Session 3 Genetics, Provenance Trials, and Vegetative Propagation

Session Papers

Provenance Trials

Problems with Range-Wide Provenance Trials of Faidherbia albida on Sandy Soils in Niger

R.J. Vandenbeldt¹

Abstract

A Faidherbia albida provenance study consisting of 32 seed origins from western Africa and 9 from eastern and southern Africa was established in Aug 1988 at the ICRISAT Sahelian Center in Niamey, Niger to compare variability between and within the two groups of seed lots. Mortality of the southern and eastern African provenances was total by the end of the second dry season. Analysis of height growth of the remaining western African provenances was continued; however, selection of superior seed origins was complicated by a high degree of site microvariability which confounded plot growth data. Caution is urged to those establishing or analyzing data from provenance studies planted on sandy soils in the Sahel.

Introduction

There has been much effort in the past 15 years to introduce and/or reintroduce Faidherbia albida in farmers' fields, especially in Sahelian Africa. Introduction of seed origins from southern Africa into the Sahel was expected to be of benefit to these efforts. A study was started in 1988 to better understand differences in early growth rates and survival of eastern and southern African versus Sahelian seed origins of F. albida.

Materials and Methods

The experiment was located at the ICRISAT Sahelian Center (ISC), 45 km south of Niamey, Niger at 13°N 2°E. Niamey receives an average of 562 mm of rainfall (standard deviation = 134 mm), most of which falls during June-September. Average potential evapotranspiration during the dry months of October-May is 1820 mm (Sivakumar 1986).

The soil was a Labucheri series of the Psammentic Paleustalf (sandy, siliceous, isohyperthermic) family. The Labucheri series is characterized by a high sand fraction (90%), moderate acidity, poor water retention (<10%), and low nutrient status. The soil has a clear depth of 3-4 m, overlying a gravelly lateritic layer (West et al. 1984). Depth to water table is about 20 m.

Seed of 32 origins of F. albida were obtained from various international and national research institutes in western and southern Africa. The 9 seed origins obtained from eastern and southern Africa covered the area from Kenya to the Limpopo River (Republic of South Africa), a N-S distance of nearly 3000 km. A similar 3000-km E-W distance was rep-resented by the western Africa origins, which ranged from eastern Chad to western Burkina Faso.

Seed were hand-scarified by nicking the edge, and sown three to a pot (100 pots per seed lot). Three-month-old seedlings were outplanted on 15 Aug 1988, irrigated with 2 L of water, fertilized with 30 g of 15-15-15, and given a 3.5 g dose of Carbofuran (3% a.i.), all mixed thoroughly with the backfill. Seedlings were weeded twice during the 1988 rainy season and once during the 1989 season.

Because of poor germination of several seed lots, not all origins could be replicated equally in the trial. Thus, a randomized incomplete block design was used, consisting of 18 core entries present in all 4 replications, with the remaining 13 seed lots assigned randomly to replications depending on the number of available seedlings. This made an average of 26 plots per replication. Plots consisted of 10 trees planted in single rows, unbordered, spaced at 2 m within plot rows and 3 m between plots.

Live height (dead tops were not measured), basal diameter, and survival were recorded monthly for all trees during the first 4 months and thereafter at half-yearly intervals. Data were analyzed using the SAS General Linear Modeling procedure for personal computers (SAS 1985). Plot averages, based on the number of surviving trees at the time of data collection, were used.

Results

The first 2-years' performance of the various provenances has been reported elsewhere (Vandenbeldt 1991). Briefly, early above-ground growth of seedlings of the southern origins exceeded that of Sahelian origins. However, as soil moisture became limiting during the first dry season, plants of the southern entry began dying. By the end of the second dry season (1990), all individuals of the southern entries had died, whereas survival of the western African accessions still exceeded 95%. Root excavations showed that, although above-ground biomass was similar, this was primarily due to a three-fold difference in root biomass between provenances of the two respective regions.

Despite this dieback, measurements were continued to compare height growth between the western African provenances on the assumption that, since plots of the eastern and southern African provenances were randomly assigned, their absence would in turn have a random effect on the experimental error. Data for height at 2.5 years .(Table 1) show that two Burkina Faso provenances (Soubaka and Zorkoum) had the greatest average height, followed by the local Sadore, Niger accession. The General Linear Model used in analysis weighs entries on the basis of the number of replications in the experiment.

However, when data from the analysis was compared with the trees in the field, it was apparent that the location of the best accessions was more important than any inherent ability to grow faster than other accessions. The field was characterized by groups of plots with exceptional growth and other groups with very poor growth. Figure 1 shows these plot clusters in relation to microrelief of the field and proximity to eroded termite mounds.

The 20 best plots (represented by '+' in Figure 1) clustered around microhighs and termite mounds, while the 20 worst plots ('-' in Figure 1) were clustered around microlows and sheet-eroded surfaces. The 'best' accessions (Soubaka and Zorkoum, Burkina Faso) happened to have both of their respective entries on the best parts of the field. Similarly, the worst accessions (Sion and Diou, Burkina Faso) were planted on low, infertile microsites. Thus, any ranking procedure of provenances was more an exercise in plot ranking rather than provenance superiority.

Table 1. Height of surviving western African F. albida provenances, (age = 2.5 years), Sadore, Niger, 1988-90.

Discussion

A more thorough discussion of effect of microsite variability on growth of F. albida on sandy Nigerien soils is presented elsewhere in these proceeding (Geiger et al. 1992; Brouwer et al. 1992). It might be argued that correctly placing blocks inside the different zones of poor, medium, and good microsite soil fertility would have eliminated the problem from the start. However, other than subtle microtopographical differences and the occasional trace of clay on the soil surface (indicative of an eroded termite mound), there is no clear indication of just where such zones are before outplanting.

Figure 1. Layout of Faidherbia albida provenance trial showing the best and the worst plots in relation to field topography and termite mounds, Sadore, Niger, 1988-90.

In such situations, one must be wary of the wisdom of establishing provenance trials and must clearly state the experimental objectives in advance. Soil microvariability occurs in many soil types of the Sahel, and this variability can severely affect results of such studies. Blocks and plots in many of these trials are so large that field microvariability is contained within them, ballooning intra-plot error and obviating effective analysis of variance.

Several possibilities exist to circumvent this problem. Among these are single-tree plots, lattice designs, and greater effort to delineate field variability. At ISC, we have tried the first two approaches without success and are now working on the third.

One method to delineate field variability is to sample soil pH—this parameter has a direct correlation with other soil parameters affecting growth of F. albida. However, one must sample to depths of over 1 m, which makes it difficult to collect sufficient samples to make the exercise worthwhile. Presently, we are investigating the use of the yield measurements from the previous seasons' millet crop to predict areas of good and poor potential F. albida growth. We also plan to remove old provenance trials and design blocks of new clonal and progeny trials around areas of good and poor tree growth as manifested by tree growth in the old trials.

References

Brouwer, J., Geiger, S.C., and Vandenbeldt, R.J. 1992. Variability in the growth of Faidherbia albida: a termite connection? These proceedings.

Geiger, S.C., Vandenbeldt, R.J., and Manu, A. 1992. Preexisting soil fertility and the variable growth of Faidherbia albida. These proceedings.

SAS. 1985. SAS Procedures Guide for Personal Computers. North Carolina, USA: SAS Institute, Inc.

Sivakumar, M.V.K. 1986. Climate of Niamey. Progress Report no. 1. B.P. 12404, Niamey, Niger: ICRI-SAT Sahelian Center.

Vandenbeldt, R.J., 1991. Rooting systems of western and southern African Faidherbia albida (Del.) A. Chev. (syn. Acacia albida Del.)—a comparative analysis with biogeographic implications. Agrofores-try Systems 14:233-244.

West, L.T., Wilding, L.P., Landeck, J.K., and Cal-houn, F.G. 1984. Soil Survey of the ICRISAT Sahelian Center, Niger, West Africa. Soil and Crop Sciences Department/TROPSOILS. College Station, Texas, USA: Texas A&M University.

Footnote__________

1 Resource Management Program (RMP), ICRISAT Sahelian Center, B.P 12404, Niamey, Niger. ICRISAT Conference Paper no. CP 764.

Vandenbeldt, R.J. 1992. Problems with range-wide provenance trials of Faidherbia albida on sandy soils in Niger. Pages 83-86 in Faidherbia albida in the West African semi-arid tropics: proceedings of a workshop, 22-26 Apr 1991, Niamey, Niger (Vandenbeldt, R.J., ed.). Patancheru, A.P. 502 324, India: International Crops Research Institute for the Semi-Arid Tropics: and Nairobi, Kenya: International Centre for Research in Agroforestry.