Ecology. 71(5), 1990, pp. 1976-1985 © 1990 by the Ecological Society of America

DECAY RATES, NITROGEN FLUXES, AND DECOMPOSER COMMUNITIES OF SINGLE- AND MIXED-SPECIES FOLIAR LITTER1 V JOHN M. BiAm,2 ROBERT W. PARMELEE,3 AND MICHAEL H. BEARE Department of Entomology and Institute of Ecology, University of Georgia, Athens, Georgia 30602 USA Abstract. Decomposition rates, N fluxes, and abundances of decomposer organisms were quantified in mixed-species litterbags (containing leaves of two or three of the following tree species: Acer rubrum, Cornus florida, and Quercus prinus) and in litterbags containing leaves of a single species. Data from single-species litterbags were used to generate predicted decay rates, N fluxes, and abundances of decomposer organisms for mixed-species litterbags, against which observed values could be compared to determine if significant interaction effects occurred when litter of different species, and different resource quality, was mixed. Decay rates of-mixed-species litterbags during the 1-yr study were not significantly different than predicted from decay rates of individual component species. However, there were significant interaction effects on N fluxes and abundances of decomposer organisms. In the C. florida-A. rubrum and C. florida-A. rubrum-Q. prinus litter combinations there were significantly greater initial releases of N and lower subsequent N immobilization than predicted. In the A. rubrum-Q. prinus and C. florida-A. rubrum-Q. prinus litter combinations, lengths of fungal hyphae were significantly less than predicted on at least half the collection dates. Bacterial numbers in the mixed-litter combinations were also generally less than predicted. Nematode abundances, especially fungivores, were generally greater than predicted in mixed-species litterbags until the last sample date. Observed mean abundances of nematodes over all dates were 20-30% greater than predicted. Microarthropod abundances were more variable, but tended to be lower than predicted. Our results indicate that measurement of N flux in single-species litterbags may not reflect actual N flux in the field, where leaves of several tree species are mixed together. The differences in N flux between single- and mixed-species litterbags can affect ecosystem-level estimates of N release or accumulation in decomposing litter. For example, estimates of ecosystem-level N fluxes at our field site, based on data from single-species litterbags, resulted in a 64% underestimate of N released by day 75 and a 183% overestimate of N accumulated in the litter by day 375, relative to estimates based on data from mixed-species litterbags. We suggest that the deviation of observed N fluxes in mixed-species litterbags from those predicted using single-species litterbags are the result of differences in the decomposer community, such as lower microbial and microarthropod densities and higher nematode densities, resulting when litter of varied resource quality is mixed together. Longer term studies will be needed to determine if the differences between observed and predicted decomposer communities in mixed-species litter combinations influence the latter stages of decomposition where invertebrate-microbial interactions may have a greater effect on decay rates and nutrient release. Key words: Acer rubrum; Cornus florida; decomposer community; decomposition; leaf litter; microarthropods; mixed-species litter; nematodes; nitrogen release; Quercus prinus; resource quality. INTRODUCTION Litter decomposition is strongly influenced by macro- and microclimatic variables, biotic activity, and resource quality. Within a site, resource quality may be the most important factor affecting microbial and 1

Manuscript received 20 March 1989; revised 8 January 1990; accepted 16 January 1990. 2 Present address: Department of Entomology, Ohio State University, Columbus, Ohio 43210 USA. 3 Present address: Center for Coastal and Environmental Studies, Rutgers University, New Brunswick, New Jersey 08903 USA.

micro- and meso-invertebrate abundances, with environmental (microclimatic) effects being secondary (Swift et al. 1979). Resource quality, by affecting the abundance, composition, and activity of the decomposer community, is a major factor controlling rates of organic matter decomposition and nutrient release in forest ecosystems. Most studies of resource quality and litter decomposition have examined the effect of some measure of resource quality on decay rates and/ or nutrient fluxes of litter of single plant species (Fogel and Cromack 1977, Berg and Staff 1980, Melillo et al. 1982, Blair 1988a, b). There are few studies of the effects of increased resource heterogeneity, such as that

October 1990

DECOMPOSITION OF MIXED-SPECIES LITTER

which occurs when litter of several tree species is mixed, on decomposer organisms and processes. Except in monoculture plantations, litter of more than one tree species is usually mixed together on the forest floor. There have been suggestions that mixing litter of different species, and different resource quality, should affect decay rates and nutrient fluxes (Seastedt 1984, Chapman et al. 1988). For example, Klemmedson (1987) proposed that rates of litter decay and nutrient release in Ponderosa pine (Pinus ponderosa) forests may be accelerated with increasing co-occurrence of Gambel oak (Quercus gambelii). He suggested that these effects may be related to improved soil fertility and/or changes in forest floor microclimate induced by the presence of the oak. Carlyle and Malcolm (1986a,b) noted enhanced growth of Sitka spruce (Picea sitchenis) in stands that included larch (Larix eurolepis), which they related to greater concentrations of N and Ca and greater net N mineralization rates in the forest floors of the mixed-species stands. Maximum differences between pure and mixed stands occurred in the 0-3 cm layer, further suggesting the influence of larch litter on decomposition and nutrient release processes in the mixed-species stands. Seastedt (1984) hypothesized that nutrient release from rapidly decaying litter types could stimulate the decomposition of adjacent recalcitrant litter types. This appears to be supported by Chapman et al. (1988), who observed greater than expected nutrient availability, forest floor leaching of N and P, litter heterotrophic activity (respiration), and numbers of selected invertebrate groups in spruce/pine mixtures, relative to monoculture spruce stands. They hypothesized that translocation of nutrients among litters of different quality may result in a more rapid and efficient utilization of litter substrates by decomposers, and that these effects are mediated by the response of litter invertebrates and microflora to increased resource heterogeneity. Conversely, inhibitory compounds such as phenolics and tannins (Harrison 1971, Dix 1979, Swift et al. 1979) could slow decomposition of some litter combinations. This may account for the lower than expected nutrient availability, respiration, and faunal numbers in spruce/alder and spruce/oak forest stands in the above study (Chapman et al. 1988). There have been few attempts to examine quantitatively the effects of mixing litter of different plant species, and hence different resource quality, on subsequent decay rates, nutrient release, and composition of decomposer organisms. In particular, there have been no quantitative comparisons of decay rates, or numbers of decomposer organisms, of well-defined mixedspecies litter assemblages with those of litter of the individual component species. This has precluded determining if there are any interactive effects of mixing litter of different species, which could be important in affecting decomposition processes in the field. For example, Thomas (1968) recovered greater numbers of

1977

microarthropods from litterbags containing both pine needles (Pinus taeda) and dogwood leaves (Cornusflorida), than from litterbags containing pine needles alone. However, he could not determine if the numbers of arthropods recovered from mixed litterbags were significantly different than would be expected from the added dogwood litter itself, because of the lack of dogwood-only litterbags. Thomas reported no effect of dogwood leaves on the Ist-yr decomposition rate of pine needles in that study. Kelly and Beauchamp (1987) investigated decay rates and nutrient dynamics in oak and mesic mixed-species litterbags, but did not compare values obtained using mixed species of litter with those from single-species litterbags. Our objectives were to compare the decay rates, nitrogen dynamics, and decomposer communities of single-species litterbags (containing only litter of one of three tree species) with those of mixed-species litterbags (containing litter of two or three of the tree species in equal proportion). Knowing the initial amount of each litter type comprising the mixed-species litterbags enabled us to use data from the single-species litterbags to calculate predicted decay rates, nitrogen fluxes, and abundances of decomposer organisms for the mixedspecies litterbags. Differences between predicted and observed values for mixed Utter could then be used to infer interaction effects due to mixing litter from different tree species. This is important because any interaction effects on decay rates or nutrient release would affect predictions of mass loss or nutrient release derived from single-species litterbags. Additionally, such interaction effects would provide a potential mechanism for the observed increase in forest floor nutrient availability reported in some comparisons of singleand mixed-species forest stands (i.e., Chapman et al. 1988). METHODS Newly senesced leaves of three tree species (Cornus florida L., flowering dogwood, Acer rubrum L., red maple, and Quercusprinus L., chestnut oak) were collected from the forest floor of a mixed-hardwood watershed (WS 2) at the Coweeta Hydrologic Laboratory, southwestern North Carolina, in mid-October 1986. These three species were chosen to represent a range of initial resource qualities and decay rates (Cromack and Monk 1975, Blair 1988a). Litterbags were filled with air-dried foliar Utter of either a single tree species or one of the following combinations of two or three species: C.florida-A. rubrum, A. rubrum-Q. prinus, and A. rubrumC. florida-Q. prinus. Litterbags were 10 x 10 cm fiberglass bags with a 1.5-mm mesh and initially contained 3 g of air-dried Utter. Mixed-species Utterbags contained equal portions of the individual species comprising the bags. We used equal amounts of Utter of component species, rather than proportional amounts based on litter inputs in the field, since our objectives were to examine the effects of mixing litter of different

1978

JOHN M. BLAIR ET AL.

species, in general, and because actual proportions of species being mixed would vary among sites, on a macroscale, and within sites on a microscale. Subsamples of litter were oven-dried at 50°C to develop conversion factors from air-dry to oven-dry. Eighteen litterbags of each species or species combination were placed in each of three replicate plots spaced «150 m apart at the same elevation on the south-facing slope of an undisturbed deciduous forest (WS 2) at the Coweeta Laboratory on 18 November 1986. Three litterbags of each type were retrieved from each replicate plot after 25,75,140,238, 325, and 378 d in the field. Microarthropods were extracted from one set of litterbags (n = 3 per litter type) using modified Tullgren funnels. Microarthropods were sorted into Collembola, suborders of Acari (Prostigmata, Mesostigmata, and Oribatei) and "other" arthropods. Following extraction, the litter was oven-dried (50°) and reweighed to determine mass loss. Subsamples of ground litter were analyzed for percent ash-free dry mass (4 h at 500°), which was used to correct Utter masses for soil infiltration (Blair 1988a). Annual decay rates were calculated using a single negative exponential decay model (Olson 1963) by regressing the natural logarithm of mean percent mass remaining vs. time. Subsamples of ground litter from these litterbags also were analyzed for N content on a Tecator flow-injection autoanalyzer following micro-Kjeldahl digestion. Mass loss and N concentration data were used to calculate changes in the absolute amount of N (net immobilization or release). Nematodes were extracted from a second set of litterbags (n = 3 per litter type) using Baermann funnels. Nematodes were separated into trophic groups based on known feeding habits or esophageal morphology (Parmelee and Alston 1986). Major trophic groups used were bacterivore, fungivore, and omnivore-predator. A third set of litterbags (n = 3 per litter type) was used to estimate numbers of total bacteria and lengths of fungal hyphae in the Utter using the methods described in Beare et al. (1989). The entire contents of the Utterbag were first coarsely chopped with scissors and mixed. A representative subsample of Utter was removed, weighed field moist, and homogenized in 60 mL of sterile water. FoUowing appropriate dilutions, bacterial cells were stained with FTTC (Babiuk and Paul 1970) and enumerated using a Zeiss IM35 photomicroscope with attached epifluorescent iUuminator. Lengths of total fungal hyphae in the Utter homogenates were measured using an agar film technique (Jones and Mollison 1948), and FDAactive fungal hyphae were measured using the methods of Soderstrom (1977) a*nd Ingham and Klein (1984). A separate subsample from each Utterbag was used to determine wet-to-dry mass conversion factors. Percent mass remaining and N concentration data from single-species Utterbags in each repUcate plot were combined to calculate predicted percent mass and N remaining for mixed-species Utterbags in individual

Ecology, Vol. 71, No. 5

plots. The values from individual plots then were averaged by date (n = 3 per date) to calculate a mean and standard error of predicted percent mass or N remaining. For example, mean predicted percent mass remaining for the C. florida-A. rubrum combination on a given sample date was calculated as: [(Cl + Al)/2 + (C2 + A2)/2) + (C3 + A3)/2]/3, where Cl = percent mass of C. florida remaining in plot 1 single-species Utterbags; Al = percent mass of A. rubrum remaining in plot 1 single-species Utterbags; C2 = percent mass of C. florida remaining in plot 2 single-species Utterbags; and so on. Similarly, observed numbers of invertebrates and microbial decomposers in single-species Utterbags in each repUcate plot were weighted for the proportion of each litter type comprising the remaining mixed-species litter and used to calculate means and standard errors of predicted abundances of decomposer organisms (n = 3) in mixedspecies Utterbags on each sample date. For example, mean predicted number of nematodes per gram in the C. florida-A. rubrum combination on a given sample date was calculated as:

[(%C x NCI + °/oA x NAY) + (%C x NC2 + %A x NA2) x NC3 + %A x NA3)\/3, where %C = percent of remaining mixed Utter that was C. florida (calculated from percent mass remaining data for single-species Utterbags); %A = percent of remaining mixed Utter that was A. rubrum (calculated from percent mass remaining data for single-species Utterbags); NCI = number of nematodes per gram of Utter in C. florida Utterbags from plot 1 ; NA 1 = number of nematodes per gram of Utter in A. rubrum Utterbags from plot 1 ; NC2 = number of nematodes per gram of Utter in C. florida Utterbags from plot 2; and so on. Observed decay rates of mixed-species Utterbags were compared with predicted decay rates using t tests comparing regression slopes (Zar 1984). Observed percent N remaining and invertebrate and microbial abundances were compared to predicted values using paired t tests over sample dates. RESULTS First-year annual decay rate constants for single- and mixed-species Utterbags were calculated from the single negative exponential model (Table 1). As expected from previous studies (Cromack and Monk 1 975, Blair 1988a), C. florida was the fastest decomposing Utter type and Q. prinus the slowest, with A. rubrum and the mixed-Utter combinations falling between the two extremes. Predicted decay rates for the mixed-species Utter combinations were calculated based on the percent mass remaining over time of component species (obtained from single-species Utterbags) and the proportion of each litter type originally contained in the mixed-species combinations. The observed decay rates

October 1990

1979

DECOMPOSITION OF MIXED-SPECIES LITTER

TABLE 1. Annual decay rates, k(n = 7), for single- and mixed-species litterbags calculated from a single negative exponential model, and measured (n = 3) percent mass remaining at the end of the study (378 d). Values are means ± 1 SE. Coefficients of determination (r2) are presented to indicate goodness of fit of the data to the model. Litter type Cornus florida Acer rubrum Quercus prinus C.florida-A. rubrum A. rubrum-Q. prinus C. florida-A. rubrum-Q. prinus

-.694 -.477 -.329 -.617 -.384 -.506

r2 .963 .783 .958 .955 .932 .942

k ± 0.061 ±0.1 12 ± 0.031 ± 0.060 ± 0.046 ± 0.056

% mass remaining 43.8 ± 0.8 52.6 ± 2.9 S 67.8 ± 3.4 46.9 ± 1.6 62.4 ± 0.6 55.6 ± 1.1

- •', r ?"*,* *•

TABLE 2. Comparison of observed decay rates of mixed-species litterbags with predicted rates based on mass loss of component species. Annual decay rates, k (X ± SE, « = 7) calculated from a single negative exponential model are presented along with mean ± SE (n = 3) percent mass remaining at the end of the study (378 d). Observed

Predicted

% mass % mass remaining Litter type P* k remaining k Cornus florida-Acer rubrum -.617 ± 0.060 46.9 ± 1.60 -.585 ± .073 48.2 ± 1.6 >.50 A. rubrum-Quercus prinus -.384 ± 0.046 62.4 ± 0.6 -.397 ± .059 60.2 ± 2.4 >.50 C. florida-A. rubrum-Q. prinus -.506 ± 0.056 55.6 ± 1.1 -.488 ± .054 54.7 ± 1.8 >.50 1 Probability values that predicted and observed decay rates are not significantly different based on t tests comparing slopes. of mixed-species litter combinations were not significantly different than the predicted rates during the 1yr study period (Table 2). Patterns of net N flux in all litter types followed the first two phases of the three-phase curve (leaching, accumulation, and release) typical of forest litter decomposition (Berg and Staaf 1981, Blair 1988a, and others). In all litter types there was an initial release of N early in decomposition followed by net N immobilization (accumulation) through the end of the study. Comparisons of observed and predicted net N fluxes are presented in Fig. 1. The initial releases of N in the C. florida-A. rubrum and C.florida-A. rubrum-Q. prinus litter combinations were significantly greater than predicted (Fig. 1 A, C). In both cases this was followed by lower net N immobilization than expected. In the A. rubrum-Q. prinus combination, the differences between observed and predicted N fluxes were less pronounced, but the maximum amount of net N immobilization observed was also less than predicted (Fig. IB). Lengths of total fungal hyphae per gram of litter remaining generally increased throughout the study in all litterbag types. Mean lengths of fungal hyphae over

FIG. 1. Comparison of observed percent N remaining over time in mixed-species litterbags with predicted percent N remaining calculated from values for component species measured in single-species litter bags. (A) Cornus florida-Acer rubrum combination, (B) A. rubrum-Quercus prinus combination and (C) C.florida-A. rubrum-Q. prinus combination. Narrow vertical bars are ±1 SE (n = 3). Asterisks indicate significant differences between predicted and observed values (P < .05).

all dates (n = 6) in the six litterbag types were positively correlated with decay rates (Table 3). FDA-active fungal hyphal lengths are not shown, but they were 1-7% of total fungal hyphae and followed the same patterns 115

A.

0---O PREDICTED • • OBSERVED

105

95 85 e>

75

Ld o: o O

O CK

100

200 300 DAYS IN THE FIELD

400

1980

JOHN M. BLAIR ET AL.

TABLE 3. Correlation matrix for annual decay rates and mean abundances over all dates of fungi, bacteria, total nematodes and total microarthropods in all single- and mixedspecies litterbag types. Data are Pearson correlation coefficients (r) with two-tailed probability values beneath (« = 6 litterbag types). MicroDecay Bac- Nema- arthrorate Fungi teria todes pods Decay rate 1.000 0.953 0.983 0.967 0.180 0.000 0.003 0.001 0.002 0.733 Fungi 1.000 0.943 0.967 0.065 0.000 0.005 0.002 0.903 Bacteria 1.000 0.916 0.306 0.000 0.011 0.556 Nematodes 1.000 -0.033 0.000 0.951 Microarthropods 1.000 0.000

at Ul t _i

"2 o Ul D. E Ul & < m 01 o

12 A. 0---0 PREDICTED -0 • • OBSERVED 10 "2*\ B /~-««Q""- --3s^ ^$ * i 6 4 2 Ql /

6 B. , 5 P-. * .-o4 • /; ? ^^ 3 2 1 •r— nl 10 8

1600

Ecology, Vol. 71, No. 5

6 4 2

ni

* o » 1 "* ' tv. •§ ^«

c.

A >^1S\ ^SK ^i £^*=^ S' 7 100

200 300 400 DAYS IN THE FIELD FIG. 3. Comparison of observed numbers of bacteria per gram of litter hi mixed-species litterbags and predicted values based on component species. (A) Cornusflorida-Acer rubrum combination, (B) A. rubrum-Quercus prinus combination, and (Q C.florida-A. rubrum-Q. ;?rzm« combination. Narrow vertical bars are ±1 SE (n = 3). Asterisks indicate significant differences between predicted and observed values (P < .05).

over time as total hyphae. Comparisons of observed lengths of fungal hyphae in the mixed-litter combinations with those predicted from individual species indicated significantly less fungal hyphae than expected in two of the three combinations (A. rubrum-Q. prinus and C. florida-A. rubrum-Q. prinus) on at least half of the sample dates (Fig. 2B, Q. Mean numbers of bacteria over all dates in each of the six litterbag types also were highly positively correlated with decay rates (Table 3). Bacterial numbers in the litter initially increased more rapidly than did fungi (Fig. 3). Comparisons of observed numbers of 200 300 100 400 bacteria in the mixed-litter combinations with those predicted from individual species generally indicated DAYS IN THE FIELD FIG. 2. Comparison of observed lengths of fungal hyphae lower numbers of bacteria than expected on most samper gram of litter in mixed-species litterbags with predicted ple dates, although the differences were statistically sigvalues based on component species. (A) Cornusflorida-Acer nificant in only the A. rubrum-Q. prinus combination rubrum combination, (B) A. rubrum-Quercus prinus combi- (Fig. 3B). nation, and (Q C.florida-A. rubrum-Q. prinus combination. Nematode and microarthropod abundances were Note differences in scale. Narrow vertical bars are ± 1 SE (n = 3). Asterisks indicate significant differences between pre- more variable than microbial abundances. Mean abundicted and observed values (P < .05). dances over all dates (« = 6) of major trophic groups

October 1990

1981

DECOMPOSITION OF MLXED-SPECIES LITTER

TABLE 4. Densities (number per gram of litter), over all dates of nematodes by major trophic group in single- and mixedspecies litterbags. Values are X ± 1 SE (n = 6 dates). Litter type Cornus florida Acer rubrum Quercus prinus C. florida-A. rubrum A. rubrum-Q. prinus C. florida-A. rubrum-Q. prinus

Fungivore 221.5 ± 109.5 72.1 ± 36.8 26.9 ± 14.4 195.1 ± 113.8 48.5 ± 23.9 147.8 ± 94.7

Bacterivore 148.6 ± 85.4 32.7 ± 15.9 24.7 ± 11.1 109.3 ± 53.7 45.8 ± 21.1 44.5 ± 24.1

Omnivore-predator 15.4 ± 8.0 0.7 ± 0.5 0.5 ± 0.5 8.6 ± 4.2 £ 0.9 ± 0.6 3.8 ± 2.1

Total 385.4 ± 198.1 105.5 ± 52.2 52.1 ± 25.1 312.9 ± 169.8 95.3 ± 44.5 196.1 ± 118.1

of nematodes are presented in Table 4. Fungivorous nematodes were the most abundant group, and bacterivores were the next most abundant. Mean abundances over all dates of total nematodes in the six litter types were highly positively correlated with decay rates and numbers of both fungi and bacteria (Table 3). Nematode numbers in the litter remained low until the 13 July sample date (238 d) and peaked on the 8 September sample date (325 d) (Fig. 4). All three mixedlitter combinations tended to support higher numbers of nematodes than predicted until the last sample date, when nematode numbers declined to less than predicted (Fig. 4). Observed mean abundances of total nematodes over all dates in the three mixed-species litter combinations were 20-30% greater than predicted from single-species litterbags. Mean abundances, over all dates, of major microarthropod groups in all six litter types are presented in Table 5. Surprisingly, overall mean abundances of total microarthropods were not significantly correlated with decay rates or with other decomposer organisms (Table 3). All three mixedlitter combinations generally supported lower peak numbers of microarthropods than predicted (Fig. 5), although the differences between observed and predicted values were generally not significant. This may be due, in part, to the high variances associated with the small sample sizes (n = 3) in this study. Observed mean abundances of total microarthropods over all 400 300 200 100 dates in the mixed-litter combinations ranged from 9% DAYS IN THE FIELD less than predicted in the C. florida-A. rubrum-Q. priFIG. 4. Comparison of observed numbers of total nemanus combination to 42% less than expected in the A. todes per gram of litter in mixed-species litterbags and prerubrum-Q. prinus combination. dicted values based on component species. (A) Cornus florida-Acer rubrum combination, (B) A. rubrum-Quercus prinus DISCUSSION combination, and (C) C. florida-A. rubrum-Q. prinus comOur results suggest no significant interaction effects bination. Narrow vertical bars are ± 1 SE (n = 3). Note difon 1 st-yr decomposition rates when litter of these par- ferences in scale. TABLE 5. Densities (number per gram of litter), over all dates of major microarthropod groups in single- and mixed-species litterbags. Values are X ± 1 SE (n — 6 dates). Litter type Prostigmata Mesostigmata Oribatei Cornus florida 14.1 ± 3.9 2.4 ± 1.3 4.1 ± 1.9 Acer rubrum 2.2 ± 1.0 19.8 ± 8.5 13.8 ± 10.6 Quercus prinus 12.1 ± 4.8 2.6 ± 1.8 5.3 ± 3.1 C. florida-A. rubrum 3.8 ± 1.9 11.3 ± 3.1 9.0 ± 2.9 A. rubrum-Q. prinus 1.4 ±0.7 . 5.5 ± 3.5 8.6 ± 4.0 C. florida-A. rubrum-Q. prinus 1.8 ± 1.2 15.8 ± 5.0 6.5 ± 3.3 * Totals include other microarthropod groups (i.e., insects, proturans, diplurans, etc.)

Collembola 4.0 ± 2.3 4.7 ± 2.3 2.1 ± 1.4 4.2 ± 2.4 2.4 ± 1.0 2.4 ± 1.4

Totals* 25.2 ± 4.9 41.9 ± 13.2 24.6 ± 7.1 29.2 ± 5.6 18.4 ± 5.8 27.2 ± 5.3

is

1

1982

JOHN M. BLAIR ET AL.

Ecology, Vol. 71, No. 5

decay rates of lower quality litter (Crossley 1977, Seastedt and Crossley 1984). Given the differences between observed and predicted abundances of inverte60 brate and microbial decomposer organisms in mixedspecies litter, noted in this study, and the potential 40 importance of these differences in affecting subsequent decomposition, it seems that longer term studies are warranted to address potential interaction effects on the latter stages of decomposition. There were significant interaction effects on 1 st-yr C9 litter N dynamics. The initial net release of N was at UJ greater and the subsequent net N immobilization was n_ generally lower in the mixed-litter combinations than VI a was predicted from individual component species (Fig. o Q_ 1). The differences in observed patterns of net N release O at and accumulation from those predicted based on sinot gle-species litterbags can affect estimates of ecosystemlevel N dynamics in litter. As an example, we have o used species-specific litterfall data from WS 2 (Risley 1987) and N concentration and decomposition data from this study to demonstrate the effects of using data o from single-species litterbags vs. data from mixedspecies litterbags on estimates of ecosystem-level litter N dynamics (i.e., Blair 19886). Estimates of total N inputs, net N release after 75 d, and net N accumulation at 375 d based on data from single-species vs. mixedspecies litterbags are presented in Table 6. Assuming 100 200 300 400 that interaction effects in the field are of the same magnitude as those observed in our C. florida-A. rubrumDAYS IN THE FIELD FIG. 5. Comparison of observed numbers of microarthro- Q. prinus mixed-species litter combination, using data pods per gram of litter in mixed-species litterbags and pre- from single-species litterbags resulted in a 64% underdicted values based on component species. (A) Cornusflori- estimate of maximum N released during the release da-Acer rubrum combination, (B) A. rubrum-Quercus prinus phase (day 75) and a 183% overestimate of net N imcombination, and (C) C. florida-A. rubrum-Q. prinus com- mobilization at the end of 1 yr, relative to estimates bination. Narrow vertical bars are ±1 SE (« = 3). based on mixed-species litterbags. Although we recognize that the composition of mixed litter in the field ticular tree species was mixed together. These results will differ from our mixed-litter combination, these should be interpreted with caution, however, since (i) results demonstrate the potential effects of litter species the design of this experiment only allowed measure- interactions on estimates of nutrient flux in the field. ment of total mass loss from mixed-litter combinations The greater increase in-initial N release and subsequent and not mass loss of individual components in the lower net N immobilization observed in mixed-species mixture, and (ii) the; experiment covered only the 1st combinations is consistent with the observed increase yr of decomposition. It is possible, for example, that in net N mineralization and availability in spruce/pine one component species decayed faster and another slower when in combination, although there would be no net effect on total mass loss. This study may also TABLE 6. Total N inputs in foliar litter of three tree species (Acer rubrum, Comus florida, and Quercus prinus) on WS have been too short to detect potential cumulative ef2 and estimates of net N release at 75 d and net N accufects of mixing litter of different qualities, since biotic mulation at 375 d based on data from single-species litteractivity may be. of relatively greater importance in the bags vs. data from mixed-species litterbags. All values are expressed as kg/ha. latter stages of decomposition (Vossbrinck et al. 1979, Seastedt et al. 1983, Blair 1988a). Many studies have N accumunoted a decline in substrate quality as decomposition N release lation at Initial proceeds (Cromack 1973, Blair 1988a, and others). For at 75 d 375 d N input this reason extrapolation of 1 st-yr decay rates to predict 7.57 -0.41 +1.16 Single-species data longer term patterns of decomposition is generally in- Mixed-species data 7.57 -1.13 +0.41 -0.72 -0.75 appropriate (Kelly and Beauchamp 1987,Blair 1988a). Difference* It also has been suggested that invertebrate-microbial * Calculated as estimate based on mixed-species data less interactions have a proportionally greater effect on the estimate based on single-species data.

80

A. 0---0 PREDICTED • • OBSERVED

October 1990

DECOMPOSITION OF MIXED-SPECIES LITTER

mixtures relative to monoculture stands (Chapman et al. 1988). It is not possible to conclude what effect the interaction among species would have on the later net release phase (Phase III of Berg and Staaf 1981), since all litter types were still in the accumulation phase at the end of our study, but further interaction effects seem likely. There were also apparent interaction effects due to mixing litter on the abundances of microbial and faunal decomposer organisms. There were significantly fewer fungal hyphae than predicted in two of the three litter combinations (Fig. 2), and numbers of bacteria in the mixed litter combinations were generally lower than expected (Fig. 3). There also were generally more nematodes and fewer microarthropods than predicted, although the invertebrate data were much more variable than the microbial data. Differences between the observed decomposer communities in mixed-litter combinations and those predicted based on data from component species imply some interaction effects on microbial and invertebrate communities when litter of different qualities is mixed. Chapman et al. (1988) also noted an interaction effect of spruce/pine mixtures on forest floor invertebrates and total heterotrophic respiration. They found greater numbers of collembola, earthworms, enchytraeids, and nematodes in the forest floor of spruce/pine mixed stands than would be expected based on abundances in single-species spruce and pine stands. They found no significant differences in numbers of Acari between mixed and single-species stands, but did observe higher total heterotrophic respiration than expected in the spruce/pine mixture. Specifying the mechanisms responsible for the apparent interaction effects on nitrogen fluxes and the decomposer community will require further research. It is possible that the lower microbial densities observed in the mixed litter combinations were due to the presence of some inhibitory compounds in one or more of the component species or effects of differences in litter water potentials among species (i.e., Dix 1984). However, this would not explain the greater than expected numbers of microbial-feeding nematodes. Instead, we suggest that the differences in observed and predicted nitrogen dynamics are a result of changes in the decomposer community caused by increased resource heterogeneity in the mixed-species Utter assemblages. Our results indicate that mixing litter of different species, and hence of different resource quality, alters the abundances of particular groups of decomposer organisms in an interactive way that cannot be predicted based on abundances in single-species litterbags alone. We suggest that these changes in the decomposer community are responsible for affecting N fluxes in mixed-litter combinations. The invertebrate component that exhibited the greatest deviation from predicted values was the nematodes (Fig. 4). Maximum total nematode numbers in the C. jlorida-A. rubrum and C.florida-A. rubrum-Q. prinus litter combinations

1983

were 2.1 times as great as predicted and maximum numbers in the A. rubrum-Q. prinus combination were 1.8 times as great as predicted. Nematode abundances were positively correlated with microbial numbers (Table 3), and the majority of these nematodes were fungivorous and/or bacterivorous, depending on the litter combination (Table 4). Fungg} grazers have been reported to affect fungal densities and activity in field and laboratory experiments (Parkinson et al. 1979, Hanlon 1981, Santos et al. 1981, Newell 1984, and others). Increased abundances of fungivorous nematodes and Acari were also implicated in reduced fungal abundances in field studies of weed litter (Parmelee et al. 1989) and crop residue decomposition (Beare et al., 1989). Fungal-feeding nematodes, by altering fungal biomass and activity, may affect patterns of nitrogen retention and release during decomposition. For example, fungal-grazing collembola were found to increase nitrogen leaching from decomposing litter (Ineson et al. 1982). This phenomenon also has been demonstrated for bacterial-feeding nematodes in soils, where these nematodes can increase N mineralization by releasing nitrogen bound in microbial biomass and/ or stimulating further microbial activity (Clarholm et al. 1981, Yeates and Coleman 1982, Ingham et al. 1985). The greater than predicted numbers of nematodes in mixed-species litterbags may have reduced fungal densities, affected microbial activity, and subsequently, altered nitrogen dynamics in the litter. Similar effects on microbial activity and nutrient release have been reported for microarthropods. The lack of correlation between microarthropod abundances and either litter decay rates or abundances of other decomposer organisms in this study are surprising, given the abundant evidence that microarthropods play an important role in forest Utter decomposition (i.e., Crossley 1977, Seastedt 1984). For example, Seastedt and Crossley (1980) reported that microarthropod activity could stimulate either nutrient release or immobiUzation, depending on the availabiUty of the nutrient from exogenous sources and the relative abundance of microarthropods. Seastedt and Crossley (1983) also reported that Utter in field microcosms treated with naphthalene to reduce microarthropod abundances retained a greater percent of N supplied in artificial throughfall. This is consistent with our observation of generally lower microarthropod densities and greater nitrogen retention than predicted in the mixed-species Utter combinations. However, other recent studies have suggested that microarthropod activity may stimulate immobilization of exogenous nitrogen, at least in the early stages of decomposition (J. M. Blair et al., personal observation). The high variability of microarthropod abundances and lack of correlation with decay rates and abundances of microorganisms in our present study precludes our making any definite conclusions regarding the effects of mixing litter of different species on microarthropod-microbial interactions.

r-

1984

JOHN M. BLAIR ET AL.

In summary, our results indicate that there are significant interaction effects on abundances of decomposer organisms and amounts of net N released and accumulated during decomposition when litter of these tree species are mixed together. Our results generally support the conclusions of Chapman et al. (1988) that (i) nutrient release from litter of a particular tree species may be influenced by the presence of litter from other species, (ii) that these effects are mediated by invertebrate-microbial interactions, and (iii) that measurement of nutrient release from single-species litterbags may not reflect actual nutrient release in situations where litter of different species are mixed together. Therefore, using N release or accumulation data from single-species litterbag studies to estimate ecosystemlevel N fluxes in the field is not appropriate for mixedspecies forest stands. While we did not observe any significant interaction effects on 1st yr decay rates in our mixed-species litter combinations, the observed differences in decomposer organisms and N fluxes in single- and mixed-species litterbags suggest that interaction effects may become important in the latter stages of decomposition, and we recommend that longer term studies be carried out to address this possibility.

Ecology, Vol. 71, No. 5

growing on a deep peat. I. Net N mineralization measured by field and laboratory incubations. Plant and Soil 93:95113. Carlyle, J. C, and D. C. Malcolm. 19866. Nitrogen availability beneath pure spruce and mixed larch + spruce stands growing on a deep peat. II. A comparison of N availability as measured by plant uptake and long-term laboratory incubations. Plant and Soil 93:115-122. Chapman, K., J. B. Whittaker, and O. W. Heal. 1988. Metabolic and fauna! activity in litters of tree mixtures compared with pure stands. Agriculture, Ecosystems and Environment 24:33-40. Clarholm, M., B. Popovic, T. Rosswall, B. Soderstrom, B. Sohlenius, H. Staaf, and A. Wiren. 1981. Biological aspects of nitrogen mineralization from a pine forest podsol incubated under different moisture and temperature conditions. Oikos 37:137-145. Cromack, K., Jr. 1973. Litter production and decomposition in a mixed hardwood watershed and in a white pine watershed at the Coweeta Hydrologic Station, North Carolina. Dissertation. University of Georgia, Athens, Georgia, USA. Cromack, K., Jr., and C. D. Monk. 1975. Litter production and decomposition in a mixed hardwood watershed and a white pine watershed. Pages 700-708 in F. G. Howell, J. B. Gentry, and M. H. Smith, editors. Mineral cycling in southeastern ecosystems. Energy Research and Development Administration Symposium Series CONF 740513. National Technical Information Service, Springfield, Virginia, USA. Crossley, D. A., Jr. 1977. The roles of terrestrial sapropha!" ACKNOWLEDGMENTS gous arthropods in forest soils: current status of concepts. We thank D. C. Coleman, D. A. Crossley, Jr., and T. R. Pages 49-56 in W. J. Mattson, editor. The role of arthroSeastedt for commenting on earlier versions of the manuscript. pods in forest ecosystems. Springer-Verlag, New York, New The manuscript was also greatly improved by the comments York, USA. of G. P. Robertson and an anonymous reviewer. Thanks to Dix, N. J. 1979. Inhibition of fungi by gallic acid in relation the U.S. Forest Service for their cooperation hi the use of the to growth on leaves and litter. Transactions of the British Coweeta facilities. This research was supported by NSF grants Mycological Society 73:329-336. BSR-8012093 and BSR-8714663 to the University of Georgia . 1984. Moisture content and water potential of abResearch Foundation. scised leaves in relation to decay. Soil Biology and Biochemistry 16:367-370. Fogel, R., and K. Cromack, Jr. 1977. Effect of habitat and LITERATURE CITED substrate quality on douglas-fir litter decomposition in Babiuk, L. A., and E. A. Paul. 1970. The use of fluorescein western Oregon. Canadian Journal of Botany 55:1632-1640. isothiocyanate in the determination of bacterial biomass of Hanlon, R. D. G. 1981. Influence of grazing by Collembola grassland soil. Canadian Journal of Microbiology 16:57on the activity of senescent fungal colonies grown on media 62. of different nutrient concentration. Oikos 36:362-367. Beare, M. H., J. M. Blair, and R. W. Parmelee. 1989. Re- Harrison, A. F. 1971. The inhibitory effect of oak leaf litter source quality and trophic responses to simulated throughtannins on the growth of fungi in relation to litter decomfall: effects on decomposition and nutrient flux in a noposition. Soil Biology and Biochemistry 3:167-172. tillage agroecosystem. Soil Biology and Biochemistry 21: Ineson, P., M. A. Leonard, and J. M. Anderson. 1982. Effect 1027-1036. of collembolan grazing upon nitrogen and cation leaching Berg, B., and H. Staaf. 1980. Decomposition rate and chemfrom decomposing leaf litter. Soil Biology and Biochemistry ical changes of Scots pine litter. II. The influence of chemical 14:601-605. composition. In 1. Perrson, editor. Structure and function Ingham, E. R., and D. A. Klein. 1984. Soil fungi: relationof northern coniferous forests-an ecosystem study. Ecologships between hyphal activity and staining with fluorescein ical Bulletins-NFR 32:373-390. diacetate. Soil Biology and Biochemistry 16:273-278. Berg, B., and H. Staaf. 1981. Leaching, accumulation and Ingham, R. E., J. A. Trofymow, E. R. Ingham, and D. C. release of nitrogen in decomposing forest litter. In F. E. Coleman. 1985. Interactions of bacteria, fungi and their Clark and T. Rosswall, editors. Terrestrial nitrogen cycles. nematode grazers: effects on nutrient cycling and plant Processes, ecosystem strategies, and management impacts. growth. Ecological Monographs 55:119-140. Ecological Bulletins-NFR 33:163-178. Jones, P. C. T, and J. E. Mollison. 1948. A technique for Blair.J.M. 1988a. Nitrogen, phosphorus and sulfur dynamthe quantitative estimation of soil microorganisms. Journal ics in decomposing deciduous leaf litter in the southern of General Microbiology 2:54-69. Appalachians. Soil Biology and Biochemistry 20:693-701. Kelly, J. M., and J. J. Beauchamp. 1987. Mass loss and . 1988b. Nutrient release from decomposing foliar nutrient changes in decomposing upland oak and mesic litter of three tree species with special reference to calcium, mixed-hardwood leaf Utter. Soil Science Society of America magnesium and potassium dynamics. Plant and Soil 110: Journal 51:1616-1622. 49-55. Klemmedson, J. O. 1987. Influence of oak in pine forests Carlyle, J. C, and D. C. Malcolm. 19860. Nitrogen availof central Arizona on selected nutrients of forest floor and ability beneath pure spruce and mixed larch + spruce stands soil. Soil Science Society of America Journal 51:1623-1628.

October 1990

DECOMPOSITION OF MIXED-SPECIES LITTER

Melillo,J. M.,J. D. Aber, andj. F. Muratore. 1982. Nitrogen and lignin control of hardwood leaf litter decomposition dynamics. Ecology 63:621-626. Newell, K. 1984. Interaction between two decomposer basidiomycetes and a collembolan under Sitka spruce: grazing and its potential effects on fungal distribution and litter decomposition. Soil Biology and Biochemistry 16:227-233. Olson, J. S. 1963. Energy storage and the balance of producers and decomposers in ecological systems. Ecology 44: 322-331. Parkinson, D., S. Visser, and J. B. Whittaker. 1979. Effects of collembolan grazing on fungal colonization of leaf litter. Soil Biology and Biochemistry 11:529-535. Parmelee, R. W., and D. G. Alston. 1986. Nematode trophic structure in conventional and no-tillage agroecosystems. Journal of Nematology 18:403-407. Parmelee, R. W., M. H. Beare, and J. M. Blair. 1989. Decomposition and nitrogen dynamics of surface weed .residues in no-tillage agroecosystems under drought conditions: influence of resource quality on the decomposer community. Soil Biology and Biochemistry 21:97-103. Risley, L. S. 1987. Acceleration of seasonal leaf fall by herbivores in the southern Appalachians. Dissertation. University of Georgia, Athens, Georgia, USA. Santos, P. F., J. Phillips, and W. G. Whitford. 1981. The role of mites and nematodes in early stages of buried litter decomposition in a desert. Ecology 62:664-669. Seastedt, T. R. 1984. The role of microarthropods in decomposition and mineralization processes. Annual Review of Entomology 29:25-46. Seastedt, T. R., and D. A. Crossley, Jr. 1980. Effects of

1985

microarthropods on the seasonal dynamics of nutrients in forest litter. Soil Biology and Biochemistry 12:337-342. Seastedt, T. R., and D. A. Crossley, Jr. 1983. Nutrients in forest litter treated with naphthalene and simulated throughfall: a field microcosm experiment. Soil Biology and Biochemistry 15:159-165. Seastedt, T. R., and D. A. Crossley, Jr. 1984. The influence of arthropods on ecosystems. BJoScience 34:157-161. Seastedt, T. R., D. A. Crossley, Jr., V. Meentemeyer, and J. B. Waide. 1983. A two-year study of leaf litter decomposition as related to macroclimatic factors and microarthropod abundance in the southern Appalachians. Holarctic Ecology 6:11-16. Soderstrom, B. 1977. Vital staining of fungi in pure culture and in soil with fluorescein diacetate. Soil Biology and Biochemistry 9:59-63. Swift, M. J., O. W. Heal, and J. M. Anderson. 1979. Decomposition .in terrestrial ecosystems. University of California Press, Berkeley, California, USA. Thomas, W. A. 1968. Decomposition of loblolly pine needles with and without addition of dogwood leaves. Ecology 49: 568-571. Vossbrinck, C. R., D. C. Coleman, and T. A. Woolley. 1979. Abiotic and biotic factors in litter decomposition in a semiarid grassland. Ecology 60:265-271. Yeates, G. W., and D. C. Coleman. 1982. Nematodes and decomposition. Pages 55-80 in D. W. Freckman, editor. Nematodes in soil ecosystems. University of Texas Press, Austin, Texas, USA. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, New Jersey, USA.