Aspen and pine leaf litter decomposition in laboratory ...

3 downloads 0 Views 835KB Size Report
TAYLOR, B. R., and PARKINSON, D. 1988. Aspen and pine leaf litter decomposition in laboratory microcosms. 11. Interactions of temperature and moisture level ...
Aspen and pine leaf litter decomposition in laboratory microcosms. H. Interactions of temperature and moisture level BARRYR. TAYLOR AND DENNISPARKINSON' Kananaskis Centre for Environmental Research, University of Calgary, Calgary, Alta., Canada 72N IN4

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

Received August 25, 1987 TAYLOR, B. R., and PARKINSON, D. 1988. Aspen and pine leaf litter decomposition in laboratory microcosms. 11. Interactions of temperature and moisture level. Can. J. Bot. 66: 1966- 1973. To explore the relative influences of substrate type, temperature, and moisture on litter decomposition rates, leaf litter of aspen (Populus tremuloides Michx.) and pine (Pinus contorta Loud. x P. banksiana Lamb.) was decomposed in laboratory microcosms for 16 weeks at 2, 10, 18, and 26°C and 15, 30, or 60 mL . week-' watering rate. Multiple regressions on mass remaining indicated temperature was a more important influence than watering rate for both species, but the degree and nature of response were different for pine compared with aspen. Decay rates of aspen were strongly affected by temperature and less so by watering rate, but pine decomposition was quite insensitive to both. For aspen. watering rate was a more important influence on decay rates at low temperatures (2 and 10°C), while for pine it was more important at high teniperatures (18 and 26°C). There was a very strong interaction of time with temperature in the determination of aspen decomposition rate, but none for pine. All these differences are attributable to the disparate chemical and physical natures of the two litter types. The time x watering rate interaction was weak for both species, and there was no temperature x watering rate interaction at all. As a consequence of these differences in response to climatic variables, aspen leaves decomposed faster than pine needles under most conditions, but under cold, dry conditions pine decomposed faster than aspen. TAYLOR, B. R., et PARKINSON, D. 1988. Aspen and pine leaf litter decomposition in laboratory microcosms. 11. Interactions of temperature and moisture level. Can. J. Bot. 66 : 1966- 1973. Pour explorer les influences relatives des types de litikre, temperature, et humidit6 sur les taux de dCcomposition de la litike, la litikre des feuilles de peuplier faux-tremble (Populus tremuloides Michx.) et de pin (Populus contorta Loud. x P. banksiuna Lamb.) fut dCcomposC dans des microcosmes de laboratoire pour 16 semaines h 2, 10, 18 ou 26°C et 15, 30 ou 60 mL - semaine-' taux d'arrosage. Des regressions multiples sur le reste de la masse ont indiquC que la tempCrature Ctait une variable plus importante que le taux d'arrosage pour les deux espkces, mais le degr@et la nature de la rCponse Ctaient different pour le pin en comparison avec le peuplier faux-tremble. Les taux de pourriture du peuplier faux-tremble Ctaient affectis fortement par la tempCrature et moins par le taux d'arrosage, mais la dCcomposition du pin Ctait trbs insensible aux deux. Pour le peuplier faux-tremble, le taux d'arrosage Ctait une influence plus importante sur les taux de pourriture au temperatures basses (2 et 10°C), tandis que pour le pin c'Ctait plus important au temperatures hautes (18 et 26°C). I1 y avait une interaction trks forte de temps avec la tempCrature pour la ddtermination du taux de dCcomposition du peuplier fauxtremble, mais aucune pour le pin. Toutes ces diffkrences peuvent &re attribuies aux natures chimiques et physiques disparates des deux types de litikre. L'interaction du temps x taux d'arrosage Ctait faible pour les deux espbces, et il n'y avait pas d'interaction de tempkrature x taux d'arrosage. En consCquence de ces diffkrences en rCponse aux variables du climat, les feuilles du peuplier faux-tremble dCcomposent plus vite que les feuilles de pin sous la plupart des conditions, mais sous les conditions froides et skches le pin dCcompose plus vite que le peuplier faux-tremble.

Introduction Decomposition is central to the normal functioning of eeosystems; it is estimated that 80-90% of all net primary production in terrestrial ecosystems is recycled by decomposers (Odum 1971). A large and diverse assemblage of decomposer organisms is dependent upon plant debris for energy, nutrients, and habitat (Swift et al. 1979). These organisms metabolize and degrade the litter enzymatically and mechanically and, along with physical actions such as leaching (Anderson 1973cr), release nutrient elements in simple chemical forms, which roots of higher plants may take up (Witkamp and Ausmus 1976). The rate and process of litter decomposition is therefore a prime determinant of nutrient cycling and ecosystem development generally (Williams 1972; Cromack and Monk 1975; Jorgensen et al. 1980). What determines how fast plant litter decays? It is generally agreed that temperature, moisture, and the physical and chemical nature of the litter ("substrate quality") are most important (Edmonds 1980). Swift et al. (1979) reviewed the literature and emphasized that consistent differences in decay rate persist between litters -

'Author to whom all correspondence should be addressed. Printed in Canada I Imprime au Canada

of different species regardless of climatic conditions, so substrate quality must b e the prime determinant of decay rate. Flanagan and Van Cleve (1983) confirmed this with extensive experiments in Alaska, as did Fogel and Cromack (19771, Howard and Howard (1980), and Edmonds (1980) in other ecosystems. On the other hand, Bunnell et al. (1977) found that temperature and moisture were more important in laboratory tests of respiration, and Mikola (1960) also claimed that climate was dominant in field studies. With respect to temperature and moisture, opinion as to which is more important is more widely spread, because the effect of either is not constant and they interact in complicated ways (e.g., Hunt 1977). The apparent ascendancy of one or the other varies according to the macroclimate of the ecosystem under consideration. The more common finding is that temperature effects dominate (Witkamp 1963, 1966a; Anderson 1973b; Phillipson et al. 1975; Edwards 1975; Reinke et al. 1981; Moore 1981; Whitford et al. 1982), but control by moisture is also frequently found (Madge 1965; Comanor and Freeman 1978; Edmonds 1979; Vogt et al. 1980; Tormala and Eloranta 1982; Rai and Srivastava 1982; Woods and Raison 1983). A third opinion is that temperature and moisture are so interdependent that their interaction is the important point, and individual effects are barely relevant (Witkamp

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

TAYLOR AND PARKINSON:

8

1966b; Reiners 1968; de Jong et al. 1974; Wildung et al. 1975; Cowling and MacLean 1981). The matter may be further confused by interactions between climatic variables and litter quality, and changes in the latter as decomposition proceeds, as well as effects of other variables (altitude, soil, aspects, animals, disturbance, season). Consequently, although decomposition has now been extensively studied in a variety of ecosystems with a variety of substrates, reliable generalizations have remained elusive. This study examined the relative importance of temperature and moisture to decay rates of aspen and pine leaf litter. This objective required that the natural covariance between temperature and moisture be disentangled, so that the effects of each could first be evaluated independently of the other. Field studies cannot resolve this issue, because weather always involves simultaneous variation in both variables; yet to be useful the results must still be applicable to the natural environment. Hence, microcosms of the forest soil were used to mimic natural forest floor conditions in the laboratory, where temperature and moisture regimes could be manipulated. In a previous study with these microcosms (Taylor and Parkinson 1988d), C 0 2 evolution was used as a measure of litter decomposition rate, but incidental measurements of mass loss suggested that mass loss would be as good a measure as C02. Measuring loss of dry mass from litter in mesh bags is widely used in field studies of decomposition (e.g., Van Cleve 1971; Suffling and Smith 1974; Lousier and Parkinson 1976) but has only occasionally been attempted in the laboratory (Witkamp and Frank 1970; Hagvar and Kjondal 1981; Day 1983; Moore 1986), perhaps because of the difficulty of maintaining natural conditions in the laboratory for a sufficient time. Earlier, it was found that forest soil microcosms, constructed of undisturbed soil, retained essential soil physicochemical structure and faunal populations for at least 6 months in the laboratory (Taylor and Parkinson 1988s). Mass loss data were complete and had low variances even in microcosms exposed to dry or cold conditions where the respiration technique was least successful, and since mass loss measures decay rate directly, it avoids the complex corrections necessary to transform gross C 0 2 evolution into litter decay rates.

Materials and methods Litter of trembling aspen (Populus tremuloides Michx.) was collected in autumn from an aspen stand in the Front Range of the Rocky Mountains, west of Calgary, Alta. (described in Crag et al. 1977, and Lousier, and Parkinson 1976). Pine needle litter was collected from a pure stand of young lodgepole -jack pine (Pinus contorta Loud. x P. banksiana Lamb.) about 45 km west of the town of Whitecourt, Alta. (Prescott and Parkinson 1985). Chemical composition and nutrient content of the leaves were measured by the methods in Taylor and Parkinson (1988e). Pine needles and aspen leaves are very similar in N content (and C:N ratio), but pine has a substantially smaller labile component and twice as much lignin as aspen (Table 1). These features have all been considered correlates of decomposition rate (Cromack and Monk 1975; Meentemeyer 1978; Schlesinger and Hasey 1981; Aber and Melillo 1982; Melillo et al. 1982). In addition, pine needles have a much lower P content and consequently a higher C:P ratio (Table I). Aspen leaves decay more rapidly in the field than pine needles. Prescott (1984) reported an annual fractional loss rate (k) based on 17 months decomposition at our collection site of 0.13; the rate for aspen at the other field site was 0.26 for the same duration (Lousier and Parkinson 1976). Hence the two species are good representatives of a fast- and a slow-decomposing litter type.

v

1967

Soil microcosms and details of experimental procedures have been described previously (Taylor and Parkinson 1988b). Samples of remoistened aspen leaves (7.00 g) or pine needles (5.00 g) were placed evenly on top of moist, undisturbed soil from the aspen or pine field sites, respectively, and incubated for 16 weeks at 2, 10, 18, or 26°C and 15, 30, or 60 mL . week-' watering rate. Every fortnight two replicates were removed from each of the 12 treatments and dry mass of litter remaining (80°C, 48 h) was determined. Original airdry masses are equivalent to 6.00 (aspen) or 4.77 g (pine) ovendry mass after corrections for air-dry moisture content and leaching loss during remoistening. Least-squares regression was used to analyze trends in mass loss of decaying leaves with temperature, moisture, and time. First, simple regressions of dry mass remaining (percent) against time (weeks) were computed for each temperature-moisture combination, after averaging duplicates to reduce variance, and with a point added at time zero (n = 9). This biased the y-intercept toward 100%; the line was not forced through 100% (as recommended by Wieder and Lang 1982) because this may induce an error when initial mass is not known exactly. because of the estimation of leaching loss and moisture content. In a comparison of the linear and negative exponential models of decay (Taylor and Parkinson 1988b), the two models produced nearly identical fits to these data. There is thus no reason to prefer the more complicated negative exponential model, so linear regression of mass loss with time have been used for these analyses. The data were checked for non-normality and for heteroscedasticity by examining plots of residuals versus estimated y values and residuals versus x values (Draper and Smith 1966). Residuals plots were also checked for outliers. Significance of regressions was tested with ANOVA (Zar 1974); in addition the Durbin- Watson test for first-order serial correlation among the residuals (Neter and Wasserman 1974) was used as a criterion for acceptance of any regression. Simple regression slopes were compared using a t-test for pairs, or a heterogeneity of slopes ANOVA (Huitema 1980) followed by Tukey's test for three or more. Forward stepwise regression (Draper and Smith 1966) was used to build successively more complex regressions of mass remaining on time, temperature, or watering rate, and their interaction, using various subsets of the full data set. Models in which serious correlations appeared between independent variables were rejected. A partial F test (Zar 1974) was used to see if variables added to any regression significantly improved the fit to the data.

Results For each of the two leaf types there are 12 simple regressions, covering the range of temperature and moisture treatments. All regressions were significant ( p < 0.81). One-way ANOVA was used to compare changes in decay rate (slopes of the linear regressions) with temperature at each moisture level, and changes with moisture level at each temperature. There are thus 4 + 3 = 7 possible comparisons for each species. Decay rates of aspen leaves were strongly affected by temperature and moisture (Table 2). There was no difference in decay rate among watering levels at 18 or 26OC ( p > 0. lo), but there was a difference at 2 and 10°C. In the latter cases it was only at the highest watering rate (60 mL . week-l) that a difference appeared (Table 2). Decay rate increased steadily with temperature at all moisture levels ( p < 0.10); at the lowest moisture level (15 mL week-') only the 26°C treatment was significantly different (Table 2). These results indicate that, over the ranges used, temperature was a more important determinant of decomposition rate than moisture. Therefore, large differences in moisture level were necessary to produce a difference in decay rate. In contrast, decay rates of pine needles were relatively

CAN. J . BQT. VQL. 66, 1988

TABLE1. Proportions of major chemical fractions and concentrations of N and P in senescent pine needles and aspen leaves

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

Component

Aspen*

TABLE3. Comparison of linear mass loss rates of pine needles and aspen leaves decomposed for 16 weeks at 2, 10, 18, or 26°C and 15, 30, or 60 mL week-' watering rate

Pine? Temp. ("C)

Labile (%) Cellulose$ (%) Lignin (%) Ash$ (%I Total N ( m g . g-') P ( m g . g-') C:N ratio C:P ratio NOTE:Values listed are rneans (SE). Aspen data are means of 1982 and 1983 levels. Analyses follow methods in Taylor and Parkinson (1988e). *n = 6; except ash, n = 20. tn = 3; except ash, n = 10. $Determined by subtraction of all other components from 100%..

Watering rate (mL - week-')

Relative rates

Probability*

Pine > aspen Pine > aspen Pine = aspen Pine = aspen Pine < aspen Pine < aspen Pine < aspen Pine < aspen Pine < aspen Pine 6 aspen Pine G aspen Pine G aspen NOTE:All regressions are significant at y < 0.01. *Significance level of t-test, df = 14.

$Determined on different salnples than organic fractions.

TABLE2. Slopes of linear regressions of mass loss (g) on time (weeks) for aspen leaves or pine needles decomposed for 16 weeks at 2, 10, 18, or 26°C and 15, 30, or 60 mL . week-' watering rate Watering (mL . weeksL)

2OC

10°C

18°C

26°C

Aspen

15 (low) 30 (medium) 60 (high)

-0.307~ -0.522cz -0.874~ - 1.771b -0.494~~ -0.650~ - 1 . 1 9 6 ~ ~-2. i 50a -0.753h - 1.000b - 1.269a -2.614a

15 (low) 30 (medium)

-0.451 -0.594 -0.667

Pine

60 (high)

-0.41 1 -0.494 -0.550

-0.468 -0.592 -0.664

-0.587 -0.625 -0.731

NOTE:All regressions are significant at p < 0.01. Estimates within one horizontal row sharing the same underscoring are not significantly different ( p > 0.10). Estimates within one vertical colulnn followed hy the same lowercase letter are not significantly different (p > 0.10). Slopes are not significantly different.

inflexible over the range of conditions tested (Table 2). No significant differences were detected among moisture levels or among temperatures (Table 2). Even if all 12 slopes are compared simultaneously, the ANOVA F ratio is not significant ( F = 0.72, df = 11, 84, p > 0.10), although there are consistent trends in the slopes (Table 2). To explore the implications of this difference in variability of decomposition rates, a series of paired t-tests was performed to compare decay rates of pine and aspen at each temperature -moisture combination. The results form an interesting pattern when arranged in order of subjectively less harsh climatic regimes (Table 3). Because of the physicochemical differences between litter of the two species, it was expected that aspen leaves would decompose more rapidly than pineneedles under all conditions. Surprisingly, this was not so. At 2°C and low or medium watering rate, pine lost weight faster than did aspen, and at 2OC and high moisture and 10°C and low moisture, there was no significant difference in rate (Table 3). Thereafter, aspen always decomposed faster

than pine, and as temperature and moisture increased, the difference became very large. Forward stepwise techniques were used to build multiple regressions of weight remaining on time and temperature or time and moisture for differently pooled data sets. Results for data pooled within each temperature treatment are shown (Table 4). Time was the first variable to enter all pine equations, but moisture levels did significantly improve the fit of the equation, except at 26OC (Table 4). However, if it is added the scedasticity of the residuals is improved. For the remaining temperatures there was a substantial increase in the proportion of the variance accounted for (R2 increase) by moisture level from 2 to 18OC, again indicating that moisture is more important at high than at low temperatures. Furthermore, the slope coefficients for moisture, which indicate the effect of watering rate when time is held constant, also increase (in absolute value) from 2 to 18OC (Table 4). Again this trend does not extend to 26OC. Quite different results were encountered for aspen leaves. Moisture entered these regressions after time and significantly improved the fit of the regression at all temperatures, including 26°C (Table 4). Increases in R2 are greatest at 2OC and decline sharply with increasing temperature (0.244 -0.027). Hence, moisture appears to be more important at low temperatures, exactly the reverse of expectations, and in marked contrast to results from needles. Moreover, there is no apparent trend in magnitude of the slope coefficients for moisture, which again contrasts with results for pine (Table 4). When data are grouped within moisture levels, the effect of temperature may be examined (data not shown). Temperature significantly contributed to pine regressions regardless of moisture level (partial F = 13.6 - 17.7, p < 0.01), but there are no trends in the improvement in R2 value (0.091 -0.120) or in slope coefficients ( -0.120 to -0.146). Thus, the importance of temperature to decay rates of pine needles is strong and equal at all moisture levels. The same is not true for aspen leaves, where there was a strong downward trend in R2 values as moisture level increased (0.347 at 15 mL . week- to 0.209 at 60 mL week- ). As watering rate increased, the relative importance of temperature to aspen decay rates declined. These results again are contrary to results from pine needles and to the

TAYLOR AND PARKINSON: V

TABLE4. Multiple regressions of mass loss (g) on time (weeks) and watering rate (mL . week- I) for pine needles and aspen leaves decoinposed for 16 weeks at 2, 10, 18, and 26°C Slopes

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

Species and temp. g°C)

Time (weeks)

Watering rate (mL . week- l)

Intercept (%'.>

R2

R2 increase*

Aspen 2 10 18 26 Pine 2 10 18 26 NOTE:Interaction terms have been excluded. *Compared to same data regressed on TIME, alone. +Increase not significant ( p > 0.05);all others signi ficant ( p < 0.01).

expectation that temperature would be more influential when moisture was not limiting. If the effect of temperature or moisture is not constant for the duration of the experiment, then interactions of time and temperature or time and moisture may appear. To explore this, we recomputed the multiple regressions, using stepwise regression, but with the potential interaction term included. If either time or moisture is already in a model to which the interaction term adds, the slope coefficients may not be reliably estimated because of interaction between the terms. The time -moisture interaction entered regressions of pine weight loss except at 26°C (Table 5), but the additional variance which it accounted for is small (1.7-0.6%) compared with the two-term regression without interaction. The interaction entered the aspen regression and there accounted for a somewhat larger part of the total variance (1.4 -7.0 %); it was more important at 2 and 10 than at 18 and 26°C. The importance of the temperature-time interaction was dramatically different between aspen and pine. There was simply no significant interaction between time and temperature in determination of pine needle decay rates, regardless of the moisture level, but the reverse is true for aspen leaves (Table 5). At all moisture levels, the decay rate of pine was a constant for a given temperature and did not change as decomposition proceeded. Here the interaction of time and temperature by itself accounted for 84% or more of total variance (Table 5). At low and medium watering rates, no other terms entered the equation; time entered at the high watering rate. The strength of this interaction is surprising: simple regressions including only the temperature -time interaction account for 7.9 - 13.5 % more of the total variance than do multiple regressions with the two terms separate. Full data sets (n = 108) were regressed against time, temperature, and watering rate to see if trends observed on subsets extended to all conditions. For both aspen and pine, time, temperature, and watering rate entered the stepwise regressions, and in that order (Table 6). All three slopes are much steeper for aspen than for pine, reflecting the more rapid decay of the former, but the relative importance of the three variables differs. Temperature is a less important term and time a more important one in the pine regressions than in aspen; this reinforces the earlier observation that decay rate of pine was rela-

tively insensitive to temperature changes. In both cases the R2 contribution of the moisture term was small and equal (Table 6). The aspen regression could be dramatically improved by replacing the temperature term with time -temperature interaction, which alone accounted for 82% of total variance (R2 of full term regression = 0.872). Interaction terms did not substantially improve the pine regression.

Discussion The outstanding feature of these results is the dissimilarity of the responses of pine needles and aspen leaves. Simple regressions of mass loss on time showed a marked effect of temperature and a lesser one of moisture on decaying aspen leaves, but no difference in pine needle decay rate was detectable over a range of 24°C. There are consistent orderings of the slopes. This illustrates the difference in substrate quality between the two litter types. Aspen is relatively easily broken down, so microorganisms are limited more by climatic factors, but pine needles are resistant both in terms of chemical composition (Table 1) and of physical structure. The thick cuticle and dense, suberized hypodermis of pine, which serve to prevent water loss during life (Esau 1977), also prevent penetration by water or microbes after death. Thus, pine decay was substrate controlled and climatic factors exerted relatively little influence. This constancy of pine decomposition rate is responsible for the surprising trend in relative rates of pine and aspen under different conditions (Table 3); for just as pine breakdown did not accelerate under benign conditions for microbial growth neither did it slow down under harsh (i.e., cold, dry) conditions. Decay of aspen leaves did respond sharply to climatic conditions, so aspen decay is much faster than that of pine at high temperatures (and moisture levels) but equal or even slower under extremely harsh conditions. These results run contrary to the assertion of Swift et al. (1979, p. 118) that each type of litter has a "characteristically different" rate of decomposition and that differences between substrates are consistent whatever the climatic conditions or the rate at which decomposition proceeds. That view is represented graphically in Fig. 1A. The y-axis represents mass loss

CAN. J . BOT. VOL. 66, 1988

TABLE5. Regressions, with interaction terms, of mass loss (g) from pine needles or aspen leaves decomposed for 16 weeks at 2, 10, 18, or 26°C and 15, 30, or 60 mL week-' watering rate (A) Regressions on time and watering rate (H20) for each temperature

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

Species and temp. ("C) Aspen 2 10 18 26 Pine 2 10 18 26

Partial Fa

First term

Second term

Time x H 2 0 Time x H 2 0 Time Time

Time Time Time x H 2 0 Time x H 2 0

Time Time Time Time

Time x H 2 0 Time x H 2 0 Time x H 2 0 H2O

R2 increaseb

R2

(B) Regressions on time and temperature for each watering rate. Slopes are given in parentheses Species and watering rate Aspen Low Medium High

Partial First term

Second term

Time x temp. (-0.0613) Time x temp. (-0.0726) Time x temp. (-0.650)

-

R2

R2 increased

0.844

0.135

-

0.896

0.131

13.1""

0.896

0.120

F"

-

Time (-0.499)

Pine Low Medium High

(No interaction) (No interaction) (No interaction)

NOTE:All regressions are significant at p < 0.01. *, significant at p 6 0.05; **, significant at p 6 0.01. 'Tests significance of addition of second variable. b~omparedwith regression on time and watering rate. 'No significant interaction. d~omparedwith regression on time and temperature.

TABLE6. Linear regressions of mass loss from decomposing pine needles and aspen leaves on time (weeks), temperature ("C), and watering rate (mL - week-') (%)

R2

R2 increase

Partial

Slope -1.133 -0.505 -0.092

99.0 106.0 109.3 99.8

0.444 0.709 0.747 0.788

0.233 0.038 0.041

95.4 15.9

98.0 99.8 101.1

0.642 0.742 0.781

0.100 0.039

Intercept Species and variable Aspen Time Temperature Watering rate arcsine transformationC Pine Time Temperature Watering rate

-0.568 -0.129 -0.039

P

Fb

-

84.7 -

-

102.6 128.8

-

190.4

40.3 18.9

123.7

-

NOTE:Interaction terms have been excluded. n = 108. 'For last term added; all significant at p 6 0.01. full model; all significant at p 6 0.01. 'To remove non-normality of residuals.

or

after some unit of time, e.g., 1 y. The x-axis represents a gradient of "environmental quality," defined as the sum of all those climatic features that contribute to growth of decomposers. Points near the origin represent cold, dry conditions, those further away represent warmer, moister conditions. For

the above hypothesis to be strictly true, lines a and b in Fig. 1A must be parallel, so that the relative difference in mass loss is maintained under all environmental conditions. Figure 1B depicts the situation realized in this experiment. Mass loss rate of aspen rises sharply with improving environmental quality,

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

TAYLOR AND PARKINSON:

-

0

Warm , moist

C o l d , dry * ENVIRONMENTAL

QUALITY "

Cold ,dry * ENVIRONMENTAL

Warm ,moist QUALITY"

FIG. 1. Graphical depiction of the response of litter decomposition rate to changes in climate and edaphic factors that favour decomposers ("environmental quality"). (A) Theoretical response, substrate u decomposes faster than substrate b under all conditions. (B) Observed response of pine needle and aspen leaf litter decomposing in the microcosms: relative rates of mass loss change with changing environmental conditions.

while that of pine is nearly constant. Consequently the difference in mass loss varies widely, and under very harsh conditions it reverses. Multiple regressions emphasize the differing behaviour of the two leaf types. For both species there can be little doubt that temperature was the more important determinant of mass loss rates than moisture content. However, the details are complicated; for pine, it appears that watering rate is even less important at cool temperatures, where temperature is limiting and evaporation is low, than at warm ones. For aspen, where watering rate exerts a relatively greater influence on mass loss rates, the effect is reversed: watering rate becomes less important as temperature increases. This could not be due entirely to saturation; excess moisture was only occasionally in evidence in the bottom layers of 2 and 10°C treatments. A closer look at these regressions reveals two distinct pairs of lines, a low-temperature pair (2 and 1O0C), where watering rate is important (accounting for 20 -24 % of total variance), and a where watering rate is high-temperature pair (18 and 26 much less important (2.7 -3.3 %). Possibly two different processes were involved. At low temperatures, where mass loss was temperature limited, the main

"h,

v

1971

effect of watering would be to leach material from the leaves. At higher temperatures leaves were drier, most of each weekly watering would be absorbed, and water would exert its influence through stimulation of microbial metabolism. Moisture contents of litter under high temperature conditions tended to be low, and the range of watering rates was not sufficient to produce consistent large differences in moisture content, so the influence of watering rate was small. Pine needles leach only reluctantly (8% of total mass in 10 days; Taylor and Parkinson 1 9 8 8 ~ )so only the second effect would matter there. The parallel result of decreasing influence of temperature on aspen mass loss rates as moisture level increased is also not shared by pine needles, which demonstrate a constant effect of temperature (accounting for about 10% of variance) regardless of watering rate. Leaching is probably also responsible for this result. Leaf leachate consists of soluble compounds, which are preferentially attacked by microbes. Hence, loss of more leachate from aspen under higher moisture levels leaves less easily metabolizable matter for microbes to degrade, and thereby reduces the effect of temperature. Pine needles, since they rarely leach, would again be immune to this effect. The interaction of time and temperature provides the most striking differentiation of aspen and pine decay processes. Whereas for pine needles temperature had the same effect on mass loss at any time in the experiment, for aspen the interaction of time and temperature was of paramount importance. This difference is almost certainly a reflection of substrate quality. In rapidly decomposing aspen leaves, there would be a marked change in chemical composition as decomposers attacked first solubles, then cellulose and related compounds, and eventually lignified tissues. The time -temperature interaction reflects the gradual change, as decay proceeds, from climatic control to substrate control of the process (Seastedt et al. 1983). Pine needles, on the other hand, are different both chemically and physically: their lignin content is high, solubles are less, and much cellulose is protected by lignified tissues (Esau 1977). In aspen leaves the waxy cuticle is the only barrier to fungal penetration, but needles are much more armoured, even in deeper layers such as the endodermis. Needles immersed in water leach at a steady rate for at least 10 days (Taylor and Parkinson 1988c), suggesting that the small pool of labile material is evenly distributed throughout the needle and penetration by water (or fungal hyphae) is slow. Edmonds (1984) found that needles of Pacific silver fir (Abies amabilis [Bougl.] Forbes) still contained 22% of the original labile material after 6 years of decomposition and 63% mass loss. Thus, decomposers attacking needles are faced with a range of labile and recalcitrant compounds and structures at all stages of decay. There is little change in substrate quality through time, and a switch from climatic to substrate control never occurs. Pine needle decay is substrate controlled from the outset. Also, pine needles lost no more than 15% of original mass during the course of this experiment, making time -temperature interactions less likely to be observed. Regressions on the full data set generally confirm earlier observations. The intriguing question is why so much less of the total variance is accounted for by the pine regression than by the aspen. It appears that some other factor that does not significantly influence aspen decomposition does influence pine. The most likely candidate is relative humidity or evaporation rate. Neither of these variables was closely controlled

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

1972

CAN. J. BOT. VOL. 66, 1988

nor were they predictable from temperature and moisture data. Humidity and evaporation rate varied largely as a function of the air exchange rate of each particular incubator and were also affected by experimental conditions. It cannot be determined from laboratory results whether evaporation rate operates as an identifiable factor in the field, but the success of predicting decay rates from actual evapotransportation (Meentemeyer 1978) strongly suggests that it does. Of course, some other factor, such as a difference in decomposer populations between soils from the aspen and pine sites could also be responsible for the unaccounted variance. Comparison of results from this experiment with those from a similar experiment measuring respiration of decomposing leaves (Taylor and Parkinson 1988d) reveals a number of similarities. The separation of respiration or decomposition rates into distinguishable low and high temperature classes was observed in the respiration experiment, and here as well, albeit less strikingly. In both experiments the boundary between the two classes was between 10 and 18"C, somewhat warmer than reported elsewhere (Waksman and Gerretson 1931; Van Cleve and Sprague 1971). In both experiments, temperature was the dominant control on decay rate, with moisture level having only a minor influence, and in both experiments it was suggested that evaporation rate may have been an important uncontrolled factor. Temperature was a decisively important determinant of final mass loss in the respiration experiment, accounting for up to 92% of total variance (Taylor and Parkinson 1988d). The greater importance of temperature and lesser importance of moisture there is to be expected since moisture levels were higher and more uniform in the former experiment. However, the two experiments together demonstrate that the dominance of temperature over moisture as a control of decomposition rates extends over a broad range of moisture contents. Only where moisture levels are extremely low (as in deserts) or extremely high (so as to produce saturation) are they likely to precede temperature in importance (Edwards 1975). In both experiments, the effects of moisture and temperature were largely independent, and their interaction was small. This is a consequence of uncoupling temperature and moisture in the experimental design and suggests that the strong interactive effects of temperature and moisture noted in field studies of respiration or mass loss (de Jong et ul. 1974; Howard and Howard 1979; Reinke et al. 1981; etc.) may actually just reflect the natural interdependence of the two factors themselves (Reiners 1968; Wildung et ul. 1975).

Acknowledgements For help with onerous field and laboratory work we thank Richard Bewley , Wade Bingle, Dan Durall , Patricia Mazier, Cindy Prescott, and Sean Sharpe. We are also grateful to Patricia Mazier for drafting the figures, and to Bill Parsons, Dr. Keith Van Cleve, and two anonymous reviewers for reviewing an earlier manuscript. Bill Parsons also did the chemical analyses of leaves. This research was funded by a Natural Sciences and Engineering Research Council of Canada (NSERC) operating grant to Dennis Parkinson. ABER,J. D., and MELILLO,J. M. 1982. Nitrogen immobilization in decaying hardwood leaf litter as a function of initial nitrogen and lignin content. Can. J. Bot. 60: 2263 -2269. ANDERSON, J. M. 1973u. The breakdown and decomposition of sweet chestnut (Custunea sutivu Mill.) and beech (Fagus sylvuticus L.)

leaf litter in two deciduous woodland soils. Oecologia, 12: 251 -274. 1973b. Carbon dioxide evolution from two temperate deciduous woodland soils. J. Appl. Ecol. 10: 361 -378. P. W., and VAN BUNNELL,F. L., TAIT, D. E. N., FLANAGAN, CLEVE,K. 1977. Microbial respiration and substrate weight loss. 11. A model of the influences of chemical composition. Soil Biol. Biochem. 9: 41-47. COMANOR, P. L., and FREEMAN, J. I. 1978. Oxygen uptake by big sagebrush litter under four temperature-moisture regimes. Soil Biol. Biochem. 10: 467-470. S. F., JR. 1981. Forest floor respiraCOWLING, J. E., and MACLEAN, tion in a black spruce taiga forest ecosystem in Alaska. Holarct. Ecol. 4: 229-237. C., PETERSON, E. B.? and CRAGG,J. B., CARTER, A., LEISCHNER, SYKES,G. N. 1977. Litter fa11 and chemical cycling in an aspen (Populus trernuloides) woodland ecosystem in the Canadian Rockies. Pedobiologia, 17: 428 -443. CROMACK, K., JR., and MONK,C. D. 1975. Litter production. decomposition and nutrient cycling in a mixed hardwood watershed and a white pine watershed. In Mineral cycling in southeastern ecosystems. Eiiited by F. G. Howell, J. B. Gentry, and M. H. Smith. ERDA Symp. Ser. CONF-7405 13. pp. 609 -624. DAY,E P., JR. 1983. Effects of flooding on leaf litter decomposition in microcosms. Oecologia, 56: 180 - 184. DE JONG,E., SCHAPPERT, H. J. V., and MACDONALD, K. B. 1974. Carbon dioxide evolution from virgin and cultivated soil as affected by management practices and climate. Can. J. Soil Sci. 54: 299307. DRAPER,N. R., and SMITH,H. 1966. Applied regression analysis. Wiley, New York. EDMONDS,R. L. 1979. Decomposition and nutrient release in Douglas-fir needle litter in relation to stand development. Can. J. For. Res. 9: 132- 140. 1980. Litter decomposition and nutrient release in Douglasfir, red alder, western hemlock, and Pacific silver fir ecosystems in western Washington. Can. J. For. Res. 10: 327-337. 1984. Long-term decomposition and nutrient dynamics in Pacific silver fir needles in western Washington. Can. J. For. Res. 14: 395 -400. EDWARDS, N. T. 1975. Effects of temperature and moisture on carbon dioxide evolution in a mixed deciduous forest floor. Soil Sci. Soc. Am. Proc. 39: 361 -365. ESAU,K. 1977. Anatomy of seed plants. 2nd ed. Wiley, New York. FLANAGAN, P. W., and VANCLEVE,K. 1983. Nutrient cycling in relation to decomposition and organic matter quality in taiga ecosystems. Can. J. For. Res. 13: 795 -817. FOGEL,R., and CROMACK, K., JR. 1977. Effect of habitat and substrate quality on Douglas-fir litter decomposition in western Oregon. Can. J. Bot. 55: 1632-1640. HAGVAR,S., and KJONDAL,B. R. 1981. Decomposition of birch leaves: dry weight loss, chemical changes and effects of artificial acid rain. Pedobiologia, 22: 232 -245. HOWARD, D. M., and HOWARD, P. J. A. 1979. Respiration of decomposing litter in relation to temperature and moisture. Microbial decomposition of tree and shrub leaf litter. 2. Oikos, 33: 457 -465. 1980. Effect of species, source of litter, type of soil, and climate on litter decomposition. Oikos, 34: 115 - 124. HUITEMA, B. E. 1980. The analysis of covariance and alternatives. Wiley, New York. HLJNT,H. W. 1977. A simulation model for decomposition in grasslands. Ecology, 58: 469 -484. JORGENSEN, J. R., WELLS,C. G., and METZ, L. J. 1980. Nutrient changes in decomposing loblolly pine forest floor. Soil Sci. Soc. Am. J. 44: 1307- 1314. LOUSIER, J. D., and PARKINSON, D. 1976. Litter decomposition in a cool temperate deciduous forest. Can. J. Bot. 54: 419-436. MADGE,D. S. 1965. Leaf fall and litter disappearance in a tropical forest. Pedobiologia, 5: 273 -288.

Can. J. Bot. Downloaded from www.nrcresearchpress.com by St. Francis X Univ on 04/30/13 For personal use only.

TAYLOR AND PARKINSON: v

MEENTEMEYER, V. 1978. Macroclimate and lignin control of litter decomposition rates. Ecology, 59: 465 -472. MELILLO, J. M., ABER.J. D., and MURATORE, J. F. 1982. Nitrogen and lignin control of hardwood leaf litter decomposition dynamics. Ecology, 63 : 62 1 - 626. MIKOLA, P. 1960. Comparative experiments in decomposition rates of forest litter in southern and northern Finland. Oikos, 11: 161 - 166. MOORE,A. M. 1986. Temperature and moisture dependence of decomposition rates of hardwood and coniferous leaf litter. Soil Biol. Biochem. 18: 427 -435. MOORE,T. R. 1981. Controls on the decomposition of organic matter in sub-arctic spruce-lichen woodland soils. Soil Sci. 131: lQ7113. NETER,J., and WASSERMAN, W. 1974. Applied linear statistical models. R. D. Irwin, Homewood, IL. ODUM,E. P. 1971. Fundamentals of ecology. 3rd ed. W. Be Saunders, Philadelphia. R. J., STED,J., and WOODELL,S. R. J. PWILLIPSON, J., PUTMAN, 1975. Litter input, litter decomposition, and the evolution of carbon dioxide in a beech woodland - Wytham Woods, Oxford. Oecologia, 20: 203 -2 17. PRESCOTT, C. E. 1984. Effects of sulphur dioxide pollution on litter decomposition in a cool temperate pine forest. M.Sc. thesis, University of Calgary, Calgary, Alta. PRESCOTT, C. E., and PARKINSON, D. 1985. Effects of sulphur pollution on rates of decomposition in a pine forest. Can. J. Bot. 63: 1436 - 1443. RAI,B., and SRIVASTAVA, A. K. 1982. Decomposition of leaf litter in relation to microbial populations and their activity in a tropical dry mixed deciduous forest. Pedobiologia, 24: 151 - 160. REINERS,W. A. 1968. Carbon dioxide evolution from the floor of three Minnesota forests. Ecology, 49: 47 1 -483. REINKE,J. J., ADRIANO, D. C., and MCLEOD,K. W. 1981. Effects of litter alteration on carbon dioxide evolution from a South Carolina pine forest floor. Soil Sci. Soc. Am. J. 45: 620-623. SSWLESINGER, W. H . , and HASEY,M. M. 1981. Decomposition of chaparral shrub foliage: losses of organic and inorganic constituents from deciduous and evergreen leaves. Ecology, 62: 762 - 774. S E A S T E J. ~ , R., CROSSLEY, D. A., JR., MEENTEMEYER, V., and WAIDE,J. B. 1983. A two-year study of leaf litter decomposition as related to macroclimatic factors and microarthropod abundance in the southern Appalachians. Holarct. Ecol. 6: 1 1 - 16. SUFFLING, R., and SMITH,D. W. 1974. Litter decomposition studies using mesh bags: spillage inaccuracies and the effect of repeated artificial drying. Can. J. Bot. 52: 2157 -2163. SWIFT,M. J., HEAL,0 . W., and ANDERSON, J. M. 1979. Decomposition in terrestrial ecosystems. Stud. Ecol. 5: 1-372. TAYLOR,B. R., and PARKINSON. D. 1 9 8 8 ~ .A new microcosm approach to litter decomposition studies. Can. J. Bot. This issue. 19886. Aspen and pine leaf litter decomposition in laboratory microcosms. I. Linear versus exponential models of decay. Can. J. Bot. This issue. 1 9 8 8 ~ Patterns . of water uptake and leaching in pine and aspen leaf litter. Soil Biol. Biochem. 20: 257-258. 1988d Respiration and mass loss rates of aspen and pine leaf

1973

litter decomposing in laboratory microcosms. Can. J. Bot. This issue. 19884. Annual differences in quality of leaf litter of aspen (Pupulus tremuloides) affecting rates of decomposition. Can. J. Bot. This issue. TORMALA, T., and ELORANTA, A. 1982. Decomposition of dead plant material in an abandoned field in central Finland. Ann. Bot. Fenn. 19: 31-38. VANCLEVE,K. 1971. Energy and weight-loss functions for decomposing foliage in birch and aspen forests in interior Alaska. Ecology, 52: 7 2 -73. VANCLEVE,K., and SPRAGUE, D. 1971. Respiration rates in the forest floor of birch and aspen stands in interior Alaska. Arct. Alp. Res. 3: 17-26. VOGT,K. A., EDMONDS, R. L., ANTOS,G. C., and VOGT,D. J. 1980. Relationships between C 0 2 evolution, ATP concentrations and decomposition in four ecosystems in western Washington. Oikos, 35: 7 2 - 7 9 . WAKSMAN, S. A., and GERRETSON, F. C. 1931. Influence of temperature and moisture upon the nature and extent of decomposition of plant residues by microorganisms. Ecology, 12: 33-60. WHITFORD, W. G., REPASS,R., PARKER, L. W., and ELKINS,W. Z. 1982. Effects of initial litter accumulation and climate on litter disappearance in a desert ecosystem. Am. Midl. Nat. 108: 1Q5 110. WIEDER,R. K., and LANG,G. E. 1982. A critique of the analytical methods used in examining decomposition data obtained from litter bags. Ecology, 63: 1636 - 1642. WLDUNG, R. E., GARFIELD, T. R., and BUSCHBOM, R. L. 1975. The interdependent effects of soil temperature and water content on soil respiration rate and plant root decomposition in arid grassland soils. Soil Biol. Biochem. 7: 373 -378. WILLIAMS, B. L. 1972. Nitrogen mineralization and organic matter decomposition in Scots pine humus. Forestry, 45: 177 - 178. WITKAMP, M. 1963. Microbial populations of leaf litter in relation to environmental conditions and decomposition. Ecology, 44: 370 377. 1966a. Rates of carbon dioxide evolution from the forest floor. Ecology, 47: 492 -494. 1966b. Decomposition of leaf litter in relation to environment, microflora, and microbial respiration. Ecology, 47: 194201. WITKAMP, M., and AUSMUS, B. 1976. Processes in decomposition and nutrient transfer in forest systems. In The role of terrestrial and aquatic organisms in decomposition processes. kidited by J. M. Anderson and A. Macfadyen. 17th Symposium of the British Ecological Society. Blackwell, London. pp. 375 -396. WITKAMP, M., and FRANK, M. L. 1970. Effects of temperature, rainfall, and fauna on transfer of '37Cs, K, Mg. and mass in consumer-decomposer microcosms. Ecology, 51: 465 -474. WOODS,P. V,, and RAISON,R. J. 1983. Decomposition of litter in subalpine forests of Eucalyptus delagatensis, E. pauciflura and E. dives. Aust. J. Ecol. 8: 287-300. ZAR,J. H. 1974. Biostatistical analysis. Prentice-Hall, New York.