Plant and Soil 231: 11–19, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands. 11 Depth distribution of cherry (Prunus avium L.) tree roots as influenced by grass root competition L.A. Dawson1,5 , E.I. Duff2 , C.D. Campbell3 & D.J. Hirst2,4 1 Plant Science Group, Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen AB15 8QH, Scotland, UK. 2 Biomathematics and Statistics Scotland, Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen AB15 8QH, Scotland, UK. 3 Soil Science Group, Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen AB15 8QH, Scotland, UK. 4 Currently at Norsk Regnesentral, P.O. Box 114, Blindern, N-0314, Oslo, Norway. 5 Corresponding author∗ Received 5 June 2000. Accepted in revised form 28 November 2000 Key words: competition, herbicide, minirhizotron, pasture, Prunus avium, root Abstract We investigated the effect of competition from grass roots (as controlled by herbicide application) on the depth distribution of white roots in cherry trees, grown with varying rates and frequency of application of N in an agroforestry system. Statistical summaries of distribution, namely mean and skewness, produced a concise and interpretable analysis of the data. There was a large increase in tree root numbers in the surface horizons after the herbicide had reduced grass root competition. Where the surrounding grass had not been reduced by herbicide, the average depth of tree roots increased with time, contrasting with a marked shift in the mode of distribution of root numbers to shallow depths when grass competition was removed. These findings are important in the understanding of plant root competition and for prescribing best practise for tree establishment in agroforestry systems. Introduction “Agroforestry is a collective name for land-use systems and technologies, where woody perennials (trees, shrubs, palms, bamboos, etc) are deliberately used on the same land management unit as agricultural crops and/or animals, either in some form of spatial arrangement or temporal sequence” (ICRAF, 1993). One of the main types of agroforestry of relevance in the UK is a silvopastoral system in which trees are planted directly into grazed pasture. These systems, therefore, force two plant species to compete for the same resources particularly when root distributions often overlap. The presence of roots of competing vegetation, such as grass, can reduce resource availability for tree growth. Roots of one plant may influence the growth and distribution of roots of neighbouring plants. For ∗ FAX No: 01224-311556. E-mail: [email protected] example, root growth of Loblolly pine (Pinus taeda L.) was reduced due to competition by crab grass (Digitaria spp.), irrespective of nitrogen or soil water status (Ludovici and Morris, 1997). Also, the presence of Eucalyptus (Eucalyptus grandis) trees reduced pasture root length densities in Queensland, Australia (Eastham and Rose, 1990). In orchard systems, grasses can compete more effectively than trees for water and nutrients, often as a result of higher root densities particularly at the soil surface (Atkinson, 1977). Also, higher application rates of N fertilizer have been shown to exacerbate competition between cherry trees and pasture in a dry year (Campbell et al., 1994b). Atkinson (1977) and Khatamian et al. (1984) showed that increasing nutrient supply could stimulate competition between grasses and trees, especially for water. On the other hand, the greater depth to which tree roots can extend, compared to grass roots, (Eastham and Rose, 1990) can allow exploitation of water and reduce interspecific competition. 12 One way to reduce belowground competition in the tree establishment phase is by maintenance of an area of reduced weed/grass around the tree (Davies, 1987; Eason et al., 1991, 1992). Work has been done on the use of herbicides in the establishment of radiata pine (P. radiata) and results demonstrate the significant effect of tree-pasture competition on water relationships and on tree growth (Connor et al., 1989), although little has been done to investigate the associated below-ground responses. In this study, we examined how the removal of grass and weed competition from around cherry (Prunus avium) influenced the depth distribution of the tree roots. Minirhizotron methods (Mackie-Dawson and Atkinson, 1991), which have the advantage of being non-destructive, were used to examine changes in the vertical distributions of ‘white’ tree roots. The application of herbicide, 4 weeks after monitoring began, to half of the trees in the study created an environment for investigating the dynamics of vertical distribution of tree roots both in the presence and absence of grass roots. This experiment, therefore, provided an opportunity to monitor the depth distribution of tree roots in a temporal fashion and to test the hypothesis that the tree root distribution is unaffected by competition from grass roots. Materials and methods The methods used to count points of root contact with minirhizotrons have been described elsewhere (Buckland et al., 1993). Similarly, the general site description and experimental design have been previously described (Campbell et al., 1994b). Therefore, only brief descriptions of both are given here. Twoyear-old cherry (Prunus avium L.) whips were planted in 1987 at a 4 × 4 m spacing in two blocks (216 trees, excluding buffer zones). Each block was divided into four plots (0.12–0.15 ha) and within each plot there were three subplots of nine trees used for measurements. Treatments consisted of two levels of nitrogen (high, 144 kg ha−1 yr−1 and low, 72 kg ha−1 yr−1 application) in two frequencies of split annual applications (4 or 12) as detailed in Campbell et al. (1994b). Actual amounts of nitrogen per application were, therefore, 36 kg ha−1 (high) and 18 kg ha−1 (low) when 4 applications were used; and 12 kg ha−1 (high) and 6 kg ha−1 (low) when 12 applications were administered. In addition, a 1m diameter circle of bare ground was maintained around each tree by treating the pasture with herbicide (glyphosate at 2 kg ha−1 ) since 1987. In 1990, however, the herbicide was applied between 24 and 29 May to the pasture around half of the trees in each treatment, on a subplot basis. The presence of clover was minor and was eliminated in 1987 by spraying with a selective herbicide. The resulting sward consisted primarily of Lolium perenne. For this study, within each subplot, two of the nine trees from each group were selected at random and minirhizotrons installed. Minirhizotron tubes were inserted vertically due to the stony and indurated nature of the soil (Campbell et al., 1994b), 3 months prior to start of measurements. They were installed at a distance of 40 cm from the base of the tree, to a depth of approximately 30 cm. The location of the tube was chosen at random to be either due North, South, East or West of the tree. A clear plastic insert, scribed with four ladders of 40 counting grids (0.7 × 0.7 cm) at 90◦ intervals, was fixed inside each tube, and the tube was placed in the soil with one ladder directly facing the tree. The method used to count the presence of tree and grass roots was that of counting first and last points of contact with the tube (including root tips) (Buckland et al., 1993). Counts were recorded for each grid square (0.7 × 0.7 cm) throughout the depth of the minirhizotrons. These detailed data, collected from each minirhizotron, were then used to derive statistics summarising the distribution of tree roots within the 30 cm zone of soil below the soil surface. At the time of recording, roots were separated into type. White roots of cherry were easily distinguished from grass roots by their generally thicker diameter, opacity, more succulent appearance and lighter coloration. Since we were primarily interested in how the distribution of tree roots might change over time, we considered it appropriate to focus on roots recorded as white as they are the best indicator of new growth (Mackie-Dawson et al., 1995). Vertical locations of tree roots, from the 48 cherry trees with minirhizotrons installed, were recorded approximately weekly throughout May and June and less frequently thereafter until the end of August in 1990. Given that tree root growth effectively ceased by the end of June (Campbell et al., 1994a), data from the sampling times during July and August were not used in the analysis. Depth distribution data were cumulated over time for each of the months May and June, respectively, resulting in depth distribution data for each of two distinct periods; time 1, the 4 monitoring times before herbicide application and time 2, the 5 13 40 cm from the tree, the same distance from the tree as the minirhizotron tubes, and at a random orientation. Cores of 7 cm internal diameter were taken from 48 trees on the 28 May and 3 July, and total root length was measured at four depth intervals (0–5 cm, 5–15 cm, 15–25 cm and 25–35 cm) for both tree and pasture roots as described in Buckland et al. (1993). Calculation of summary statistics For the minirhizotron data, the distributions of cherry roots through depth were summarised by deriving two statistics: mean and skewness of depth of roots. Both summary statistics were derived from the pooled distributions (e.g. Figure 1), resulting in one value for before and one value for after the time of herbicide application for each of the 48 trees in the study. Skewness is defined as follows: P (xi − m)3 Skewness = i (n − 1)s 3 Figure 1. Distribution of ‘white’ tree roots through the soil profile for a single minirhizotron tube. monitoring times after herbicide application. Because of the need to retain information on the location of roots, this involved summing counts at the level of individual grid squares. As an example, Figure 1 shows the root distribution through the soil profile for one tree (G7), a low N, high frequency treatment, before and after application of herbicide. Soil cores were also collected from the site, using a power driven system, and washed free of soil with a hydropneumatic elutriation method (MackieDawson and Atkinson, 1991). Trees were selected at random from all treatment plots, with the constraint that no cores were taken from the trees monitored by minirhizotrons, to avoid affecting subsequent growth. A single core was taken from each sampled tree at where xi is the depth at which a root is recorded, m is mean depth of roots in the sample, s is standard deviation of depth distribution of roots in the sample and n is number of roots in the sample. Negative skewness indicates that the majority of roots were recorded at lower depths with few recorded towards the surface, whilst positive skewness indicates that most roots were recorded nearer the surface with few recorded at greater depths. For the two cases where there were less than three roots observed, skewness was treated as missing (as was mean depth) and estimated in the subsequent statistical analysis. In addition to the statistics summarising the distribution of tree roots through depth, counts of white roots (before and after herbicide) were summed over the entire depth of the minirhizotron. Counts per minirhizotron were standardised by dividing by the appropriate number of ladders read per period, giving the number of white roots per ladder. Whilst root counts, summed over depth, do not contribute to our understanding of the distribution of root depths, analysis of this variable was included to set the scene as to the overall treatment effects and to highlight the extra information gained by the more sensitive approach of examining depth distributions. Statistical analysis The summary statistics were analysed by analysis of variance using Genstat 5.3 software (NAG Ltd., Oxford, UK). Analysis of variance was used to assess 14 Figure 2. Average densities by depth of ‘white’ tree roots in May (—) and June (- - - -). For (a) Low Frequency of N; no herbicide, (b) Low Frequency of N; with herbicide, (c) High Frequency of N; no herbicide (d) High Frequency of N; with herbicide. The data were smoothed by fitting cubic smoothing splines. the effects of herbicide, level of nitrogen, frequency of nitrogen application and time on each of mean, skewness and total number of white roots per ladder. Since herbicide was applied to groups of trees (each group comprising of two trees with minirhizotrons) and minirhizotron data was analysed at two time periods sampled, the structure of the analysis was a nested design (tree groups, trees within groups, time within tree groups and time within trees). Root density data from the minirhizotrons were log transformed prior to statistical analysis. To allow for differences due to operator variation, operator identifiers were included as covariates in the analysis of counts of white roots per ladder. These covariates were derived by calculating, for each minirhizotron, the proportion of ladders read per period by each of the three operators recording counts of roots. However, operator identifiers were not statistically significant and so were omitted from the final analysis. Root density data from soil coring was log transformed prior to being analysed using an analysis of variance in the same nested design as for the minirhizotron data. As there were no significant effects of nitrogen treatment, only means across all N treatments are presented. Results Minirhizotron data Number of white roots per ladder Treatment means relating to root counts, summed over the depth of the minirhizotron, are shown in Table 1. Trees receiving the high level of nitrogen (averaged over the two time periods), showed higher counts 15 Table 1. Effects of Nitrogen and Herbicide on differences in Counts per ladder (log transformed) of ‘white’ cherry roots in 2 time periods (pre-herbicide application and post herbicide application). H+ refers to herbicide applied, H− refers to no herbicide. The N levels refer to a Low N total supply (72 kg ha−1 yr−1 ), High N total supply (144 kg ha−1 yr−1 ), Low Frequency of application (4 applications) and a High Frequency of application (12 applications) Variable Treatment PreHerbicide (Period 1) PostHerbicide (Period 2) Difference Counts per ladder Low N, H+ Low N, H− High N, H+ High N, H− 0.93 0.58 1.08 1.12 1.07 0.86 1.58 1.40 0.14 0.28 0.50 0.28 0.14 Low Frequency, H+ Low Frequency, H− High Frequency, H+ High Frequency, H− 1.09 1.23 0.92 0.47 1.21 1.30 1.45 0.96 0.12 0.07 0.52 0.49 0.14 Counts per ladder Standard error of Difference (P=0.05) than trees under low N (log (counts) were 1.29 and 0.86 for high and low levels of nitrogen, respectively). Numbers of tree roots increased with time, the period after herbicide application (June) showing much higher (P<0.001) counts of tree roots than the period prior to herbicide application (May). Log (counts) were 0.93 and 1.23 for May and June, respectively. Increased numbers in the post-herbicide period (June) were more evident with the high frequency of N application than with the low frequency (Table 1), as indicated by the significant interaction (P<0.05) between frequency of nitrogen application and time. Total numbers of white roots, summed over the depth of the minirhizotron, showed no significant effect of herbicide. Mean: Distribution of tree roots Treatment means relating to mean of depth distribution of tree roots are presented in Table 2. Where herbicide was applied, mean depth of tree roots was less after herbicide application (June) than before (May), whereas in the absence of herbicide, mean depth of tree roots was greater in June than in May, as indicated by the significant interaction (P<0.01) between herbicide treatment and time. This effect was more pronounced with high level of N application than with low level of N application (Table 2), although the evidence for this was weak (P=0.086). Figure 3. Average densities by depth of ‘white’ grass roots in May (—) and June (- - - -). For (a) With herbicide and (b) No herbicide. The data were smoothed by fitting cubic smoothing splines. Skewness: Distribution of tree roots Treatment means relating to skewness of depth distribution of tree roots are given in Table 2. Changes in skewness over time were positive (from −0.122 to 16 Table 2. Effects of Nitrogen and Herbicide on differences in Mean of depth distribution (cms) of ‘white’ cherry roots and on Skewness of depth distribution of ‘white’ cherry roots in 2 time periods (pre herbicide application and post herbicide application). H+ refers to herbicide applied, H− refers to no herbicide. The N levels refer to a Low N total supply (72 kg ha−1 yr−1 ), High N total supply (144 kg ha−1 ), Low Frequency of application (4 applications) and a High Frequency of application (12 applications) Variable Treatment Mean of depth distribution Low N, H+ Low N, H− High N, H+ High N, H− 15.00 12.86 14.54 12.65 14.59 −0.41 13.28 0.42 12.71 −1.83 14.02 1.37 0.65 Low Frequency, H+ Low Frequency, H− High Frequency, H+ High Frequency, H− 14.51 13.85 15.03 11.66 13.64 −0.87 14.01 0.16 13.65 −1.38 13.29 1.63 0.65 Mean of depth distribution Skewness of depth distribution Skewness of depth distribution PreHerbicide (Time 1) PostHerbicide (Time 2) Difference Standard error of Difference Low N, H+ Low N, H− High N, H+ High N, H− −0.218 0.210 −0.026 0.234 −0.139 0.079 −0.093 −0.303 0.384 0.410 0.026 −0.208 0.123 Low Frequency, H+ Low Frequency, H− High Frequency, H+ High Frequency, H− −0.138 0.029 −0.106 0.415 −0.125 0.013 −0.056 −0.085 0.370 0.476 −0.011 −0.426 0.123 0.123) for trees receiving herbicide, but negative (from 0.22 to −0.034) for trees which did not receive herbicide (P<0.001 for interaction of herbicide treatment and time). This change in skewness reflects an upward shift in the mode of the root distribution for trees where herbicide was applied and a downward shift for trees where herbicide was not applied. Positive and negative changes in skewness over time (for trees receiving or not receiving herbicide, respectively), were more evident with high frequency of N application than with low frequency of N (P<0.01 for herbicide by time by N frequency interaction). Depth distributions Average distributions of depths of ‘white’ cherry roots, for particular treatment combinations, are shown in Figure 2. This illustrates the shift in tree root density profile nearer to the surface with time where herbicide had been applied, whereas it moves into deeper zones with time in the absence of herbicide. It also illustrates how this trend is similar for both frequencies of N application, but that the effect is most marked at the higher frequency of application. Figure 3, derived in a similar way to Figure 2, illustrates the temporal change in grass root density profile for trees where herbicide was applied and also for those where no herbicide was applied. Root coring data Grass root length densities were concentrated in the top 5 cm in May (Figure 4a). In July, the pasture roots were still concentrated at the surface but reduced growth generally at the surface and increased growth at other depths meant the trend was less pronounced (Figure 4a). Herbicide significantly (P<0.05), reduced surface grass roots in July, 70 days after the herbicide application. In July, tree root distributions were greatest (P<0.05) in the 5–15 cm depth (Figure 4b), contrasting with the grass root distributions, and they were unaffected by herbicide application. 17 Figure 4. Effect of soil depth on root length density from soil coring of (a) grass roots (circle) and (b) ‘white’ tree roots (triangle) on 2 sampling dates, with (open symbols) and without (solid symbols) herbicide treatment. Data were pooled over all N treatments. Bars represent s.e.d. for herbicide, ∗ represents a significant difference, p<0.05 in ANOVA. Discussion and conclusions Methodology Numbers of tree roots increased over time, reflecting the increasing size of the trees in the establishment phase and the time of year when the trees were actively growing (Campbell et al., 1994b). The increase in number of white roots was greatest with the greatest rate of addition of N, due to a faster plant growth rate. The novel approach adopted in this study highlighted changes in root location over time, observed from the minirhizotron data, with differences due to both herbicide and nitrogen, which was not evident from the root coring data. Both mean depth and skewness of depth distribution showed effects of herbicide treatment over time, with temporal effects more pronounced where numbers of tree roots increased in the post-herbicide period. Co-existence of root systems The different root distribution between trees and grass (Campbell et al., 1994a) reflects that root systems of plants can avoid being in the same competitive space (Eastham and Rose, 1990; Mou et al., 1995) as is suggested by the results of our study. Results suggest that the tree roots were avoiding the surface soil when grass roots were present, leading to a rejection of the original null hypothesis. Caldwell et al. (1996) found that at the small scale, shrub and grass roots tended to avoid each other. Bould and Jarrett (1962) found that Timothy (Phleum pratense) and ryegrass (Lolium perenne) restricted apple (Cox’s orange pippin) tree growth due to competition for nitrogen, which was reflected in a reduced N concentration in the tree leaves. Watson (1988), working with maple, ash and oak, in an urban environment showed that the presence of grass competition in the root zone of trees inhibits fine root development of trees. Evidence from the distribu- 18 tion of forest trees suggests that tree roots will, without understory competition, colonise the surface layers of soil (Nambiar, 1983). Also, in orchards, where herbicide is applied in large areas, fruit tree roots often colonise the surface soil, and the vertical distribution can be similar to our study with the greatest density of rooting being found in the 10–20 cm soil depth (Atkinson, 1977). The altered root distribution of the ‘white’ cherry tree roots due to the herbicide application, as demonstrated in this study, therefore, has many implications for future tree growth and management and interacts with factors such as N supply, soil moisture and stage of tree development. Acknowledgements Root depth distribution In our study, changes through time in the depth distribution of tree roots were dominated by the herbicide treatment but also influenced by the availability of N. When grass was untreated with herbicide, ‘white’ tree roots grew deeper with time, particularly for trees receiving an increased availability of N. When the grass was reduced, ‘white’ tree roots tended to grow nearer the surface, again particularly for trees with increased N availability. The increase in ‘white’ tree root numbers in the surface, and the change in the pattern of root distribution towards the soil surface after herbicide had partially removed grass root competition, demonstrates the ability of tree roots to grow in shallow depths after competition has been removed. This change in root distribution may reflect a proliferation response to the increased N availability in that soil zone once the grass was removed (Caldwell et al., 1996), reflect a greater availability of soil water (Watson et al., 1992) or as a response to the combination of both water and nitrogen (Goode and Hyrycz, 1976). Where the surrounding grass was not removed, the average ‘white’ tree root depth increased with time as the competition from the grass roots remained in the surface horizons. Reflecting greater tree root growth at depths when grass is present, Atkinson and White (1980) showed a greater uptake of 32 P from 90 cm depth by trees under grass, than trees which had received a herbicide treatment. Other workers have also found deeper tree root distribution when under a cover crop, e.g. Bjorkman and Lundeberg (1971). However, this reduction in deep rooting as a consequence of herbicide application may not be totally advantageous (Conner et al., 1987) as the altered root spatial distribution can reduce tree stability (Helliwell, 1989; Stokes et al., 1995) and increase susceptibility to drought (Stuckey, 1961). The presence of the grazing animal has also been shown to increase soil compaction around the base of the tree (Eason et al., 1992; Wairiu et al., 1994), which would have more pronounced effects if the grass is removed. This work was funded by the Scottish Executive Rural Affairs Department. The authors would like to thank S Pratt and E Reid for technical assistance. References Atkinson D 1977 Some observations on the root growth of young apple trees and their uptake of nutrients when grown in herbicide strips in grassed orchards. Plant Soil 46, 459–471. Atkinson D and G C White 1980 Some effects of orchard soil management on the mineral nutrition of apple trees. In The Mineral Nutrition of Fruit Trees. Eds. D Atkinson, JE Jackson, RO Sharples and WM Waller. pp 241–254. Butterworths, UK. Bjorkman E and Lundeberg G 1971 Studies of root competition in a poor pine forest by supply of labelled nitrogen and phosphorus. Stud. For. Succ . 91, 16. Bould C and Jarrett R M 1962 The effect of cover crops and NPK fertilizers on growth, crop yield and leaf nutrient status of young dessert apple trees. J. Hortic. Sci. 37, 58–82. Buckland S T, Campbell C D, Mackie-Dawson L A, Horgan G W and Duff E I 1993 A method for counting roots observed in minirhizotrons and their theoretical conversion to root length density. Plant Soil 153, 1–9. Caldwell M M, Manwaring J H and Durham S L 1996 Species interactions at the level of fine roots in field – influence of soil nutrient heterogeneity and plant size. Oecologia 106, 440–447. Campbell C D, Mackie-Dawson L A, Reid E J, Pratt S M, Duff E I and Buckland S T 1994a Manual recording of minirhizotron data and its application to study the effect of herbicide and nitrogen fertilizer on tree and pasture growth in a silvopastoral system Agrofor. Syst. 26, 75–87. Campbell C D, Atkinson D, Jarvis P G and Newbould P 1994b Effects of nitrogen fertilizer on tree/pasture competition during the establishment phase of a silvopastoral system. Ann. Appl. Biol. 124, 83–96. Connor D J, Sands R and Strandgard M 1989 Competition for water, light and nutrients in agroforestry associations of Pinus radiata and pasture. In Meteorology and Agroforestry. Eds. WS Reifsnyder and TO Darnhofer. pp 451–462. Proceedings of International Workshop, Nairobi, Kenya, February 0–13, 1987. ICRAF. Davies R J 1987 Trees and weeds. Weed control for successful tree establishment. Forestry Commission Handbook 2. HMSO. London. pp 36. Eason W R, Newman E I and Chuba G H 1991 Specificity of interplant cycling of nutrients: the role of mycorrhizas. Plant Soil 137, 267–274. Eason W R, Tomlinson H F and Hainsworth C 1992 Effect of ground vegetation on root distribution of ash trees. Asp. Appl. Biol. 29, 225–231. 19 Eastham J and Rose C W 1990 Tree/pasture interactions at a range of tree densities in an agroforestry experiment.1.Rooting patterns. Aust. J. Agric. Res. 41, 683–695. Goode J E and Hyrycz K J 1976 The effect of nitrogen on young, newly-planted apple rootstocks in the presence and absence of grass competition. J. Hort. Sci. 51, 321–327. Helliwell D R 1989 Tree roots and the stability of trees. Arboricul. J. 13, 243–248. ICRAF 1993 International Centre for Research in Agroforestry: Annual Report 1993. Nairobi, Kenya. 208 pp. Khatamian H, Pair J C and Carrow R 1984 Effects of turf competition and fertilizer application on trunk diameter of nutrient comosition of Honeylocust. J. Arboricul. 10, 156–159. Ludovici K H and Morris L A 1997 Competition-induced reductions in soil water availability reduced pine root extension rates. Soil. Sci. Soc. Am. J. 61, 1196–1202. Mackie-Dawson L A and Atkinson D 1991 Methodology for the study of roots in field experiments and the interpretation of results. In Plant Root Growth, An Ecological Perspective. Ed. D Atkinson. pp 25-47. Blackwell, Oxford. Mackie-Dawson L A, Pratt S M, Buckland S T and Duff E I 1995 The effect of nitrogen on fine root persistence in cherry (Prunus avium). Plant Soil 173, 349–353. Mou P, Jones R H, Mitchell R J and Zutter B 1995 Spatial distribution of roots in sweetgum and loblolly pine monocultures and relations with above-ground biomass and soil nutrients. Funct. Ecol. 9, 689–699. Nambiar E K S 1983 Root development and configuration in intensively managed radiata pine plantations. Plant Soil 71, 37–47. Stokes A, Fitter A H and Coutts M P 1995 Response of young trees to wind and shading – effects on root architecture. J. Exp. Bot. 46, 1139–1146. Stuckey I H 1961 Root growth of Taxus. Am. Nurseryman 114, 117– 114. Wairiu M, Mullins C E and Campbell C D 1994 Soil physical factors affecting the growth of sycamore (Acer pseudoplatanus L.) in a silvopastoral system on a stony upland soil in North-East Scotland. Agrofor. Sys. 24, 295–306. Watson G W 1988 Organic mulch and grass competition influence tree root development. J. Arboricul. 14, 200–203. Watson C A, Hooker J E, Atkinson D, Birley M A and Newbould P 1992 Effects of vegetation management on nitrate loss to ground water from agroforestry systems. Asp. Appl. Biol. 29, 407–412. Section editor: R. Aerts
© Copyright 2026 Paperzz