effects on extractable P were inconsistent across soil samples and the inconsistencies are unexplained.

The results demonstrated the sensitivity of the Bray I procedure to modification. Results for each laboratory differed from those of the original Bray I procedure, with the greatest deviations due to the modifications of ISAR-Rubona laboratory. The P values obtained using the modifications of the ISAR-Rubona laboratory were 30-55% higher than those from the laboratories of the Agronomic Faculty and PASI, and 75% higher than those obtained using the original procedure for Bray I (Fig. 5). Glaser and Drechsel (1992) recommended against the use of Bray I because of the occurrence of such discrepancies with modifications of the procedure.

Methodological errors in the preparation of the standards or in the calculation of Bray I values occurred in two of the three laboratories. The effects of the errors were dramatic with P values ranging from 2.5 to 300% (0.1 to 25 ppm P) of the correctly determined values (Fig. 6). Only one of the laboratories was regularly checking its performance against control samples of known P test levels.

Based on the above findings and those of Neel et al. (1976), we recommend the following:.

1. Do not exceed 12 to 16 samples per extraction series to better control the total time that the soil is in contact with the extraction solution.

Laboratories currently handle 12 to over 50 samples per set.

2. Use an extraction ratio of 1:10, or 1 g soil per 10 ml of Bray I solution.

3. Use a 0.5 mm sieve.

4. Shake for five minutes at 175 rpm.

5. Use direct filtration with Whatman No. 42 or S&S 5893 (blue ribbon). Filtration should not exceed 10 minutes. If the filtrate is not clear, filter it again.

6. Compare laboratory results with known control samples. Exchange samples with other laboratories.

Reference values for the interpretation of foliar phosphorus. Bean yield was significantly related to foliar P (R2 = 0.68), but even better related with DRIS index values for foliar P (R2 = 0.83). This agrees with Wortmann et al. (1992) who found interpretation by DRIS to be more accurate in predicting nutrient needs of bean than use of critical foliar nutrient levels. The data suggest a critical P level of 0.40% which agrees with that reported by Reuter and Robins (1986), but is higher than that estimated by Wortmann et al. (1992). DRIS norms for the interpretation of foliar analyses are expected to be widely applicable (Walworth and Sumner, 1987), but Wortmann et al. (1992) found significant differences between norms estimated from a broad-based data set and those estimated from East African data alone, suggesting accuracy of DRIS may be further improved by having sets of norms for specific production environments.

The DRIS norms for N, P, Ca and Mg estimated in the current study in Southern Rwanda (Drechsel and Hagedorn, 1994) differ from those estimated for Eastern Africa by Wortmann et al. (1992) (Table 3). N and P values were relatively high, while Ca and Mg values were relatively low. The differences may be due to several factors. Yield levels were high in the current study and the lower limit of the DRIS data set was 2000 kg ha^-1, while Wortmann et al. (1992) used a lower limit of 1100 kg ha^-1, which may account for the higher N and P values. The relatively low concentration of Ca and Mg may have been caused by reduced uptake due to moisture stress at sampling time. In the current study, a single bean variety (cv. RWR 221 which is known for tolerance to some soil fertility related problems) was used, while the earlier study used data collected from numerous varieties. Confirmation studies will be needed to verify that the locally generated norms are more accurate for Southern Rwanda.


The results illustrate the importance of standardisation of soil and foliar test procedures. Small modifications of an analytical procedure can significantly affect the results. Time of sampling soil is important to interpretation of the results. Response to applied P is expected if soil P values by Bray I are less than 6 or 9 ppm, when soil is sampled in January or July, respectively. Interpretation of foliar test results by DRIS is expected to be superior to use of critical nutrient levels, but the results indicate that DRIS norms estimated for a specific environment may be more accurate for that environment than broad-based norms.


The investigations have been carried out with the assistance of ISAR-Rubona, the National University in Butare, as well as PASI and Projet dÕIntensification Agro-Sylvo-Pastoral. We are especially indebted to Dr. K.G. Steiner (GTZ/ISAR) for support of this study. Finally, we would like to express our sincere respect for all members of our research staff who were killed or otherwise affected by the civil war in Rwanda.


Agrar-und Hydrotechnik GmbH. 1986. Etude d'amenagement du marais Mwogo. Rapport deuxieme phase et documents cartogra-phiques. Annexes 1- 4. Republique Rwandaise, MINAGRI. Essen (RFA).

Bingham, F.T. 1973. Phosphorus. In: Diagnostic Criteria for Plants and Soils. Chapman, H.D. (Ed.), pp. 324-361. Second Printing. Abilene.

Bray, R.H. and Kurtz, L.T. 1945. Determination of total, organic and available forms of phosphorus in soils. Soil Science 59:39-45.

Cope, J.T. and Evans, C.E. 1985. Soil testing. Advances in Soil Science 1:201-228.

Drechsel, P. 1994. DRIS (Diagnosis and Recommendation Integrated System) in Theorie und Praxis. Eine SPSS-orientierte Anleitung. Bayreuther Bodenkundliche Berichte 34, Bayreuth, 72 pp.

Drechsel, P. and Hagedorn, F. 1994. DRIS-Normen fur Phaseolus vulgaris L. In: Drechsel, P. 1994. DRIS (Diagnosis and Recommendation Integrated System) in Theorie und Praxis. Eine SPSS-orientierte Anleitung. Bayreuther Bodenkundliche Berichte 34, Bayreuth, 72 pp. Drechsel (1994), p: 65-66.

Glaser, B. and Drechsel, P. 1992. Beziehungen zwischen verfugbarem Bodenphosphat und den Phosphatblattgehalten von Tectona grandis (Teak) in Westafrika. Zeitschrift Pflanzenernahrung Bodenkunde 155:115-119.

Hagedorn, F. 1991. Evaluation des effects des engrais verts et du fumier sur les variations de l'azote et du phosphore dans les sols de la station experimentale de Rubona. Document de travail No. 11, GTZ/ISAR, Rubona, 29 pp.

Hagedorn, F. 1995. Organische Dungung in Sud-Rwanda. Einfluss auf Kohlenstoff. und Nahrstoffdynamik. Unpubl. Master's thesis, University of Bayreuth, Germany.

Mutwewingabo, B. 1989. Genese, caracteristiques et contraintes d'amenagement des sols acides e horizon sombre de profondeur de la region de haute altitude du Rwanda. In: Actes du Premier Seminaire franco-africain de pedologie tropicale, 6-12 Febr. 1989 in Lome (SOLTROP 89). University du Benin/ORSTOM, pp. 353-385.

Mutwewingabo, B. and Rutunga, V. 1987. Etude des sols des stations dÕessai du projet PIA (PDAG). Service Agricole, Gikongoro.

Neel, H., de Prins, H. and Vindevogel, E. 1976. Etude de la fertilite des sols e l'aide des vases de vegetation. ISAR, Note Technique 1, Rubona.

Page, A.L., Miller, R. and Keeney, D. (Eds.) 1982. Methods of Soil Analysis. Agronomy No. 9, Part II. Sec. Ed., pp. 416-418, Am. Soc. Agr. and Soil Sci. Soc. Am., Madison, Wisconsin.

Pietrowicz, P. 1983. Ausgewahlte Methoden der Nutzungsoptimierung im Rahmen des standortgerechten Landbaus in Rwanda. Giessener Beitrege zur Entwicklungs-forschung, Reihe I, Vol. 9:143-162.

Pietrowicz, P. 1985. Les sols de la region d'action du Projet Agro-Pastoral de Nyabisindu, Rwanda. PAP, Etudes et Experiences No. 7. Nyabisindu.

Randall, G.W. and Grava, J. 1971. Effect of soil: Bray no.1 ratios on the amount of phosphorus extracted from calcareous Minnesota soils. Soil Science Society of America Proceeedings. 35:112-114.

Reuter, D.J. and Robinson, J.B. 1986. Plant Analysis - An Interpretation Manual. Inkata Press, Melbourne, pp. 151-153.

Rutunga, V. 1991. Essai de synthese des connaissances acquises sur la fertilite des terres et la fertilisation des cultures au Rwanda (+1960 - 1990). MINAGRI/PNUD, Kigali, 124 pp.

Singh, B.R. 1979. Nutrient availability indices and optimum nutrient levels in soils for crops under agroforestry conditions. In: Mongi, H.O. and Huxley, P.A. (Eds.). Soil Research in Agroforestry. ICRAF, Nairobi, p. 439-470.

Steiner, K., Drechsel, P., Sekayange, L. and Nzabonihankuye, G. 1995. Chances et limites des engrais verts pour la gestion de la fertilite des sols du Centre-Sud du Rwanda. Les essais de l'ISAR 1988-1993, ISAR Note technique (116 pp., in press), Rubona.

Vander Zaag, P. 1982. La fertilite des sols au Rwanda. Bulletin Agricole du Rwanda 1:3-24.

Walworth, J.L. and Sumner, M.E. 1987. The Diagnosis and Recommendation Integrated System (DRIS). Advances in Soil Science 6:149-188.

Wortmann, C.S., Kisakye, J. and Edje, O.T. 1992. The Diagnosis and Recommendation Integrated System for dry bean: determination and validation of norms. Journal of Plant Nutrition 15:2369-2379.

Copyright 1996 The African Crop Science Society

The following images related to this document are available:

Line drawing images

[cs96054c.gif] [cs96054h.gif] [cs96054g.gif] [cs96054d.gif] [cs96054b.gif] [cs96054i.gif] [cs96054a.gif] [cs96054e.gif] [cs96054f.gif]
Home Faq Resources Email Bioline
© Bioline International, 1989 - 2022, Site last up-dated on 11-May-2022.
Site created and maintained by the Reference Center on Environmental Information, CRIA, Brazil
System hosted by the Internet Data Center of Rede Nacional de Ensino e Pesquisa, RNP, Brazil