When the current level of carbohydrates produced by photosynthesis is not enough to meet the C demand for maintenance, growth or metabolism, trees use stored carbohydrates. In rubber trees (Hevea brasiliensis Muell. Arg.), however, a previous study (Silpi U., A. Lacointe, P. Kasemsap, S. Thanisawanyangkura, P. Chantuma, E. Gohet, N. Musigamart, A. Clement, T. Améglio and P. Thaler. 2007. Carbohydrate reserves as a competing sink: evidence from tapping the rubber tree. Tree Physiol. 27:881–889) showed that the additional sink created by latex tapping results not in a decrease, but in an increase in the non-structural carbohydrate (NSC) storage in trunk wood. In this study, the response of NSC storage to latex tapping was further investigated to better understand the trade-off between latex regeneration, biomass and storage. Three tapping systems were compared to the untapped Control for 2 years. Soluble sugars and starch were analyzed in bark and wood on both sides of the trunk, from 50 to 200 cm from the ground. The results confirmed over the 2 years that tapped trees stored more NSC, mainly starch, than untapped Control. Moreover, a double cut alternative tapping system, which produced a higher latex yield than conventional systems, led to even higher NSC concentrations. In all tapped trees, the increase in storage occurred together with a reduction in trunk radial growth. This was interpreted as a shift in carbon allocation toward the creation of reserves, at the expense of growth, to cover the increased risk induced by tapping (repeated wounding and loss of C in latex). Starch was lower in bark than in wood, whereas it was the contrary for soluble sugars. The resulting NSC was twice as low and less variable in bark than in wood. Although latex regeneration occurs in the bark, changes related to latex tapping were more marked in wood than in bark. From seasonal dynamics and differences between the two sides of the trunk in response to tapping, we concluded that starch in wood behaved as the long-term reserve compartment at the whole trunk level, whereas starch in bark was a local buffer. Soluble sugars behaved like an intermediate, ready-to-use compartment in both wood and bark. Finally, the dynamics of carbohydrate reserves appears a relevant parameter to assess the long-term performance of latex tapping systems.
Trees store a large amount of carbohydrates in parenchymatous tissues of their wood and bark, mainly as starch (see e.g., Glerum 1980, Kozlowski 1992, Lacointe et al. 1993, 1995, Witt and Sauter 1994, Barbaroux and Bréda 2002, Hoch et al. 2003), with a few exceptions like coconut palm, that stores soluble sugars too (Mialet-Serra et al. 2005). These stored carbohydrates play an important role in tree functioning, as they can be used when current photosynthesis is not enough to meet the carbon needs for maintenance and growth. Particularly after leaf shedding in deciduous species, they are the only source of carbohydrate during bud-break and early spring growth (Lacointe et al. 1993, Barbaroux and Bréda 2002). When trees experience a stress such as water stress or pest damage, stored carbohydrates are mobilized too (Dunn et al. 1990, Canham et al. 1999), either to compensate for reduced photosynthesis or as a response to additional carbohydrate demand. A more specific definition of carbohydrate storage in plants differentiates non-competitive accumulation, occurring only when supply exceeds demand for growth and maintenance, from actual reserve formation, i.e., synthesis of storage compounds that might otherwise directly promote growth (Chapin et al. 1990). When plants develop specialized reserve organs, such as tubers, this competitive reserve formation seems obvious (Milford and Thorne 1973). On the other hand, carbohydrate storage in tree wood parenchyma has long been viewed as only a passive accumulation process. However, such a view has been challenged in a recent study by Silpi et al. (2007) demonstrating in rubber tree (Hevea brasiliensis Muell. Arg.) that an increased C demand does not necessarily result in a depletion of carbohydrate concentration in wood. In the rubber tree, latex yield by tapping induces a strong, extra sink: regeneration of the exported latex, which consumes assimilates derived from the other sinks (Templeton 1969, Wycherley 1976, Jacob et al. 1998). Despite latex regeneration being carbon-costly, Silpi et al. (2007) showed that tapped trees experienced not a decrease, but an increase in the carbohydrate concentration of trunk wood, as compared to untapped trees. Such results backed the hypothesis that carbohydrate storage in tree wood, rather than passive accumulation, is actually a competitive reserve formation process (Cannell and Dewar, 1994, Lacointe 2000, Le Roux et al. 2001, Lacointe et al. 2004). Moreover, as concurrently to this increase in storage trunk radial growth was reduced (Silpi et al. 2006), the authors hypothesized that trees tend to adapt the level of stored carbohydrate to current metabolic demand, at the possible expense of other sinks (Silpi et al. 2007). However, another explanation could be that competition for assimilate was not the direct cause of growth limitation in tapped trees, as recent studies showing limited seasonal variations in non-structural carbon in several tree species (Hoch et al. 2003, Würth et al. 2005) suggest that growth of trees is not often carbon limited. To better understand carbon allocation among functional sinks – growth, secondary metabolites and reserves – long-term studies enabling the comparison of contrasting levels of assimilate availability are required. Therefore, the first objective of this study was to corroborate over 2 years that latex tapping induces an increase in NSC in wood and whether different tapping systems, inducing different carbon sinks, affect the level of storage.
In rubber tree, latex is synthesized in specialized cells, namely laticifers, organized in vessels located within the phloem tissue (Hébant and De Fay 1980). As sucrose is both the precursor of the rubber molecule (cis-polyisoprene, representing about 90% of latex dry mass), and the source of energy for latex metabolism, tapping induces a strong additional sucrose sink in trunk bark (Jacob et al. 1998). The origin of latex sucrose can be either the direct flow of photo-assimilates through phloem or starch pools located in the bark itself and in the wood (Gohet 1996), connected to laticifers through vascular rays (Hébant and De Fay 1980). Therefore, trunks of tapped trees represent an interesting model to study the respective role of the different fractions of NSC (soluble sugars and starch), which are compartmented in different tissues (wood and bark) within tree trunks. Is there a specific role for each compartment (wood and bark) and for the different combinations of compartment and chemical fraction? Due to difficulty of sampling, few studies have assessed such an issue (Egger et al. 1996, Barbaroux et al. 2003). Comparing dynamics of carbohydrates in wood and bark as related to tapping would help understanding of their role in reserve dynamics in trees, beyond the specific case of rubber trees. However, the previous study (Silpi et al. 2007) analyzed carbohydrates in wood only. Thus, our second objective was to unravel the relative contribution to trunk carbohydrate dynamics of soluble sugars and starch in both wood and bark and the relationships between these different compartments.
More specifically, in rubber trees, the ability of different tapping systems to mobilize carbohydrate toward latex regeneration is considered a key for their productivity and durability (Wycherley 1976, Gohet 1996). Therefore, measuring carbohydrate concentration in trunk wood has been proposed as a complementary tool to help forecast long-term performance of tapping systems (Silpi et al. 2007), in addition to latex diagnosis (Jacob et al. 1995), which is a good indicator of the metabolic status of the laticifer cells. In this study, dynamics of carbohydrates were compared between three tapping systems, including the new double cut alternative (DCA) tapping system (Gohet and Chantuma 2003) specifically designed for smallholders. It was shown that trees under this system have a favorable latex metabolic profile (Chantuma et al. 2006), but the physiological bases of this system are not well enough established yet, so that its long-term potential remains uncertain. Thus, a third, more applied objective of this study was to test whether DCA performance relies on the differences in carbohydrate mobilization for latex regeneration and conversely if carbohydrate dynamics could help forecasting the long-term impact of such a tapping system.
Materials and methods
Experimental site and plant material
The experiment was conducted in a H. brasiliensis (rubber tree) monoclonal plot, clone RRIM 600, at the Chachoengsao Rubber Research Center (CRRC-DOA, 13.41° N and 101.04° E, 69 m above the mean sea level), Thailand. Trees were planted in 1992 in a 2.5 × 7.0 m planting design (571 trees ha−1). The soil type was Kabin Buri soil series (sandy clay loam to clay loam). The climate is considered as tropical monsoon climate (Am, Köppen classification), although the dry season lasts about 5 months, from December to April. Temperature ranged from 17.6 to 36.5 °C, and the mean relative air humidity was 63.5%. Annual rainfall averaged 1291 mm year−1 (15 years mean).
Latex tapping treatments (Figure 1)
Latex tapping involves periodically cutting bark on the trunk, and hence severing latex vessels. One cut is performed on half the circumference of the trunk and regularly reopened by excising at each tapping a new, thin shaving of bark from the sloping cut. As a result, latex flows immediately along the cut into a cup attached to the trunk. The flow progressively diminishes, and eventually stops after 1–3 h as severed vessel ends become plugged by caps of latex coagulum. In this experiment, three tapping systems (D/2, D/4 and DCA) were compared to untapped trees (Control).
Conventional mono-cut systems, half spiral D/2 and D/4
Trees were tapped (i.e., the cut was reopened) either every 2 days (D/2) or every 4 days (D/4). High frequency (D/2) is considered the standard tapping system in Thailand (the world’s first natural rubber producer). Lower frequency (D/4) is widely used in large estates to increase labor productivity, as it allows higher production per tapping day, linked to longer regeneration time between two tapping dates (review in Jacob et al. 1998). However, this system requires the use of Etephon (2-chloroethylphosphonic acid; Ethrel® 5 LS, Bayer Cropscience, Monheim, Germany) stimulation to compensate for the low tapping frequency. Etephon was applied with a brush on the tapped panel, close to the cut, six times per year. We used 0.6 g of 2.5% stimulant paste per tree and per stimulation (i.e., 15 mg of active ingredient per tree per stimulation). The effects of Etephon, which releases ethylene in the tissues, on latex physiology are widely documented (Coupé and Chrestin 1989). Mainly, it prolongs latex flow and enhances metabolism. Stimulant applications were evenly distributed from May to December.
In Thailand, most farms are too small in size to benefit from low-frequency systems and use, on the contrary, very high tapping frequencies, that allow tapping of the full plantation within 1 day (Anekachai 1989) but do not allow enough time for the tree to fully regenerate latex between two tapping times (Jacob et al. 1998). In such a context, the new DCA (Gohet and Chantuma 2003) was specifically designed for smallholders, to fit better to the physiological potential of trees, without reducing the income per tapping day or per surface area. As compared to the reference system (D/2, tapping every 2 days), DCA increased the latex yield by 20% during the first 5 years, without additional cost. The principle of this system is to enlarge the area involved in latex regeneration and to increase the regeneration time by splitting tapping between two cuts, located on opposite faces (named panels) of the trunk (Figure 1). Each cut is tapped every 4 days alternately, so that the tapping frequency at tree scale is D/2, whereas it is D/4 for each panel. DCA was not stimulated. Details of the system and its results over 5 years of tapping are provided in Chantuma et al. (2006). According to International Nomenclature, tapping systems are 1/2S d/2 (D/2), 1/2S d/4 ET 2.5% 6/y (D/4) and 2 × 1/2S d/4 (DCA). Tapping started in May 2000, when the trees were 8 years old. It was stopped every year in February, March and April (dry season), allowing 9 months of tapping and 3 months of resting period (standard schedule in this area). Each tapping treatment comprises 10 trees (treatment replications).
Sampling for analyses of carbohydrates
Non-structural carbohydrates (NSC) were measured during 2 years (2003 to 2004 or years 4 and 5 of tapping). Sampling schedule was designed to assess the key periods according to season, climate, the annual growth cycle and the latex production cycle (Table 1).
|5 February||19 January||Dry||Leaf shedding, no radial growth||Yes|
|6 March||20 February||Dry||Re-foliation, first flush completed||No|
|2 May||11 May||Rain starting||Mature leaf, radial growth starting||Starting (low yield)|
|–||6 August||Rainy||High radial growth||Yes|
|28 October||18 October||Rainy||Low radial growth||Yes (peak)|
|5 February||19 January||Dry||Leaf shedding, no radial growth||Yes|
|6 March||20 February||Dry||Re-foliation, first flush completed||No|
|2 May||11 May||Rain starting||Mature leaf, radial growth starting||Starting (low yield)|
|–||6 August||Rainy||High radial growth||Yes|
|28 October||18 October||Rainy||Low radial growth||Yes (peak)|
Along the trunk, samples were taken at 20, 50, 80, 110, 140, 170 and 200 cm from the ground on one side of the tree in Control and on both sides (panels A and B) in tapped treatments (seven samples on each panel, including renewing bark area in tapped panels). Additional samples at 300, 400, 500 and 600 cm from the ground and at 10 and 30 cm deep in taproots were also analyzed, but this study focuses on the unbranched section of the trunk. Moreover, these extra samples would unbalance the design as they comprise only one sample per height, not considering the side of the tree.
At each sampling date, groups of three trees from each treatment were sampled among the 10 replications. Sample trees were alternated, to reduce the metabolic perturbation and necrosis hazard due to core-sampling from one period to the next one. Samples consisted of 0.5 cm diameter, 5 cm long cores, including around 1 cm of bark and 4 cm of wood. They were made with a wood auger. Wood and bark were separated. After each core was sampled, it was soaked immediately in liquid nitrogen and was kept in a cryo-tube immersed in liquid nitrogen until transfer to the laboratory and stored at −80 °C, before freeze-drying using a −50 °C freeze-dryer (Telstar Cryodos, Spain). Thereafter, the samples were individually blended using ball-blender MM200 (Retsch, Germany), ball diameter 7 mm. Storage after this step until extraction and chemical analysis was at −80 °C.
Starch, sucrose, glucose and fructose concentrations were analyzed enzymatically (Boehringer 1984). The powder was re-dried in the oven for 2 h at 65 °C. Soluble sugars were extracted from 20 mg samples with 80% ethanol during 30 min at 80 °C and then centrifuged. This step was repeated twice, first with 80% ethanol and then with 50% ethanol, and all the supernatants were pooled. The sediment, which contained starch, was filled with 80% ethanol and kept at −80 °C until analysis. The supernatant was filtered in crushed glass mini-columns added with a mixture of polyvinyl polypyrrolidone (Sigma-Aldrich, St. Louis, MO) and activated charcoal to eliminate pigments and polyphenols. Ethanol was evaporated using a vacuum dryer (Maxi Dry Plus®, Heto, Denmark). Soluble sugars and starch were quantified by enzymatic analysis (enzymes from Roche Diagnostics®, Meylan, France). Sucrose was hydrolyzed into glucose and fructose by invertase (β-fructofuranosidase). Glucose and fructose were quantified using hexokinase, glucose-6-phosphate dehydrogenase and phosphoglucose isomerase followed by spectrophotometry (Shimadzu, Kyoto, Japan) of the resulting NADPH at 340 nm. For starch analysis, after the ethanol was evaporated, the sediment was hydrolyzed with NaOH 0.02 N for 1.5 h at 90 °C, digested with α-amyloglucosidase for 1 h at 50 °C and then glucose was quantified as described above. The results were expressed as milligram glucose equivalent per gram of structural dry matter (mgGlu/gSDM), i.e., the dry matter after extraction of the NSC. Sum of starch and soluble sugar (sucrose, fructose and glucose) represented the NSC.
Complementary measurements: latex production and girth measurements
Latex production and girth were measured from the onset of tapping in May 2000. Latex was collected from the field and weighted every 4 weeks for each tree, and its dry weight was estimated according to 85% mean total solid concentration. Trunk girth was measured yearly at 170 cm from the ground with a tape. Mean tree girth at the beginning of the experiment (May 2000) was 49.2 cm, ranging from 46.5 to 51.6 cm. Latex yield and trunk girth increment were calculated from May to May according to the tapping calendar.
The significance of tapping treatment (Control, D/2, D/4 and DCA) and season (nine sampling dates) on starch, soluble sugars and NSC concentrations was assessed through a crossed-factor analysis of variance (ANOVA), with sampling height above ground level (0–200 cm) as an additional, replication block factor to improve test power and panels (sides of the tree) merged together.
Effect of panel (A versus B) was assessed on a sub-sample made of treatments D/2 and D/4 only, as the tapping effect was not differentiated in DCA, where both panels were tapped. Thereafter, combined treatments (panel × tapping system combinations) were compared as independent treatments. Statistical analyses were conducted using SAS (SAS Institute, Cary, NC) or Xlstat (Addinsoft SARL, France).
Effect of latex tapping on trunk carbohydrate content (Figure 2)
Differences between treatments were much larger in wood than in bark.
Mean NSC concentration on a trunk scale was significantly higher (P < 0.001) in tapped treatments (D/2, D/4 and DCA) than in untapped Control. This was a result of higher starch concentrations. Starch and NSC concentrations were the highest in DCA, medium in D/2 and D/4, and the lowest in Control. The soluble sugars fraction was much lower (ca. 19% of NSC) and less responsive to treatments than starch, although D/4 had a higher soluble sugars concentration than the others (P = 0.001). Measurements within the crown (300–600 cm from the ground, data not shown) had the same trend. In roots (data not shown), there was no difference between treatments.
NSC was more than twice lower in bark than in wood in all treatments. Soluble sugars accounted for about 70% of NSC in bark and tapping did not change much concentration in both starch and NSC. D/4 only had a significantly lower NSC (P < 0.001), due to lower soluble sugars, whereas starch was slightly higher in DCA than in other treatments (P < 0.001).
Effect of panel on trunk carbohydrate content (Figure 3)
In conventional systems D/2 and D/4, the untapped side of the tree (panel B) had significantly higher starch (+4 mgGlu/gSDM and P = 0.001) than the tapped one (panel A). However, the latter had still higher concentration than Control (+6 mgGlu/gSDM and P = 0.001). Considering NSC, the difference between panels was significant only in D/2 (P < 0.001). In DCA, both panels, which were tapped, had the same concentration in starch and NSC as the untapped panel of D/2 and D/4, and therefore higher concentration than the tapped panel of these conventional treatments. Difference in starch concentration between DCA and Control was high (+11 mgGlu/gSDM and P < 0.001).
Bark of D/2 and D/4 had significantly higher starch (P = 0.006), soluble sugars (P < 0.001) and NSC (P < 0.001) in tapped panel A than in untapped panel B. Thus, contrary to what was recorded in wood, carbohydrate in bark was higher in the area closer to the tapping cut. Among all panel × treatment combinations, tapped panel A of D/2 had the highest NSC. In DCA, there was no difference in NSC between panels, both tapped.
When compared to Control, starch in bark was slightly higher in tapped panel A and slightly lower in untapped panel B of D/2 and D/4. The balance in the whole trunk was not different from Control. Only DCA had a slightly higher starch concentration than Control in bark in both panels and in the whole trunk.
Starch, soluble sugars and NSC seasonal dynamics in Control
Effect of sampling date was highly significant (P < 0.001) on starch, soluble sugars and NSC concentrations in both wood and bark. The overall carbohydrate concentration pattern (Figures 4 and 5) had the same seasonal trends in wood and bark but the range of change was much larger in wood than in bark. In wood (Figure 4), during the first year of observation, the highest NSC concentration was recorded at leaf fall (February 2003) followed by a huge drop just after complete re-foliation (March 2003). A net deposition occurred mainly from May 2003 to next leaf fall (January 2004), i.e., the period, including the rainy season, when radial growth occurred. During the second year of observation (2004), the trend was the same, although in wood the drop in starch and NSC concentration after re-foliation was of lower extent than the previous year. During the following vegetative season (from May to October 2004), the increase in starch and NSC was not as clear as the previous year. In bark (Figure 5), seasonal changes in starch and NSC were similar in both years.
In wood, the concentration of soluble sugars was lower than the concentration of starch all year long and the seasonal pattern was different, nearly the opposite (Figure 4). In the late season and defoliation period (from October to February of both years), starch was high and decreasing, whereas soluble sugars were low and stable or increasing. Conversely just after growth resumption, in March, starch concentration was low and re-increasing, whereas the concentration of soluble sugars was high and decreasing. However, seasonal variations in NSC were mainly accounted for by variations in starch.
In bark (Figure 5), soluble sugars and starch exhibited more parallel patterns. In particular, both fractions peaked during the defoliated period and were low just after re-foliation (between March 2003 and February 2004). However, soluble sugars re-increased immediately after re-foliation even if starch re-deposition was delayed by several weeks. Both changes in soluble sugars and starch contributed significantly to seasonal variations in bark NSC.
Effect of latex tapping on starch, soluble sugars and NSC dynamics
Although the overall pattern was similar for all treatments, the date × treatment interaction was significant (P < 0.001) for all carbohydrate fractions, indicating that dynamics differed among treatments at times in both wood and bark.
Most of the time, starch and NSC were higher in tapped treatments (D/2, D/4 and DCA) than in untapped Control (Figure 4), although there were some variations during the 2 years. For example, at the second leaf fall (January 2004), starch peaked for all tapped treatments, whereas it decreased for Control. At that time, the difference between tapped treatments and Control was the highest recorded during the 2 years for both starch and NSC. The additional sampling date in August 2004 showed that during the time of high growth for tapped trees (from May to August), starch decreased slightly in tapped trees, whereas it increased in Control. During the time of high growth and high latex production (rainy period, from May to October 2003), soluble sugars decreased much more for Control than for tapped treatments, which exhibited no or very slight decrease.
Differences between tapped treatments were not the same over the year. Particularly, the concentration of soluble sugars changed more over time for D/4 than for D/2 and DCA.
Differences between tapped and untapped treatments were not as clear as in wood. Starch dynamics were very close for D/2 and Control, except in October 2004. As a whole, variations of starch in DCA were more pronounced and not synchronized with other treatments. Particularly, its concentration dropped at leaf fall in January 2004, whereas it peaked for all other treatments. At leaf fall in both years, Control had the highest soluble sugars, whereas differences between treatments were the lowest after re-foliation, when the soluble sugars concentration was low for all treatments. In August 2004, the soluble sugars concentration dropped in D/4 contrary to other treatments.
Trunk girth and latex yield (Figure 6)
Trunk girth increment was higher in the untapped Control (19.9 cm cumulated over 5 years) than in the tapped treatments, but the difference was not significant in 2003 to 2004, when the precipitations were lower than the mean. Within tapped treatments, D/2 had a higher cumulated girth increment (13.5 cm) than D/4 (10.7 cm) and DCA (11.4 cm).
DCA had the highest cumulated latex yield (22.39 kg tree−1) during the first 5 years of tapping, followed by D/2 (20.44 kg tree−1) and D/4 (14.62 kg tree−1). However, in year 5 (2004 to 2005) DCA had a lower latex yield than D/2.
This study confirmed the results obtained previously in the same agronomic conditions (same location and same rubber tree clone) during a 1-year-long study in 2002 (Silpi et al. 2007). Mainly, NSC in trunk wood was higher in the tapped treatments than in the untapped Control. This was found to be true at almost all the sampling periods over 2 years of experiment. Thereby, we confirmed that the additional carbohydrate demand created by the regeneration of latex did not deplete wood carbon storage, but, on the contrary, resulted in an increase of such storage.
Trade-off between latex regeneration, biomass increment and storage
The concurrent decrease in radial growth under tapping raises questions. Recent studies on both tropical (Würth et al. 2005) and temperate (Hoch et al. 2003) forest species suggested that tree growth is not often carbon limited, but that water limitation is more probable, even in wet tropical conditions. The very fast decrease in the growth rate of rubber trees following the onset of tapping (Silpi et al. 2006) backs such a hypothesis, as it is not likely that carbohydrate reserves could be completely depleted so fast. Therefore, we can wonder whether the trade-off between latex production and growth is based on the carbon availability or on the availability of other elements. Actually, Häring and Körner (2004) showed that latex synthesis in Euphorbia lathyris L. is not carbon limited. However, this situation was comparable to that of our untapped rubber trees. Without tapping there is no loss of latex, and therefore low metabolic activity in laticifers (Jacob et al. 1998). Conversely, there is no doubt that tapping creates an additional carbon sink, as it drives latex out of the trees, and this latex has to be regenerated. Such a regeneration uses sucrose that drops in laticifers drastically, following tapping (Gohet 1996, Jacob et al. 1998). Moreover, the biomass equivalent of exported latex is significant as compared to trunk biomass increment (Gohet 1996). On the other hand, even if the protein content of fresh latex is not negligible, N and other elements are exported in very low amounts in latex (0.09% of N, 0.02% of P and 0.06% of K of latex dry weight, d’Auzac and Jacob 1989) as compared to the amount of carbon (about 80% of latex dry weight). Similarly, very little water is exported in fresh latex (0.2–0.6 l day−1 in our conditions) as compared to the amount of transpired water (120–180 l day−1, Pakianathan et al. 1989). Hence, although our data are not sufficient to completely dismiss this hypothesis, we do not think that tapping limits the availability of necessary elements other than carbon.
An alternative hypothesis to explain enhanced storage together with reduced growth refers to the economic analogy of storage proposed by Chapin et al. (1990) who stated that ‘The greater the risk (high probability of a large or a frequent loss), the more a firm or a plant should save’. Tapping actually creates a risky situation, as the tree is submitted to repeated, almost continuous, wounding and loss of carbon. Hence, as an ‘insurance’ to face this increased risk, the tree may adjust carbon allocation to store more carbohydrates. Moreover, as latex synthesis in reaction to wounding is a defense mechanism, tapped trees could devote larger amounts of carbon resource to sustain this defense (Christiansen et al., 1987), at the possible expense of growth. It has been recently demonstrated that the carbon found in the reaction zones around bark infection by fungi comes mainly from reserves (Guérard et al., 2007). Therefore, the concomitant increase in carbon storage and decrease in trunk radial growth do not necessarily mean that carbon is not limiting, but rather that growth is not always the priority sink.
NSC concentration as related to tapping systems
Differences in NSC concentration between tapping treatments can be explained by this sink effect induced by latex tapping. In conventional systems, with only one side of the tree under tapping, starch tended to accumulate in the wood tissue of the side opposite to tapping cut. This could be understood as follows: tapping panel A created a sink effect on both panels, but as panel B was not directly involved in latex regeneration and associated C consumption, this resulted in higher starch accumulation rate creating a reserve pool that may be available later. In DCA, tapping on each panel created a sink effect on the other one. As a whole, the increase in demand (latex regeneration plus stronger wounding stress-associated metabolism due to two tapping cuts instead of one) resulted in a higher storage of carbohydrates. As expected, this higher reserve level combined with higher yield resulted in DCA having the lowest radial growth rate among all treatments.
As the exploited part of the rubber tree is trunk bark, where rubber biosynthesis actually occurs, it was important to assess carbohydrate dynamics in this tissue, in addition to wood. Similarly to results obtained in beech and oak species (Barbaroux et al. 2003), soluble sugars were more concentrated than starch in bark. But contrary to what was found by these authors, lower total carbohydrate (NSC) concentrations were recorded in bark than in wood. Thus, despite the occurrence of specific laticifer tissue and the related metabolic activity, trunk bark of rubber trees is not particularly concentrated in carbohydrate. Tapping induces a strong decrease in sucrose concentration within laticifer cells, due to regeneration of the exported latex (Gohet 1996, Jacob et al. 1998). Therefore, we could expect clear differences in bark carbohydrate concentration between tapped and untapped trees. On a whole trunk scale, this was not marked but, more locally, carbohydrate in bark was higher in the area closer to the tapping cut (tapped panel) showing a local sink effect for carbohydrate in bark.
Variability in seasonal dynamics
In the year 2003, we confirmed the seasonal variations observed in 2002 by Silpi et al. (2007). However, as we investigated two consecutive years, we found that inter-annual variability was large. Particularly, peak starch concentration was similar in 2003 and in 2004 to that recorded in Silpi et al. (2007), but drop following re-foliation was larger in 2003 and lower in 2004. Mean starch concentration in Control in year 2, 56.5 mgGlu/gSDM, was much higher than in Silpi et al. (2007), 38.5 mgGlu/gSDM. Such annual differences may be related to climate. The lower decrease in NSC recorded at re-foliation during the dry year (2004) may indicate that growth and consequently carbohydrate demand were limited by drought (Würth et al. 2005). High variability for treatment D/4 in year 2 may be related to stimulation in this tapping system, as after application of stimulant (six times a year), steep but transitional changes are known to occur in trunk metabolism (Jacob et al. 1998). Additional sampling in August 2004, in the middle of the rainy season, showed that starch and NSC concentration increased only in Control at that time. Results by Silpi et al. (2006) showed that radial growth of tapped trees was almost stopped in August, whereas it was steady in Control. Thereby, a steady radial growth may be beneficial to accumulation of starch in trunk wood, as newly formed tissues have a high proportion of parenchyma (De Fay 1999), providing ‘room for storage’ (Wargo 1979, Lacointe et al. 1993, Lacointe 2000).
Role of the different carbohydrate compartments
Comparison of seasonal dynamics of carbohydrates in wood and bark of Control also provided information on the relative role of the four-trunk carbohydrate pools: soluble sugars in wood, soluble sugars in bark, starch in wood and starch in bark. The major trait was a clear opposition between soluble sugars in wood and the other three pools, particularly the increase in soluble sugars in wood during re-foliation when soluble sugars in bark dropped together with starch in both wood and bark. As (i) starch in wood was by far the major component of NSC at trunk scale and (ii) its variation range was wider than that of the other components, we consider that it constituted the long-term reserve tank, dispatched all along the trunk, even far from the main sinks (i.e., in the untapped panel). Starch in bark acted more as a local buffer, less variable and with higher concentration in the vicinity of the sink (tapped panel). Soluble sugars in wood did not play a major role in reserves but their dynamics were more related to transport patterns. In bark, the concentrations of soluble sugars were slightly higher than in wood, with a similarly narrow range of absolute variation. Although we have no information on how much soluble sugars were located in active phloem, parenchyma and the laticifer vessels, respectively, we can infer that soluble sugars in bark were a ready-to-use compartment.
Application to the assessment of tapping systems
Analysis of carbohydrate dynamics proved a relevant tool to help understand the impact of different tapping systems on the tree functioning and to provide information for tapping management. As starch accumulated in the untapped panel B of conventional treatments (D/2 and D/4), we can forecast that tapping one panel for several years consecutively would result in a high pool of starch available in the opposite panel, explaining the increase in latex yield obtained after shifting to this side (Lacote et al. 2004). It would be relevant to repeat the same study with systems shifting tapping panel every year or every 2 years (Lacote et al. 2004). The new and more productive DCA, could benefit from the observed positive interaction between the two tapped panels resulting in higher starch concentration on both panels and in the whole trunk. Therefore, as more carbon resources were available in the vicinity of tapping cuts, not only latex regeneration could be higher, but the metabolic profile was also better. Chantuma et al. (2006) demonstrated that DCA latex had both high Pi (high metabolic activity) and high sucrose concentration (high substrate and energetic resource) for rubber biosynthesis. Such a favorable pattern is likely to ensure sustainable advantage to DCA as compared to conventional—single cut—systems. Next development of this applied research field will be to go further in studies of relationships between latex metabolism and trunk carbohydrates to better understand the relative contributions of the different areas of the tree to latex regeneration. However, as latex yield is also related to trunk girth (Gohet 1996) long-term production could be hampered in DCA, because this treatment has a lower growth. Moreover, the possible negative consequences on leaf growth and thereafter on carbon assimilation have to be assessed too. Future research should aim at extrapolating from carbohydrate concentrations to total carbon pools, through assessment of the size of the different compartments, to better evaluate the relative importance of carbon allocated to biomass increment, storage and latex.
This work was funded by the Action Thématique Programmée CIRAD ‘Les réserves carbonées chez le cocotier, le palmier à huile, l’hévéa et le manguier: origines, dynamiques et conséquences pour la gestion des plantations’. Bilateral fellowships were funded by the Thai-French Committee for Higher Education and Research, ‘Improving the rubber tree productivity’ project. We thank CRRC-DOA for hosting the experiment and for laboratory facilities. Grateful thanks to CRRC staff for both laboratory and field work.