tree orchard vigour

Scientia Horticulturae 106 (2005) 76–90 www.elsevier.com/locate/scihorti

A set of vegetative morphological variables to objectively estimate apple (Malus domestica) tree orchard vigour Thomas Nesme a,*, Daniel Plenet a, Bruno Hucbourg b, Georges Fandos c, Pierre-Eric Lauri d a

Unite´ Plante et Syste`mes de Culture Horticoles (PSH), Institut National de la Recherche Agronomique (INRA), Domaine St. Paul, Site Agroparc, 84914 Avignon Cedex 9, France b Groupement Re´gional des Centres d’Etude des Techniques Agricoles (GR-CETA) de Basse Durance, 2 Route de Molle`ges, 13210 Saint Re´my de Provence, France c Cofruid’Oc, 286 Route de Saint Nazaire de Pe´zan, 34400 Saint Just, France d Unite´ Mixte de Recherche Biologie du De´veloppement des Espe`ces Pe´rennes Cultive´es (UMR BEPC), Institut National de la Recherche Agronomique (INRA), 2 Place Pierre Viala, 34060 Montpellier Cedex 1, France Received 26 May 2004; received in revised form 24 February 2005; accepted 25 February 2005

Abstract Orchard vigour, defined as the intensity of vegetative growth, is an important indicator for crop management in fruit tree cropping systems. It is often evaluated in commercial plots by experts on a non-formalised basis or measured with a single variable known as trunk cross-sectional area (TCSA). In this article, we proposed a set of 11 tree or plot morphological vegetative variables for apple orchards and applied it on 117 farm plots in south-eastern France. Relationships between variables were studied by component analysis (CA) and plots were classified into four clusters according to the first two factors of the CA. These modelled vigour marks were compared to expert vigour marks on 14 plots. Plot modelled vigour classification was re-estimated with only three morphological variables and compared to the original classification. These morphological variables were: TCSA, number of water sprouts on the trunk and length of annual shoot at the distal part of fruiting branches at the bottom of the tree. The first three factors of the CA correspond to vegetative growth intensity, opposition between annual and cumulative growth and vigour balance, respectively. Modelled and * Corresponding author. Present address: ENITA de Bordeaux, BP 201, 33175 Gradignan Cedex, France. Tel.: +33 5 57 35 07 07; fax: +33 5 57 35 07 59. E-mail address: [email protected] (T. Nesme). 0304-4238/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.scienta.2005.02.017

T. Nesme et al. / Scientia Horticulturae 106 (2005) 76–90

77

expert plot vigour classifications were generally in agreement, except in the case of heterogeneous plots. Re-estimated and original modelled classifications were also in agreement, except in the case of older and more vigorous orchards. Results showed that plot vigour modelling based on these three morphological variables may be relevant. TCSA thus did not appear to be sufficient. Results are discussed in relation to plant architecture features. # 2005 Elsevier B.V. All rights reserved. Keywords: Apple tree; Orchard vigour; Component analysis; Trunk cross-sectional area; Vegetative growth; Expert vigour assessment

1. Introduction Tree vigour can be defined as the intensity of vegetative growth. Therefore, it has various expressions depending on the studied scale: morphological (e.g. long lateral branching; Gunkel et al., 1949), anatomical (e.g. ratio of primary phloem to primary xylem; Kurian and Iyer, 1992) or physiological (e.g. efficiency of the balance between photosynthesis and respiration; Way et al., 1983). Although the concept of ‘vigour’ lacks common and quantitative definition, it is employed, in many fields of plant research, e.g. forestry (Kyto¨ et al., 1999; Duchesneau et al., 2001; Vincent et al., 2002), ecology (Moravie et al., 1999) and horticulture (Crabbe´ , 1987; Neilsen et al., 1997; Lo Bianco et al., 2003). Different attempts to define tree vigour were realised based on morphological measurements (Waring et al., 1980; Moravie et al., 1999) but, to our knowledge, they never concerned cultivated species such as apple tree. Moreover, vigour definition was always given a priori, based on a few variables, and has never been based on expert vigour assessment. In addition to pedoclimatic factors which act in any conditions, various cropping techniques in horticulture may influence tree vigour: rootstocks may have sizecontrolling effects on above-ground vegetative growth (Barden and Marini, 1997); fruit load influences ratio between vegetative and reproductive growth (Berman and DeJong, 2003); pruning affects leaf area development (Lakso, 1984); irrigation shortage can influence various components of vegetative growth (Li et al., 1990); and N fertilisation affects seasonal vegetative growth patterns (Weinbaum et al., 1992; Lobit et al., 2001). Moreover, tree vigour control plays an important role in orchard management as it may influence within-tree microclimate and thus disease development (Penrose and Nicol, 1996), or fruiting pattern (Lauri et al., 1997; Lauri, 2002). Technical advice is sometimes based on tree vigour: for instance, Spring et al. (1993) propose reducing nitrogen fertiliser amounts in the case of vigorous orchards. Thus, apple tree vigour is a widely used crop indicator by farmers, in relation to both pruning and nitrogen fertilisation (Nesme et al., 2003). Orchard vigour evaluation is a complete and integrative but non-formalised process when carried out by farmers or experts such as technical advisers. When attempts are made to objectively evaluate orchard vigour, a single variable such as trunk cross-sectional area (TCSA) is often used (Khatamian and Hilton, 1977; Marsh et al., 1996; Quamme et al., 1998; Webster and Hollands, 1999; Ro and Park, 2000; Mataa, 2000; Barden et al., 2002).

78


However, due to their interactions, all morphological variables should be considered as a whole. Thus, there is a need for an objective and non-destructive method based on the observation of several morphological variables to quantify and model apple tree vigour, according to expert vigour assessment. The objectives of this study were: (i) to propose a set of vegetative growth variables to evaluate apple orchard vigour, simple enough to be used in cropping conditions; (ii) to compare modelled vigour marks resulting from tree growth measurements with expert vigour marks; and (iii) to simplify the set of tree growth variables in order to make it easier to use with the smallest loss of information possible. The method was applied on farmers’ apple tree plots in order to encounter a wider range of pedoclimatic conditions, orchard designs and crop management sequence combinations than would otherwise be possible in an experimental situation.

2. Materials and methods 2.1. Building a set of morphological variables Eleven variables were chosen from among all possible vegetative growth variables, on the basis of: (i) their relevance according to scientific literature or expert knowledge; and (ii) their ease of measurement. They were: TCSA, grafting point height, scion-rooting, number of water sprouts on the trunk or on fruiting branches, annual shoot length at the top or bottom of the tree, distribution of laterals on the fruiting branch, sum of fruiting branch sectional areas, row height and width. TCSA (Khatamian and Hilton, 1977; Quamme et al., 1998; Webster and Hollands, 1999; Ro and Park, 2000; Mataa, 2000; Barden et al., 2002), or TCSA increase (Marsh et al., 1996), is the most common variable used to estimate cumulative growth over long periods. It provides integrative information about whole tree growth. We measured trunk girth at 20 cm above the grafting point and converted it into trunk sectional area. The grafting point height is often found to be inversely correlated to vegetative growth intensity (Parry, 1986; Quamme et al., 1998), mainly when rootstocks have a dwarfing effect. Grafting point height was the height between the soil and the graft union. When scionrooted, the effect of dwarfing rootstock is suppressed because of above graft union adventive roots, and above-ground tree growth is considerably increased. Thus, trees were defined as either being scion-rooted or not. Water sprouts often express an excessive vegetative growth. Their number was evaluated on the trunk—notably on the upper, often bent part—and on fruiting branches. On the trunk, their number was converted into classes of 0, 1–4, 5–10 or more than 10 water sprouts per tree. On fruiting branches, the number of water sprouts was expressed as the proportion of fruiting branches bearing at least one water sprout. Four classes were determined: 0%, 1–25%, 25–50% or >50%. Elongation shoot growth, when cumulated for the whole tree, is often used to estimate tree vigour (Khatamian and Hilton, 1977; Parry, 1986; Lehman et al., 1990; Webster and Hollands, 1999; Ro and Park, 2000). Annual shoot length was recorded at the distal part of fruiting branches at the top and the bottom of trees, respectively. It was converted into three classes: 20 cm. Three classes of distribution of laterals on the fruiting


79

branch were defined, according to whether the laterals were in a distal or proximal position, or evenly distributed along the branch. This rather subjective indicator provides information concerning the equilibrium of within-tree growth. The basal diameter of fruiting branches was measured because of its importance for fruit-set and inflorescence leaf development (Lauri et al., 1996) and because it may be used to determine the leaf area of the fruiting branch (as proposed by Sinoquet et al. (2001) at the shoot level). Thus, the diameter of all branches measuring over 1 cm in diameter growing out of the trunk was measured with a calliper, at 5 cm from the trunk to avoid branch butt, and then converted into sectional area. Sectional areas were totalled in order to compute the sum of fruiting branch sectional areas per tree. Finally, mean row height and width were recorded to estimate the volume taken up by the aerial parts of trees. The last two variables were measured once per plot. The other nine variables were measured on six trees, chosen as being representative of the plot: in a single row, three adjacent trees were chosen on each side, located at least 30 m from each plot border. Annual shoot length was observed on at least four fruiting branches per tree. Since yield data were not available for all of the orchards, fruit production was not integrated into this study although it has considerable consequences on vegetative growth, particularly in the case of alternate fruit bearing (Singh, 1948). But the studied varieties are not subject to alternate bearing (Trillot et al., 1993). Moreover, since the studied plots belonged to farmers, adapted cropping operations (pruning, thinning and nitrogen fertilisation) were aimed at limiting pluriannual crop load variation. 2.2. Application of the set of morphological variables The set of morphological variables was applied in 117 non-regrafted apple orchards ranging from 4 to 42 years old and belonging to 21 farmers. All plots were located in southeastern France, between Nıˆmes and Montpellier (43.68N, 4.18E), with a Mediterranean climate. Apple trees were mainly located on four types of soils according to the classification of Baize and Girard (1995), which typically correspond to the location of apple tree production in France. Varieties were Golden Delicious, Gala (Royal-Gala, Imperial-Gala, Obro-Gala and Galaxy), Pink Lady1 and Granny Smith. The first two are type III varieties, according to the classification of Lespinasse (1977), with fruiting on spurs and crowned brindles. Pink Lady1 and Granny Smith are type IV varieties, i.e. with fruiting in the terminal position. They tend to develop branching on the upper third of the branch and to localise the production on the periphery of the tree. Type III or IV varieties of the classification proposed by Lespinasse (1977) had similar vegetative growth patterns (Lauri et al., 1995) and could thus be compared. Trees were grafted onto M9 (including Pajam 1 and Pajam 2) (94 plots), M106 (8 plots), M26 (7 plots), M7 (4 plots) or other rootstocks (4 plots). All trees were trained as vertical axis, i.e. central axis with renewal pruning of the fruiting branches, or Solaxe, i.e. central axis with bending of the upper part of the trunk and of branches and no renewal pruning (Lauri and Lespinasse, 1999; Lauri, 2002). Older orchards were often former palmettes restructured as vertical axis trees. All cropping operations were at the discretion of the individual farmers. Measurements were made by the same observers in late August and September 2002, at the end of the growing season.

80


2.3. Vigour assessment The previous set of morphological variables was used to give a vigour mark to each studied plot. Therefore, variables measured at tree level were restricted to plot values: the mean of each tree-measured quantitative variable (TCSA, grafting point height and sum of fruiting branch sectional areas) was computed. Qualitative variables (scion-rooting, water sprouts on the trunk or on fruiting branches, annual shoot length at the distal part of fruiting branches at the top or bottom of the tree, distribution of laterals) were transformed into quantitative variables: their classes were converted into ordinal variables (Table 1) and their mean per plot was computed. All 11 variables were scaled and submitted to component analysis (Escofier and Page`s, 1988). Quality representation of variables was assessed by the proximity of the variable segment to the correlation circle. Plots were then described by their coordinates on the first two factors of the component analysis and classified using the K-mean method (MathSoft, 1999), based on calculation of the centroid of each cluster. Plot vigour is usually classified into five groups but the studied sample did not contain any very low vigorous plots. Therefore, four clusters were built, representing plot vigour marks, and projected on the factorial plane. They were represented by variance ellipses that included 95% of the plots of the considered cluster. Vigour assessment based on morphological variable measurement could be considered as vigour modelling. The homogeneity of variance for plot age between vigour marks was checked with a nonparametrical test of multiple comparison (Sprent and Ley, 1992). Classical analysis of variance was performed, followed by a multiple comparison Tukey test. 2.4. Test and simplification of vigour modelling Vigour marks were compared to expert vigour marks. In July 2003, eight independent technical advisers or researchers, hereafter referred to as experts, were asked to give vigour marks (from 1 to 4) to 14 orchards chosen from among the 117 studied orchards and covering a wide variety of growth conditions. Experts were also asked to evaluate the balance between vegetative and reproductive growth and the reasons for their decision. The mean of expert marks was computed and rounded off. Both modelled and expert vigour marks were compared by a non-parametric Wilcoxon signed rank sum test (Sprent and Ley, 1992). Table 1 Conversion from qualitative to ordinal values of tree-measured qualitative variables Variable

Qualitative classes

Ordinal classes

Scion-rooted Number of water sprouts on the trunk Number of water sprouts on fruiting branches Annual shoot length at the distal part of fruiting branches at the top of the tree Annual shoot length at the distal part of fruiting branches at the bottom of the tree Distribution of laterals

No/yes 0/1–4/5–10/more than 10 per tree 0%/1–25%/25–50%/more than 50% of tree fruiting branches affected 20 cm

1/2 1/2/3/4 1/2/3/4

20 cm

1/2/3

Homogeneous/mean/heterogeneous

1/2/3

1/2/3


81

To simplify vigour modelling, a new plot vigour classification was built, based on the following three variables: TCSA, number of water sprouts on the trunk and annual shoot length at the bottom of the tree. They were chosen because of their ease of measurement and their representation on the first three factors of the component analysis. These three variables were scaled and submitted to component analysis. As previously realised, plots were described by their coordinates on the first two factors of the component analysis and classified using the K-mean method. Comparison between original and re-estimated modelled vigour marks was carried out with the Wilcoxon signed rank sum test. Data analysis was conducted with S+2000 software (MathSoft, 1999).

3. Results 3.1. Relationships between morphological variables No clear relationship between couples of variables could be observed, except between plot mean TCSA and the plot mean of fruiting branch sectional area per tree (Fig. 1). Both seemed to be linearly related for TCSA lower than 150 cm2, i.e. for plots less than 20 years old (data not shown). For older plots, the sum of fruiting branch sectional areas no longer paralleled TCSA increase because of the influence of pruning and tree restructuring (topping and cutting of primary branches). The projection of variables on the factorial plane formed by the first two factors of the component analysis is given in Fig. 2. The first factor (40% of the total variance) opposed the grafting point height, to the left, to all other variables (except distribution of laterals), to the right. The second factor (17% of the total variance) opposed annual shoot length and water sprout intensity, at the bottom, to scion-rooting, TCSA, row height, row width and

Fig. 1. Relationship between the plot mean trunk cross-sectional area (TCSA, in cm2) and the plot mean of fruiting branch sectional areas per tree (cm2).

82


Fig. 2. Projection of morphological variables on the factorial plane made of the first two factors of the component analysis. TCSA = trunk cross-sectional area; graft.pt.height = grafting point height; scion-root. = scion-rooted; wat.sprout trunk = number of water sprouts on the trunk; wat.sprout FB = number of water sprouts on fruiting branches; ann.growth FB top = annual shoot length at the distal part of fruiting branches at the top of the tree; ann.growth FB bott. = annual shoot length at the distal part of fruiting branches at the bottom of the tree; later.distr. = distribution of laterals; area FB = sum of fruiting branch sectional areas; height = row height.

sum of fruiting branch sectional areas, at the top. Distribution of laterals was poorly represented on this factorial plane, perhaps because of the difficulty of measuring such a qualitative variable. The projection of variables on the plane formed by the second and third factors of the component analysis is given in Fig. 3. The third factor (10% of the total variance) opposed annual shoot length, at the bottom, to all other variables except TCSA and scion-rooting, at the top. Most variables were poorly represented on this factorial plane. Annual shoot length at the distal part of fruiting branches at the top or at the bottom of the tree was associated on both factorial planes, as water sprout number on the trunk or on fruiting branches. Factorial planes were unchanged when only younger than 20 years old orchards were considered or when TCSA was divided by orchard age (data not shown). Since all variables, except distribution of laterals and grafting point height, were grouped according to factor 1, this factor can be interpreted as a growth factor. It may therefore express plot vigour. The second factor can be interpreted as opposition between annual growth (annual shoot length and number of water sprouts), at the bottom, and cumulative growth (scion-rooting, TCSA, height, row width and sum of fruiting branch sectional areas), at the top. Thus, it may be a factor of tree age, meaning that young trees often have considerable annual growth. Finally, since the third factor opposed annual shoot length to number of water sprouts or distribution of laterals, it showed that plots can be trained with high annual shoot growth and few water sprouts, which is the case of well-balanced orchards. Therefore, it may represent the vigour balance. But plot classification based on this factor was not correlated to expert balance marks (data not shown).


83

Fig. 3. Projection of morphological variables on the factorial plane made of the second and third factors of the component analysis. TCSA = trunk cross-sectional area; graft.pt.height = grafting point height; scion-root. = scion-rooted; wat.sprout trunk = number of water sprouts on the trunk; wat.sprout FB = number of water sprouts on fruiting branches; ann.growth FB top = annual shoot length at the distal part of fruiting branches at the top of the tree; ann.growth FB bott. = annual shoot length at the distal part of fruiting branches at the bottom of the tree; later.distr. = distribution of laterals; area FB = sum of fruiting branch sectional areas; height = row height.

3.2. Classification of tree vigour Fig. 4 represents the projection of clusters of plots on the factorial plane formed by the first two factors. The size of clusters 1–4 was 23, 28, 31 and 25, respectively. The fourth cluster variability appeared to be larger than the one for other clusters. The expert vigour marks are given in Table 2. For each plot, it can be observed that expert mark dispersion is rather small, except for plots 4, 9, 11, 13 and 14 (at least one difference equal or greater than 1.5 units between two different expert marks). For three of these plots, the experts ascertained that there was a high degree of heterogeneity between trees, usually related to scion-rooting. The comparison between modelled and mean expert vigour marks is shown in Table 3. According to the Wilcoxon rank sum test, both plot vigour classifications were not significantly different (P < 0.05). Eight plots were given the same marks by modelling vigour or by experts. On the contrary, for the six poorly marked plots, a difference of just one unit occurred between both classifications. Three of these plots correspond to heterogeneous orchards according to expert comments. 3.3. Simplification of vigour modelling To make vigour evaluation easier, plot vigour modelling was re-estimated with only three easy-to-measure morphological variables (TCSA, number of water sprouts on the

84


Fig. 4. Projection of plots on the plane formed by the first two factors. Plot clusters based on plot coordinates on the first two factors are represented by variance ellipses (confidence level: 95%). The numbers located next to the ellipses refer to the cluster numbers, and thus to plot vigour marks, used in the text. Bold characters indicate plots used for expert vigour marking.

trunk and annual shoot length at the bottom of the tree). The comparison between original (11 variables) and re-estimated modelled plot vigour mark distributions is shown in Table 4. According to the Wilcoxon rank sum test, both modelled plot vigour classifications were significantly different (P < 0.05). As suggested by Table 4, Table 2 Expert vigour marks and rounded-off means Plot number

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Expert number

Rounded-off mean

1

2

3

4

5

6

7

8

1 2 3 3 2.5 4 4 3 2.5 2 2.5 2 3 2.5

2 2 3 3 2 4 4 4 2 2 3 2 3 3

1 2 3 4 2 4 4 3 3 2 2.5 2 3 3

1 3 3.5 3.5 2 3.5 4 3 2 2 2 2 – –

1.5 2.5 3 2.5 2.5 3.5 3.5 3.5 2 2 2.5 2.5 1.5 2

1.5 2 3.5 2.5 1.5 4 4 3.5 1.5 2 2.5 1.5 2.5 1

2 3 3.5 3.5 1.5 4 4 4 2 1.5 1.5 2.5 2.5 1.5

– – – – 2 4 4 4 2 2 2.5 2 3 1.5

1 2 3 3 2 4 4 4 2 2 2 2 3 2


85

Table 3 Contingency table between modelled and mean expert vigour marks Modelled vigour marks

Expert vigour marks

1 2 3 4

1

2

3

4

1 0 0 0

2 3 2 0

0 0 1 2

0 0 0 3

Numbers in italics indicate plots for which both markings do not fit.

Table 4 Contingency table between original and re-estimated plot vigour marks Original classification

1 2 3 4

Re-estimated classification 1

2

3

4

19 5 0 3

4 29 5 2

0 4 26 5

0 0 0 15


classifications differed mainly for the original cluster 4. When cluster 4 plots were removed, both classifications did not differ (P > 0.05). Since variances were homogeneous concerning age between original plot clusters, the analysis of variance was performed and showed significant differences: cluster 4 gathered older plots (Fig. 5). On the contrary,

Fig. 5. Boxplot representing plot age according to the original modelled vigour marks. The horizontal white layer, the extremities of the box, the square bracket and the extreme strokes represent the mean, the upper and lower quartile, the upper and lower extreme (excluding outliers) and the outliers of the variable, respectively. Anything farther than 1.5 times the interquartile range is considered to be an outlier. Different letters in front of the mean indicate significant differences at the P = 0.01 level with a Tukey test.

86


Table 5 Contingency table between original and TCSA-based plot vigour marks Original classification

TCSA-based classification 1

2

3

4

1 2 3 4

23 34 25 0

0 0 0 10

0 0 0 7

0 4 6 8


re-estimation of modelled vigour marks for plots in clusters 1–3 was fairly good since 75% of the plots were identically classified by both modelling methods. When replacing TCSA by the sum of fruiting branch sectional areas, modelled vigour mark re-estimation was not much better (data not shown). Therefore, vigour modelling for older plots is difficult because trees are often scion-rooted or affected by tree shape restructuring (Nesme et al., 2003). A re-estimation of modelled vigour mark distribution based only on TCSA was highly different from the original modelled vigour mark distribution, as shown in Table 5 (P < 0.01).

4. Discussion The first factor of the component analysis was interpreted as plot vigour factor. This implies that all variables grouped according to factor 1 (i.e. all morphological variables except distribution of laterals and grafting point height) provide information about vegetative growth. On the contrary, the grafting point height, typically opposed to scionrooting, was rather antagonist to vegetative growth, as mentioned by other authors (Parry, 1986; Quamme et al., 1998). The second factor was interpreted as the opposition between annual and cumulative growth. Therefore, any variable among scion-rooting, TCSA, row width and sum of fruiting branch sectional area can be chosen as a representative of cumulative growth. On the contrary, any variable indicating the annual shoot length or the number of water sprouts is a representative of annual growth. The third factor represents the vigour balance, but it is poorly correlated to expert balance marks. This may be due to the lack of data about crop load as vigour balance may be related to the balance between vegetative and reproductive growth. Therefore, variables such as crop load, pruning type or branch description should have been recorded in order to provide a more accurate vigour balance evaluation. Original modelled and expert vigour classification appeared in relative good agreement (Table 3). However, cases of disagreement may be due to the fact that measurements and expert evaluation were carried out in 2002 and 2003, respectively. Although cumulative growth may not vary greatly from one year to the next, annual growth may vary, especially if orchards bear fruit alternately or if farmers change their practices or carry out traumatic operations such as renewal pruning during winter (Lakso, 1984). Classification disagreement may also be due to plot heterogeneity or to expert mark dispersion. Finally,


87

classification disagreement may be due to factors that experts take into account to evaluate plot vigour. In the reasons they gave to justify their decision, experts mentioned their assessment was mainly based on the number of water sprout on branches, tree height, canopy porosity, annual shoot length, scion-rooting, fruiting branch structure, etc. but also pruning practice, variety, fruit load or shoot length distribution within tree canopy. They also mentioned they integrated the prognosis of plot behaviour or the different ease of tree management between varieties. Therefore, expert vigour assessment appeared as a complex subjective process, not only related to botanic features, as mentioned by Navarrete et al. (1997), which could not be entirely simulated by morphological plant measurements. Although TCSA is a widely used indicator, related to above-ground (Barden and Marini, 2001) or whole tree (Westwood and Roberts, 1970; Strong and Azarenko, 2000) vegetative biomass, it was not sufficient to evaluate plot vigour in our study (Table 5). In particular, TCSA was not related to the sum of fruiting branch sectional areas (Fig. 1) for older plots and, therefore, not to leaf weight or to vegetative development pattern either (Lauri et al., 1997). Other authors had already mentioned that a single criterion could not be used to model crop vigour and give the same results as expert evaluation (Navarrete et al., 1997). However, a combination of three easy-to-measure variables (TCSA, number of water sprouts on the trunk, annual shoot length) provided a modelled plot vigour classification closer to the original classification, particularly for younger plots. Although some variables (such as the number of water sprouts on the trunk or on fruiting branches and the length of annual shoot at the distal part of fruiting branches at the top or the bottom of tree) appeared to be redundant, considering all 11 morphological variables as a whole gave more information about model tree vigour. This is due to interactions between morphological traits such as between the number of water sprouts on the trunk and tree height, for example.

5. Conclusion The proposed set of 11 morphological variables made it possible to model plot vigour. To our knowledge, it is the first attempt to identify morphological variables enabling to evaluate orchard vigour. Testing this set of variables in various cropping conditions showed an opposition between indicators of annual and cumulated growth and gave initial elements for estimating the balance between vegetative vigour and fruiting. Although TCSA did not appear to be a sufficient indicator of tree vigour in such a diverse orchard population, the set of variables can be simplified for younger plots to make field measurements easier by extension services for example. The three morphological variables used to re-estimate plot vigour marks (TCSA, number of water sprouts on the trunk and annual shoot length at the distal part of fruiting branches at the bottom of tree) can be measured on ten representative trees per plot within 15 min. Plot vigour mark re-estimation agrees in 75% of the cases with the more rigorous vigour modelling. This type of tree measurement method is important to formalise a non-formal widely used indicator by farmers and technical advisers for irrigation or N fertilisation adjustment. Establishing such a correspondence may lead to the sharing of common crop indicators for crop management by researchers and technicians (Meynard et al., 2002). This may be helpful for research aimed at understanding farmers’ cropping practices.

88


Acknowledgments The authors would like to thank Claude-Eric Parveaud for preliminary work on apple tree vigour evaluation, Pierre Rouet, Re´ gis Laurent and Yacine Ikhlef for help in collecting morphological data, Gail Wagman for improving the English, the technical advisers for playing the role of experts and the farmers for allowing us to make field measurements. This work was funded by the nationwide INRA programme on Integrated Fruit Production ‘‘Action Transversale 67, Production Fruitie`re Inte´ gre´ e’’. References Baize, D., Girard, M.-C., 1995. Re´ fe´ rentiel pe´ dologique. INRA, Paris. Barden, J.A., Cline, J.A., Kushad, M.M., Parker, M.L., 2002. Various measures of tree vigor, yield, and efficiency of apple trees in the 1990 NC-140 systems trial as influenced by location, cultivar and orchard system. J. Am. Pomol. Soc. 56, 208–214. Barden, J.A., Marini, R.P., 1997. Growth and fruiting of a spur-type and a standard strain of ‘Golden Delicious’ on several rootstocks over eighteen years. Fruit Varieties J. 51, 165–175. Barden, J.A., Marini, R.P., 2001. Comparison of methods to express growth, size and productivity of apple trees. J. Am. Pomol. Soc. 55, 251–256. Berman, M.E., DeJong, T.M., 2003. Seasonal patterns of vegetative growth and competition with reproductive sinks in peach (Prunus persica). J. Hortic. Sci. Biotechnol. 78, 303–309. Crabbe´ , J., 1987. Aspects particuliers de la morphogene`se caulinaire des ve´ ge´ taux ligneux et introduction a` leur e´ tude quantitative. IRSIA, Bruxelles. Duchesneau, R., Lesage, I., Messier, C., Morin, H., 2001. Effects of light and intraspecific competition on growth and crown morphology of two size classes of understory balsam fir saplings. Forest Ecol. Manage. 140, 215– 225. Escofier, B., Page`s, J., 1988. Analyses factorielles simples et multiples. Bordas, Paris. Gunkel, J.E., Thimann, K.V., Wetmore, R.H., 1949. Studies of development in long and short shoots of Ginkgo biloba L.: IV. Growth habit, shoot expression and the mechanism of its control. Am. J. Bot. 36, 309–316. Khatamian, H., Hilton, R.J., 1977. The relationship between shoot growth and area of trunk cross-section in several woody plant species. HortScience 12, 255–257. Kurian, R.M., Iyer, C.P.A., 1992. Stem anatomical characters in relation to tree vigour in mango (Mangifera indica L.). Sci. Hortic. 50, 245–253. Kyto¨ , M., Niemela, P., Annila, E., Varama, M., 1999. Effects of forest fertilization on the radial growth and resin exudation on insect-defoliated Scots pines. J. Appl. Ecol. 36, 763–769. Lakso, A.N., 1984. Leaf area development patterns in young pruned and unpruned apple trees. J. Am. Soc. Hortic. Sci. 109, 861–865. Lauri, P.-E., 2002. From tree architecture to tree training: An overview of recent concepts developed in apple in France. J. Korean Soc. Hortic. Sci. 43, 782–788. Lauri, P.-E., Lespinasse, J.-M., 1999. De l’axe vertical au solaxe, vers un renouvellement des concepts. Le fruit belge 477, 25–31. Lauri, P.-E., Te´ rouanne, E., Lespinasse, J.-M., 1996. Quantitative analysis of relationships between inflorescence size, bearing-axis size and fruit-set: An apple tree case study. Ann. Bot. 77, 277–286. Lauri, P.-E., Te´ rouanne, E., Lespinasse, J.-M., 1997. Relationship between the early development of apple fruiting branches and the regularity of bearing: An approach to the strategies of various cultivars. J. Hortic. Sci. Biotechnol. 72, 519–530. Lauri, P.-E., Te´ rouanne, E., Lespinasse, J.-M., Regnard, J.-L., Kelner, J.-J., 1995. Genotypic differences in the axillary bud growth and fruiting pattern of apple fruiting branches over several years: An approach to regulation of fruit bearing. Sci. Hortic. 64, 265–281.


89

Lehman, L.J., Young, E., Unrath, C.R., 1990. Apple tree vigor influences flowering and dry weight after Paclobutrazol application. HortScience 25, 933–935. Lespinasse, J.-M., 1977. La conduite du pommier: Type de fructification, incidence sur la conduite de l’arbre. INRA, Paris. Li, S.H., Huguet, J.-G., Schoch, P.-G., Bussi, C., Orlando, P., L’Hoˆ tel, J.-C., 1990. Re´ ponse de jeunes peˆ chers cultive´ s en pots a` diffe´ rents re´ gimes d’alimentation hydrique: II. Effets sur la croissance et le de´ veloppement. Agronomie 10, 353–360. Lo Bianco, R., Policarpo, M., Scariano, L., 2003. Effects of rootstock vigour and in-row spacing on stem and root growth, conformation and dry-matter distribution of young apple trees. J. Hortic. Sci. Biotechnol. 78, 828– 836. Lobit, P., Soing, P., Ge´ nard, M., Habib, R., 2001. Effects of timing of nitrogen fertilization on shoot development in peach (Prunus persica) trees. Tree Physiol. 20, 35–42. Marsh, K.B., Volz, R.K., Cashmore, W., Reay, P., 1996. Fruit colour, leaf nitrogen level, and tree vigour in ‘Fuji’ apples. N. Z. J. Crop Hortic. 24, 393–399. Mataa, M., 2000. Performance of some apple cultivars under tropical Zambian conditions. J. Hortic. Sci. Biotechnol. 75, 346–349. MathSoft, 1999. S-Plus 2000: Guide to Statistics. MathSoft, Seattle. Meynard, J.-M., Cerf, M., Guichard, L., Jeuffroy, M.-H., Makowski, D., 2002. Which decision support tools for the environmental management of nitrogen? Agronomie 22, 817–829. Moravie, M.-A., Durand, M., Houllier, F., 1999. Ecological meaning and predictive ability of social status, vigour and competition indices in a tropical rain forest (India). Forest Ecol. Manage. 117, 221–240. Navarrete, M., Jeannequin, B., Sebillotte, M., 1997. Vigour of greenhouse tomato plants (Lycopersicon esculentum Mill.): analysis of the criteria used by growers and search for objective criteria. J. Hortic. Sci. 72, 821–829. Neilsen, G.H., Parchomchuk, P., Berard, R., Neilsen, D., 1997. Irrigation frequency and quantity affect root and top growth of fertigated ‘McIntosh’ apple on M9, M26 and M7 rootstock. Can. J. Plant Sci. 77, 133–139. Nesme, T., Lescourret, F., Bellon, S., Ple´ net, D., Habib, R., 2003. Relevance of orchard design issuing from growers’ planting choices to study fruit tree cropping systems. Agronomie 23, 651–660. Parry, M.S., 1986. The effects of budding height on the field performance of two apple cultivars on three rootstocks. J. Hortic. Sci. 61, 1–7. Penrose, L.J., Nicol, H.I., 1996. Aspects of microclimate variation within apple tree canopies and between sites in relation to potential Venturia inaequalis infection. N. Z. J. Crop Hortic. 24, 259–266. Quamme, H.A., Hampson, C.R., Brownlee, R.T., 1998. Use of planting depth and budding height to modify vigour control of Ottawa 3 apple rootstock. Can. J. Plant Sci. 78, 353–355. Ro, H.-M., Park, J.-M., 2000. Nitrogen requirements and vegetative growth of pot-lysimeter-grown ‘Fuji’ apple trees fertilized with three nitrogen rates. J. Hortic. Sci. Biotechnol. 75, 237–242. Singh, L.B., 1948. Studies in biennial bearing: IV. Bud-rubbing, blossom-thinning and defoliation as possible control measures. J. Hortic. Sci. 24, 159–177. Sinoquet, H., Le Roux, X., Adam, B., Ameglio, T., Daudet, F.A., 2001. RATP: a model for simulating the spatial distribution of radiation absorption, transpiration and photosynthesis within canopies: Application to an isolated tree crown. Plant Cell Environ. 24, 395–406. Sprent, P., Ley, J.P., 1992. Pratique des statistiques non parame´ triques. INRA, Paris. Spring, J.L., Chapuis, P., Eve´ quoz, C., Girardet, G., Ryser, J.P., Schmid, C., Terrettaz, R., Thentz, M., Vanetti, R., 1993. La fertilisation des arbres fruitiers, kiwis et des arbustes a` baies. Rev. Suisse Vitic. Arboric. Hortic. 25, 189–199. Strong, D., Azarenko, A.N., 2000. Relationship between trunk cross-sectional area, harvest index, total tree dry weight and yield components of ‘Starkspur supreme delicious’ apple trees. J. Am. Pomol. Soc. 54, 22–27. Trillot, M., Masseron, A., Tronel, C., 1993. Pomme, les varie´ te´ s. Ctifl, Paris. Vincent, G., de Foresta, H., Mulia, R., 2002. Predictors of tree growth in a Dipterocarp-based agroforest: a critical assessment. Forest Ecol. Manage. 161, 39–52. Waring, R.H., Thies, W.G., Muscato, D., 1980. Stem growth per unit of leaf area: a measure of tree vigor. Forest Sci. 26, 112–117.

90


Way, R.D., Sanford, J.C., Lakso, A.N., 1983. Fruitfulness and productivity. In: Moore, J.N., Janick, J. (Eds.), Methods of Fruit Breeding. Purdue University Press, West Lafayette, pp. 353–367. Webster, A.D., Hollands, M., 1999. Orchard comparisons of ‘Cox Orange Pippin’ grown on selections of the apple rootstock M9. J. Hortic. Sci. Biotechnol. 74, 513–521. Weinbaum, S.A., Johnson, R.S., DeJong, T.M., 1992. Causes and consequences of overfertilization in orchards. HortTechnology 2, 112–121. Westwood, M.N., Roberts, A.N., 1970. The relationship between trunk cross-sectional area and weight of apple trees. J. Am. Soc. Hortic. Sci. 95, 28–30.