Retranslocation of foliar nutrients in evergreen tree species planted in a Mediterranean environment

Tree Physiology 28, 187 196 2008 Heron Publishing Victoria, Canada Retranslocation of foliar nutrients in evergreen tree species planted in a Mediterranean environment D. N. FIFE, 1 E. K. S. NAMBIAR 2,3 and E. SAUR 4 1 Ensis, CSIRO Plantation Forest Research Centre, P.O. Box 946, Mount Gambier, SA 5290, Australia 2 Ensis, CSIRO Forestry and Forest Products, P.O. Box E4008, Kingston, ACT 2604, Australia 3 Corresponding author (sadu.nambiar@csiro.au) 4 École Nationale d Ingénieurs des Travaux Agricoles de Bordeaux, 1, cours du Général de Gaulle, CS 4021, F. 33175 Grandigan Cedex, France Received November 23, 2006; accepted September 3, 2007; published online December 3, 2007 Summary Internal nutrient recycling through retranslocation (resorption) is important for meeting the nutrient demands of new tissue production in trees. We conducted a comparative study of nutrient retranslocation from leaves of five tree species from three genera grown in plantation forests for commercial or environmental purposes in southern Australia Acacia mearnsii De Wild., Eucalyptus globulus Labill., E. fraxinoides H. Deane & Maiden, E. grandis W. Hill ex Maiden and Pinus radiata D. Don. Significant amounts of nitrogen, phosphorus and potassium were retranslocated during three phases of leaf life. In the first phase, retranslocation occurred from young leaves beginning 6 months after leaf initiation, even when leaves were physiologically most active. In the second phase, retranslocation occurred from mature green leaves during their second year, and in the third phase, retranslocation occurred during senescence before leaf fall. Nutrient retranslocation occurred mainly in response to new shoot production. The pattern of retranslocation was remarkably similar in the leaves of all study species (and in the phyllodes of Casuarina glauca Sieber ex Spreng.), despite their diverse genetics, leaf forms and growth rates. There was no net retranslocation of calcium in any of the species. The amounts of nutrients at the start of each pre-retranslocation phase had a strong positive relationship with the amounts subsequently retranslocated, and all species fitted a common relationship. The percentage reduction in concentration or content (retranslocation efficiency) at a particular growth phase is subject to many variables, even within a species, and is therefore not a meaningful measure of interspecific variation. It is proposed that the pattern of retranslocation and its governing factors are similar among species in the absence of interspecies competition for growth and crown structure which occurs in mixed species stands. Keywords: Acacia, Eucalyptus, interspecies variation, pine. Introduction Internal nutrient recycling by retranslocation (resorption) from leaves is an important factor in the supply of nitrogen, phosphorus and potassium for new growth in tree species at different phases of foliage development (Nambiar and Fife 1991, Millard 1994, Fife and Nambiar 1995, Aerts 1996, Saur et al. 2000, van Heerwaarden et al. 2003). Calcium is largely immobile within the plant. Nutrient retranslocation studies help to quantify the net amounts of nutrients available for reuse by new growth and provide information for nutrient management in production forests (Fife and Nambiar 1997). Understanding retranslocation efficiency is important because retranslocation provides a mechanism for nutrient conservation at the species level and does not differ strongly among growth forms of perennials (Aerts 1996). Most soil nutrients taken up by trees are used in annual production of foliage, which serves as a reservoir of reusable nutrients. Nutrients in one generation of foliage can be retranslocated to support the production of the next generation of foliage, irrespective of the rate of soil nutrient supply (Nambiar and Fife 1987). In this process, nutrient retranslocation occurs not only from senescing leaves after the stand has fully occupied the site, as has been commonly reported, but also from young leaves in open stands, including those of transplanted seedlings (Nambiar and Fife 1991). At a nitrogen-deficient site in a 13-year-old Pinus radiata D. Don stand, application of nitrogen increased wood production by up to 45%, and the amount of nitrogen retranslocated from needles was increased 4.5-fold, showing that the higher the growth rate, the greater the amount of nutrients retranslocated (Fife and Nambiar 1997). Several studies have reported interspecies variation in nutrient retranslocation from foliage. Hawkins and Polglase (2000) compared retranslocation during senescence among 15 species of eucalypts, each species grown in separate plots supplied with non-limiting water and nutrients, and found that about 50% of the nitrogen was retranslocated during senescence, in contrast to 11 24% of the phosphorus. Kumar and Singh (2005) examined variation among eight tropical species grown over mine spoils and reported retranslocation efficiency during leaf senescence from 38 to 69% for nitrogen and 50 to 71% for phosphorus. Studies have shown that leaf lifespan is a

188 FIFE, NAMBIAR AND SAUR key determinant of retranslocation efficiency in several species (Escudero et al. 1992, Niinemets and Tamm 2005). In mixed species stands of deciduous trees, variation in nutrient retranslocation efficiency and resorption kinetics among species was related to differences in leaf longevity (lifespan based on timing of leaf fall) and the nature of the biochemical pool of nutrients (Niinemets and Tamm 2005). However, these studies have two potential limitations. First, substantial amounts of retranslocation occur from young foliage well before senescence (Fife and Nambiar 1991, 1995, Millard 1994). Second, leaf mass is lost during senescence that affects mass-based concentration values and affects the estimation and interpretation of retranslocation (van Heewaarden 2003). Based on studies of individual species, Pinus radiata (Nambiar and Fife 1991, Fife and Nambiar 1997) and Eucalyptus globulus Labill. (Saur et al. 1997), we have suggested that, despite the contrasting genetic and morphological characteristics of these two species, retranslocation from both young and senescing leaves is governed by the same set of tree variables when the trees are grown in the same environment. The range of species being planted for commercial or environmental benefits in southern Australia has increased in recent years, and there is value in understanding their nutritional physiology. In the research reported here, we extended our nutrient retranslocation studies by focusing on interspecies variation. Our three main objectives were as follows. First, to compare the nature and extent of retranslocation from leaves of genetically diverse species planted at the same time at the same site, but in separate stands. This experimental design provided a more robust basis for interspecies comparison and avoided the confounding effects of interspecies competition on growth and crown structure that occur in mixed species stands. Second, to follow nutrient accumulation and retranslocation continuously during the full lifespan of leaves rather than during senescence (leaf fall) alone, as has been done in other studies. Third, to examine nutrient retranslocation efficiency calculated on the basis of both nutrient concentrations and contents in leaves. With this information we tested the hypothesis that the pattern of retranslocation and the factors governing it are similar among species in the absence of interspecies competition on growth and crown structure. Materials and methods Site The experimental site was located 10 km north of Mount Gambier (the Green Triangle region of southeast South Australia and southwest Victoria; 37 45 S, 140 47 24 E, 63.0 m a.s.l.), Australia. The growing environment is Mediterranean (characterized by winter rainfall and dry summers) with a 5-m deep sandy soil. The principal profile form is Uc 2.2 (Northcote 1979) and described as a Spodic Quartzipsamment (Soil Survey Staff 1975) or a bleached vertic subplastic yellow chromosol (Isbell 1996). Before establishing the experimental stand, the site was under pasture that had improved soil fertility. Characteristics of the top 30 cm of soil are presented in Table 1. Long-term mean annual rainfall and evaporation are 710 and 1350 mm, respectively. On average, 39% of rain falls in winter (June to August), 26% in spring, 13% in summer (December February) and 23% in autumn. During our study, rainfall was 700 mm year 1. Stand The experimental stand was planted with six tree species: Acacia mearnsii De Wild. (a nitrogen fixing species); Eucalyptus globulus; E. fraxinoides H. Deane Maiden; E. grandis W. Hill ex Maiden; Pinus radiata; and Casuarina glauca Sieber ex Spreng. Nursery-grown seedlings of all species, about 10 cm high, were planted in May after autumn rains in plots containing 120 trees 2.4 m apart (1700 trees ha 1 ), equally spaced within and between rows. Three replications of each species were arranged in a randomized block design. Weeds were controlled with herbicide before planting, and subsequently by mowing until weeds were suppressed by the tree canopy. Tree sampling The study began when the trees were 27 months old. Within each plot (species), two groups of 16 trees (representing the size classes but avoiding those that were heavily suppressed or had damaged crowns) were marked. Within each group, five trees were marked for regular foliar sampling. Thus measurements within species were treated as six replications. The results for C. glauca have been reported separately (Saur et al. 1997) and are discussed in this paper in relation to the results for the other species. Growth measurements Height and diameter at breast height over bark (DBH) of all trees were measured every month over one year. Foliar sampling Morphology and growth form of leaves differed considerably among genera and species (Figure 1). Acacia mearnsii leaves Table 1. Selected soil properties. Mineralizable nitrogen (N) measured after 60 days of incubating undisturbed soil cores in the laboratory. Soil depth (cm) ph in water Total C (%) Total N (%) Mineralizable N (mg kg 1 ) Total P (mg kg 1 ) Exchangeable Ca (cmol kg 1 ) 0 5 6.2 1.8 0.13 54.4 112.7 1.68 5 10 5.9 1.0 0.06 18.7 97.7 0.84 10 15 5.7 0.7 0.04 9.4 89.8 0.61 15 30 5.7 0.6 0.03 6.8 77.5 0.51 TREE PHYSIOLOGY VOLUME 28, 2008

RETRANSLOCATION OF FOLIAR NUTRIENTS IN EVERGREENS 189 branches and so could not be collected. On the five trees marked for foliar sampling, branches at 1.3 m height were selected. From each branch, a set of young leaves (recently initiated) and a set of fully expanded mature leaves were marked by attaching a label to their petioles. Sufficient numbers of leaves were marked to provide samples at monthly intervals for 12 months. At each sampling, two leaves per tree were removed from each age class and bulked for a composite sample. The leaves in these composite samples were recounted, dried at 70 C for 48 hours and weighed. Figure 1. Mature green leaves of Acacia mearnsii, Eucalyptus globulus, Eucalyptus fraxinoides, Eucalyptus grandis and Pinus radiata. in adult form are described as bipinnate with 8 21 pairs of pinnae and 15 70 pairs of leaflets (Boland et al. 1985). Eucalypt leaves are described as ovate to broad-lanceolate in E. fraxinoides, elliptic-ovate in E. globulus and ovate in E. grandis (Boland et al. 1985). All eucalypt leaves sampled were of adult form. The foliage of P. radiata consisted of fascicles with three or four needles. The results for P. radiata are expressed on a needle basis because the number of needles per fascicle varies. Hereafter, the unit of foliage for all species is referred to as a leaf. From all species, leaves at three phases of development were sampled regularly. Phase 1: young, still expanding, leaves were sampled from initiation to about 12 months of age. Phase 2: mature leaves that had attained full size at the time of the first sampling (at an age of about 12 months) were followed as long as they remained green (to the age of about 22 months). Phase 3: senesced leaves still attached to branches were collected just before leaf fall together with samples of adjacent live green foliage. This allowed reliable comparison of leaf mass and nutrient contents between pairs of green and senesced leaves. However, in A. mearnsii, pinnae were too fragile at the onset of senescence and rarely remained on Plant analysis A subsample of each dried leaf sample was digested in sulfuric acid. Nitrogen and phosphorus were measured colorimetrically, potassium by flame photometry and calcium by atomic absorption spectroscopy. Foliar nutrient concentrations and contents (mass leaf 1 ) were calculated. Relationships investigated were: (1) the maximum nutrient contents in young and mature leaves and the amounts subsequently retranslocated; (2) the nutrient contents in mature live leaves and the amounts retranslocated from senescing leaves; and (3) height, diameter increments and branch elongation and the amount of nutrients retranslocated over the same interval. Data analysis To test the potential bias from the location of two sample groups within plots, we analyzed the data (see Tables 2 and 3) in a split plot design, treating each sampling group as a sub plot. Results were not significantly affected by sampling method. Therefore, we used analysis of variance (ANOVA) or regression analysis for interpreting data based on six replications (sample groups) for each species. Results Tree growth When the study began, there were large differences among species in height and stem diameter of the 27-month-old trees (Table 2). Height ranged from 2.2 m for P. radiata to 5.1 m for A. mearnsii, and DBH ranged from 2.0 cm for E. grandis to 5.3 cm for A. mearnsii. Between age 27 and 39 months, height and diameter increments differed significantly between species, but the seasonal patterns of growth were similar. For all species, height and diameter growth peaked during summer Table 2. Mean height and diameter at breast height over bark (DBH) of trees at 27 and 39 months after planting. Increment equals growth between 27 and 39 months. The main effect of species was significant (P < 0.001) in all cases. Within a column, values followed by different letters differed significantly at P < 0.05. Species Height (m) DBH (cm) 27 months 39 months Increment 27 months 39 months Increment Acacia mearnsii 5.1 7.8 2.7 ab 5.3 8.6 a 3.3 Eucalyptus globulus 3.5 a 6.9 a 3.4 c 3.2 a 7.2 ab 4.1 b Eucalyptus fraxinoides 3.3 a 6.7 a 3.4 c 2.8 a 7.8 a 4.9 a Eucalyptus grandis 2.6 b 5.4 2.8 b 2.0 a 6.1 b 4.1 b Pinus radiata 2.2 b 4.4 2.2 a 2.8 a 7.3 ab 4.4 ab TREE PHYSIOLOGY ONLINE at http://heronpublishing.com

190 FIFE, NAMBIAR AND SAUR from January to March (data not shown) and subsequently decreased. Height growth over the 12 months was greatest for E. fraxinoides and E. globulus and least for P. radiata, whereas diameter growth was greatest for E. fraxinoides and least for A. mearnsii (Table 2). Leaf growth Leaf morphology varied considerably among species (Figure 1), and there were major differences in crown configuration among the genera and among eucalypt species. The change in mass of young leaves of all species followed a similar pattern from the first sampling in early spring (September, Figure 2). Mass per leaf increased rapidly over the first four months, with little change subsequently. During the peak growing season (October to December), leaf mass increased 35-fold for A. mearnsii and 3-fold for P. radiata. Individual young leaves of A. mearnsii reached 1.03 g leaf 1. Young leaves of E. globulus, E. fraxinoides, E. grandis and P. radiata attained mean biomasses of 0.80, 0.44, 0.42 and 0.03 g leaf 1, respectively. The mass of mature leaves changed little during the 12 months after the first sampling; A. mearnsii, E. globulus, E. fraxinoides, E. grandis and P. radiata weighed 1.20, 0.76, 0.37, 0.31 and 0.04 g leaf 1 respectively. Beginning in June (early winter when trees were 36 months old), some of the old leaves began to senesce. There were no clear interspecific differences in leaf longevity (lifespan) or timing of senescence. Furthermore, senescence occurred unevenly and not all leaves of the same age in the same species senesced at the same time. The estimated lifespan of leaves ranged from 20 to 23 months for all species, but the longevity of E. globulus leaves may be less than that of other species. There was a net reduction in dry mass of leaves at senescence, 9 10% for each of the three eucalypt species and 32% for P. radiata. Foliar nutrient concentrations Young leaves The pattern of changes in foliar concentrations of nitrogen, phosphorus, potassium and calcium in young leaves (Figure 3A) was similar among species. Nitrogen concentrations were consistently higher in A. mearnsii than in the other species, and decreased in all species from October to January as leaves increased in size. Similar patterns were found for Figure 2. (A) Changes in mean biomass of young leaves of Acacia mearnsii ( ); Eucalyptus globulus ( ); E. fraxinoides ( ); E. grandis ( ); and Pinus radiata ( ). (B) Changes in mean biomass of young leaves of P. radiata on an enlarged scale. foliar concentrations of phosphorus and potassium (Figure 3), and there were no significant differences among species. In contrast to other nutrients, calcium concentrations increased continuously, and there were significant differences among species. The ranking in increasing order of foliar calcium concentration was P. radiata, E. fraxinoides, A. mearnsii, E. globulus and E. grandis. By the final sampling, there was a more than 4-fold difference in foliar calcium concentration between P. radiata (0.5%) and E. grandis (2.2%) (Figure 3B). There were also significant differences in calcium concentrations among eucalypt species. Mature leaves The concentrations of nitrogen, phosphorus and potassium in mature leaves decreased gradually from August to April (Figure 3B). The nitrogen concentration in A. mearnsii was significantly higher than in the other species (Figure 3B) but differed little among species despite the differences in leaf size (Figure 1). Foliar phosphorus concentrations decreased continuously over the 12 months with no significant differences among species. Senescing leaves At senescence, there was a further decrease in the concentrations of nitrogen, phosphorus and potassium in all species (Figure 3B). Nitrogen concentrations in senesced leaves were highest in A. mearnsii (1.39%) and lowest in E. fraxinoides (0.28%). Interspecific differences in nutrient concentrations in senesced leaves did not parallel those observed in live mature leaves, because the relative decline in concentration during senescence varied among species. Furthermore, the coefficients of variation of foliar concentrations in green leaves were between 1.7 and 22.7% compared with between 5.1 and 98.2% in senesced leaves. Consequently, the correlations between foliar nutrient concentrations of green and senesced leaves was low (r 2 = 0.13 for phosphorus; r 2 = 0.05 for potassium). The correlation was higher for nitrogen, because the value for A. mearnsii dominated the regression. Leaf nutrient contents and retranslocation Young leaves Contents of nitrogen (Figure 4A), phosphorus (Figure 4B) and potassium (Figure 4C) in young leaves increased rapidly during the study in all species. Rates of accumulation were highest between October (spring) and January (midsummer), after which, the pattern of change varied depending on species and nutrients. The amounts of nitrogen in young leaves of A. mearnsii and E. globulus (Figure 4A) decreased between February (late summer) and May June (early winter). In contrast, there was little change in foliar nitrogen contents in E. fraxinoides, E. grandis and P. radiata. Phosphorus content increased in young leaves of all species, reaching maximum values in summer (January February, Figure 4B), followed by significant net reductions in A. mearnsii, E. globulus and P. radiata but not in E. fraxinoides or E. grandis. The seasonal changes in potassium content for each species (Figure 4C) were similar to those observed for nitrogen. Calcium content of young leaves of all species increased progressively over the study period (data not shown), but the rate of accumulation differed among TREE PHYSIOLOGY VOLUME 28, 2008

RETRANSLOCATION OF FOLIAR NUTRIENTS IN EVERGREENS 191 species. At the final sampling in August, the calcium content in young foliage ranged from 0.1 mg leaf 1 in P. radiata to 14.1 mg leaf 1 in A. mearnsii. There was great variation in calcium content among the three eucalypt species (3.03, 9.02 and 13.4 mg leaf 1 in E. fraxinoides, E. grandis and E. globulus, respectively). Mature leaves The contents of nitrogen (Figure 5A), phosphorus (Figure 5B) and potassium (Figure 5C) in mature leaves declined significantly in all species during the study, typically from September to May of the following year, when there were no changes in unit leaf mass and all sampled leaves (Figure 5) were green. The net reduction in nutrient content reflects nutrient retranslocation from green leaves. These amounts represented a large proportion of those present initially when mature leaves were first sampled (Table 3). The proportions retranslocated ranged as follows: nitrogen from Figure 3. Changes in nutrient concentrations in (A) young and (B) mature leaves of Acacia mearnsii ( ); Eucalyptus globulus ( ); E. fraxinoides ( ); E. grandis ( ); and Pinus radiata ( ). Separate symbols on the right are for senesced foliage collected at the time of the last collection of live foliage. Vertical bars are least significant differences at P = 0.05 for the interaction between time and species. 31% in E. globulus to 60% in P. radiata; phosphorus from 54% in E. globulus to 63% in P. radiata; and potassium from 18% in E. globulus to 38% in A. mearnsii and P. radiata. There were further decreases in nutrient contents during leaf senescence (Figure 5). The net amounts retranslocated during leaf senescence as a proportion of the pre-senescent values ranged as follows: nitrogen 76% in P. radiata and 56% in E. globulus; phosphorus 78% in P. radiata and 37% in E. globulus; and potassium 88% in P. radiata and 54% in E. globulus. From first sampling of mature needles until after senescence (the sum of the above two processes; see Figure 5) there were significant differences among species in both amount and proportion of the initial pool of both nitrogen and phosphorus that were withdrawn. The total amount of nitrogen withdrawn ranged from 0.67 mg leaf 1 (or 87% of the initial pool) for P. radiata to 7.90 mg leaf 1 (or 70% of the initial pool) for TREE PHYSIOLOGY ONLINE at http://heronpublishing.com

192 FIFE, NAMBIAR AND SAUR E. globulus. The total amount of phosphorus withdrawn ranged from 0.06 mg leaf 1 (or 89% of the initial pool) for P. radiata to 7.90 mg leaf 1 (or 63% of the initial pool) for E. globulus. Figure 4. Changes in (A) nitrogen, (B) phosphorus and (C) potassium contents of young foliage during their first growing season for Acacia mearnsii ( ); Eucalyptus globulus ( ); E. fraxinoides ( ); E. grandis ( ); and Pinus radiata ( ). Arrows mark maximum and minimum values. Vertical bars are least significant differences for changes with time (P = 0.05). Relationships between initial nutrient contents and net nutrient retranslocation In both young and mature leaves, there were significant positive linear relationships between maximum nutrient contents of both young and mature leaves while they remained green during Phases 1 and 2 and the amounts subsequently retranslocated (Figure 6). All species fitted within the common relationships for nitrogen (Figure 6A), phosphorus (Figure 6B) and potassium (Figure 6C). Similarly, there were strong positive linear relationships between the amounts of nitrogen (Figure 7A), phosphorus (Fig- Figure 5. Changes in (A) nitrogen, (B) phosphorus and (C) potassium contents of mature foliage during their first growing season for Acacia mearnsii ( ); Eucalyptus globulus ( ); E. fraxinoides ( ); E. grandis ( ); and Pinus radiata ( ). Arrows mark beginning and end of retranslocation calculations. Vertical bars are least significant differences for changes at P = 0.05. Symbols on the right are for dead leaves sampled during the last collection of live needles. TREE PHYSIOLOGY VOLUME 28, 2008

RETRANSLOCATION OF FOLIAR NUTRIENTS IN EVERGREENS 193 Table 3. Net amounts of nitrogen (N), phosphorus (P) and potassium (K) retranslocated from green leaves between about 12 and 22 months of age (marked by arrows in Figure 5). Leaves remained green throughout the sampling period. The main effect of species was significant (P < 0.001) in all cases except for P (% initial) where the effect was not significant. Within a column, values followed by different letters differed significantly at P < 0.05. Species N (mg leaf 1 ) N (% initial) P (mg leaf 1 ) P (% initial) K (mg leaf 1 ) K (% initial) Acacia mearnsii 11.72 34.2 a 0.73 59.8 3.03 38.2 a Eucalyptus globulus 3.53 a 31.0 a 0.46 53.8 0.74 a 18.3 Eucalyptus fraxinoides 3.53 a 51.9 b 0.22 a 58.1 0.43 ab 30.8 a Eucalyptus grandis 3.43 a 54.0 bc 0.26 a 60.7 0.73 a 36.8 a Pinus radiata 0.45 59.9 c 0.04 62.9 0.12 b 37.8 a ure 7B) and potassium (Figure 7C) in pre-senescent mature leaves and the amounts of those nutrients retranslocated at senescence. This significant relationship existed despite a significant loss in leaf mass during senescence. Thus, at all stages of the leaf lifespan, the net amounts of nutrients retranslocated increased as the amounts in the corresponding pre-withdrawal stage increased, indicating that the sizes of the initial nutrient pools determined the amounts retranslocated in all species. Relation between tree growth and net nutrient retranslocation Relationships between growth increments (measured as height or diameter growth or branch extension) and nutrient retranslocation for the corresponding growth period were examined. We observed that decreases in nitrogen, phosphorus and potassium contents, in both young and mature leaves of all species, coincided with the period of maximum growth rates. However, there were no consistent relationships between amounts retranslocated and growth increments (data not shown). Discussion Leaf forms Our study species represent wide genetic diversity. Acacia mearnsii (a nitrogen-fixing species) and the three eucalypts are indigenous to Australia, but they do not occur in natural stands in the region where this experiment was conducted. Pinus radiata is an exotic species but has been grown in the region for more than 120 years and has been subjected to several Figure 6. Relationship between maximum nutrient contents of (A) nitrogen, (B) phosphorus and (C) potassium in young and mature leaves and the amounts retranslocated in Acacia mearnsii ( ); Eucalyptus globulus ( ); E. fraxinoides ( ); E. grandis ( ); and Pinus radiata ( ). For each species, one value represents retranslocation during Phase 1 (Figure 4) and the other during Phase 2 (Figure 5). Values are means of six replicates. Broken lines are 95% confidence limits of each regression. Figure 7. Relationship between contents of (A) nitrogen, (B) phosphorus and (C) potassium in mature leaves before senescence and amounts retranslocated during senescence in Eucalyptus globulus ( ); E. fraxinoides ( ); E. grandis ( ); and Pinus radiata ( ). Broken lines are 95% confidence limits of each regression. TREE PHYSIOLOGY ONLINE at http://heronpublishing.com

194 FIFE, NAMBIAR AND SAUR cycles of genetic selection. This set of species has a wide range of leaf forms (Figure 1), crown characteristics and rates of leaf (Figure 2) and tree growth (Table 2), all of which influence nutrient retranslocation. Despite the large variation in attributes among species, the seasonal pattern and duration of leaf growth were remarkably similar and leaves of all species did not increase in mass after 4 5 months of age (Figure 2). A similar pattern of growth was found in studies with P. radiata, even when nitrogen application increased basal area growth by 40% (Fife and Nambiar 1997). Furthermore, based on sequential sampling (Figure 2) and other observations, the estimated leaf lifespan of all species in our study was typically 20 23 months after initiation, a narrow range given the species diversity. These results suggest that the growing environment strongly affected the timing and duration of foliage growth across species. Three phases of foliar nutrient retranslocation Most studies have examined nutrient retranslocation and its efficiency based on comparison of green and abscised leaves. This approach yields an incomplete picture, because retranslocation occurs copiously from young green leaves as well as senescing leaves (Fife and Nambiar 1984, Sheriff et al. 1986, Nambiar and Fife 1987, Saur et al. 1997). We examined interspecific variation in the dynamics of changes in nutrient concentration and content in three distinct phases during the lifespan of leaves. Phase 1 occurred during the first year of leaf development when leaves were still growing (Figures 2, 3A and 4). Initially, foliar nutrient concentrations declined (Figure 3) as leaves expanded and gained mass in the first 5 months (Figure 2A). However, nutrient contents per leaf increased (Figure 4) until net nutrient retranslocation began from these young (6 7 months old) leaves (Figure 4). Phase 2 was measured during a 10-month period in fully expanded (still green) leaves aged from about 12 months to 22 months. Net amounts of nutrients retranslocated differed among species and were largely determined by leaf size (Figure 2), but the patterns and processes were similar across species. We have previously found similar results in P. radiata (Fife and Nambiar 1997) and E. globulus (Saur et al. 2000). Phase 3 occurred during leaf senescence, which we measured using matched adjacent pairs of green and senesced leaves. This provides a more accurate measure of retranslocation than a comparison of nutrient concentrations in fallen leaves and in leaves of the green crown. The latter approach is especially difficult in evergreen species that simultaneously bear leaves of different age classes. Senescence also leads to mass loss. In pine needles, this was three times greater (32%) than in eucalypt leaves and would confound the comparison of nutrient retranslocation between species. Changes in leaf mass at senescence due to retranslocation of carbohydrates have been found in many species, including P. radiata (Fife and Nambiar 1997), Alaskan birch (Betula papyrifera var. humilis (Reg.); Chapin and Moilanen 1991) and mangroves (Kandelia candel (L.); Lin and Wang 2001). One factor contributing to the decline in foliar nutrients, especially potassium, is leaching by throughfall. However, we can exclude a role of leaching in our study, because significant declines in nutrient contents occurred from December to March when rainfall (as light showers) ranged from 18 to 38 mm month 1 and potential evaporation ranged from 170 to 200 mm month 1 and there was no throughfall or stem flow. Furthermore, we have previously demonstrated significant reductions in foliar nutrient contents (including potassium) during dry rainless summers in studies with P. radiata (Fife and Nambiar 1982, 1984), C. glauca (Saur et al. 1997) and E. globulus (Saur et al. 2000). The key finding is the close similarity between contrasting species in foliar nutrient dynamics, especially the cycles of nutrient accumulation, depletion and retranslocation. This leads to the question: what factors determine the amounts of nutrients retranslocated during the leaf lifespan and are they common across species? The nurient content of a leaf is strongly influenced by its mass which varies among species. Results show that the amount of retranslocated nutrients correlates strongly with the amount present just before retranslocation. This is consistent with an earlier observation in P. radiata (Nambiar and Fife 1991) and is supported by evidence from other species, including Picea sitchensis (Bong.) Carr. (Millard and Proe 1992, 1993) and E. globulus (Saur et al. 2000). In contrast, Niinemets and Tamm (2005) studied nutrient retranslocation in mixed species stands by comparing nutrient concentrations in green leaves from the mid-canopy with freshly fallen leaves and concluded that mid-seasonal foliar nutrient concentrations were unrelated to nutrient retranslocation. Several methodological factors contribute to this difference in conclusions. Niinemets and Tam (2005) used a single mid-season sample representing the mid-crown from mixed species stands, and therefore, the match between green leaves and fallen leaves may not be as close as in our study. Our method is based on monthly sequential sampling of groups of leaves from the same set of branches to define the phases of nutrient accumulation and retranslocation as a continuous process, throughout the leaf lifespan. In mixed species stands, interspecific variation in dominance influences nutrient uptake, growth rates and the relative shading in the crown. Leaf shading stimulates leaf abscission and significant retranslocation even when shading is only partial and leaves are young (Saur et al. 2000). These considerations are important in interpreting results of different studies and of interspecific difference in retranslocation. We found a common relationship between the pre-retranslocation amount of leaf nutrients and the amount of nutrients retranslocated across a range of contrasting species. This common relationship holds for all phases of retranslocation (representing different leaf ages) and for nitrogen, phosphorus and potassium (Figures 6 and 7). A similar pattern of nutrient movement has been found in the phyllodes of C. glauca grown at this site (Saur et al. 1997). We conclude that the amount of nutrients retranslocated during any phase in the life of leaves is strongly dependent on the amount (content) present at the start of the retranslocation phase, and that this is a generic relationship applicable across species when grown in plantations. In TREE PHYSIOLOGY VOLUME 28, 2008

RETRANSLOCATION OF FOLIAR NUTRIENTS IN EVERGREENS 195 addition to leaf size, several studies have drawn attention to the role of various nutrient pools in retranslocation (Chapin and Moilanen 1991, Hawkins and Polglase 2000, Wang et al. 2003, Niinemets and Tamm 2005). Retranslocation efficiency Nutrient retranslocation efficiency has been expressed as a proportional change in foliar nutrient concentration or content from the pre-retranslocation value. Niinemets and Tamm (2005) estimated retranslocation efficiency during leaf senescence in deciduous trees and attributed variation among species to differences in leaf longevity (lifespan based on timing of leaf fall) and the nature of nutrient pools, both leading to differences in the kinetics of retranslocation. In an ecological context, it has been shown that leaf lifespan is related to leaf mass per area; i.e., plants with low leaf mass per unit area have leaves with a short lifespan and high nitrogen concentration (Wright and Westoby 2004). This relationship would be important in mixed species stands containing species differing in leaf longevity. However, we found no evidence of a ranking for species by nutrient retranslocation efficiency, and there was no reliable variation in leaf longevity among our study species. In the Australian environment, leaf longevity is strongly dependent on site factors including soil water and nutrient availability. These considerations raise the question of whether the notion of retranslocation efficiency, expressed as a proportion of a somewhat arbitrarily defined initial value, is a useful explanatory measure for ranking species? To be so, the ratio must have a degree of stability within species. In our study, percentage retranslocation (content based) for the mature green leaves of P. radiata were 60% nitrogen and 63% phosphorus. In an earlier study applying the same method of sampling and calculations for comparable needle age, the percentage for nitrogen varied between 26 and 41% and percentage phosphorus between 19% and 40% (Nambiar and Fife 1991). Importantly, the rate of nitrogen applied to the stand, and the resulting response to growth, had a strong influence on percent retranslocation of nitrogen, phosphorus (Fife and Nambiar 1997) and potassium (Fife and Nambiar, unpublished data). Clearly, the percentage retranslocation varies considerably depending on the phase of retranslocation. For example, Table 3 shows no significant difference in percentage retranslocation for phosphorus in mature green leaves among species, but the corresponding species values were further apart when based on estimates during leaf senescence. Saur et al. (2000) studied retranslocation in E. globulus growing at another site using identical methods to those used here and found the following percentage retranslocation from mature green leaves: nitrogen 51%; phosphorus 54%; and potassium 31% (compared to 31% nitrogen and 18% potassium in our study). Thus the percentage retranslocation, even within a species, is highly dependent on site and stand characteristics, age of leaves during their life cycle and factors such as leaf shading. Consequently, it is unlikely to be a reliable measure for interspecies comparisons of retranslocation efficiency. Caution is required in relating resorption efficiency and the resorption proficiency derived from limited datasets to tree growth of different species in mixed stands (cf. May et al. 2005). Such analyses do not provide a reliable basis for exploring interspecific divergence in nutrient retranslocation. In conclusion, when genetically diverse species are grown in the same environment without interspecies competition for resources, there is a remarkable similarity in the pattern of, and the mechanism governing, nutrient retranslocation. Retranslocation occurred in three sequential phases during the life of leaves in all species. The interspecific variation in net amounts retranslocated during these phases was primarily due to the size of the foliage unit, and secondarily to the difference in foliar nutrient concentrations. There was no consistent difference between, or ranking in, retranslocation efficiency among species. Retranslocation was largely governed by seasonal effects on the nutrient requirement for shoot growth and factors that determine nutrient concentrations and contents of leaves. Acknowledgments We thank David Gritton for assistance with experimental work, Julian Mattay for advice on statistical analysis and presentation of the paper and Dr. Chris Beadle and Dr. Barrie May for constructive comments on the manuscript. References Aerts, R. 1996. Nutrient resorption from senescing leaves of perennials: are there general patterns? J. Ecol. 84:597 608. Boland, D.J., M.I.H. Brooker, G.M. Chippendale, N. Hall, B.P.M. Hyland, R.D. Johnston, D.A. Kleinig and J.D. Turner. 1984. Forest trees of Australia. Thomas Nelson Australia and CSIRO Publishing, Melbourne, Australia, 687 p. Chapin, F.S., III and L. Moilanen. 1991. Nutritional controls over nitrogen and phosphorus resorption from Alaskan birch leaves. Ecology 72:709 715. Escudero, A., J.M. del Arco, I.C. Sanz and J. Ayala. 1992. Effects of leaf longevity and retranslocation efficiency on the retention time of nutrients in the leaf biomass of different wood species. Oecologia 90:80 87. Fife, D.N. and E.K.S. Nambiar. 1982. Accumulation and retranslocation of mineral nutrients in developing needles in relation to seasonal growth in young radiata pine trees. Ann. Bot. 50: 817 829. Fife, D.N. and E.K.S. Nambiar. 1984. Movement of nutrients in radiata pine needles in relation to growth of shoots. Ann. Bot. 54:303 314. Fife, D.N. and E.K.S. Nambiar. 1991. Nutrient retranslocation in temperate conifers. Tree Physiol. 9:185 207. Fife, D.N. and E.K.S. Nambiar. 1995. Effect of nitrogen on growth and water relations of radiata pine families. Plant Soil. 168/169: 279 285. Fife, D.N. and E.K.S. Nambiar. 1997. Changes in the canopy and growth of Pinus radiata in response to nitrogen supply. For. Ecol. Manage. 93:137 152. Hawkins, B. and P.J. Polglase. 2000. Foliar concentrations and resorption of nitrogen and phosphorus in 15 species of eucalypts grown under non-limited water and nutrient availability. Aust. J. Bot. 48:597 602. Isbell, R.F. 1996. Australian soil and land survey handbook. The Australian Soil Classification, CSIRO Publishing, Melbourne, Australia, 143 p. TREE PHYSIOLOGY ONLINE at http://heronpublishing.com

196 FIFE, NAMBIAR AND SAUR Kumar, P. and A. Singh. 2005. N and P resorption efficiency in certain tropical tree species planted on mine spoil. Ind. J. For. 28:371 375. Lin, P. and W. Wang. 2001. Changes in leaf composition, leaf mass and leaf area during leaf senescence in three species of mangroves. Ecol. Eng. 16:415 424. May, J.D., S.B. Burdette, F.S. Gilliam and M.B. Adams. 2005. Interspecific divergence in foliar nutrient dynamics and stem growth in a temperate forest in response to chronic nitrogen inputs. Can. J. For. Res. 35:1023 1030. Millard, P. 1994. Measurement of remobilization of nitrogen for spring leaf growth of trees under field conditions. Tree Physiol. 14:1049 1054. Millard, P. and M.F. Proe. 1992. Storage and internal cycling of nitrogen in relation to seasonal growth of Sitka spruce. Tree Physiol. 10:33 43. Millard, P. and M.F. Proe. 1993. Nitrogen uptake, partitioning and internal cycling in Picea sitchensis (Bong.) Carr. as influenced by nitrogen supply. New Phytol. 125:113 119. Nambiar, E.K.S. and D.N. Fife. 1987. Growth and nutrient retranslocation in needles of radiata pine in relation to nitrogen supply. Ann. Bot. 60:147 156. Nambiar, E.K.S. and D.N. Fife. 1991. Nutrient retranslocation in temperate conifers. Tree Physiol. 9:185 207. Niinemets, Ü and Ü. Tamm. 2005. Species differences in timing of leaf fall and foliage chemistry modify nutrient resorption efficiency in deciduous temperate forest stands. Tree Physiol. 25:1001 1014. Northcote, K.H. 1979. A factual key for the recognition of Australian soils. Rellim, Adelaide, Australia, 123 p. Saur, E., D.N. Fife and E.K.S. Nambiar. 1997. Seasonal changes in accumulation and retranslocation of nutrients in Casuarina glauca cladodes. In Plant Nutrition for Sustainable Food Production and Environments. Eds. T. Ando et al. Kluwer Academic Publishers, Dortrecht, The Netherlands, pp 171 172. Saur, E., E.K.S. Nambiar and D.N. Fife. 2000. Foliar nutrient retranslocation in Eucalyptus globulus. Tree Physiol. 20:1105 1112. Sheriff, D.W., E.K.S. Nambiar and D.N. Fife. 1986. Relationship between nutrient status, carbon assimilation and water use efficiency in Pinus radiata (D. Don) needles. Tree Physiol. 2:73 88. Soil Survey Staff. 1975. Soil taxonomy. U.S. Dept. Agric. Hand. No. 436. U.S. Gov. Printing Office, Washington, DC, 754 p. van Heerwaarden, L.M, S. Toet and R. Aerts. 2003. Current measures of nutrient resorption efficiency lead to a substantial underestimation of real resorptrion efficiency: facts and solutions. Oikos 101: 664 669. Wang, W., M. Wang and P. Lin. 2003. Seasonal changes in element contents in mangrove element retranslocation during leaf senescence. Plant Soil 252:187 193. Wright, I. and M. Westoby. 2004. The economics of leaves. Austr. Sci. 25:33 37. TREE PHYSIOLOGY VOLUME 28, 2008