Bitter problems in ecological feeding experiments: Commercial tannin preparations and common methods for tannin quantifications

In the field of plant–herbivore interactions, research methods where plant secondary metabolites are manipulated are becoming more and more popular. Among the most commonly used is tannic acid. However, recent studies have shown that different tannic acid preparations are not comparable in their tannin structures. While tannic acids are meant to contain only gallotannins (GTs), some commercial preparations compose mainly of more simple galloylglucoses (that have, e.g. much lower protein precipitation capacity than GTs) or even of gallic acid (the hydrolysis product of GTs). Another group of tannins used in feeding trials is condensed tannins (CTs), usually in the form of quebracho tannin. Quebracho, however, contains different CT structures than, e.g. leaves of many deciduous trees. Additionally, when analysed with the common acid-butanol assay for total CTs, quebracho tannins give even 30-fold lower absorbance than the CTs of those deciduous trees. In addition to above problems it has been shown that different tannins can give different response even within the same herbivore species, and that the same tannin structure can cause different response in different herbivores. Below we review these problems, as well as some means to deal with them.     

Ecological tannin assays: a critique

Summary The concept of protein precipitation potential has recently been introduced by Wisdom et al. (1987) as a means to combine chemical and protein precipitation assays of tannins for ecological studies. The definition of protein precipitation potential was not theoretically rigorous, and data analysis was obscure. Our attempts to repeat the tannin extraction procedure gave incomplete recovery (24% loss of quebracho) of condensed tannins, the only type considered by Wisdom et al. In contrast we found their method efficient for hydrolysable tannins (80% recovery of tannic acid) which were undetected by their chemical assay of tannins. Their protein precipitation assay was confounded by chemical interference from both types of tannins. We conclude with recommendations for this type of analysis.     

Microbial colonisation of tannin-rich tropical plants: Interplay between degradability,methane production and tannin disappearance in the rumen

Condensed tannins in plants are found free and attached to protein and fibre but it is not known whether these fractions influence rumen degradation and microbial colonisation. This study explored the rumen degradation of tropical tannin-rich plants and the relationship between the disappearance of free and bound condensed tannin fractions and microbial communities colonising plant particles using in situ and in vitro experiments. Leaves from Calliandra calothyrsus, Gliricidia sepium, and Leucaena leucocephala, pods from Acacia nilotica and the leaves of two agricultural by-products: Manihot esculenta and Musa spp. were incubated in situ in the rumen of three dairy cows to determine their degradability for up to 96 h. Tannin disappearance was determined at 24 h of incubation, and adherent microbial communities were examined at 3 and 12 h of incubation using a metataxonomic approach. An in vitro approach was also used to assess the effects of these plants on rumen fermentation parameters. All plants contained more than 100 g/kg of condensed tannins with a large proportion (32–61%) bound to proteins. Calliandra calothyrsus had the highest concentration of condensed tannins at 361 g/kg, whereas Acacia nilotica was particularly rich in hydrolysable tannins (350 g/kg). Free condensed tannins from all plants completely disappeared after 24-h incubation in the rumen. Disappearance of protein-bound condensed tannins was variable with values ranging from 93% for Gliricidia sepium to 21% for Acacia nilotica. In contrast, fibre-bound condensed tannin disappearance averaged ～ 82% and did not vary between plants. Disappearance of bound fractions of condensed tannins was not associated with the degradability of plant fractions. The presence of tannins interfered with the microbial colonisation of plants. Each plant had distinct bacterial and archaeal communities after 3 and 12 h of incubation in the rumen and distinct protozoal communities at 3 h. Adherent communities in tannin-rich plants had a lower relative abundance of fibrolytic microbes, notably Fibrobacter spp. whereas, archaea diversity was reduced in high-tannin-containing Calliandra calothyrsus and Acacia nilotica at 12 h of incubation. Concurrently, in vitro methane production was lower for Calliandra calothyrsus, Acacia nilotica and Leucaena leucocephala although for the latter total volatile fatty acids production was not affected and was similar to control. Here, we show that the total amount of hydrolysable and condensed tannins contained in a plant govern the interaction with rumen microbes affecting degradability and fermentation. The effect of protein- and fibre-bound condensed tannins on degradability is less important.     

Cottonwood hybridization affects tannin and nitrogen content of leaf litter and alters decomposition

Cottonwoods are dominant riparian trees of the western United States and are known for their propensity to hybridize. We compared the decomposition of leaf litter from two species (Populus angustifolia and P. fremontii) and their hybrids. Three patterns were found. First, in one terrestrial and two aquatic experiments, decomposition varied twofold among tree types. Second, backcross hybrid leaves decomposed more slowly than those of either parent. Third, the variation in decomposition between F₁ and backcross hybrids was as great as the variation between species. These results show significant differences in decomposition in a low-diversity system, where >80% of the leaf litter comes from just two species and their hybrids. Mechanistically, high concentrations of condensed tannins in leaves appear to inhibit decomposition (r ²=0.63). The initial condensed tannin concentration was high in narrowleaf leaves, low or undetectable in Fremont leaves, and intermediate in F₁ hybrid leaves (additive inheritance). Backcross hybrids were high in condensed tannins and were not different from narrowleaf (dominant inheritance). Neither nitrogen (N) concentration nor the ratio of ash-free dry weight to N (a surrogate for carbon:nitrogen ratio) were significantly correlated with decomposition. The N content of leaf material at the end of each year’s experiment was inversely correlated with rates of litter mass loss and varied 1.6- to 2.1-fold among tree classes. This result suggests that hybrids and their parental species are used differently by the microbial community. Received: 7 April 1999 / Accepted: 2 November 1999     

A genetic polymorphism for tannin production in Lotus corniculatus and its relationship to cyanide polymorphism

Summary A study of 172 unnamed populations and 22 cultivars of Lotus corniculatus showed: (1) that all plants of most strains of both categories contained leaf tannins (total 172 strains); (2) that 6 strains were tannin-negative; and (3) that 16 strains were polymorphic. Because of the small number of tested plants per strain, the above frequency of polymorphism is probably underestimated. Tannin-negative or polymorphic strains are frequent in Iran and Turkey. Leaftannin production is inherited as a monogenic dominant with tetrasomic inheritance. Repeated scores suggest that some individuals always, others sometimes and yet others never produce leaf tannins. Mean tannin content of 6 cultivars was strongly negatively associated with mean cyanide content.     

A comparison of two strategies to modify the hydroxylation of condensed tannin polymers in Lotus corniculatus L

A full-length sense Antirrhinum majus dihydroflavonol reductase (DFR) sequence was introduced into birdsfoot trefoil (Lotus corniculatus L.) in experiments aimed at modifying condensed tannin content and polymer hydroxylation in a predictable manner. Analysis of transgenic plants indicated lines that showed enhanced tannin content in leaf and stem tissues. In contrast to previous data from root cultures, levels of propelargonidin units were not markedly elevated in lines with enhanced tannin content. RT-PCR analysis of four selected lines indicated a correlation between enhanced tannin content and expression of the introduced DFR transgene. Using a contrasting approach we introduced a flavonoid 3'5' hydroxylase (F3'5'H) sequence derived from Eustoma grandiflorum into Lotus root cultures. Expression of the transgene was associated with increased levels of condensed tannins and in this case there was also no alteration in polymer hydroxylation. These results suggest that additional mechanisms may exist that control the hydroxylation state of condensed tannins in this model species.     

Methane emission by goats consuming different sources of condensed tannins

Twenty-four yearling Boer × Spanish wethers (7/8 Boer; initial body weight (BW) of 37.5 ± 0.91 kg) were used to assess effects of different condensed tannin (CT) sources on methane (CH₄) emission. Diets were Kobe lespedeza (Lespedeza striata; K), K plus quebracho providing CT at 50 g/kg dry matter (DM) intake (KQ), Sericea lespedeza (Lespedeza cuneata; S), and a 1:1 mixture of K and S (KS). Forages harvested daily were fed at 1.3 times the maintenance metabolizable energy requirement. The experiment was 51 days divided into two phases. In phase A forage diets were fed alone, and in phase B, 25 g/day of polyethylene glycol (PEG) was given mixed with 50 g/day of ground maize grain. Adaptation periods were 28 and 7 days in phases A and B, respectively. After adaptation there were 8 days for feces and urine collections, with gas exchange measured on the last 2 days. Ruminal fluid was collected at the end of the experiment via stomach tube for microbiology assays. The N concentration was 22.8 and 23.6 g/kg DM, in vitro true DM digestibility was 0.698 and 0.648, and the level of CT was 140 and 151 g/kg DM for S and K, respectively. DM intake was similar among treatments in both phases (phase A: 720, 611, 745, and 719 g/day (S.E. = 59.0); phase B: 832, 822, 867, and 880 g/day (S.E. = 55.3) for K, KQ, S, and KS, respectively). N digestibility was affected by treatment in phase A (P<0.05) but not in phase B (phase A: 0.514, 0.492, 0.280, and 0.413 (S.E. = 0.0376); phase B: 0.683, 0.650, 0.638, and 0.662 (S.E. = 0.0203) for K, KQ, S, and KS, respectively). Gross energy digestibility was similar among treatments in phase A (0.475, 0.407, 0.393, and 0.411 (S.E. = 0.0353)) but differed among treatments in phase B (0.449, 0.373, 0.353, and 0.409 for K, KQ, S, and KS, respectively (S.E. = 0.0221)) CH₄ emission was 9.6, 6.8, 10.6, and 8.9 l/day (S.E. = 1.44) in phase A and 19.0, 16.6, 21.8, and 19.2 l/day (S.E. = 1.51) in phase B for K, KQ, S, and KS, respectively (S.E. = 1.25). When data of both phases were pooled, supplementation with PEG in phase B markedly increased (P<0.05) CH₄ emission (9.0 l/day versus 19.1 l/day). In accordance, there was a substantial difference (P<0.05) between phases in in vitro CH₄ emission by ruminal fluid incubated for 3 weeks in a methanogenic medium and with other conditions promoting activity by methanogens (11.5 and 22.9 ml in phases A and B, respectively). Counts of total bacteria and protozoa were similar among treatments in both phases, but values were greater (P<0.05) in phase B versus phase A. In summary, CT from different sources had a disparate influence on N digestion, but similar effects on ruminal microbial CH₄ emission by goats, possibly by altering activity of ruminal methanogenic bacteria though change in actions of other bacteria and/or protozoa may also be involved.     

Methane emission by goats consuming diets with different levels of condensed tannins from lespedeza

Twenty-four yearling Boer × Spanish wethers (7/8 Boer; initial body weight (BW) of 34.1 ± 1.02 kg) were used to determine effects on methane (CH₄) emission of dietary levels of a condensed tannin (CT)-containing forage, Kobe lespedeza (Lespedeza striata; K), and a forage very low in CT, sorghum-sudangrass (Sorghum bicolor; G). Treatments were dietary K levels (dry matter (DM) basis) of 1.00, 0.67, 0.33, and 0 (100, 67, 33, and 0 K, respectively). Forages were harvested daily and fed at approximately 1.3 times maintenance metabolizable energy requirement. The experiment lasted 21 days, with most measures on the last 8 days. The CT concentration was 0.3 and 151 g/kg DM in G and K, respectively. DM intake was similar among treatments (i.e., 682, 675, 654, and 648 g/day; S.E. = 30.0) and gross energy (GE) digestibility increased linearly (P<0.05) with decreasing K (0.472, 0.522, 0.606, and 0.666 for 100, 67, 33, and 0 K, respectively). CH₄ emission changed quadratically (P<0.05) with decreasing K (10.9, 13.8, 17.6, and 26.2 l/day; 32, 42, 57, and 88 kJ/MJ GE; 69, 81, 94, and 133 kJ/MJ digestible energy for 100, 67, 33, and 0 K, respectively). In vitro CH₄ emission by incubation of ruminal fluid for 3 weeks with a medium for methanogenic bacteria and other conditions promoting activity by methanogens also was affected quadratically (P<0.05) by K level (7.0, 8.1, 9.2, and 16.1 ml for 100, 67, 33, and 0 K, respectively). The total bacterial count of ruminal samples was similar among K levels, but the number of total protozoa increased linearly (P<0.05) as K declined (8.3, 11.8, 15.6, and 27.1 × 10⁵ ml⁻¹ for 100, 67, 33, and 0 K, respectively). The CT-containing forage K decreased CH₄ emission by goats regardless of its feeding level, although the effect per unit of K increased with decreasing K. Forage type (i.e., legume versus grass) may have contributed to the effect of K on CH₄ emission, but most of the change appeared attributable to CT, which appeared to directly impact activity of methanogenic bacteria, although alterations of protozoal activity could have been involved. These findings suggest that relatively low dietary levels of CT could be employed to lessen CH₄ emission without a marked detrimental effect on other conditions such as total tract protein digestion.     

棉花植株中的单宁测定方法研究

通过 3种方法测定棉花组织中单宁含量比较表明,Folin酚还原法测定的 4个品种不同组织和不同生育期顶叶的含量显著高于正丁醇盐酸法 (即花色素反应,专门用于缩合单宁的测定)近 2倍,说明这种方法测定出的是相对总酚含量,用于表达棉花缩合单宁的含量是不合适的,而香草醛法测定结果与正丁醇盐酸法差异不显著,可用于棉花组织单宁含量的测定.在棉花各个组织中,花萼、铃皮和叶片缩合单宁含量较高,陆地棉中一般达 5%～10%;花瓣、花柱子房和铃心中含量较低 (2%左右).顶端嫩叶缩合单宁含量从苗期 (1%以下)起不断增加,至吐絮期达最高 (10%左右),表明缩合单宁含量与植物组织成熟衰老和木质化程度密切相关.     

A method to determine tannin concentration by the measurement and quantification of protein — tannin interactions

Summary Protein determinations on protein-tannin complexes after protein isolation (gel filtration and trichloroacetic acid [TCA] precipitation) or phenolic extraction (polyvinyl pyrrolidone [PVP] and organic solvent precipitation) were unsuccessuful. Kjeldahl determinations of the amount of unprecipitated protein bovine serum albumin [BSA] showed a sigmoid relationship with increasing concentrations of tannins. A similar relationship was found for the reduced viscosity of BSA and plant protein, and the concentration of tannin. Non-linear regression and curve normalization allowed three variable (k ₁, k ₂ and T _1/2) to be defined for the quantification of the protein-tannin interaction/s. Such a treatment may be useful in studies of the role of tannins in plant-herbivore interactions.     

Seasonal changes in the concentrations of plant secondary metabolites and their effects on food selection by Microtus oeconomus

Using cafeteria trials conducted during June–September 2008 in Qinghai Province, China, we investigated the selection of 20 plant species by root voles (Microtus oeconomus). It was found that both favored and edible plant groups of root voles comprised 6 species, while the remaining 8 species were anorectic plants. Three plant secondary metabolites (PSMs): flavonoids, condensed tannins, and total phenols, exhibited seasonal changes in concentration; being lowest at June and gradually increasing from July to August/September. Total phenols was the only factor included in the best model of generalized linear models, indicating that total phenols was the most important factor deterring food selection by root voles. In contrast, tannins had a weak effect on food selection by root voles. This study indicated that PSMs play an important role in food selection by root voles; however, the effects of PSMs depend on the type of PSMs. Furthermore, this finding partly verifies the hypothesis that PSMs contribute to the defense strategy of plants, significantly influencing plant selection by root voles.     

Anthocyanidin reductases from Medicago truncatula and Arabidopsis thaliana

Anthocyanidin reductase (ANR), encoded by the BANYULS gene, is a newly discovered enzyme of the flavonoid pathway involved in the biosynthesis of condensed tannins. ANR functions immediately downstream of anthocyanidin synthase to convert anthocyanidins into the corresponding 2,3-cis-flavan-3-ols. We report the biochemical properties of ANRs from the model legume Medicago truncatula (MtANR) and the model crucifer Arabidopsis thaliana (AtANR). Both enzymes have high temperature optima. MtANR uses both NADPH and NADH as reductant with slight preference for NADPH over NADH. In contrast, AtANR only uses NADPH and exhibits positive cooperativity for the co-substrate. MtANR shows preference for potential anthocyanidin substrates in the order cyanidin>pelargonidin>delphinidin, with typical Michaelis-Menten kinetics for each substrate. In contrast, AtANR exhibits the reverse preference, with substrate inhibition at high concentrations of cyanidin and pelargonidin. (+)-Catechin and (+/-)-dihydroquercetin inhibit AtANR but not MtANR, whereas quercetin inhibits both enzymes. Possible catalytic reaction sequences for ANRs are discussed.     

Relationship between intake of tannin-containing tropical tree forage,PEG supplementation,and salivary haze development in hair sheep and goats

The objective of this study was to estimate the relationship between tannin binding salivary protein (TBSP) and condensed tannins (CT) intake in hair sheep and creole goats. Foliage was obtained from trees with different levels of CT content; animals were offered foliage ad libitum, with or without polyethylene glycol (PEG). Saliva haze development (SHD) was evaluated as evidence for TBSP. PEG consumption did not affect dry matter intake (DMI) (P > 0.05). Lignin (R = −0.714, P < 0.001) and Crude Protein (CP) (R = 0.622, P < 0.001) contents had a stronger association with DMI than CT (R = 0.622, P < 0.011) in sheep; no significant association was found in goats. The positive relationship between tannin intake and SHD (P < 0.05) was not confirmed after PEG supplementation in sheep (P > 0.09), but remained significant for goats (P < 0.01), except for those fed Lysiloma latisiliquum (P = 0.07). Foliage lignin or CP contents are better predictors of foliage intake than CT. Sheep and goats fed with tropical tree forages containing different levels of tannins exhibited differences in intake behavior; moreover, individual variations in TBSP expression helps explaining foliage DMI.     

Effect of different dietary tannin extracts on lamb growth performances and meat oxidative stability: comparison between mimosa,chestnut and tara

Little information is available on the effects of different sources of tannins on ruminant product quality. Nowadays several tannin-rich extracts, produced from different plants, are available and contain tannins belonging to different chemical groups, but most of these have not been used so far as feed supplements. The present study aimed at comparing the effects of feeding three tannin extracts (one containing condensed tannins and two containing hydrolysable tannins) to lambs on growth performances and meat oxidative stability. Comisana male lambs were divided into four groups (n=9 each) and were fed for 75 days: a concentrate-based diet (CON), or CON supplemented with 4% tannin extracts from either mimosa (MI; Acacia mearnsii, De Wild; condensed tannins), chestnut (CH; Castanea sativa, Mill; hydrolysable ellagitannins) or tara (TA; Cesalpinia spinosa, (Molina) Kuntze; hydrolysable gallotannins). Only CH reduced growth rate, final weight, carcass weight and feed intake (P<0.05). Tannins did not affect the concentration of the main fatty acid classes and the peroxidability of the intramuscular fat (P>0.05). The TA diet increased (P<0.001) the concentration of γ-tocopherol in muscle and tended to increase that of α-tocopherol (P=0.058). Oxidative stability of raw and cooked meat, or of meat homogenates incubated with pro-oxidants, was not affected by the extracts. These results, compared with those reported in the literature, highlight that some effects of tannins cannot be easily generalized, but may strictly depend on their specific characteristics and on conditions inherent to the basal diet and the metabolic status of the animals.     

Palatability in sheep and in vitro nutritional value of dried and ensiled sainfoin (Onobrychis viciifolia) birdsfoot trefoil (Lotus corniculatus), and chicory (Cichorium intybus)

Abstract

Three temperate forages, sainfoin, birdsfoot trefoil, and chicory, characterized by elevated contents of plant secondary compounds, were compared to a ryegrass-clover mixture (control) in dried (Experiment 1) and ensiled form (Experiment 2) in their palatability and nutritional value. Palatability was measured in adult wethers (n = 6) allowed to choose between the familiar control forage and one of the three test plants. Palatability index was calculated from differences in intake of control and test plants measured after given times. Generally at first contact, palatability of the unfamiliar plants was low. Lag time until palatability index approached or exceeded a value of 100 was 2 – 5 d, but could not be related to the content of condensed tannins. Sainfoin had a high palatability, the highest content of condensed tannins (77.4 ± 10.23 g/kg DM), a high content of duodenally utilisable crude protein (94.7 ± 16.87 g/100 g CP), and a high content of metabolizable energy (9.5 ± 0.38 MJ ME/kg DM), making this plant most promising for various purposes including anthelmintic action.     

Effects of simulated and insect herbivory on nitrogen and protein precipitable phenolic concentrations of two legumes

Protein-precipitating polyphenolics (PPPs) serve as a plant defense against herbivory, increasing with stress. We studied how varying intensities of simulated and Melanoplus differentialis herbivory affected (1) PPP concentration; (2) protein bound by PPP (PB); and (3) N concentration of panicled tick-clover (Desmodium paniculatum; PTC) and sericea lespedeza (Lespedeza cuneata; SL) leaf regrowth. Leaves of PTC that were submitted to simulated herbivory had lesser (p?≤?.05 for all significant differences) PPP concentration than the control for most treatments. For PTC, PPP concentration decreased with increasing herbivory intensity for both herbivory types. For SL, PPP was similar between herbivory types for Harvest 1 but not for 2, decreasing as herbivory intensity increased for both herbivory types. Simulated herbivory resulted in lower PB concentrations for PTC and SL compared to the grasshopper herbivory. Nitrogen concentration was similar for PTC and SL between herbivory types but variable among degree of herbivory. Herbivory type affects PPP.     

Litter dynamics of low and high tannin sericea lespedeza plants under field conditions

Sericea lespedeza [Lespedeza cuneata (Dum. Cours) G. Don.] is a perennial legume with high polyphenol (tannin) concentration. The decomposition and nutrient release of sericea lespedeza residues with two tannin levels (high = 22% or low = 16%), which were either placed on the soil surface or buried at a depth of 5 cm were studied. An experiment was carried out at two sites for 2 yr. Litter mass loss and nutrient (N, P, K, Ca, Mg, Cu, Fe, Mn, Zn, B, Mo, Co) release were recorded. Litter mass loss was higher for the low-tannin residues and for those buried. From the beginning of the experiments, the C : N ratio was lower than 25 : 1 and N was mineralized for most of the studied cases; this means that N increased in availability to plants. The release of the rest of the studied nutrients was higher at the depth of 5 cm than on the soil surface and higher for low-tannin residues than for high-tannin. The observed accumulation of some nutrients in the first months presumably depended on microbial activity. The rate of release was higher for the macronutrients than for the micronutrients, with the exception of B, which had the same behaviour as the macronutrients. This revised version was published online in June 2006 with corrections to the Cover Date.     

Effects of persimmon tannin on survival and reproduction of Halyomorpha halys (Hemiptera: Pentatomidae)

Halyomorpha halys (Hemiptera: Pentatomidae) occurs in sweet persimmon orchards from late July to early September and its damage level is negligible before mid‐July. It rarely appears in persimmon orchards after mid‐September, but is easily found in soybean fields. From the phenological observation of H. halys, it is assumed that seasonal migration of H. halys to sweet persimmon orchards may be related to the content of soluble tannin in sweet persimmon fruits. The feeding preference of H. halys was evaluated on sweet and astringent persimmon fruits, and the effects of persimmon fruits and purified persimmon tannin were tested on the survival of H. halys. Numbers of visiting bugs and feeding spots on the fruits were significantly higher on sweet persimmon than those on astringent persimmon. There was no significant difference in survivorships of H. halys fed sweet persimmon and soybean. However, the number of eggs laid by sweet persimmon fed adult females was significantly lower than those fed soybean. The higher concentration of persimmon tannin caused higher mortality of H. halys. Male and female H. halys fed with 3% persimmon tannin showed the lowest survival. Longevities of male H. halys fed with 0, 0.01, 0.1 and 1% persimmon tannin were 14.9, 12.7, 14.5 and 16.3 days, respectively. However, males fed with 3% tannin lived only 6.4 days. The longevities of female H. halys were similar to those of male (10.1, 14.4, 13.8, 12.0 and 6.1 days for 0, 0.01, 0.1, 1 and 3% tannin concentration, respectively). These results suggest that H. halys invades sweet persimmon orchards during late July and early September when more suitable foods, e.g. soybean, are in short supply and sweet persimmon fruits become suitable for feeding (e.g. lower tannin content).     

Allocation to defense or growth in dipterocarp forest seedlings in Borneo

We quantified the allocation of net production to plant secondary metabolites (especially condensed tannins and lignins) to evaluate the investment into defense mechanisms of three tropical forest canopy species in Borneo vis-à-vis the resource availability hypothesis. In particular, we focused on Borneo ironwood (Eusideroxylon zwageri, Lauraceae), which seemed to employ an extreme defensive strategy. The wood of this species is extremely durable and has a high specific gravity with a very slow growth rate. The allocation to defense by Borneo ironwood was compared to two emergent species of Dipterocarpaceae, the dominant family in this forest community. We conducted shade-house experiments on seedlings under four controlled conditions (two light levels × two nutrient levels) and showed that the growth rate of E. zwageri was much lower than those of the other two species, and it allocated more of its net production to leaves and roots than to stems. The concentrations of condensed tannins and lignins were very high in the leaves and stems of this species, at about 20 and 30%, respectively. In total, E. zwageri allocated a maximum of about 35% of its net production to defensive substances (i.e., condensed tannins + lignins). In contrast, the two dipterocarp species allocated about 20–25% of their net production to defensive substances. The condensed tannins in E. zwageri help to prolong the lifespan of the leaves, and the lignins in the stems enhance the durability of the wood. Thus, although E. zwageri grows very slowly, the allocation to defensive substances seems to be an advantageous strategy for survival under dark conditions.An erratum to this article can be found at