Food Sec. (2012) 4:103–114 DOI 10.1007/s12571-012-0165-4 ORIGINAL PAPER Effects of light environment on maize in hillside agroforestry systems of Nepal Thakur Prasad Tiwari & Robert M. Brook & Paul Wagstaff & Fergus L. Sinclair Received: 22 October 2011 / Accepted: 2 January 2012 / Published online: 28 January 2012 # Springer Science+Business Media B.V. & International Society for Plant Pathology 2012 Abstract Maize (Zea mays L.) is the most important staple food in the mid-hills region of Nepal. The mid-hills are characterized by steeply sloping land and complex farming systems where crops, livestock and trees are inseparable components, and maize has to compete with trees grown for fodder, fuel wood, building materials and other purposes in a landscape severely constrained for agricultural purposes. This paper reports the effects of the presence of trees growing on crop terrace risers on bari (upper-slope, rainfed) land on growth and yield of maize grown on terrace benches. Maize performance was compared with and without tree and artificial shade to determine its responses above and below ground to such limiting factors. Mean photosynthetic photon flux density (PPFD) incident on maize in farm conditions was lower than 700 μmol m-2 s-1, well below the light saturation point for maize (1,500 μmol m−2 s−1). Grain yield was reduced by 33% under tree shade and by 43% under artificial shade compared with natural (unshaded) conditions. As the light environment is sub-optimal for maize, the crop rarely achieved maximum rates of T. P. Tiwari (*) CIMMYT, Dhaka, Gulshan, P.O. Box 6057, Bangladesh e-mail: [email protected] R. M. Brook : F. L. Sinclair School of the Environment, Natural Resources and Geography, College of Natural Sciences, Bangor University, Bangor, Gwynedd LL57 2UW, UK P. Wagstaff Concern Worldwide 52/55 Lower Camden Street, Dublin 2, Ireland F. L. Sinclair World Agroforestry Centre, Nairobi, Kenya photosynthesis. Farmers claim that local landraces are better adapted to shade than station-bred genotypes, but there was no evidence of varietal effects upon rates of photosynthesis. However, there was some evidence that there were varietal adaptations to shade for other factors such as greater numbers of leaves and more competitive rooting patterns. Maize varieties with deeper root systems and adapted to low light conditions are required if productivity in these complex systems is to be improved. The findings of this study should be useful to breeders in developing maize genotypes suitable for the complex hillside systems of Nepal, thereby improving food security. Keywords Maize . Zea mays . Agroforestry . Photosynthesis . Hillside systems . Shade Introduction Maize (Zea mays L.) is a key staple in the mid-hills of Nepal, second only to rice (Oryza sativa L.) in terms of national food security. Finger millet (Eleusine coracana Gaertn.) is relay cropped with maize on the terraces where fodder trees grow on the risers of bari land (upper slopes, rain-fed, often terraced with fodder trees on the risers) in the hillside systems of Nepal. More than 90 (Thapa 1994) and 111 (Carter 1992) species of trees for fodder and fuel have been recorded on the farmlands of the eastern and central mid-hills of Nepal, respectively. These authors recorded a number of positive as well as negative aspects concerning the hillside bari land system where maize is grown. The negative aspects identified include competition between trees and crops for nutrients, water and light. The positive aspects include increased productivity of the whole system, largely through fodder and fuel wood provision and meeting 104 farmers’ multiple objectives of erosion control and maintenance of soil organic matter through leaf litter decomposition and the application of farm compost. Previous workers have reported that shade affects maize growth and development negatively (Earley et al. 1966; Midmore et al. 1988; Reed et al. 1988; Andrade et al. 1993; Ephrath et al. 1993; Chamshama et al. 1998); and that there were varietal differences in adaptation to shade (Duncan 1980) and rate of photosynthesis (Duncan and Hesketh 1968; Gaskel and Pearce 1981). Characterization of the climatic and agronomic context in which farmers grow maize in the hillside systems revealed a complex and heterogeneous environment, dominated by variation in farmers’ practices (Tiwari et al. 2004). Farmers in the mid-hills of Nepal appear to be prepared to trade off maize and finger millet yield penalties against tree products such as fodder and fuel wood. This study formed a component of a programme designed to improve maize production in the mid-hills zone of Nepal. Previous papers have described farmers’ agronomic growing practices which are peculiar to this zone (Tiwari et al. 2004) and the participatory variety selection programme conducted with farmers, using farmers’ own selection criteria (Tiwari et al. 2009b). In response to farmers reporting that some maize landraces were more tolerant of shade than newer, research station bred genotypes (Thapa 1994; Tiwari et al. 2009a,b), this paper reports research which sought to characterise the light environment experienced by maize on bari terraces and test the responses of differing maize genotypes to shade, both induced by artificial shade material and by trees growing along terrace risers. Materials and methods Site description Three locations were selected for on-farm micro-climatic studies, at Marga and Patle in Dhankuta district and Fakchamara in Tehrathum district. They were chosen as representative of the range of conditions used for maize cultivation in the eastern mid-hills of Nepal and were located between 1,200 and 1,500 m altitude (Table 1). Measurements were carried out in five farmers’ fields at each site in 1999. The sample fields were widely distributed throughout the villages and were of variable sizes, ranging from 500 m2 to 2,000 m2. They were located on terraces with trees being grown on the risers and captured the full range of biophysical environments that farmers use to grow maize. Six quadrats measuring 3×3 m within the selected fields were selected immediately after maize was sown to represent the full range of light environments experienced by maize in this farming system. T.P. Tiwari et al. To discriminate between effects of shade alone and that of trees, which would include shade but confounded with below ground competition, both artificial and natural shade experiments were established in 2000. These were conducted at the Agricultural Research Station, Pakhribas (ARSP) in 2000, at 1,740 m altitude, located 27° 17′ N and 87° 17′ E. For the artificial shade experiment, a split-plot design was adopted with two shade regimes: no shade (unshaded) and artificial shade (shaded) as main plots, and three Nepalbred maize varieties (PM-1, PM-3, Manakamana-1) and Madi white (a local landrace) in sub-plots, randomized within the main plots and replicated four times. The whole terrace was sown to maize to create a continuous canopy, within which treatment plots were demarcated. Sub-plot size was 2×2 m, and main plots were 8×2 m. Seeds were primed to help ensure good germination (Harris et al. 1999) and were sown on 4 April, 2000 at 50×50 cm spacing, giving 16 maize plants in each sub-plot (40,000 plants ha −1 ). Manakamana-1 maize was first introduced as a maize variety for the mid-hills zone in 1986, and is widely cultivated in this region, although much of the crop is established from seed saved on the farms, which has experienced several generations of cross-pollination with local varieties (Tiwari et al. 2004). Madi white is one such local ‘variety’, and this genotype was regarded as being the local check. PM1 and PM3 were preliminary selections as part of a programme to develop new composites for testing by farmers. Chemical fertilisers, N as urea and P as P2O5, each at the rate of 30 kg of element ha−1 plus 15 tha−1 farm yard manure (FYM) were applied as a basal dressing. Another 30 kg ha−1 N was added as a top dressing on 31st May, 2000 (when the maize was at about knee high). Shade was imposed using green, loose-woven polypropylene, suspended over a bamboo frame 10 m long×4.5 m wide, which allowed 60% transmission of photosynthetically active radiation (PAR). The frame was suspended from vertical poles at just above crop canopy maximum height and was raised as the crop grew, up to a maximum height of 2.5 m. To limit ingress of lateral solar radiation, shade material was also hung down each side as a 500 mm curtain. Finger millet was transplanted at a rate of 100 seedlings m−2 on 5 July under the maize to simulate the relay cropping used by farmers, but no recordings were made on this crop. This experiment was established on a wide, treeless terrace, which at the steeply sloping ARSP is at a premium, in constant use for agronomic experiments and thus subject to regular fertilization. For the natural shade experiment, narrower terraces more representative of hillside bari land and the maize agroforestry systems (Plate 1) were used, one bounded by mixed fodder tree species and two adjacent terraces without trees. The same maize varieties, degree of replication and fertilizer Effects of light environment on maize in hillside agroforestry 105 Table 1 Description of study sites Biophysical parameters\sites Marga Patle Fakchamara On-station (ARSP) Altitude (m asl) Aspect 1,250 to 1,400 South-east 1,300 to 1,500 South-east 1,200 to 1,400 North-east 1,740 South Major cropping patterns Maize/finger millet followed by fallow, or Maize/finger millet followed by fallow or Maize-finger millet followed by fallow or Maize/finger millet followed by fallow or Maize/finger millet followed by mustard Maize/finger mille followed by mustard Maize-finger millet followed by mustard or Maize/finger millet followed by wheat or mustard Maize-finger millet intercrop followed by potato Size of terraces Tree species on terrace rises Soils Contour terraces across Contour terraces across Contour terraces across 9–10 m wide and slopes typically 5–7 m slopes typically 6–8 m slopes typically 7–10 m 40–50 m long wide and 50–70 m wide and 50–70 m wide and 70–100 m long long long Patches or lines of mixed timber (Alnus nepalensis, Schima wallichii, Pinus wallichiana) and fodder tree species (typically Bauhinea perpurea, Ficus nerifolia, Ficus semicordata, Albizia julibrissin, Celtis australis, Litsea polyantha, Bambusa spp.) The top soils are medium to The top soils are brown The soils are deep on moderately steep slopes and strongly acidic to deep and dark brown to developed on sloping terraces; dark to reddish medium, the pH ranges reddish brown and sandy brown in colour and the texture sandy loam from 5.0 to 6.0. The majority loam to loam in texture. to loam whilst the sola are deep and clay loam of soils are shallow to Soil pH ranges from 5.9 in texture. The organic matter is at medium medium in depth developed to 6.5. The organic matter level. on bench terraces. They content varies from 1.6 to are sandy loam and well 3.14%. The soils are drained. Organic matter underlain by reddish soil and total nitrogen content with sandy loam to sandy are at medium level. clay loam. The available phosphorus (Olsen-P) and potassium (Ammonium Acetate Extraction) varies from medium to very high and high to very high at all the sites, respectively. In the cropping patterns sign (/) denotes relay crop applications were used as for the artificial shade trial. Due to the land constraints on the research station, blocks were nested within sites. On-farm light environment Photosynthetic photon flux density (PPFD) was measured on five of the sample farms at each of the Patle and Marga sites using a Sunscan Canopy Analyser (Delta-T Devices, Burwell, Cambridge, UK). Measurements were taken from the central part of each 9 m2 quadrat, above and at the base of the maize canopy. Measurements started 1 month before maize sowing and then at 30 day intervals until the maize reached physiological maturity, with five measurement times during each sampling day at two-hourly intervals commencing at 8 a.m. The fractional interception of PPFD by maize (ƒ) was calculated using the following formula: f ¼ 1 ðI=OÞ Plate 1 Typical hill terraces where maize is commonly grown Where: I0PPFD measured below the maize canopy and O0 PPFD measured above the canopy. In order to determine whether the light quality environment was similar under tree canopies and under artificial shade, the red:far-red ratio (r:fr) was measured using an 106 SKR 110; 660:730 nm sensor (Skye Instruments, Llandrindod Wells, Powys, UK). The measurement schedule and procedure were the same as those used for the PPFD measurements in farmers’ fields in 1999, and the same procedures were used to measure PPFD and r:fr ratio in the artificial shade experiment in 2000. Photosynthesis Farmers contend that local maize varieties are more stress tolerant (e.g. tolerant to competition from trees, etc.) than introduced varieties. To determine whether there was a physiological basis for this, net photosynthesis was measured on maize plants grown at ARSP using a portable infrared gas analyser (IRGA) model LCA3 (Analytical Development Co. Ltd., Hoddesden, Herts., UK) at a wide range of natural photosynthetic photon flux densities, resulting from the mixed sunny and cloudy conditions that often prevail during the rainy season. Due to equipment and time constraints, recordings of photosynthesis were concentrated on the artificial shade experiment in 2000, supplemented by photosynthesis data collected from maize grown on a treelined terrace in 1999. A Parkinson leaf chamber with a cuvette window of dimensions 20 mm × 60 mm was clamped over the uppermost fully expanded maize leaf, taking care to fill the leaf chamber window but to avoid the mid-rib. The leaf chamber was held horizontally. Air to both the IRGA and the leaf chamber was supplied by an integral air supply unit which drew its reference air down a tube with its orifice positioned above the shades and crop canopies, well away from the zone of measurement. The air was dried by passing it through two columns of silica gel. Once the leaf chamber was clamped over the leaf, the differential mode of the IRGA was used, which compared the volumetric concentrations of CO2 and H2O in the reference and measurement streams of air (in volumes per million). When a steady reading for the differential in CO2 concentrations was obtained (normally after about 1 min), it was recorded. Readings were taken when the maize in the experiment was at approximately the seven leaf stage, on 28 and 29 June, 1999 and 16 to 18 May, 2000 (due to severely delayed onset of rains, the crop was established 6 weeks later in 1999 than in 2000). In 1999, two maize genotypes, Madi white (local landrace) and Manakamana-1, growing on a 10 m wide terrace with mixed fodder tree species established along one side, were recorded. Forty randomly selected plants of each genotype, divided equally into shaded and non-shaded sample groups, were used for measurements; thus, 80 plants were sampled. Recordings were made between 11.30 am and 2.10 pm, when the sun was at or close to its zenith. In the event, conditions were predominantly overcast and photon flux densities rarely approached those close to light saturation. T.P. Tiwari et al. In 2000, plants from the artificial shade experiment only were measured, because shading conditions were less variable than natural tree shade. Eleven plants of four varieties were sampled with and without shade; thus 88 plants in all were measured. Recordings were made between 11.30 a.m. and 2.10 pm. The Mitscherlich model, an exponential curve rising to an asymptote often fitted to photosynthesis curves (Peek et al. 2002) was fitted to the PPFD response curve obtained from the non-shaded plants, using Sigma Plot v. 8 (SPSS Inc.). This took the form: y ¼ a 1 ebx where a is the maximum rate of net photosynthesis (the asymptote), b is the initial slope of the curve (the initial quantum efficiency) and x is PPFD. In the case of overcast conditions in 1999 and shaded plants in 2000, PPFD never approached light saturation for maize, a plant with the C4 pathway of carboxylation. Thus, where appropriate, a linear response was fitted instead. Maize growth From both the artificial and natural shade experiments, time from sowing to 50% tasseling, 50% silking and 95% maturity were recorded. At maturity, plant height, number of nodes per plant, number of ears per plant and grains per ear, degree of turcicum blight (Exserohilium turcicum (Pass) Leonard and Suggs) infection, lodging and barrenness were recorded. Plot grain yields were weighed at 15% moisture content. Residuals from data for degree of turcicum blight infection, lodging and barrenness were not normally distributed, so were square root transformed prior to statistical analysis. In the artificial shade experiment five randomly selected plants per sub-plot were tagged for recording of number of leaves, Aleaf (leaf area) and plant height at 10 day intervals starting from 45 DAS until vertical extension ceased. For consistency plant height was measured from the ground level to the tip of the longest leaf. Leaf numbers 4, 7, 10, 13 and 16 from the five randomly selected plants were measured. Aleaf was calculated using the following formula: Aleaf ¼ 0:75 L B where L0length from base to the tip of a leaf, B0breadth from the widest point of a leaf, and 0.750constant (Maddoni and Otegui 1996). Maize root systems Maize root development was monitored only in the artificial shade experiment. Due to resource limitations, roots were Effects of light environment on maize in hillside agroforestry sampled only twice, at 48 DAS and at 68 DAS using a 6.3 cm internal diameter mild steel corer, made locally using 2 m long plumbing pipe sharpened at one end and reinforced with a wooden anvil at the other end for hammering. A hole was drilled towards the upper end of pipe to insert a rod used to remove it from the soil with a twisting motion. Soil cores were pushed out from the corer using a 1.5 m length of bamboo in increments of 10 cm for four rooting zones 0–10, 10–20, 20–30 and 30–40 cm at 48 DAS. The sampling depth was extended to 40–50 cm depth at 68 DAS. The soil cores were stored at 4°C until they could be washed. The roots were separated from the soil cores by the flotation technique using a root washer. Organic material washed out of the soil cores was preserved in dilute acetic acid and refrigerated at 4°C until it could be analysed. The roots were then hand sorted to separate maize roots from those of other plants and organic material, with the help of a binocular, ×10 dissecting microscope. The number of intersections between lines on a grid and living roots were counted. Tennant’s Modified Line Intersect Method (Tennant 1975) was used to determine the total root length of each sample, where Root length ¼ 11 grid size number of intersections 14 where 11 and 14 are constants. Two grid sizes were used, depending on number of roots in the sample, 4.5×4.5 mm and 10.5×10.5 mm. Statistical analyses Analyses of variance were performed using Genstat v. 3.2 (Lawes Agricultural Trust, Harpenden, UK), data from the artificial shade experiment being treated as a split plot design, whilst for the natural shade experiment, blocks were nested within each site (shaded or non-shaded). Thus, it was not possible to conduct a combined analysis of the two shading experiments. To compare treatment means, the standard error of the difference (s.e.d.) was computed. Results and discussion Light environment Mean diurnal PPFD incident (IPPFD) throughout the day upon the maize averaged over the season was 693 μmol m−2 s−1 onfarm in 1999 (Fig. 1) and 788 μmol m−2 s−1 in the open and 455 μmol m−2 s−1 under artificial shade on-station in 2000 (Fig. 2). In all cases, about 50% of IPPFD was intercepted by the crop. IPPFD values were well below the light saturation point of maize leaves of about 1,500 μmol m−2 s−1 (Stirling et al. 1993), and were similar to those recorded by Prasad and 107 Brook (2005) in the rainy season in the central mid-hills of Nepal. Even in the open, mean incident light (IPPFD) levels on farms was well below the saturation point for maize, due to cloudy conditions during the rainy season, and was further reduced by farmers growing fodder trees on crop terrace risers. It is the red:far red ratio (r:fr) rather than total radiation that controls photomorphogenetic responses in plants (Morgan and Smith 1981; Smith 1986). However, studies of r:fr per se upon maize are very few, and none has been reported from field conditions. The mean r: fr ratio recorded in the 1999 season under on-farm conditions was 1.07 in 1999. In the 2000 season, it was 1.11 in the open and 1.06 under artificial shade. Thus, the environment created by the netting was close to that created by the natural shade with respect to quality of the light. McCalla et al. (1939) cited by Earley et al. (1966) reported that shade treatments alter the quality of incident light and this in turn might alter hormonal balances which could alter plant height and other attributes. In this experiment, artificial shade significantly reduced plant height (Table 2), so there was no evidence of etiolation. Therefore the evidence suggested that the reduction of total incident PPFD exerted a greater effect upon maize growth than any change in r:fr ratio. In addition to effects upon PPFD, shade, either artificial or natural, exerts other effects upon the crop including humidity within the canopy and temperature of leaf surfaces. During significant dry periods, tree roots will also compete with the crop for water, although in the seasons during which the present work was conducted, rainfall was abundant. Photosynthesis The net photosynthesis rate of maize leaves rose with increasing PPFD, reaching a plateau at approximately 1,500 μmol m−2 s−1 for those unshaded. All regressions, fitted with the Mitscherlich model were highly significant (P<0.001) and yielded a coefficient for maximum rate of photosynthesis of 29.3 μmol CO2 m−2 s−1 for maize growing in association with natural shade in 1999, and 28.2 μmol CO2 m−2 s−1 for non-shaded maize in the artificial shade experiment in the year 2000 (Fig. 3). This model yielded low values for initial quantum efficiency (0.0013 and 0.0011 μmol CO2/μmol photons for 1999 and 2000 respectively). CO2 exchange rates measured on leaves of shaded maize plants never reached light saturation. For the sake of consistency, the Mitscherlich model was fitted to data from shaded maize leaves in 1999, but in the case of artificial shade in 2000, a linear regression provided a better fit (r2 0 0.91) and the quantum efficiency for the linear phase of the response to PPFD was estimated to be 0.024 μmol CO2/ μmol photons (Fig. 3). Ephrath et al. (1993) also reported similar findings. When PPFD was recorded in farmers’ 108 T.P. Tiwari et al. Fig. 1 Incident photosynthetic photon flux density (PPFD) recorded in on-farm condition under mixed fodder tree species (natural shade) and level of interception at different stages of maize growth, mean of ten farms, 1999. BS before sowing fields, mean light levels were sub-optimal for maize, regardless of whether or not they were growing under tree shade. The implications of this finding, interpreted alongside the light response curves of maize, are that even small reductions due to tree shading would have a large impact on photosynthesis because the maize leaves are operating in the linear portion of their light response curve. The results also showed that there were no varietal differences in photosynthesis in maize, so any indication of adaptation to shade, such as occurs in some species (Johnson et al. 1995; Peek et al. 2002) was absent. However, others have reported evidence of genotypic variation in rates of photosynthesis at light saturation in maize (Duncan and Hesketh 1968; Gaskel and Pearce 1981), but this variation was not apparently exhibited in the genetic material tested in the experiments described here. Therefore, any varietal adaptation to shade as reported by farmers must be due to other factors. Maize growth Leaf area (or its integral over land area, leaf area index) is an important determinant of productivity in maize (Prasad and Brook 2005), particularly in the light limited environment Fig. 2 Incident photosynthetic photon flux density (PPFD) recorded in open and under shade (artificial shade) and level of interception at different stages of maize growth, mean of four varieties, 2000. BS before sowing experienced during the rainy season in Nepal and especially under tree canopies. Therefore, the development of leaf area to capture solar radiation is important. The components of leaf area per plant are area per leaf (Aleaf) and number of leaves per plant. Despite the lack of shade adaptation in photosynthetic response in maize, the development of leaf area was not affected by natural shade (Table 3) (P>0.05), but was by artificial shade (Table 4) (P<0.05). Under artificial shade, mean Aleaf per leaf was about 10% lower than in plants grown in the open. Except for a few cases (var. PM-1, leaf numbers 13 and 16) unshaded plants always produced larger leaves (data not shown). Ephrath et al. (1993) found that shading decreased leaf size, but the effect was minimal. Wilson et al. (1998) also found a similar result in sorghum (Sorghum bicolor), where green leaf and stem mass of shaded plants were significantly lower than those of semi-shaded or those in the open. Leaf size increased until leaf 13 in all varieties and above this point leaf size decreased. Varietal differences were observed in mean Aleaf and this demonstrated that variety and shade affected leaf size, as also found by Maddoni and Otegui (1996). In the artificial shade experiment, number of leaves until 85 DAS were fewer under shade but they caught up by 95 DAS and shaded plants had more leaves at 105 DAS (P<0.05) Effects of light environment on maize in hillside agroforestry Table 2 Effect of artificial shade on plant height (cm) in the year 2000. Shade consisted of green, loose-woven polypropylene, which allowed 60% transmission of light, hung over a bamboo frame 10 m long×4.5 m wide. Shade material hung down each side in a 50 cm curtain, and the shade was raised as the crop grew Statistical results Days after sowing (DAS) 45 DAS 55 DAS 65 DAS 75 DAS 85 DAS 95 DAS 105 DAS No shade (Unshaded) 93 151 207 250 292 316 318 Artificial shade (Shaded) 68 112 159 200 243 274 296 Mean F-test 81.0 *** 131.0 *** 183.0 ** 225.0 ** 268.0 ** 295.0 ** 307.0 * s.e.d. 1.7 2.5 4.7 6.1 8.2 6.9 5.1 86 81 139 132 192 184 237 227 285 271 316 299 334 313 V30Mana-1 V40M/local 70 86 115 139 161 194 196 238 233 282 249 316 263 319 F-test ** *** *** *** *** *** *** s.e.d 4.0 5.4 6.4 6.5 6.5 8.2 7.9 Variety effect, V10PM-1 V20PM-3 ***0P<0.001; **0P<0.01; *0P<0.05. Interaction effects were not significant, so no data presented 109 (Fig. 4). Conversely, Midmore et al. (1988) found that the leaf numbers in maize were not modified by shade. There were significant variety effects (P < 0.001) (as also found by Duncan 1980; Robertson 1994), where Manakamana-1, an early maturing variety had fewer leaves in both no shade and shade treatments, followed by PM 3 (Fig. 5). Early maturing varieties generally have fewer leaves and are shorter plants (Duncan and Hesketh 1968; Maddoni and Otegui 1996). In the context of maize culture in the mid-hills of Nepal, the stover of maize has the significance of being an important fodder supply for livestock (Tiwari et al. 2004), so identification of any varietal traits to maintain a higher leaf area under shade may be important and might have been one criterion used by farmers to assess shading tolerance of maize (Tiwari et al. 2009a,b). Unlike many less determinate crops, maize does not exhibit plasticity in development of leaf area in response to population density (Prasad and Brook 2005), so Fig. 3 Relationship between net assimilation rate and photosynthetic photon flux density (PPFD) in maize, 1999 and 2000 establishment of a denser population at harvesting time appears to be one means to increase total maize productivity per unit land area. However, in the hill farming systems of Nepal, maize is usually under-cropped with finger millet or soybean, and denser maize establishment has a negative effect upon these (Prasad and Brook 2005). Therefore, proper or suitable plant populations for each crop combination needs further investigation. Nepali hill farmers also progressively thin maize crops throughout the season to provide livestock fodder and to optimize plant population (Tiwari et al. 2004). Agronomic attributes The two main attributes of interest to Nepali farmers are grain yield and fodder production, the stover being fed to livestock after grain harvest. Mean grain yield under natural 0.82 s.e.d. † original data transformed before analysis of variance, therefore s.e.d. presented are from transformed data. Transformed data are in parentheses. There were no significant interaction effects ns 16.5 1.7 ns *** 9.5 11.7 ns ns 0.07 0.344 ns *** 11.0 17.07 ns ns 0.686 0.0867 ns ns 0.94 ns F-test ns 146 82 M/local (V4) 90 143 88 81 Mana-1 (V3) 1.6 407 395 415 409 12.1 333 211 0.90 2.76 285 5,625 (62.1) 8.0 (2.17) 2.5 (1.577) 11.4 269 235 0.95 2.78 227 3,438 (53.5) 4.0 (1.38) 2.4 (1.556) 10.9 15.1 316 297 224 230 0.93 0.89 2.51 2.75 289 268 6,875 (68.3) 3,438 (49.8) 7.0 (2.45) 6.5 (1.73) 2.6 (1.596) 2.1 (1.435) 144 144 92 82 90 82 PM-3 (V2) Varieties PM-1 (V1) ns 11.6 1.2 ns ns 6.7 8.3 *** * 0.05 0.243 *** ns 7.7 12.1 ** ns 0.0434 ns *** *** 0.58 1.2 *** s.e.d. 149 94 85 140 87 78 No shade (Unshaded) Tree shade (Shaded) F-test Shades 0.67 12.1 298 199 0.86 2.16 261 6,719 (76.2) 7.6 (1.21) 309 251 0.97 3.24 273 2,969 (40.7) 5.0 (1.63) 2.4 (1.546) 2.4 (1.536) 416 12.7 Biomass t ha−1 1,000 grain weight (g) No. of grains ear−1 No. of ears plant−1 Grain yield t ha−1 Plant height (cm) † barren plants ha−1 † lodging (%) †Turcicum blight (1–5 scale) 95% maturity (DAS) 50% silking (DAS) 50% tasseling (DAS) Characters treatments Table 3 Effect of natural shade on maize varieties and their agronomic attributes in the year 2000 397 T.P. Tiwari et al. Leaf size plant−1 (cm2) 110 tree shade both on-farm in 1999 (Tiwari et al. 2004) and onstation in 2000 were the same (2.16 tha−1), and represented yield decreases of 37% in 1999 and 33% in 2000, compared to maize plants growing further away from trees (Tables 3 and 4). On the other hand, whereas there was a large penalty in biomass production under trees on-farm in 1999 (Tiwari et al. 2004), the difference on-station in 2000 was small and non-significant (Table 3). Farmers trade-off lower maize yield for the fodder supplied by trees. To compensate for this loss, other work as part of this research programme showed that farmers exhibited a preference for higher yielding maize, even though these types were longer in duration, as long as they exhibited a degree of tolerance to tree shade (Tiwari et al. 2004). Grain yield and biomass production in unshaded treatments in the artificial shade experiment were much greater (Table 4) than the natural shade experiment (Table 3), both conducted in 2000. The only terraces on the research station wide enough for an artificial shade experiment away from the influence of trees had been used intensively in previous seasons for horticultural experiments, and had a long history of heavy fertilization, whereas narrower terraces influenced by tree shade were used less and consequently fertilized moderately. Therefore, it was perhaps more appropriate to compare proportional response to shade than absolute grain yields. Artificial shade reduced grain yield by 43%, a figure in proportion to the 40% of light intercepted by the shade netting, whilst shade due to trees reduced on-station maize yields by 33%. However, leaving aside the issue of below ground interactions, the nature of tree shade is rather different from that of artificial shade cloth, so explaining such differences in proportional response is difficult. Natural shade on-farm (except at Marga) in 1999 (data presented in Tiwari et al. 2004) and artificial and natural shade on-station in 2000 significantly delayed tasseling (P< 0.001), silking (P<0.001) and maturity (P<0.001) (Tables 3 and 4). There were no varietal effects in the 2000 season on time to 50% tasseling, neither was there an interaction effect, showing that varieties displayed the same trend in both the open and under shade. Unlike tasseling, variety did have an effect on time to 50% silking (P<0.05) and maturity (P<0.01) under artificial shade. Manakamana-1 was the earliest, but this only differed significantly from PM-1 (Table 4) with respect to days to silking. There were no interaction effects for these traits either. Shade, both natural and artificial, resulted in fewer grains per ear. There might have been floret or kernel abortion due to insufficient plant assimilates under shade after fertilization. Tollenaar (1977, cited by Andrade et al. 1993) reported that the amount of intercepted radiation at flowering stage is critical for kernel set. If it is limited at flowering, ear abortion occurs immediately, whereas kernel abortion can continue up to 20 days after pollination. Effects of light environment on maize in hillside agroforestry 111 Table 4 Effect of artificial shade on maize varieties and their agronomic attributes, 2000 Characters treatments 50% tasseling (DAS) 50% silking (DAS) 95% maturity (DAS) Grain yield (t ha−1) Plant population (ha−1) No. of ear plant−1 No. of grain ear−1 1,000 grain weight (g) Biomass (t ha−1) Leaf size plant−1 (cm2) Main plot (shade) No shade (Unshaded) Under shade (Shaded) F-test 82 88 142 5.31 39,800 1.2 319 332 26.0 521 92 99 153 3.01 39,800 0.9 242 310 16.9 470 *** *** *** ** NS * ** ns *** ** s.e.d. 0.32 0.64 0.33 0.381 255.2 0.05 12.0 10.8 0.66 8.69 506 Sub-plot (variety) PM-1 (V1) 87 95 147 4.21 40,000 1.11 268 327 22.5 PM-3 (V2) 86 94 148 4.38 39,688 1.12 287 326 21.8 501 Mana-1 (V3) 86 92 145 3.99 39,688 1.03 295 285 16.6 479 M/local (V4) 87 93 149 4.08 40,000 1.08 274 346 25 498 F-test ns * ** ns ns ns ns *** *** ns s.e.d. 0.74 0.99 0.90 0.279 312.5 0.07 15.9 7.5 1.22 16.2 ns not significant Maize root systems The non-shaded treatment had greater root length than the artificial shade treatment on both recording dates (48 DAS and 68 DAS) (Fig. 6), which is consistent with findings of other researchers (Ephrath et al. 1993; Huxley et al. 1994). Found that under light limited conditions (50% of the normal light, which is almost equivalent to the artificial shade treatment in the present study) the growth of roots was reduced, which led to a decrease in the root/shoot ratio. Mean root length density (RLD) was greatest in the 10– 20 cm horizon, followed by 0–10 cm at the first recording (48 DAS), and 10–20 cm, followed by 20–30 at the second (68 DAS) (data not shown). This pattern of higher root biomass in upper soil layers has been well documented in maize by Jose et al. (2001). A common problem with root coring studies are large errors arising from sampling, combined with the considerable Fig. 4 Mean number of leaves at specific time intervals as affected by artificial shade. Error bars smaller than the point size are not visible amount of effort and time needed to acquire data by this means, which necessarily reduces sample size. Therefore, trends will be discussed here as indicators of genotype and shading effects, in line with what is known from local knowledge. Although not significant, Manakamana-1 had greater root length than other varieties (Madi local, PM-3 and PM-1) at 48 DAS, but less at 68 DAS. The higher root length at 48 DAS was mainly due to the fact that Manakamana-1 matured earlier and therefore roots grew more quickly than in other later maturing counterparts. In contrast, later maturing varieties showed slow initial root development but overtook Manakamana-1 by 68 DAS. The root distribution patterns confirmed the results obtained in 1999 from local knowledge acquisition that local maize varieties have deeper rooting systems than more recently released varieties, and in Bangor, UK where five Nepalese varieties (Manakamana-1, Local white, Local yellow, Manakamana-3 (formerly called Population-22, Tiwari et al. 2009b) and Gadbade (a mixed 112 T.P. Tiwari et al. Fig. 5 Effect of artificial shade and genotype upon average maximum number of leaves of maize, Pakhribas 2000 type)) were grown in a rhizotron experiment in a glasshouse over an 8 week period in 1998 showed similar trends (data not shown). In summary, this work has shown that there was no evidence for a difference in photosynthetic adaptation to shade amongst the varieties tested but there were indications that local germplasm developed a longer (possibly deeper) root system early in the growth cycle under shade than more recently released varieties. This could confer drought tolerance, which might explain farmers’ perceptions that local varieties are drought tolerant. However, this hypothesis requires further testing. Additionally, there was evidence that some genotypes can compensate for sub-optimal light conditions by growing more leaves. This work reinforced findings that environments in which maize is grown in the hillside systems are variable and very different from the conditions on research stations where new germplasm is usually selected and recommendations are generated. Conventional maize breeding programmes are often based in non-stressed environments and therefore give results that are frequently not representative of target enviFig. 6 Effect of artificial shade and genotype on root length density of maize verieties, at two growth stages, Pakhribas 2000 ronments. Given the heterogeneity in conditions on farms, flexibility in both the germplasm available to farmers and the recommendations that accompany it are necessary to cater for the diverse environments in which the crop is grown. The selection of better root morphology, if feasible, could improve both lodging and drought tolerance. As there appears not to be any capacity of the maize photosynthetic apparatus to compensate for low levels of incident light then selection for a larger leaf area per plant, so increasing the opportunity to capture solar radiation, appears to be a more promising strategy. This also has the benefit of producing more fodder for livestock. Conclusions Maize is the principal staple food of the population of the mid-hills of Nepal. Yet, this study has shown that farmer practices and the marginal environment of the hillside systems impose significant constraints on maize productivity, and indicate that presently available maize germplasm is Effects of light environment on maize in hillside agroforestry poorly adapted to them. The implications for food security in the mid-hills in particular is that because the growing environment is so different from that which is optimal for maize production, breeding and agronomy programmes need to be conducted in the agro-climatic conditions experienced by the crops if locally adapted and robust planting material is to be developed. Acknowledgements We thank the farmers of Marga, Patle and Fakchamara for their co-operation and inputs in running on-farm trials. Help from Messrs P. B. Baruwal, P. P. Poudel, R. B. Katuwal, D.P. Sherchan and P. P. Khatiwada is gratefully acknowledged. We also benefited from interactions with Dr Joel Ransom, a CIMMYT Agronomist based in Nepal at the time of this research. We would also like to thank Drs. Kevin Pixley, and Marianne Banziger CIMMYT, for their encouragement and guidance in preparing this paper. The authors appreciate the assistance of Mr. Surath Pradhan in editing figures. Thanks are also due to SDC for the support of maize research and development in the hills of Nepal. This publication is an output from a research project partly funded by the United Kingdom Department for International Development (DFID) for the benefit of developing countries (Project No. R7281, Plant Sciences Research Programme). The views expressed are not necessarily those of DFID. References Andrade, H. F., Uhart, S. A., & Frugone, M. I. (1993). Intercepted radiation at flowering and kernel number in maize: shade versus plant density effects. Crop Science, 33, 482–485. Carter, E. J. (1992). Tree cultivation on private land in the middle hills of Nepal: lessons from some villages of Dolakha district. Mountain Research and Development, 12(3), 241–255. Chamshama, S. A. O., Mugasha, A. G., Kloustad, A., Haveraaen, O., & Maliond, S. M. S. (1998). Growth and yield of maize alley cropped with Leucaena leucocephala and Faidherbia albida in Morogoro, Tanzania. Agroforestry Systems, 40, 215–225. Duncan, W. G. (1980). Maize. In L. T. Evans (Ed.), Crop physiology (pp. 23–50). UK: University Press Cambridge. Duncan, W. G., & Hesketh, J. D. (1968). Net photosynthetic rates, relative leaf growth rates and leaf numbers of 22 races of maize grown at eight temperatures. Crop Science, 8, 670–674. Earley, E. B., Miller, R. J., Reichert, G. L., Hageman, R. H., & Seif, R. D. (1966). Effects of shade on maize production under field conditions. Crop Science, 6, 1–7. Ephrath, J. E., Wang, R. F., Terashima, K., Hesketh, J. D., Huck, M. G., & Hummel. (1993). Shading effect on soybean and corn. Biotronics, 22, 15–24. Gaskel, M. L., & Pearce, R. B. (1981). Growth analysis of maize hybrids differing in photosynthetic capability. Agronomy Journal, 73, 817–821. Harris, D., Joshi, A., Khan, P. A., Gothkar, P., & Sodhi, P. S. (1999). On-farm seed priming in semi-arid agriculture: development and evaluation in maize, rice and chickpea in India using participatory methods. Experimental Agriculture, 35, 15–29. Huxley, P. A., Pinney, A., Akunda, E., & Muraya, P. (1994). A treecrop interface orientation experiment with a Grevillea robusta hedgerow and maize. Agroforestry Systems, 26, 23–45. 113 Johnson, I. R., Riha, S. J., & Wilks, D. S. (1995). Modelling daily net canopy photosynthesis and its adaptation to irradiance and atmospheric CO2 concentration. Agricultural Systems, 50, 1–35. Jose, S., Gillespie, A. R., Seifert, J. R., & Pope, P. E. (2001). Comparison of minirhizotron and soil core methods for quantifying root biomass in a temperate alley cropping system. Agroforestry Systems, 52, 161–168. Maddoni, G. A., & Otegui, M. E. (1996). Leaf area, light interception and crop development in maize. Field Crops Research, 48, 81–87. Midmore, D. J., Roca, J., & Berrios, D. (1988). Potato (Solanum spp) in the hot tropics IV. Intercropping with maize and the influence of shade on potato microenvironment and crop growth. Field Crops Research, 18, 141–157. Morgan, D. C., & Smith, H. (1981). Non-photosynthetic responses to light quality. In O. L. Lange, P. S. Nobel, C. B. Osmond, & H. Ziegler (Eds.), Encyclopedia of plant physiology: physiological plant ecology. I. Responses to the physical environment (pp. 108– 134). Berlin: Springer Verlag. Peek, M. S., Russek-Cohen, E., Wait, D. A., & Forseth, I. N. (2002). Physiological response curve analysis using nonlinear mixed models. Oecologia, 132, 175–180. Prasad, R. B., & Brook, R. M. (2005). Effect of varying maize densities on intercropped maize and soybean in Nepal. Experimental Agriculture, 41, 365–382. Reed, A. J., Singletary, G. W., Schussler, J. R., Williamson, D. R., & Christy, A. L. (1988). Shading effects on dry matter and nitrogen partitioning, kernel number and yield of maize. Crop Science, 28, 819–825. Robertson, M. J. (1994). Relationships between internode elongation, plant height and leaf appearance in maize. Field Crops Research, 38, 135–145. Smith, H. (1986). The perception of light quality. In R. E. Kendrick & G. H. M. Kronenberg (Eds.), Photomorphogenesis in plants (pp. 187–216). Dortrecht: Martinus Nijhoff. Stirling, C. M., Rodrigo, V. H., & Emberru, J. (1993). Chilling and photosynthetic productivity of field grown maize (Zea mays); changes in the parameters of the light-response curve, canopy leaf CO2 assimilation rate and crop radiation-use efficiency. Photosynthesis Research, 38, 125–133. Tennant, D. (1975). A test of a modified line intersect method of estimating root length. Journal of Ecology, 63, 995–1001. Thapa, B. (1994). Farmers’ ecological knowledge about the management and use of farmland tree fodder resources in the middle hills of eastern Nepal. Ph.D. Thesis. Bangor: University of Wales. Tiwari, T. P., Brook, R. M., & Sinclair, F. L. (2004). Implications of hill farmers’ agronomic practices in Nepal for crop improvement in maize. Experimental Agriculture, 40, 397–417. Tiwari, T. P., Virk, D. S., & Sinclair, F. L. (2009). Rapid gains in yield and adoption of new maize varieties for complex hillside environments through farmer participation. I. Improving options through participatory varietal selection (PVS). Field Crops Research, 111, 137–143. Tiwari, T. P., Ortiz-Ferrara, G., Urrea, C., Katuwal, R. B., Koirala, K. B., Prasad, R. C., Gurung, D. B., Sharma, D., Hamal, B., Bhandari, B., & Thapa, M. (2009). Rapid gains in yield and adoption of new maize varieties for complex hillside environments through farmer participation. II. Scaling-up the adoption through community-based seed production (CBSP). Field Crops Research, 111, 144–151. Wilson, T. D., Brook, R. M., & Tomlinson, H. F. (1998). Interaction between Nere (Parkia biglobosa) and under-planted sorghum in a parkland system in Burkina Faso. Experimental Agriculture, 34, 85–98. 114 Dr. Thakur Prasad Tiwari gained his PhD in Agronomy from the University of Bangor, UK, in 2001. He started his professional career in 1982 and worked for many years in the hills of Nepal. He is currently working as a Cropping Systems Agronomist for CIMMYT, based in Bangladesh. He is an output oriented person and believes in participatory research and development and enjoys working closely with farmers to understand their social and biophysical circumstances in order to focus on targeted research and development interventions. He has published several papers in national and international journals. Dr. Robert M. Brook has taught agroforestry, crop science and rural development in the School of Environment, Natural Resources and Geography, Bangor University, Wales, United Kingdom since 1993. Research on agroforestry and tropical farming systems conducted at Bangor University has been conducted in Nepal, India, Bangladesh, Kenya, Nigeria, Costa Rica and Guatemala. He started his career in tropical agriculture working for the Papua New Guinea government agricultural research service for 6 years, and for shorter periods in South Sudan and Indonesia. His PhD was awarded in 1980 from Dundee University whilst conducting research at the Scottish Horticultural Research Institute. T.P. Tiwari et al. Paul Wagstaff studied tropical agroforestry at Bangor University in 1999–2000, and carried out research at the Pakhribas Agricultural Research Centre, Nepal. Paul started his career in tropical agriculture as a Crop Protection Officer with the Tanzanian Ministry of Agriculture and Livestock Development in 1989 and has since worked in smallholder agriculture and tropical forestry for a wide range of local and national NGOs, government departments and bilateral development agencies in Africa and South Asia. He is currently the Agricultural Advisor for Concern Worldwide, an Irish international development NGO, supporting agricultural programs in 27 countries in Africa, Asia and the Caribbean. Dr. Fergus Sinclair is the Global Research Leader in Production Ecology, World Agroforestry Centre (ICRAF), Nairobi, Kenya and the Director, Centre for Advanced Research in International Agricultural Development (CARIAD), College of Natural Sciences, Bangor University, Wales, UK. He has published several scientific papers in peerreviewed international journals and books.
© Copyright 2026 Paperzz