Trees (2004) 18: 608–613 DOI 10.1007/s00468-004-0354-7
ORIGINA L ARTI CLE
David Sommerville . Robert Bradley . Daniel Mailly
Leaf litter quality and decomposition rates of yellow birch and sugar maple seedlings grown in mono-culture and mixed-culture pots at three soil fertility levels Received: 28 January 2004 / Accepted: 1 June 2004 / Published online: 30 July 2004 # Springer-Verlag 2004
Abstract Seedlings of yellow birch (Betula alleghaniensis Britton) and sugar maple (Acer saccharum Marsh.) were grown for 2 years in mono-culture and mixed-culture and at three fertility levels. Following the second growing season, senescent leaves were analysed for N concentration, acid hydrolysable substances (AHS), and nonhydrolysable remains (NHR). A litter sub-sample was then inoculated with indigenous soil microflora, incubated 14 weeks, and mass loss was measured. Litter-N was significantly higher at medium than at poor fertility, as well as in yellow birch than in sugar maple litter. The species effect on litter-N increased with increasing fertility. At medium fertility, litter-N of sugar maple litter was lower in mixed-culture than in mono-culture. AHS, NHR as well the NHR/N ratio were significantly higher in yellow birch than in sugar maple litter. At medium fertility, the NHR/N ratio of sugar maple litter was significantly lower in mono-culture than in mixed-culture. Mass loss was significantly greater at medium and rich fertility than at poor fertility, and in yellow birch than in sugar maple litter. At poor fertility, mixed-litter decomposed at a rate comparable to yellow birch, whereas at medium and rich fertility, mixed-litter decomposed at a rate comparable to sugar maple. There was a significant positive relationship between litter-N and mass loss. A similar positive D. Sommerville Department of Natural Resource Sciences, Macdonald College of McGill University, 21 111 Lakeshore Road, Ste-Anne de Bellevue, Québec, PQ H9X 3V9, Canada R. Bradley (*) Département de biologie, Centre de recherche en biologie forestière (CRBF), Université de Sherbrooke, Sherbrooke, Québec, J1K 2R1, Canada e-mail:
[email protected] Tel.: +1-819-8218000 Fax: +1-819-8218049 D. Mailly Direction de la recherche forestirè, 2700 rue Einstein, Ste-Foy, Québec, G1P 3W8, Canada
relationship between NHR and mass loss was presumed to be a “species” effect on decomposition. Results support the hypothesis that species × fertility and species × mixture interactions can be important determinants of litter quality and, by implication, of site nutrient cycling. Keywords Acer saccharum . Betula alleghaniensis . Decomposition . Litter quality . Mixed species
Introduction The competitive success of tree species during forest succession depends on their ability to cope with environmental constraints, their efficiency to acquire and use resources, and how they modify resource supply for themselves and their competitors. A key mechanism by which tree species can control nutrient supply is through the chemical quality of their litter, which determines decomposition rates. The natural forests of southeastern Québec’s temperate region typically include hardwood species, such as sugar maple (Acer saccharum Marsh.) and yellow birch (Betula alleghaniensis Britton), that may occur in finely mixed assemblages (i.e. high alpha diversity). Soil nutrient supply depends, therefore, on the chemical quality of each litter type in each annual litter cohort. Inter-specific differences in the chemical quality and decomposition rate of single litters have been abundantly studied (e.g. Trofymow et al. 1995), but there is some evidence that decomposition rates of mixed litters may not be additive (Rothe and Binkley 2001). The empirical evidence on the nutritional interactions of tree species grown in mixtures have focused exclusively, however, on comparing the decomposition rate of litter mixtures to that of pure litters. To our knowledge, no study has compared leaf litter quality and decomposition rates of a given species growing in mixed-culture to that of the same species growing in mono-culture. Yet, it is well established that inter-specific rhizosphere interactions may be antagonistic (Yamasaki et al. 1998) or beneficial (Malcolm 1987;
609
Bradley and Fyles 1996) in terms of nutrient acquisition, which could affect litter quality and decomposition rates. Numerous studies have derived indices of litter quality to predict decomposition rates of individual litter types and their subsequent effect on nutrient cycling and site productivity. Melillo et al. (1982) were among the first to report an inverse relationship between lignin/N ratios and decay rates, an observation confirmed by subsequent studies (e.g. Moore et al. 1999). The interpretation of this relationship is that lignin, measured as the nonhydrolysable litter fraction, represents a C pool which is chemically recalcitrant to decomposition, whereas N represents an important microbial nutrient which may enhance metabolic efficiency of decomposers. Other studies have reported independent relationships between lignin content (Austin and Vitousek 2000), or N content (Taylor et al. 1989), and litter mass loss. Yet another chemical characteristic of litter that may control decomposition rates is its “holocellulose” (i.e. acid hydrolysable carbohydrate) content, which presumably represents the principal available C pool to decomposing microorganisms (Berg et al. 1982; Melillo et al. 1989). We report on a experiment in which we compared concentrations of N, of holocellulose, and of lignin, as well as the initial rate of mass loss, of leaf litter produced by seedlings of two companion species, sugar maple and yellow birch, grown in mono-culture and mixed-culture pots. Seedlings were also grown at three fixed fertility levels, as there is evidence that rhizosphere interactions are dependent on soil nutrient status (Bradley et al. 1997). Our objective was to provide evidence supporting the hypothesis that inter-specific rhizosphere interactions occurring in mixedwood stands could affect leaf litter quality and decomposition rates.
Materials and methods Treatments Mineral soil was collected in early May, 2000, from an 80-year-old sugar maple-yellow birch dominated stand near the city of Sherbrooke, Québec (45°26′N, 71°41′W). The location has been classified by foresters as a medium quality site based on stand characteristics. The dominant soil order is classified as Orthic Dystric Brunisol, according to the Canadian System of Soil Classification (Soil Classification Working Group 1998). Approximately 200 kg of Bm1 mineral soil horizon material (0–15 cm) were collected over a 0.25 ha area, coarse sieved, bulked and brought back for potting at the Agriculture and Agri-Food Canada Research Station in Lennoxville, Québec. The experimental soil was amended so as to provide three fertility levels. The “poor” fertility level consisted of one part soil mixed with three parts acid-washed silica sand. The “medium” fertility level consisted of nonamended soil. The “rich” fertility level consisted of soil amended with the fertiliser equivalent of 300 kg ha−1 year−1 (i.e. assuming a 20 cm deep rooting zone) of N/P/K fertiliser (20:20:20), which also contained small amounts of Ca, Mg and S as well as micronutrients. The soil of each fertility level was assigned to nine 8 l pots (ca. 6 kg dry wt equiv. per pot). In the rich fertility level, one-third of the fertiliser was applied at the beginning of the growing season (mid-May), one third in late-June, and one third in early August.
Sugar maple and/or yellow birch seedlings were planted in each pot and grown for two complete growing seasons (summers 2000 and 2001). The seedlings were winter-hardened bare-root stock (1–2 years old) obtained in spring 2000 from the Ministère des ressources naturelles, at their Berthierville nursery. Only seedlings weighing 20 ±1 g were selected for the assay. Within each fertility level, three pots were each assigned two sugar maple seedlings, three pots were each assigned two yellow birch seedlings, and three pots were each assigned one seedling of each species. The 27 pots were randomised in three complete blocks inside a greenhouse with a retractable roof. Each block contained a complete set of all nine treatments shrouded by shade-cloth intercepting 80% of incident light. The shade-cloth simulated a light level similar to that found in the sub-canopy of deciduous stands of southern Québec and northeastern USA (Beaudet et al. 2000; Finzi and Canham 2000). Seedlings were watered daily during the growing season to avoid moisture stress. Water was applied over the soil surface and below the foliage, to avoid leaching of foliar nutrients into the soil. In November 2000, following leaf litterfall, all potted seedlings were transported to the University of Sherbrooke and placed in a dark cold-room set at 2°C. Seedlings were returned to the greenhouse the following spring (May 2001) where they were grown and fertilised in a similar manner as the previous year.
Litter-N and proximate C fractions During the months of October and November, senesced leaves fallen around each pot, or still dangling from branches, were hand-picked on a daily basis and sorted according to the pot from which they were collected. In mixed-culture pots, leaf litter was further sorted according to species. Leaves were dried in a forced-air oven and ground to pass through a 0.1 mm mesh. A sub-sample from each of the 36 ground litter samples was oven-dried at 101°C, weighed, burned in a muffle furnace at 550°C, and re-weighed to calculate its ash content. A second sub-sample was analysed for acid hydrolysable substances (AHS) and nonhydrolysable remains (NHR) according to the sequential extraction method described by Ryan et al. 1990. A third subsample was digested by the microKjeldahl method, and the digests were analysed colorimetrically (salicylate–nitroprusside) for NH4+-N. For each litter sample, we computed the NHR/N ratio (i.e. lignin/N ratio).
Initial decomposition rates A 250 mg (dry weight) subsample of litter material from each monoculture pot, and a mixture of 125 mg litter material of each species from mixed-culture pots, were weighed (±10 μg) in 10 ml glass vials. Each vial was inoculated with 750 μl aqueous solution containing the original soil’s microflora that had been obtained by cold centrifugation [i.e. 3 g soil + 30 glass beads (4 mm) + 30 ml of 1% Na-pyrophosphate solution; the mixture shaken for 5 min and centrifuged for 5 min at 1,000g and 4°C; the suspension diluted 1:100 with 0.85% NaCl solution]. The vials were covered with a polyethylene film to prevent desiccation and allow gas exchange, and were incubated at 25°C for 14 weeks. Every 15 days, vials were weighed and gravimetric moisture content was adjusted to about 300%. Following the incubation, polyethylene films were removed, vials were dried for 24 h in an air-draft oven (101°C) and weighed. The initial decomposition rate of litter samples was calculated as the percent mass loss following the 14 week incubation.
Data analysis In the analysis of litter chemistry, senescent leaves in mixed-culture pots were separated according to species. Thus, treatments comprised a factorial array of three experimental factors, namely fertility (poor, medium, and rich), species (yellow birch, and sugar
610 maple) and mixture (mono-culture, and mixed-culture), with a total sample size of n=36. In the decomposition trial, litters from each species in the mixed-culture pots were pooled prior to the assay such that data were analysed according to only two experimental factors, namely fertility and litter-type (yellow birch, sugar maple, and mixture), with a total sample size of n=27. The chemical quality of mixed-litter samples was calculated as the average chemical quality value of the two litter sub-samples. Data were analysed by three-way, two-way and one-way analysis of variance (ANOVA) using general linear models, and means were compared using Tukey’s studentized range test (TSRT). Because the low number of replicates (n=3) available in this study implies a low power of the ANOVA and TSRT tests, we considered α=0.10 (rather than the conventional α=0.05) to be significant and, in so doing, accepted a greater risk of committing type-I errors (i.e. rejecting H0 when H0 is true). Simple linear regression was used to test relationships between decomposition rate (dependent variable) and each of the four independent variables describing litter chemistry (i.e. litter-N, AHS, NHR, and NHR/N ratio). All analyses were performed using SAS statistical software (SAS 1998).
Results Litter quality Three-way ANOVA revealed that both fertility (P=0.01) and species (P=0.06) had a significant effect on litter-N concentrations. More specifically, litter-N concentrations were significantly higher at medium fertility than at poor fertility, and significantly higher in yellow birch than in sugar maple leaf litter. There was, however, a significant fertility × species interaction (P=0.05) controlling litter-N, which required that we perform two-way ANOVA tests within each fertility level. Subsequently, two-way ANOVAs (Table 1) and TSRTs showed that the significant species effect was confined mainly within the rich fertility level. Two-way ANOVA and TSRT also showed a significant effect of mixture (P=0.07), and of species × mixture interaction (P=0.08), controlling litter-N within Fig. 1 Mean values of litter-N, acid hydrolysable substances (AHS), nonhydrolysable remains (NHR), and of the NHR/ N ratio for each species × mixture combination within each fertility level. Means within the same fertility level bearing different lower-case letters differ significantly (PF
F
P>F
0.35 0.64 1.23 8.84 0.25 1.17 10.72 0.61 2.36 7.75 0.62
