ASSESSMENT OF FOUR NEEM (AZADIRACHTA INDICA A. JUSS.) INTERNATIONAL PROVENANCE TRIALS IN TANZANIA³⁴

by
Peter Iversen³⁵, Jan Svejgaard Jensen³⁶ and John Mtika³⁷

INTRODUCTION

An assessment of the four international provenance trials of neem in Tanzania took place in October 2001 as recommended by the International Neem Network (INN) at the workshop on data analysis held at the Arid Forest Research Institute (AFRI), in Jodhpur, India, March 2001³⁸.

The objective of the assessment was to evaluate the trials and give recommendations for the future management and assessment of the international provenance trials both for Tanzania and for the INN. Important questions in this case are: which provenances are the best suited to a particular environment and a particular purpose? How is the performance of the local provenance compared to introduced material, and could better reproduction material improve the utilization of neem as a contribution to development?

MATERIAL

Four provenance trials of neem (Azadirachta indica A. Juss.) were established in Tanzania in 1996 as randomized complete block design trials with 21 provenances and seed sources from 11 countries, with five or six replications at each site (table 1).

Tabel 1. Provenances in the international neem provenance trials in Tanzania (Anonymous 1998)

All trial sites have been described and documented (Anon. 1998). The trials were established with 16 (4x4) trees per plot, planted at 3.5 x 3.5m spacing.

Table 2. Geographical key data for the Tanzanian trials. Data is based on Hansen et al. (2000).

Figure 1. Map of Tanzania with the trial sites.

ASSESSMENTS

Field measurements were carried out in October 2001, after six growth seasons, at the end of the dry season. At this time the growth was apparently slow as the crown leaf mass was scarce or even missing. Slashing of grass under the trees was done before the quantitative characteristics were measured. The characters tree height and diameter of all stems above one cm at breast height (DBH) were measured, the number of stems originating from below 1.3 m were counted, and straightness of the main stem(s) on a scale from one to nine were measured on all trees in the trials, where one was given for a tree with straight stem(s) and nine for very crocked stem(s).

Health status was recorded when possible, but no attempt was made to quantify this character. Leaf mass is frequent used as a health indicator, but as the present assessment carried out during the dry season, many trees have dropped their leaves. Several trees at the Kwalukonge trial were seriously attacked by termites. This damage might be secondary following an unknown stress factor, may be water deficiency. Top-dying of trees was recorded if observed.

STATISTICAL METHODS

Calculations were based on plot means. Plots with less than five living trees were left out of the analysis. Only a few plots from the Ubena trial and Chamwino trials were left out because of low survival. A general linear model was used for data analysis of each trial.. The principles of 2-way analysis of variance, univariate test of normality and test for outliers are described in details by Hansen (2000), and have been suggested to be used for data analysis in the International Neem Network.³⁹

RESULTS

There is consistency between trials exhibit a good consistency between growth values among the trials (Table 3). However it should be noted that the highest mortality is found in the second and third best growing trials, whereas the typical dry area of Chamwino has quite low mortality. There seems to be a negative relation between missing trees and flowering frequency. It is noted that maintenance of the trials is carried out locally and it is likely that there is some variance in how it is done and this can be a source of differences in survival rates.

Table 3. Average tree parameters (arithmetic mean) for four neem provenance trials.

Provenance means for all four Tanzanian provenance trials are presented in table 4.

Table 4. Provenance mean data for four Tanzanian provenance trials with deviations from average means.

DISCUSSION

The provenances from Thailand were among the best provenances in regards to growth rate in general (height and diameter). The Thai and the Lao provenances almost always had one stem only, and the stem was much straighter for these provenances. These provenances were significantly different in the characteristics assessed. This has lead to the suggestion that they are a different subspecies A. indica ssp. siamensis (Cutshall 1998) often referred to as the "Thai neem". This subspecies differs from the spp. indica by being single stemmed and by having different leaf morphology. The other provenances seem to belong to the more common, A. indica ssp. indica, which is native to Northeast India and Myanmar (Schmidt & Jøker, 2000).

The Thai provenances will be the best ones if production of timber is the main objective of planting neem, on the other hand, survival of these provenances is lower than average, and at the Kwalukonge trial also quite large trees were suffering dieback.

Among the A. indica var. indica provenances, the provenances of Ghaati and Rammanaguda showed good growth, high survival and an intermediate number of stems, and could possibly be grown for many purposes. The provenances of Multan, Tibbi Laran and Mandore, all from the dry region around the Thar Desert, had the poorest growth. Except for the Multan provenance, these provenances also had low survival. Notably, the Mandore provenance failed in three trials but showed good survival in Chamwino - the most typical dry site locality. The National Tree Seed Program in Tanzania have noted that these three provenances were performing poor at the nursery stage and generally looked weaker than other provenances during the raining season, which could indicate that they are better adapted for more arid conditions.

CONCLUSION AND RECOMMENDATIONS

Provenances from central and eastern India seem to be superior: they grow well and have low mortality. They perform better than the South Indian and the Nepalese provenances and especially the provenances originating from very arid conditions in Western India and Pakistan.

Most important is probably the finding that the local seed source, Chamwino, performs below average, and that there could be gains from introducing other provenances. The provenance trials give indications on regions from which introductions should take place. At this stage, after six years of growth, it is recommended to import seeds from seed zone which have shown promising in this series of trials e.g. central and eastern India, and to establish demonstration plots in different parts of the country.

For the future management of the international neem provenance trials in Tanzania it is recommended to maintain and monitor the trials for at least 5 more years and carry out selective thinning when the tree canopy closes. This will provide a better idea of the phenotypes, when grown without story competition from neighboring trees, in the same way as neem trees are grown in Tanzania.

REFERENCES

Anon., 1998. Description of neem seed sources. International Neem Network. FAO Forest Resources Division. Rome, Italy. Also available at: http://www.fao.org/forestry/foris/webview/forestry2/index.jsp?siteId=2021&langId=1&30392167

Cutshall, B. K. 1998. A comparison of the potential Benefits from Two Types of Small Scale Neem Tree Farming in Thailand. Abstract. http://www.chiangmai.ac.th/abstract1998/eco/abstract/eco980418.html .

Hansen, C.P. 2000. Preliminary analysis report. Preliminary results of provenance trials of neem (Azadirachta indica) in Myanmar. DFCS, Denmark. Internal document. 16 pp + annexes.

Hansen, C.P., Lunde, G. & Jørgensen, M. (Compilers) 2000. International provenance trials of neem. Description of international trials of Neem (Azadirachta indica A. Juss) established by collaborators of the International Neem Network. FAO, Forest Resources Division. Rome. 230 pp.

Schmidt, L. & Jøker, D. 2000. Seed leaflet Azadiractha indica A. Juss. 12 Sept. 2000. Danida Forest Seed Centre. 2 pp.

³⁴ Received August 2002. Original language: English
³⁵ Forest Resources Division, FAO,Viale delle Terme di Caracalla, 00100 Rome.
³⁶ Danish Forest and Nature Research Institute, Hørsholm Kongevej 11, 2970 Hørsholm, Denmark
³⁷ National Tree Seed Programme, P.O. Box 373 Morogoro, Tanzania.
³⁸ The report of the workshop is available online at http://www.fao.org/forestry/FOR/FORM/FOGENRES/Inn/publications/wjodhfr.stm
³⁹ A more comprehensive report will be published later, including results on Genotype x Environment interaction.