Annals of Botany 105: 607 –616, 2010 doi:10.1093/aob/mcq006, available online at www.aob.oxfordjournals.org

Insights into secondary growth in perennial plants: its unequal spatial and temporal dynamics in the apple (Malus domestica) is driven by architectural position and fruit load ´ . Lauri1,*, J. J. Kelner1, C. Trottier2 and E. Costes1 P. E 1

UMR DAP, INRA-SUPAGRO-CIRAD-UM II. Equipe ‘Architecture et Fonctionnement des Espe`ces Fruitie`res’, CIRAD Lavalette, Avenue Agropolis, TA A-96/03, 34398 Montpellier Cedex 5, France and 2Universite´ Montpellier II, UMR I3M, #5149, Equipe ‘Probabilite´s et Statistique’, 34095 Montpellier Cedex 5, France * For correspondence. E-mail [email protected] Received: 26 October 2009 Returned for revision: 27 November 2009 Accepted: 14 December 2009 Published electronically: 12 March 2010

† Background and Aims Secondary growth is a main physiological sink. However, the hierarchy between the processes which compete with secondary growth is still a matter of debate, especially on fruit trees where fruit weight dramatically increases with time. It was hypothesized that tree architecture, here mediated by branch age, is likely to have a major effect on the dynamics of secondary growth within a growing season. † Methods Three variables were monitored on 6-year-old ‘Golden Delicious’ apple trees from flowering time to harvest: primary shoot growth, fruit volume, and cross-section area of branch portions of consecutive ages. Analyses were done through an ANOVA-type analysis in a linear mixed model framework. † Key Results Secondary growth exhibited three consecutive phases characterized by unequal relative area increment over the season. The age of the branch had the strongest effect, with the highest and lowest relative area increment for the current-year shoots and the trunk, respectively. The growth phase had a lower effect, with a shift of secondary growth through the season from leafy shoots towards older branch portions. Eventually, fruit load had an effect on secondary growth mainly after primary growth had ceased. † Conclusions The results support the idea that relationships between production of photosynthates and allocation depend on both primary growth and branch architectural position. Fruit load mainly interacted with secondary growth later in the season, especially on old branch portions. Key words: Branch age, fruit load, growth phase, Malus domestica (apple), primary growth, secondary growth, tree architecture.

IN T RO DU C T IO N The relationships between sources of assimilates and metabolic sinks are a function of the carbon assimilation capacity of source organs, growth rate and size of the sink, and transport between the two (Ho, 1988, 1996; Finazzo et al., 1994; Marcelis-Van Acker, 1994; Farrar, 1996; Pallas et al., 2008). They are thus a property of the entire plant system (Minchin and Lacointe, 2005) and evolve dynamically through the season (Preston, 1998; Novoplanski, 2003; Orians et al., 2005). Factors affecting these relationships are external and internal to the plant. In the first case the status of the whole plant in its environment significantly affects growth with, as demonstrated in forest trees, an earlier, faster and longer growth in the dominant tree as compared with the suppressed tree (Kozlowski, 1963). Furthermore, the removal of leaf area may lead to the compensation of photosynthetic activity of the remaining leaf area at least up to a certain threshold (Layne and Flore, 1992). At the within-tree scale, the growth of an organ depends on the microclimatic context, e.g. light vs. shade, in which it develops with regard to other tree parts (Umeki and Seino, 2003; Lacointe et al., 2004; Umeki et al., 2006). However, the priority of an organ for assimilates (carbon, water) also depends on its position relative to the other organs on the same tree (Sprugel, 2002) and with

respect to tree ontogeny (Sterck and Bongers, 2001). A sink hierarchy is usually considered among organs or tissues of perennial species and summarized by the following priority rank ordering from high to low: seeds . fleshy fruit part . shoot apices . mature leaves . cambium . roots . storage (Oliveira and Priestley, 1988; Wardlaw, 1990; Wolstenholme, 1990; Obeso, 2002; McQueen et al., 2004). However, as pointed out by Minchin and Lacointe (2005) priority does not mean strength and if seeds have high priority they generally have a low strength because of their small size. This ordered sink hierarchy is not sufficient in itself to explain the complex and dynamic interactions between organs and tissues. Indeed, the priority of each potential sink evolves during the growing season. On a large scale, the competitions between generative and vegetative sinks are well illustrated by the abscission of flowers and fruits to match fruit and seed number with the available resources over a wide range of environmental conditions (Stephenson et al., 1981; Corelli-Grappadelli et al., 1996). Within the vegetative compartment, i.e. between primary and secondary (cambial) growth, the latter one is usually not considered to behave as a high priority sink as exemplified by its median position among all potential sinks. However, due to the large biomass of secondary tissues at the whole-tree scale it has a high

# The Author 2010. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. For Permissions, please email: [email protected]

608

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics

sink strength and it has been stressed that it is an active component of tree growth with documented competition with epicormic shoot production (Ishii et al., 2007) and fruit growth (Forshey and Elfving, 1989). A major part of the increase in dry weight associated with secondary growth occurs after cessation of primary growth, suggesting that the former cannot occur before the latter has occurred (Forshey and Elfving, 1989). However, the causal relationships between the two are not univocal and it has been suggested that secondary growth may contribute to shoot primary growth cessation (Barnola and Crabbe´, 1993; Costes et al., 2000). Ancient studies have demonstrated the positive relationship between secondary growth and the development of leaves (e.g. Knudson, 1916). Indeed, there is a positive relationship between the diameter at the base of a branch composed of one or more units of extension and the leaf area above (Briand et al., 1998; Brouat et al., 1998; Mansour and de Fay¨, 1998), and between the diameter of the trunk and of scaffold branches, and the number of distally situated branches (Oosthuyse and Jacobs, 1995). These relationships have been formalized in the ‘pipe model’ which stresses that stems and branches may be considered as the assemblage of unit pipes each supporting a unit amount of photosynthetic organs (Shinozaki et al., 1964). Such relationships result from a dynamic trade-off between the optimization of water transport and the maintenance of mechanical stability (Ma¨kela¨, 2002; McCulloh and Sperry, 2005; Niklas et al., 2006; Normand et al., 2008). However, allometric relationships between the basal diameter of a stem and the leaf mass it bears in distal position are dependent on the age of the stem. As shown by Suzuki and Hiura (2000) on a range of forest species, allometric relationships between the two variables are compatible with the ‘pipe model’ for old branches directly inserted on the trunk but not for current-year shoots which may have a more plastic behaviour related to their light interception function. These results would suggest that the overlap in time between primary growth, encompassing shoot extension and fruit growth, and secondary growth plays an important role in shaping these allometric relationships. Fruit trees offer a unique opportunity to address the question of the hierarchy between the processes which compete with secondary growth and may impact on its dynamics through a growing season. Two main processes should be considered: (1) leaves produced during primary vegetative growth as the carbohydrate source; (2) fruit growth as a competitive sink. In this framework the architectural position of the branch portion affects the distance between sources and sinks and should be considered as a factor affecting the relationship between these processes. In this paper, branch portion age will be considered as a proxy for architectural position, the older the branch portion the greater the distance from the source. The study was developed in the apple tree on which the relationships between fruiting and primary growth have been extensively studied at the whole-tree scale in relation to the fruit-load concept, i.e. the amount of fruit per unit of vegetative biomass (Forshey and Elfving, 1989; Bound, 2001). There is, however, a lack of knowledge on the dynamic relationships through the growing season between primary and secondary growth, and fruit load, either at the whole-tree scale, or at the annual shoot scale (i.e. considering the presence

or absence of an adjacent fruit). The following questions were addressed. (a) How is the secondary growth dynamics over a growing season: regular or with distinct phases? (b) What are the relationships between secondary growth and the dynamics of primary shoot and fruit growth? (c) What are the respective effects of the factors which are likely to affect secondary growth, namely branch portion age and fruit load? M AT E R I A L S A N D M E T H O D S Plant material and treatments

The experiment was located at the INRA research field, near Montpellier (438350 N, 38520 E). Six-year-old ‘Golden Delicious’ trees grafted on M9 (Pajam 1) rootstock and planted in adjacent rows at distances of 6 m between rows and 1.8 m between trees in a same row were used in this study. Tree training included minimal pruning strategies and were trained as Solaxe trees with a vertical trunk bent at a height of approx. 2.5 m and regularly spaced fruiting branches (Lauri and Lespinasse, 2000; Lauri, 2009; Fig. 1A). Agricultural practices including irrigation with microsprinklers, fertilization and spraying against pests and diseases were done according to standard practices in the area. At full bloom (12 April 1999), 30 full-flowering trees were chosen for their uniformity of shape and volume. On the 15 A

5 4 3 2

1

6

B

Bourse-shoot

Adjacent fruit F I G . 1. (A) Solaxe tree with branch portion age (from ‘6’, at trunk bottom, to ‘1’, current year shoot). (B) Detail of a current-year bourse-shoot with an adjacent fruit.

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics May, following physiological fruit drops, trees were subjected to fruit-load treatments expressed as the ratio of the number of fruit per unit of trunk cross sectional area (TCSA; Lombard et al., 1988). Although this estimation of fruit load is subjected to criticism especially for old trees (Corelli-Grapadelli and Lakso, 2004) it is still an easy-to-use and reliable measure on young and lightly pruned trees as was the case here. Fruit thinning was carried out by hand so as to obtain low (L), medium (M) and high (H) fruit load set at 5, 15 and 25 fruits cm22 of TCSA, respectively, on ten randomly chosen trees per treatment. This range of fruit load was set so as to include the generally accepted 7 – 8 fruits cm22 of TCSA considered as an optimized fruit load to balance the number of fruits per tree and individual fruit size (Roper, 1995). As TCSA ranges from 11.5 to 22.9 cm2 (average, 17.0 cm2) trees were grouped according to initial TCSA and equally distributed among the three fruit-load levels. Fruit thinning was carried out uniformly around the tree canopy to avoid any possible localized effect of fruit load on primary and secondary shoot growth and fruit growth (Palmer et al., 1991). As a rule of thumb only the king fruit was kept on the conserved flower clusters. Data collection

Trees were subjected to the following measurements. First, shoot primary growth, i.e. growth resulting from the functioning of the terminal meristem, was monitored on four shoots per tree distributed around the tree crown at breast height and developing from 20- to 30-cm-long shoots grown in the previous year. Since there were almost no vegetative shoots the year of study, all the shoots measured belonged to the bourseshoot category, i.e. they develop in axillary position on a flower cluster which set fruit (Fig. 1B). For brevity they will be hereafter referred to as shoots only. To study the effect of an adjacent fruit on primary growth of the bourse-shoot further, fruitlets of half of the flower clusters, i.e. two out of four per tree, were removed by hand on 26 April. Twenty shoots were then studied per fruit load  presence of an adjacent fruit combination. On each shoot, the length (cm) was monitored to the nearest millimetre. Leaf area per shoot (cm2) was also measured by counting the number of leaves per shoot, and using an allometric relationship between the product of the length by the width of each leaf blade and the actual leaf area obtained from a destructive sub-sample of leaves (data not shown). Measurements were made every week from 14 April to the end of shoot primary growth, i.e. until mid-July for the later growing leaves. Secondly, fruit size, estimated by the equatorial diameter, was monitored on four fruits per tree selected on flower clusters in the terminal position on shoots distributed around the tree, yielding up to 40 fruits per fruit-load treatment. Measurements were made every week, beginning on 14 April, and ending on 21 July, i.e. after the end of primary growth of all shoots. A final measurement was carried out at harvest on a sub-sample of fruits (15 September). Measurements were made in the morning and approximately at the same time for a given fruit to limit the introduction of likely within-day variations. Fruit diameter was further transformed in fruit volume (cm3) using the following relationship

609

obtained from a sub-sample of fruits collected at various stages of fruit growth: volume ¼ 0.0018  diameter265 (n ¼ 30; r 2 ¼ 0.99; P , 0.001). Eventually, branches were divided in consecutive portions, each one being characterized by its age. On each portion, secondary growth was monitored using an electronic calliper (Codiam Scientific) to the nearest 1/100 mm. One- (currentyear shoot) to 5-year-old portions, on three branches per tree, and 6-year-old portions (trunk, age 6) at 20 cm above the grafting point, were measured yielding up to 16 points per tree and 160 points per fruit-load treatment (Fig. 1A). Measurements were made at weekly intervals from 14 April to 15 September, the harvest date. As for fruit, measurements were made in the morning and approximately at the same time for a given branch age to limit the introduction of the likely daily variations of branch diameter in the study (Daudet et al., 2005). Diameters were then transformed to area assuming that the cross-section of the branch portions was a disc. The analyses were done on the relative daily area increment (RDAI), computed as follows: [(Akþ1 – Ak)/A0]/(Dkþ1 – Dk), in day21, where A0 is the initial area value of secondary growth, D is the date, and k and k þ 1 are the indices of consecutive measurements. ln(RDAI) was used to fit with the normality of residuals hypothesis. For brevity RDAI will be used hereafter. Data analysis

Current-year shoot primary growth was assessed in two ways: (1) by plotting the number of growing shoots; (2) by plotting the length (cm) and the total leaf area (cm2) of all growing shoots at each measurement date. Then an ANOVA was carried out to determine the respective effects on the final length and leaf area of all shoots of two factors, fruit load and the presence of an adjacent fruit. Similarly fruit volume (cm3) was plotted against date from the beginning of April to harvest. At each date the effect of fruit load on fruit volume was assessed through an ANOVA. In both cases, when needed, the ANOVA was followed by an honestly significant difference (HSD) multiple mean comparison Tukey test at P , 0.05. The analysis of secondary growth of current-year shoots (‘age 1’), and of 2- to 6-year-old branch portions (namely ‘age 2’ to ‘age 6’) was carried out on branch portions with active secondary growth according to two consecutive steps. First, consecutive growth phases were visually assessed based on both the dynamics of cumulated secondary growth increment (Fig. 4A) and the proportion of branch portions with active secondary growth (Fig. 4B). Secondly, the effects on secondary growth of the three factors, namely growth phases, branch portion age (from 1 to 6), and fruit load, were investigated through an ANOVA-type analysis in a linear mixed model framework. The model includes a branch and a branch portion age as random effects and phase of growth and fruit load as fixed effects. Introducing random effects into the model enables the dependency of the repeated measurements on the same branch portion to be taken into account and also the identification of inter-branch portion variability. The random effect structure was chosen using the Akaike’s information criterion (AIC). The AIC is a

610

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics

measure of the goodness-of-fit of an estimated statistical model taking into account the maximized value of the likelihood function penalized by the number of parameters in order to reach a most parsimonious model. A parsimonious model for the fixed effect part was selected starting with the most complex model, and then removing step by step the less significant terms (for each factor level and order 2 and 3 interactions), keeping the AIC always decreasing. This led to a 33-parameter model. Finally, a specific analysis was carried out on the effects on current-year shoot RDAI (age 1) of two factors – fruit load and the presence of an adjacent fruit. As for the previous analysis a parsimonious model was selected starting with the most complex model, and then removing step by step the less significant order 2 and 3 interaction terms. The models were investigated using R software, package Lme4 with the function lmer, version 2.7.2 (http://www.rproject.org). RES ULT S Dynamics of primary shoot growth: stem length and leaf area

After a first phase of primary growth until mid-May, there was a fast drop in the proportion of growing shoots by the beginning of June, with ,30 % of growing shoots on the 9 June except for the H-NF (high fruit load–no presence of an adjacent fruit) treatment (Fig. 2A). All shoots had stopped their growth in the beginning of July (Fig. 2A). Final shoot length reached 130–210 mm (Fig. 2B). Leaf area continued to grow after shoot elongation had stopped, reaching a final leaf area per shoot of 240–350 cm2 (Fig. 2C). There were no significant effects on shoot length or shoot leaf area of either the fruit load or the presence of an adjacent fruit, or of their interactions (Table 1). Dynamics of fruit growth

Whatever fruit load, fruit volume followed a classical sigmoid curve with an exponential phase until the beginning of June, a linear phase until the beginning of August and, although not precisely assessed, a decrease in volume increment thereafter (Fig. 3). A first significant effect of fruit load was observed from the 27 May onwards with a significantly higher fruit volume on L-trees compared with H-trees, M-trees being in intermediate position (Fig. 3). At harvest, L-trees maintained their higher fruit volume compared with fruit on both M- and H-trees (Fig. 3). Dynamics of secondary growth

Since the cumulated secondary growth area (Fig. 4A) and the proportion of branch portions with active secondary growth (Fig. 4B) were more strongly affected by branch portion age than by fruit load (data not shown) only the effects of age are represented here (Fig. 4A and B). The progression of cumulated secondary growth area over the growing season exhibited three main consecutive growth phases with low (from budburst to 5 May; phase I), maximal (from the 14 May to 3 June; phase II) and eventually intermediate (from the 9 June to 15 September; phase III) values

(Fig. 4A). Paralleling these variations in cumulated growth area, the proportion of branch portions with active secondary growth was highest in phase II with low to intermediate values in phases I and III, respectively. However, various patterns were observed depending on age: compared with ages 2 – 5 with similar variations, age 1 generally had a lower number of shoots with secondary growth in phase III, whereas age-6 portions were more numerous in phase I and with more irregular variations in phase III (Fig. 4B). Variations of RDAI presented similar patterns on the three tree fruit load (data not shown, only L-fruit load being presented here; Fig. 4C): whatever branch portion age, RDAI had low to intermediate values in phases I and III and the highest values in phase II. The six branch-portion ages were well discriminated all through the growing season with decreasing values from age 1 – 5, and low values with a more variable pattern for age 6. The selected model showed that RDAI was mainly affected by age with, whatever fruit load and growth phase, a significant decrease of RDAI between age 1 and all the other ages, and within the latter group ordered decreasing effects from age 2 to age 6 (Table 2, lines 1 – 5). The growth phase had a lower effect with, compared with phase I and whatever fruit load and age, higher values for phase II (Table 2, line 6, effect: þ0.87, P , 10220) and lower values for phase III (Table 2, line 7, effect: 20.75, P , 10220). Eventually, fruit load had the lowest effect with, whatever growth phase and age, higher values on the L- and, to a lesser extent, the M-fruit load, as compared with the H-fruit load (Table 2, lines 8 and 9). These main effects were modulated by interactions. Strong positive interactions were observed on RDAI between ages 2– 6 and phase II. This meant that the global positive effect of phase II on RDAI was further enhanced on age 2 to 6, compared with age 1 (Table 2, lines 10 – 14; see the higher increase of RDAI of age 2 – 6 compared with age 1 between phase I and phase II in Fig. 4C). Strong positive interactions were also observed on RDAI between age 2 –6 and phase III, meaning that the global negative effect of phase III was counterbalanced on age 2 – 5 compared with age 1 (Table 2, lines 15– 18; Fig. 4C; see the lower values of RDAI of age 1 in phase III compared with phase I, whereas for the other ages phase III values were intermediate between phases I and II). This interaction was especially true for age 6 in phase III with similar values to those in phase II (Table 2, line 19, effect: þ1.60, P , 10220). Furthermore, in phase III, the L- and, to a lesser extent, the M-fruit load increased RDAI compared with the H-fruit load (Table 2, lines 23 and 24). This was especially true for age 3 and over (Table 2, lines 27– 30). The effects on age 1-RDAI of the growth phase, fruit load and the presence of an adjacent fruit were analysed through a specific model integrating factors and interactions (Table 3). The growth phase had the highest effect, with positive values for phase II, and negative values for phase III compared with phase I (effects: þ1.08 and – 0.41, P , 10220 and P ¼ 2  1024, respectively). In the latter case, the negative values were partially compensated by an interaction with L-fruit load, with an increase of RDAI when there was no fruit on the adjacent inflorescence (effect: þ0.25, P ¼ 0.04). There was a strong interaction between the presence of an adjacent fruit and growth phase, with a decrease of RDAI of shoots

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics

Proportion of growing shoots

1·0

611

A

0·9

H-NF

0·8

H-F

0·7

M-NF

0·6

M-F

0·5

L-NF L-F

0·4 0·3 0·2 0·1 0·0

Length of growing shoots (mm)

250

B

200

150

100

50

0

Leaf area growing shoots (cm2)

400

C

350 300 250 200 150 100 50 0 10/4

22/4

4/5

16/5

28/5

9/6

21/6

3/7

15/7

Date F I G . 2. Shoot primary growth. Evolution of (A) the proportion of growing shoots, and (B) length and (C) total shoot leaf area of these growing shoots when more than four shoots, depending on tree fruit load (H, high; M, medium; L, low fruit load, respectively) and the presence (F, fruit) vs. absence (NF, no fruit) of an adjacent fruit.

with an adjacent fruit in phases II and III (effects: 20.44 and 20.60, P ¼ 5  10220 and P ¼ 2  1024, respectively) compared with phase I. Fruit load only had an effect for L- compared with H-fruit load, with higher RDAI on the former compared with the latter (effect: þ0.28, P ¼ 0.02). There was no effect of either M-fruit load (effect: 20.05, P ¼ 0.74) or presence of an adjacent fruit (effect: 20.02, P ¼ 0.89) but both these factors interacted positively, resulting in a slight increase of RDAI on shoots with an adjacent fruit on M-fruit load trees (effect: þ0.46, P ¼ 0.02).

DISCUSSION A hierarchy of factors affects RDAI of secondary growth

The dynamics of secondary growth at the whole-tree scale depended on a hierarchy of factors. The strongest effect was branch portion age, with an ordered decreasing RDAI from current-year shoot to a group composed of all older branch portions. Although the way the present statistical model was parameterized made it impossible to obtain significant information on the branch portion age effects within the group of

612

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics

TA B L E 1. Shoot length (mm) and total leaf area (cm2) of shoots at the end of shoot primary growth depending on two factors – fruit load (H, high; M, medium; L, low; see text) – and the presence (F) vs. absence (NF) of an adjacent fruit, and their interactions (P-values of ANOVA are given for each factor and their interactions) Factors Fruit load

responses of xylogenesis to air temperatures (Panyushkina et al., 2003; Rossi et al., 2008; Gruber et al., 2009) but also to tree water status and carbon balance (Zweifel et al., 2006). The third factor, with the lowest effect, was the wholetree fruit load with a lower RDAI for high compared with low, and to a lesser extent medium, fruit load.

Studied variables Fruit presence vs. absence

H

F NF M F NF L F NF Fruit load effect, P value Fruit presence effect, P value Fruit load : fruit presence effect, P value

Length (mm)

Leaf area (cm2)

113.79 + 12.97 139.50 + 15.90 155.56 + 16.01 138.25 + 14.11 142.50 + 15.05 137.05 + 16.83 0.553 0.919 0.366

180.17 + 13.56 220.92 + 17.83 209.33 + 17.11 201.51 + 17.56 211.16 + 15.83 225.86 + 24.07 0.469 0.144 0.534

Values are mean + s.e. for all shoots.

ages 2 – 6, the ordered decreasing values within this group suggest an increasing negative effect on RDAI of ages 2 – 6 (Table 2, lines 1 – 5). This result is consistent with previous studies showing age-dependent effects on xylogenesis with cell numbers correlated to cambial age (Panyushkina et al., 2003; Rossi et al., 2007). The second factor, with a lower effect, was the period in the growing season, with three consecutive phases characterized by low, maximal and eventually intermediate to low values of RDAI, respectively. The a priori decomposition of secondary growth in the three phases was thus a posteriori confirmed by the significant effect the growth phase factor had on RDAI. In forest species, the temporal dynamics of wood formation also exhibits consecutive phases with a maximum daily increment peak reached before summer solstice and increment continuing at a lower rate until the end of the growing season (Zweifel et al., 2006; Gruber et al., 2009). These phases are usually interpreted as 250

Fruit volume (cm3)

200 150 100

The growth phase affects the relationships between primary and secondary growth

The first growth phase ending approx. 3 weeks after full bloom presented a unique pattern with, whatever branch portion age, a marked reduction in the number of shoots with active secondary growth on the 28 April (Fig. 4B). This phenomenon was especially true for current-year shoots which were also characterized by a contemporaneous reduction in RDAI (Fig. 4C). At this date, current-year shoots had developed 5.3 + 0.1 metamers (mean + standard error; data not shown). According to literature, the carbohydrate and hydric status of the shoot evolves during its development. Indeed, the transition from pre-formation with the concomitant burst of several leaves, to neo-formation with the sequential development of individual leaves typically illustrates the transition from a ‘parasitic’ (Watson and Casper, 1984) to an autonomous state of the shoot (Kozlowski and Clausen, 1966; Lakso and Correlli-Grapadelli, 1992; Lauri and Kelner, 2001). Bijhouwer (1924) (also see Pratt, 1988) stated that the mean number of metamers of axillary shoots developed in a fruit bud, i.e. giving rise to bourse-shoots which was quasi-exclusively the case here, varies between 3 and 5. It is therefore suggested that the low secondary growth of all branch portion ages in phase I, and especially the decrease of RDAI characterizing this phase, was causally related to the heterotrophic expansion of preformed metamers. This phase ended when neoformed metamers began to expand with the likely beginning of autotrophy of current-year shoots. Although based on secondary growth dynamics, phase II and phase III were also characterized by specific patterns of primary shoot growth. The maximum values of RDAI,

High Medium Low Fruit volume depending on fruit load (27 May): L: 18·8a H: 16·9ab M: 16·2b Fruit volume depending on fruit load (harvest): L: 218.7a M: 202·0ab H: 180·5b

50 0 30/4

15/5

30/5

14/6

29/6

14/7

29/7

13/8

28/8

12/9

Date F I G . 3. Fruit growth depending on tree fruit load from 30 April until harvest (15 September). Fruit load adjustment was made on 15 May. Fruit volume is estimated from the equatorial diameter (Volume ¼ 0.0018  Diameter265). Values are means (for clarity, the standard error is not shown). Significant differences between fruit loads were observed from the 27 May (black arrow) onwards. The last values (grey symbols, 15 September) were measured on a sub-sample of fruits. On 27 May and 15 September, different letters indicate significant differences in fruit volume according to the HSD Tukey test at P , 0.05.

Proportion of branch portions with active secondary growth

400

A I

350 300 250 200 150 100

90

II

80

III

70 Branch portion age: 6 5 4 3 2 1

60 50 40 30 20

50

10

0

0

1·0

613

Cumulated secondary growth area (mm2) Age 1 to 5

Cumulated secondary growth area (mm2) Age 6

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics

B

0·9 0·8 0·7 0·6 0·5 0·4 0·3 0·2 0·1 0·0 –2 –3

C

In[RDAI (d–1)]

–4 –5 –6 –7 –8 –9 –10 10/4

25/4

10/5

25/5

9/6

9/7

24/6

24/7

8/8

23/8

7/9

22/9

Date

F I G . 4. Secondary growth from full bloom to harvest (15 September) as a function of branch portion age: (A) cumulated secondary growth, (B) proportion of shoots with active secondary growth, for the three merged fruit load values, and (C) ln(RDAI) as a function of branch portion age for L- (low fruit load) trees [similar patterns of variation throughout the growing season and order between branch portion ages were observed for H (high) and M (medium) fruit loads]. Branch portion age varies from 1 (current-year shoot) to 6 (trunk). The growth phases ( phases I, II and III) are noted in (A). Data are means when more than four growing branch portions are included. Negative values of ln(RDAI) result from mean RDAI values ,1. Statistics of the effects on RDAI of age, phase and fruit load and their interactions are given in Table 2.

characterizing phase II, lasted around 3 weeks and were concomitant to active primary shoot lengthening and leaf area expansion. phase II would then well illustrate the tight and positive relationship between active primary growth and strong secondary growth of all branch ages with leaves first feeding their own axes and then the rest of the tree (Wilson, 1990). RDAI of current-year shoots was unequivocally the highest during the three growth phases denoting a stronger relative secondary growth of these shoots compared with older branch portions. However, current-year shoots in phase III presented a unique pattern with a low proportion of shoots with active secondary growth compared with the other branch ages (30 – 50 % vs. 40– 90 % of shoots with

active secondary growth, respectively; Fig. 4B), and decreasing values through this phase, whereas values for the other branch portion ages remained stable (Fig. 4C). Both phenomena would indicate a shift of secondary growth through the season from the outer situated leafy shoots towards older branch portions including the trunk. The various patterns of primary and secondary growth throughout the growing season with phase II characterized by a positive relationship between the two, and phase III characterized by a decrease followed by an arrest of primary growth and a still active secondary growth bring more strength to the idea that sink strengths and their distribution within tree architecture significantly evolve during the growing season.

614

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics

TA B L E 2. Effect of branch portion age, growth phase in the growing season and tree fruit load on relative daily area increment (lnRDAI) in the parsimonious model with interactions Factors/levels of factors

Effects

P

Line no.

Common intercept

23.56

,10220

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Age* 2 Age 3 Age 4 Age 5 Age 6 Phase† II Phase III Fruit load‡ M Fruit load L Phase II: age 2 Phase II: age 3 Phase II: age 4 Phase II: age 5 Phase II: age 6 Phase III: age 2 Phase III: age 3 Phase III: age 4 Phase III: age 5 Phase III: age 6 M: age 5 L: age 2 L: age 4 Phase III: M Phase III: L L: phase II: age 2 L: phase II: age 5 L: phase III: age 3 L: phase III: age 4 L: phase III: age 5 L: phase III: age 6

22.61 22.97 23.47 23.92 24.22 0.87 20.75 0.09 0.22 0.90 0.67 0.67 0.66 0.99 0.86 0.47 0.58 0.49 1.60 0.14 0.26 0.20 0.14 0.31 20.28 0.23 0.36 0.41 0.81 0.41

,10220 2  1024 ,10220 ,10220 ,10220 ,10220 ,10220 0.16 2  1023 ,10220 ,10220 ,10220 1.99  10220 2.29  10210 5.75  10211 9.11  1024 5.34  1025 7.25  1024 ,10220 0.12 0.05 0.04 0.06 4.23  1023 0.06 0.05 0.02 0.02 1.23  1026 0.06

* Branch portion ages 2 –6 (trunk). Age 1 (current-year shoot) is the reference. † Growth phases II and III in the growing season. Phase I is the reference. ‡ Fruit load at the whole-tree scale: M, medium; L, low. High (H) fruit load is the reference. Lines are given numbers to facilitate comments in the text.

TA B L E 3. Effect of growth phase, tree fruit load and presence of an adjacent fruit on relative daily area increment (lnRDAI) of the current-year shoot (age 1) in the parsimonious model with interactions Factors/levels of factors

Effects

P

Common intercept Phase* II Phase III Fruit load† M Fruit load L Fruit presence‡ Phase II: fruit presence Phase III: fruit presence Phase III : L M : fruit presence

23.62 1.08 20.41 20.05 0.28 20.02 20.44 20.60 0.25 0.46

,10220 ,10220 2  1024 0.74 0.02 0.89 5  10220 2  1024 0.04 0.02

* Growth phases II and III in the growing season. Phase I is the reference. † Fruit load at the whole-tree scale: M, medium; L, low. High (H) fruit load is the reference. ‡ Presence of an adjacent fruit. No fruit (NF) is the reference.

Therefore the ranking of sinks commonly stated in literature should be conceived as a dynamic process. As shown in the present study, the increasing biomass of secondary tissues over the growing season is likely to shift the sink strength of secondary growth from low in the beginning of the season to high during phases II and III. As will be commented below, this seasonal change in sink strength not only concerns secondary growth but also includes the increasing sink strength of fruit. The present results extend previous statements showing a basipetal secondary growth movement preceding bud opening (Yoda et al., 2003) which is still active during the growth season (Pratt, 1990). This secondary growth dynamic pattern is caused by the formation of new conducting wood which is essential to support the expansion of new leaves and the beginning of the transpiration flux (Zweifel et al., 2006). It is shown here that the relative increment of secondary growth depends on branch age with a different pattern of secondary growth between current-year shoot and older branch portions. It also lasts for a longer time on older branch portions compared with current-year shoots (see the positive interaction between phase III and branch portion age for ages more than 2 which almost overrode the negative effect of phase III on RDAI; Table 2). Based on measurements made in midsummer, Suzuki and Hiura (2000) stated that the allometric relationships between stem area and leaf area supporting the pipe model theory hold true for old branches but not for current-year shoots. In the light of the present results in the apple showing the dynamic pattern of adjustment between primary and secondary growth we suggest that the timing of measurement is of paramount importance. Indeed, it is likely that only measurements carried out at the end of the vegetative season, i.e. after both primary and secondary growth has stopped, are needed to really assess these relationships. Moreover, the present results show that even though the pipe model initially proposed by Shinozaki et al. (1964) is a consistent concept for simulating organ allometry at the end of a growing season, formulation changes must be considered for simulating the dynamics of cambial growth within a growing season in functional – structural plant models (Costes et al., 2008). The extent to which fruit growth affects secondary growth

The effect of fruit on secondary growth was analysed at two scales: (1) the whole-tree with fruit load; (2) the shoot with the presence of an adjacent fruit. At the whole-tree scale, fruit load had a main significant effect on individual fruit growth (Fig. 3). It had no significant effect on primary growth which was likely to be related to the low vigour of the experimental trees, with an average of 20 cm for the longer current-year shoots whatever fruit load (Fig. 2). Furthermore, it only discriminates RDAI between extreme fruit loads (low vs. high). Transitions to phase II and then to phase III corresponded to specific phases of fruit growth. Indeed, phase II with high RDAI values was contemporaneous to the exponential growth of fruit volume (Fig. 3; Corelli-Grappadelli and Lakso, 2004). Transition to phase III, occurring at the

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics beginning of June in this experiment, was simultaneous with the transition between the exponential and the linear phases of fruit growth and the first difference in fruit volume between fruit-load treatments. It should be mentioned that, although fruit load was adjusted at the latest on 15 May at the beginning of phase II (natural fruit drop followed by hand-thinning), the differences of RDAI between fruit-load treatments occurred only after RDAI began to decrease, i.e. during phase III (see interactions between phase III and L; Table 2, line 24). Apart from the main difference between extreme fruit loads, i.e. low vs. high, the interesting point was the strong order 3 interactions between low fruit load, phase III and ages 3 –5 (Table 3, lines 27– 29). This would indicate that, compared with high and, to a lesser extent, medium fruit load, low fruit load increased secondary growth of old branch portions mainly during phase III. It also suggests a plasticity of secondary growth during the last growth phase, as previously found in forest trees grown in different climatic sites (Zweifel et al., 2006). According to literature, there is a negative followed by a positive effect of the bourse-shoot on fruit retention (Quinlan and Preston, 1971) and growth (Zhang and Tanabe, 2008). The present study showed that, at the current-year shoot scale and over the whole season, the presence of an adjacent fruit had a negligible effect on RDAI. However, the negative interaction between the presence of an adjacent fruit and phase III, and to a lesser extent phase II (Table 3), showed that the fruit may in return be detrimental to secondary growth of the adjacent bourse-shoot only later in the season. The two aspects of fruit effects on secondary growth, i.e. at both global and local scales, converge to the common statement that during phase III and, to a lesser extent, phase II intense competitions take place between the fruit compartment known to be a strong physiological sink late in the season due to its increasing biomass (Hansen, 1977; Kozlowski, 1992; Corelli-Grappadelli et al., 1996) and secondary growth. Although the hypothesis of direct competition between the two compartments may be assumed, an alternative interpretation can be suggested using the principles of ‘coordination theory’ for which shoot growth parallels root growth (Ge´nard et al., 2008). In this case, considering a relatively weak primary growth not affected by fruit load, competition would mainly exist between fruits and roots, reducing the growth of both compartments. The decrease in root growth would subsequently result in a decrease in secondary growth. In conclusion, this study has demonstrated that the withintree secondary growth dynamics throughout a growing season may be broken down into three consecutive phases characterized by unequal relative area increment over the season. These phases are affected by a hierarchy of factors with a strong effect of the branch portion age, a proxy for the architectural position, and a lower effect of the whole-tree fruit load. As underlined by Zweifel et al. (2006), studies developed on secondary growth have generally focused on one or several of the five following mechanisms: (1) carbon as a source of energy; (2) nutrients as limiting growth factors; (3) auxin as a plant hormone determining cambial activity; (4) mechanical stress; (5) water relations. In the current debate on the complex relationships between photosynthesis and production of biomass (Olesen et al., 2008; Smith,

615

2008; Murchie et al., 2009) the contribution of the present study is 3-fold: (1) primary growth characteristics, e.g. heterotrophic vs. autotrophic growth pattern or short vs. long shoots (Lauri and Kelner, 2001), are likely to play a crucial role in the dynamics of secondary growth at tree level; (2) tree architecture appears to be a driving force for secondary growth partitioning at both spatial and temporal levels; (3) the increasing competition between fruit load and secondary growth throughout the growing season mainly affects older branch portions.

ACK NOW LED GE MENTS We thank Stephan Benzing for his contribution to collection and a first analysis of data for his Master thesis at the Fachbereich Gartenbau und landespflege, Fachhochschule Wiesbaden/Geisenheim, Germany. We also thank two anonymous reviewers for valuable comments and suggestions.

L I T E R AT U R E CI T E D Barnola P, Crabbe´ J. 1993. L’activite´ cambiale, composante active ou passive dans les re´actions de croissance de l’arbre ? Acta Botanica Gallica 140: 403– 412. Bijhouwer J. 1924. De periodiciteit van de knopontwikkeling bij den appel. Mededeelingen van de Landbouwhoogeschool 27: 1– 69. Bound SA. 2001. Managing crop load. In: Dris RNR, Niskanen R, Jain SM. eds. Crop management and postharvest handling of horticultural products. I. Quality management. Enfield, NH, USA: Science Publishers, 89–110. Briand CH, Daniel AD, Wilson KA, Woods HE. 1998. Allometry of axis length, diameter, and taper in the devil’s walking stick (Aralia spinosa; Araliaceae). American Journal of Botany 85: 1201– 1206. Brouat C, Gibernau M, Amsellem L, McKey D. 1998. Corner’s rules revisited: ontogenetic and interspecific patterns in leaf-stem allometry. New Phytologist 139: 459 –470. Corelli-Grappadelli L, Lakso AN. 2004. Fruit development in deciduous tree crops as affected by physiological factors and environmental conditions. Acta Horticulturae 636: 425– 441. Corelli-Grappadelli L, Ravaglia G, Asirelli A. 1996. Shoot type and light exposure influence carbon partitioning in peach cv. Elegant Lady. Journal of Horticultural Science 71: 533– 543. Costes E, Fournier D, Salles JC. 2000. Changes in primary and secondary growth as influenced by crop load effects in ‘Fantasmew’ apricot trees. Journal of Horticultural Science & Biotechnology 75: 510 –519. Costes E, Smith C, Renton M, Gue´don Y, Prusinkiewicz, Godin C. 2008. MAppleT: simulation of apple tree development using mixed stochastic and biomechanical models. Functional Plant Biology 35: 936– 950. Daudet FA, Ame´glio T, Cochard H, Archilla O, Lacointe A. 2005. Experimental analysis of the role of water and carbon in tree stem diameter variations. Journal in Experimental Botany 56: 135– 144. Farrar JF. 1996. Sinks: integral parts of a whole plant. Journal of Experimental Botany 47 (Special Issue): 1273–1279. Finazzo SF, Davenport TL, Schaffer B. 1994. Partitioning of photoassimilates in avocado (Persea americana Mill.) during flowering and fruit set. Tree Physiology 14: 153 –164. Forshey CG, Elfving DC. 1989. The relationship between vegetative growth and fruiting in apple trees. Horticultural Reviews 11: 229 –287. Ge´nard M, Dauzat J, Franck N, et al. 2008. Carbon allocation in fruit trees: from theory to modelling. Trees 22: 269– 282. Gruber A, Baumgartner D, Zimmermann J, Oberhuber W. 2009. Temporal dynamics of wood formation in Pinus cembra along the alpine treeline ecotone and the effect of climate variables. Trees 23: 623–635. Hansen P. 1977. Carbohydrate allocation. In: Landsberg JJ, Cutting CV. eds. Environmental effects on crop physiology. London: Academic Press, 247–258.

616

Lauri et al. — Effects of tree architecture and fruit load on secondary growth dynamics

Ho LC. 1988. Metabolism and compartmentation of imported sugars in sink organs in relation to sink strength. Annual Review of Plant Physiology and Molecular Biology 39: 355– 378. Ho LC. 1996. The mechanisms of assimilate partitioning and carbohydrate compartmentation in fruit in relation to the quality and yield of tomato. Journal of Experimental Botany 47 (Special Issue): 1239–1243. Ishii HT, Ford ED, Kennedy MC. 2007. Physiological and ecological implications of adaptive reiteration as a mechanism for crown maintenance and longevity. Tree Physiology 27: 455–462. Knudson L. 1916. Cambial activity in certain horticultural plants. Torrey Botanical Club Bulletin 43: 533–537. Kozlowski TT. 1963. Growth characteristics of forest trees. Journal of Forestry 61: 655– 662. Kozlowski TT. 1992. Carbohydrate sources and sinks in woody plants. The Botanical Review 58: 107– 222. Kozlowski TT, Clausen JJ. 1966. Shoot growth characteristics of heterophyllous woody plants. Canadian Journal of Botany 44: 827 –844. Lacointe A, Dele´ens E, Ameglio T, et al. 2004. Testing the branch autonomy theory: a 13C/14C double-labelling experiment on differentially shaded branches. Plant, Cell & Environment 27: 1159– 1168. Lakso AN, Corelli-Grappadelli L. 1992. Implications of pruning and training practices to carbon partitioning and fruit development in apple. Acta Horticulturae 322: 231– 239. ´ . 2009. Developing a new paradigm for apple training. Compact Lauri PE Fruit Tree 42: 17–19. ´ , Kelner JJ. 2001. Shoot type demography and dry matter partitionLauri PE ing: a morphometric approach in apple (Malus  domestica Borkh.). Canadian Journal of Botany 79: 1270–1273. ´ , Lespinasse JM. 2000. The vertical axis and solaxe systems in Lauri PE France. Acta Horticulturae 513: 287–296. Layne DR, Flore JA. 1992. Photosynthetic compensation to partial leaf area reduction in sour cherry. Journal of the American Society for Horticultural Science 117: 279– 286. Lombard PB, Callan NW, Dennis Jr FG, et al. 1988. Towards a standardized nomenclature, procedures, values, and units in determining fruit and nut tree yield performance. HortScience 23: 813–817. McCulloh KA, Sperry JS. 2005. Patterns in hydraulic architecture and their implications for transport efficiency. Tree Physiology 25: 257–267. McQueen JC, Silvester WB, Green TGA, Minchin PEH. 2004. Carbohydrate allocation in apple stems can be altered by fruit load. Acta Horticulturae 636: 267–273. Ma¨kela¨ A. 2002. Derivation of stem taper from the pipe theory in a carbon balance framework. Tree Physiology 22: 891– 905. Mansour A, de Fay¨ E. 1998. Rhythmic growth rings of wood and their relationship with the foliage in oak seedlings grown in a favourable environment. Annals of Botany 82: 89–96. Marcelis-Van Acker CAM. 1994. Effect of assimilate supply on development and growth potential of axillary buds in roses. Annals of Botany 73: 415–420. Minchin PEH, Lacointe A. 2005. New understanding on phloem physiology and possible consequences for modeling long-distance carbon transport. New Phytologist 166: 771– 779. Murchie EH, Pinto M, Horton P. 2009. Agriculture and the new challenges for photosynthesis research. New Phytologist 181: 532–552. Niklas KJ, Spatz HC, Vincent J. 2006. Plant biomechanics: an overview and prospectus. American Journal of Botany 93: 1369–1378. ´ . 2008. Hydraulic and mechanNormand F, Bissery C, Damour G, Lauri PE ical stem properties affect leaf-stem allometry in mango cultivars. New Phytologist 178: 590–602. Novoplansky A. 2003. Ecological implications of the determination of branch hierarchies. New Phytologist 160: 111– 118. Obeso JR. 2002. The costs of reproduction in plants. New Phytologist 155: 321–348. Olesen T, Robertson D, Muldoon S, Meyer R. 2008. The role of carbohydrate reserves in evergreen tree development, with particular reference to macadamia. Scientia Horticulturae 117: 73–77. Oliveira CM, Priestley CA. 1988. Carbohydrate reserves in deciduous fruit trees. Horticultural Reviews 10: 403–430.

Oosthuyse SA, Jacobs G. 1995. Relationship between branching frequency, and growth, cropping and structural strength of 2-year-old mango trees. Scientia Horticulturae 64: 85– 93. Orians CM, Smith SDP, Sack L. 2005. How are leaves plumbed inside a branch? Differences in leaf-to-leaf hydraulic sectoriality among six temperate tree species. Journal of Experimental Botany 56: 2267–2273. Pallas B, Louarn G, Christophe A, Lebon E, Lecoeur J. 2008. Influence of intra-shoot trophic competition on shoot development in two grapevine cultivars (Vitis vinifera). Physiologia Plantarum 134: 49– 63. Palmer JW, Cai YL, Edjamo Y. 1991. Effect of part-tree flower thinning on fruiting, vegetative growth and leaf photosynthesis in ‘Cox’s Orange Pippin’ apple. Journal of Horticultural Science 66: 319–325. Panyushkina IP, Hughes MK, Vaganov EA, Munro MAR. 2003. Summer temperature in northeastern Siberia since 1642 reconstructed from tracheid dimensions and cell numbers of Larix cajanderi. Canadian Journal of Forest Research 33: 1905–1914. Pratt C. 1988. Apple flower and fruit: morphology and anatomy. Horticultural Reviews 10: 273–308. Pratt C. 1990. Apple trees: morphology and anatomy. Horticultural Reviews 12: 265 –305. Preston KA. 1998. The effects of development stage and source leaf position on integration and sectorial patterns of carbohydrate movement in an annual plant, Perilla frutescens (Lamiaceae). American Journal of Botany 85: 1695–1703. Quinlan JD, Preston AP. 1971. The influence of shoot competition on fruit retention and cropping of apple trees. Journal of Horticultural Science 46: 525 –534. Roper T. 1995. Estimating appropriate crop loads. Annual report of the secretary of the State Horticultural Society of Michigan for the year 1995, 176: 4. Rossi S, Deslauriers A, Anfodillo T, Carrer M. 2008. Age-dependent xylogenesis in timberline conifers. New Phytologist 177: 199–208. Shinozaki K, Yoda K, Hozumi K, Kira T. 1964. A quantitative analysis of plant form: the pipe model theory. I. Basic analysis. Japanese Journal of Ecology 14: 97– 105. Smith AM. 2008. Prospects for increasing starch and sucrose yields for bioethanol production. The Plant Journal 54: 546–558. Sprugel DG. 2002. When branch autonomy fails: Milton’s law of resource availability and allocation. Tree Physiology 22: 1119–1124. Stephenson AG. 1981. Flower and fruit abortion: proximate causes and ultimate functions. Annual Review of Ecology and Systematics 12: 253– 279. Sterck FJ, Bongers F. 2001. Crown development in tropical rain forest trees: patterns with tree height and light availability. Journal of Ecology 89: 1 –13. Suzuki M, Hiura T. 2000. Allometric differences between current-year shoots and large branches. Tree Physiology 20: 203–209. Umeki K, Seino T. 2003. Growth of first-order branches in Betula platyphylla saplings as related to the age, position, size, angle, and light availability of branches. Canadian Journal of Forest Research 33: 1276–1286. Umeki K, Seino T, Lim EM, Honjo T. 2006. Patterns of shoot mortality in Betula platyphylla in northern Japan. Tree Physiology 26: 623–632. Wardlaw IF. 1990. The control of carbon partioning in plants. New Phytologist 116: 341– 381. Watson MA, Casper BB. 1984. Morphogenetic constraints on patterns of carbon distribution in plants. Annual Review of Ecology and Systematics 15: 233–258. Wilson BF. 1990. The development of tree form. HortScience 25: 52–54. Wolstenholme BN. 1990. Resource allocation and vegetative-reproductive competition: opportunities for manipulation in evergreen fruit trees. Acta Horticulturae 275: 451 –459. Yoda K, Wagatsuma H, Suzuki M, Suzuki H. 2003. Stem diameter changes before bud opening in Zelkova serrata saplings. Journal of Plant Research 116: 13– 18. Zhang C, Tanabe K. 2008. Partitioning of 13C-photosynthates from different current shoots neighboring with fruiting spur in later-maturing Japanese pear during the period of rapid fruit growth. Scientia Horticulturae 117: 142–150. Zweifel R, Zimmermann L, Zeugin F, Newberry DM. 2006. Intra-annual radial growth and water relations of trees: implications towards a growth mechanism. Journal of Experimental Botany 57: 1445–1459.