COMPETITION BETWEEN SIX HEDGEROW TREE SPECIES AND MUNG BEAN (VIGNA RADIATA (L.) WILCZEK) IN THE MID-COUNTRY INTERMEDIATE ZONE

: Tree and annual crops in agroforestry systems compete for growth resources. The trhjective of this investigation was to partition the overall tree-crop competition into above-and helow-ground components in contour hedgerow intzrcropping systems involving six tree species (C~~llirln.tlra c:olotlryrsrrs, Uesnrotliirrrr rcmsortii, FLrrn~,irt,gia ni.acro~~li.,ylla, Gliricicbic~. S P ~ I ~ I L ~ I . , Cassia. sl~ectc~l~ibis a~i~rl'7l).tlro7~.ir~ diricrsifolia) and mung hean (Vig71.u rc~diatci) grown at Pallekelle in the mid-country inteimediate zone (IMJ of Sri Lanka. Lateral extension of tree roots to the area occupied hy mung bean was prevented hy cutting 1.5 m deep trenches at 25 cm from hedgerows. Below-ground (root.) competition was estimated as the yield difference hetween trenching ( T R ) and non-trenching ( N T R ) treatments. Ahcrve-ground (shoot) competition was estimated as the yield difference hetween the yields of' a sole crop control and TR. Mung bean yield reductions due t;o a1,ove-ground competition was greater. than those due to helow-ground con~petition in all hedgerow intercrops except that involving Gliricirlin in which the two coml~onents were appr~~ximately eil~lal. Total competition (i.e. yield difference hetween control and NTR) and nhove-ground competition was highest by C:ussiu and lowest hy Oliricidia. 112 terms or absolute yield reductions, helow-ground competition was highest in CJollirrlzdra and lowest in Cassin. Wlien computed as percentages of total competition, the higl~cst and lowest helow-ground competition were shown by C:liricirlin and Cr~ssiu respect,ively. Variations Iwtween l~edgerow species in biomass p~.oducl;ion capacity, the amount trf nut.rients added to the soil t111.ougli pluniags and the syncl.lrony of their decc)rnposition with periods of nutrient demai1.d of' Lhe crop could he possible reasons f o ~ the ul~served inter-species variation of' nhove-and helow-g~ound competition.


INTRODUCTION
Contour hedgerow intercropping is a simultaneous agroforestry system1 which has been recommended to sustain annual crop yields in sloping lands of the central highlands of Sri Lanlra. It irlvolves incorporation of fast-growing, coppiceable, multi-purpose tree species as hedgerows along contours and growing annual crops in the alleys inbetween. In a previous paper" it was shown that there was significant competition between the tree hedge and the annual crop for essential resources such as light, water and nutrients and that significant variation existed between different hedgerow tree species in the degree of competition exerted. The present study reports a continuation of the above work with the following objectives: (a) To partition the overall tree-crop competition into above-ground (shoot) and below-ground (root) competition; (b) To determine whether the degree of overall competition exerted by different hedgerow species followed the same pattern as in the previous experiment; and (c) To examine whether the associated mung bean (Vigna racliata. L. Wilczek) yields have been sustained in the second season.

MATERIALS AND METHODS
Experinzentnl location: The experiment was conducted from June to October, 1998 at Kundasale (367 m above sea level) in Sri Lanlza (7-8"N & 80-81"E). The location had a distinctly bi-modal rainfall pattern of 1400 mm/ yr." The soil is a moderately well-drained sandy clay loam belonging to Rhodudult~.~ The pH, CEC and C:N ratio are 6.5,16 me/100g and 10 respectively. Dominant minerals are quartz and lraolinite with Ca and Mg being the dominant cations. The sub-soil base saturation is 25%. The respective soil water contents at field capacity and permanent wilting point are 26% and 12%." Experimcr7.tc1.t trcatrncn,ts: The expe~imental treatment structure was a two-factor factorial with tree species and trenching as the two factors laid out in a split-plot design. There were six t,ree or perennial shrub species (Callicrndm caloth.yrrsr~ Meissn., Dcsrnodiunz rurasonii (L.) DC., Fle~n.ingin macroph.ylln (Willd.) Merr., Gltrbcidia scl~iunz (Jacq.) Steud., Cassia spectn.bilis (L.) DC. and Tithonia rlivcrsifolitr.
(Ilemsl.)A. Gray in hedgerow intercrops with mull:: bean (Vigrra radinlu (L.) Wilczelr) plus a sole crop of mung bean used as the control. Within each 11~:Jgerow intercrop (i.e. main plot factor), there were trenching (TR) and no-trenching (NTR) treatments as the sub-plot factor (Fig. 1). In the trenching treatment, a 1.5 m deep trench was cut at 25 cm from the hedgerow to prevent lateral growt,h of tree roots t,o the area planted with mung bean. This treatment structure was replicated three times.
Estu,blishnlent of hedgerows aracl anrzl~nl crops: The tree hedges were established in 1992 as double hedgerows (15 cm apart) along contours in a land wit11 an even slope of lof%. The distance between two adjacent double hedgerows was 4-5 m. Mung bean (variety MI-5) was established in the alleys between hedgerows at an inter-row spacing of 30 cm and an jntra-row spacing of 10 cm. Each experimental sub-plot consisted of a hedgerow length of 5 m and its associated mung bean plot area spanning the wl~ole width of the alley.  Mungbean was sown on 23 June 1998. All hedgerows were pruned to aheight of 0.5 m on 20 June 1998 and the prunlngs were deposited between the double l~edgerows. Thereafter, the hedges were allowed to grow freely during the cropping season. Mung bean crops were grown without any additions of chemical fertilizers or organic manure. Weeds were controlled manually and pests and diseases were controlled by periodic spraying of appropriate pesticides.
Mea,suremen.ts: Vegetative growth of mung bean was measured at 50% flowering (i.e. 50 days after sowing). Measurements were done separately on samples of five randomly-selected plants from each row within the alley. Leaf area was measured by an automatic leaf area meter and total dry weight was measured after oven-drying at 80°C to a constant weight. Mung bean yield was obtained by harvesting the pods in three different picks at 70, 80 and 90 days after sowing (DAS). Yields of all crop rows were measured separately by harvesting all plants in a predetermined row length of 1 m from the central area of each plot. In addition to pod yield, yield components (i.e. number of pods per plant, number of seeds per pod and hundred-seed dry weight) were measured in each crop row. All corresponding growth and yield measurements were done on sole crop control plots also. The central 4 m length of hedgerow in each plot was pruned to a height of 0.5 m at final harvesting of mung bean. Fresh weight of pruned biomass was measured in the field immediately after pruning. A sub-sample of prunings of each plot was oven-dried at 80°C to a constant weight to obtain the dry weight of pruned biomass.
Data analy.sis: Leaf area index (L) of mung bean was calculated as total leaf area per unit land area. Total dry weight (W) of mung bean and dry biomass of hedgerows were expressed on a per ha basis. In the calculations of L and W, adjustments were made to account for the land area occupied by hedgerows.
Mung bean yield (Y) was expressed as pod dry weight per ha. Analysis of variance was used to test the significance of the effects of different tree species, trenching and their interaction. Whenever a given effect was significant at p=0.05 or below, standard error of mean was used to separate the means.
The magnitudes of shoot ( above-ground) and root (below-ground) competitions were estimated on the basis of mung bean pod yields (Y) using the following equations : where YC is sole mung bean yield; YH, and YH, , are respective mung bean yields in hedgerow intercrops with and without trenching. The above absolute estimates of-competition were also expressed as percentages of total competition (i.e. YC -YH,,,).
Inter-relationships between pod yield and yield components were estimated by multiple correlation analysis."

Meteorological conditions during the experimental period:
The total rainfall received during the experimental period (Table I), i.e. 574.8 mm, is lower than that received during the previous season, i.e. 1237 mm.2 However, the monthly rainfall exceeded pan evaporation in all months except August. Therefore, the crops did not experience any significant drought. Temperature and radiation levels did not show much variation during the experimental period.

Leaf area index (L) and total biomass (W) of mung bean:
The main effects of tree species and trenching on both L and W were highly significant (p<0.001). However, species x trenching interaction effects were not significant at p=0.05. Hedgerows of all tree species tested caused significant  growth reductions in mung bean in both trenching (TR) and non-trenching (NTR) treatments ( Table 2). The reductions in L and W due to the presence of hedgerows were greatest in mung bean grown with Cassia and Flemingia. In NTR treatments, the highest growth was shown by mung bean grown with Gliricidia and Desmodium.
With trenching, both L and W of mung bean increased under all hedgerow tree species with the highest increases being shown in mung bean with Callian.dra and Tithon,ia ( Table 2). On the other hand, the positive responses of L and W to trenching were lower in mung bean crops with Gliricidia and Desmodium,.
Mung bean yield (Y) and estimates of competition: Mung bean yields varied significantly with the associated hedgerow tree species (p<0.001) and trenching (p<0.05) treatment. The effect of tree species x trenching interaction on mung bean yield was not significant at p=0.05. Presence of hedgerows caused significant reductions in Y in both TR and NTR treatments ( Table 3). The highest and the lowest yield reductions were shown by mung bean crops with Cassia and Gliricidia respectively. Within each tree species, mung bean crops grown with trenching had greater Y as compared to those grown without trenching. The greatest positive response to trenching in terms of Y was shown in mung bean with Callian.dra whereas mung bean with Cassia and 2Ith.on.ia showed the lowest positive response (Table 3).  when considered as a percentage of the total yield reduction relative to the control, the percentage root competition was highest in mung bean with Gliricidia ( Table 3). When the total yield reduction due to the presence of hedgerows was partitioned as being due to shoot and root competition, only Gliricidia and Calliandra showed percentage root competition levels which exceeded 25%. Shoot competition was the dominant component in all hedgerow species except Gliricidia where the two components were approximately equal. The last column of Table 3 shows the corresponding pod yields obtained in the previous season2 in which trenching was not practised. Despite the lower rainfall, the sole mung bean yield has nearly doubled in the present season. Yields of some hedgerow intercrops (NTR treatments) have increased (e.g. with Calliandra and Tithonia) whereas the others have decreased. The most drastic yield reduction has occurred in mung bean with Desmodium which had the highest yield in the previous season.

Mung bean yield components:
Variation of the number of pods per plant (NP), number of seeds per pod (SP) and hundred seed weight (SW) in mung bean crops under different hedgerow tree species and trenching treatments are shown in Table 4. There were highly significant (p<0.0001) differences between tree species in all yield components.

Correlations between yield and yield components:
In both TR and NTR treatments, mung bean seed yields showed highly significant (pc0.0001) positive correlations with all yield components ( Table 5). The strongest correlation was with the number of pods per plant with r' exceeding 0.9. Moreover, all yield components showed highly significant positive correlations among themselves. Biomass production of hedgerow tree species: There were highly significant (p<0.001) differences between tree species in their above-ground biomass production during the 90-day duration of the mung bean crop. In both TR and NTR treatments, Cassia and Flenzilzgia had the highest and lowest biomass productjon levels respectively (Table 6). Tithonia, Gliricidia and Colliandra also showed higher levels of biomass which were significantly (p<0.05) greater than those of Flemilzgia and Desmodium, but significantly lower than that, of Cassia. There was a slight decrease in above-ground biomass production due to trenching in all tree species (Table 6). Despite the lower rainfall, biomass production of all tree species in the present season (in NTR treatments) was higher than in the previous season.

DISCUSSION
The main finding of this experiment was the dominance of above-ground (shoot) competition over below-ground (root) competition in the present environment. This contrasted with the findings of Singh et al.',Rao et al?,Rao et 1 . and Ong et a1.I0 who observed significant below-ground competition. However, dominance of below-ground competition could vary with the availability of below-ground resources such as soil nutrients1l and moist~re.~~,~~ These aspects of the present experiment will be reported in De Costa and Chandrapala.14 It is also possible that root systems of hedgerow species and mung bean may have occupied complementary zones in the soil profile which would have minimized the cornpetitlion for absorption of water and nutrients. Such a situation has been observed by HuxIey et al.'?n a Grevillea rohusta x maize hedgerow intercropping system. Observed variation in below-ground conlpetition between tree species could be because of the variation in the extent and depth of their root systems.l"17 Teubig (unpublished) who studied the morphology of hedgerow root systems in the present, site observed that Gliricidia had a shallow and spreading root system whereas the other tree species had deeper and less-spreading root systems. This may explain the greater percentage below-ground competition by Gliricidia. Experiments will be carried out in the next phase of this programme to provide detailed, quanti.tative infoformati.on on the root systems of hedgerows and the annual crop at the present site.
The significant inter-species variation in above-and below-ground conlpetition observed in the present experiment could indicate possible variation i11 the balance between negative competi.tion effects and positive 'fertility' effects1", I!' of different tree species on associated mung bean. The significant variation in biomass production between the different tree species could determine tjhe magnitude of overall competjtion exerted.'K2" For example, the highest overall, conlpetition effect was shown in mung bean grown with Ca,.ssia, which had the highest biomass production. The demand, for both above-and be1.o~-ground resources is likely to be greater with higher biomass production, thus intensifying competition on the annual crop. However, among the other tree species, similar magnitudes of overall competition have been observed in mung bean crops grown with tree species which had large differences in biomass prod.uction such as ZIth.o7zia and Flen~t7zgicr.. This could be d.ue to the differences in the fertility effect of the above two species. Firstly, the large amount of prunirigs added to mung bean crops grown with nthonia (which had a higher biomass production) would have added a greater amount of nutrients on decomposition. In contrast, the fertility effect of mung bean associated withFlemi7;lgio (which had a lower biomass production) would be con~paratively low. Secondly, De Costa and AtapattuZ1 observed in a separate study that the rate of decomposition of ?Zth.onia mulch is much faster than that of Flemingia. Hence, both the amount of prunings added as mulch and the degree of synchrony of nutrient release with the nutrient demand of the annual crop"p 23 can determine the fertility effect of a given hedgerow species on the annual crop.
A secondary objective of this experiment was to compare the degree of competition exerted by different hedgerow species in the present experiment with that observed in the previous season as reported in De Costa and Chandrapala." Comparison of the estimates of overall tree-crop interaction showed some notable differences between the two seasons such as Desmodiunl changing from a substantially positive to negative interaction. However, in terms of absolute mung bean yield, except for the crops with Desmodium, the comparative ranking of hedgerow species did not differ much between the two seasons.
The magnitude of tree-crop competition and the level of annual crop (i.e, mung bean) yield could change from season to season because of several factors. These include variation jn growth of different tree species, the level of resources (e.g. nutrients) removed for tree and crop growth in the previous season and the level of resources added to the system (e.g. tree prunings and crop residues) from the previous For example, plots with Desmodium had an extremely high mung bean yjeld in the previous season and removed a high amount of nitrogen"; while providing only a small amount of prunings from Desnzodir~nz. This could be the reason for the substantial decrease of mung bean yield in plots with Desnaodiun~ in the present season. Therefore, long-term observations of crop yields are needed to ascertain the sustainability of a hedgerow intercropping system involving a specific tree species.
The dominance of above-ground compet~tion observed in the present experiment nleans that appropriate hedgerow management practices should be formulated to ~ninimize shoot competitio11. These could include pruning of hedgerows in the mlddle of the cropping season instead of waiting until the harvestmg of mung bean.