Temporal and Spatial Patterns of Specific Leaf Weight in Successional Northern Hardwood Tree Species'

Amer. J. Bot. 73(8): 1083-1092. 1986.

TEMPORAL AND SPATIAL PATTERNS OF SPECIFIC LEAF WEIGHT IN SUCCESSIONAL NORTHERN HARDWOOD TREE SPECIES'

THOMAS W. JURIK2 University ofMichigan Biological Station, Ann Arbor, Michigan 48109-1048

ABSTRACT Temporal and spatial patterns of specific leaf weight (SLW, g/m 2) were determined for de ciduous hardwood tree species in natural habitats in northern lower Michigan to evaluate the utility of SLW as an index of leaf photosynthetic capacity. No significant diurnal changes in SLW were found. Specific leaf weight decreased and then increased during leaf expansion in the spring. Most species, especially those located in the understory, then had relatively constant SLW for most of the growing season, followed by a decline in SLW during autumn. Specific leaf weight decreased exponentially down through the canopy with increasing cumulative leaf area index. Red oak (Quercus rubra), paper birch (Betula papyri/era), bigtooth aspen (Populus grandidentata), red maple (Acer rubrum), sugar maple (A. saccharum), and beech (Fagus gran difoliaygenerally had successively lower SLW, for leaves at anyone level in the canopy. On a given site, comparisons between years and comparisons ofleaves growing within 35 em ofeach other showed that differences in SLW among species were not due solely to microenvironmental effects on SLW. Bigtooth aspen, red oak, and red maple on lower-fertility sites had lower SLW

than the same species on higher-fertility sites. Maximum CO 2 exchange rate, measured at light

saturation in ambient CO 2 and leaf temperatures of 20 to 25 C, increased with SLW. Photo synthetic capacities of species ranked by SLW in a shaded habitat suggest that red oak, red maple, sugar maple, and beech are successively better adapted to shady conditions.

LEAVES ARE KNOWN to be highly plastic in re Smith, 1975; Nobel, 1977; Chabot and Cha sponse to their growth conditions, i.e., they can bot, 1977; Chabot, Jurik and Chabot, 1979; exhibit wide ranges ofblade morphology, me Jurik, Chabot and Chabot, 1982), although in sophyll development, and physiological char tracellular effects of other factors such as nu acteristics. Knowledge of how leaf character trient supply also influence photosynthetic ca istics vary over time and space is required for pacity. Thus, SLW can potentially be used as investigations ranging from studies of intra an indirect measure ofthe photosyntheticchar cellular components up to studies of whole acteristics of a leaf, while also giving a direct canopy photosynthesis. For such larger scale measurement of allocations of biomass in a studies dealing with canopies, it is generally plant. The study ofSLW provides a means of impractical to do extensive measurements of integrating various factors such as canopy leaf physiological characteristics, since such structure, leaf area index, light environment, measurements are usually difficult and time andleafphotosyntheticperformance. Forthese consuming. For such reasons, a simple, indi reasons, the determination ofpatterns ofSLW rect index of leaf physiological characteristics within and among plants and over space and would be a useful tool for studies dealing with time is very useful to studies of the carbon large numbers ofleaves. balance and growth of plants, as well as to One possible index is Specific Leaf Weight studies of basic adaptive processes. . (SLW), the ratio of blade mass to blade area, This study evaluated temporal and spatial which is in general an indicatorofleafthickness variability in SLW ofseveral majortree species and the degree of mesophyll development of northern hardwood successional forests. within a leaf blade. The extent of mesophyll Specific LeafWeight was related to photosyn development largely determines the photosyn thetic capacity to determine ifSLWcan be used thetic capacity of a leaf (Nobel, Zaragoza and as an index ofphysiological characteristics and to determine ifspecies can be differentiated on 1 Received for publication 30 August 1985; revision ac cepted 20 November 1985. the basis of their photosynthetic characteris I thank J. Axelrod, B. Gaynor, K. Kandt, J. Simons, A. tics. Parman, and D. Kirschtel for assistance with the field work and D. M. Gates and J. A. Weber for providing the op portunity for this research. METHODS-Field sites were located at the 2 Present address: Dept. ofBotany, Iowa State Univer Univ. of Michigan Biological Station near sity, Ames, IA 50011. Pellston, Michigan (45°33'N, 84°42'W). The

1083 1084 AMERICAN JOURNAL OF BOTANY [Vol. 73 original forest was completely removed by log Leaves for determination of seasonal pat ging and burning by 1909. Some ofthe succes terns of SLW were usually collected in late sional regrowth forests on the sites have since morning; the interval between sample dates been subjected to cutting or burning. Two I-ha varied from one to four wks. For a study of plots on soils of relatively high and low soil diurnal patterns of SLW, five leaves of each fertility (see Koerper and Richardson, 1980) species were collected for each offive sample were clearcut in 1972 (the "HIFE"and "LOFE" times spaced at three-hr intervals from 6:30 to sites, respectively). Regrowth has been rapid 18:30 local solartimeon 30 August 1984. Time on HIFE, with trees reaching six to ten m height ofday for collection ofsamples in other studies by 1984; a dense, diverse understory has de varied. Leaf areas were measured with a LI veloped. Regrowth on LOFE has been rela COR Model 3000 leaf area m; leaf dry mass tively slow, with most trees only three to six was determined after drying to constant wt in m in height by 1984; there was only a sparse an oven at 70 C. understory. Onehundred m southeastofLOFE Leaf Area Index (LAI) was determined for was a site burned in 1954 ("54 BURN"), one the HIFE, LOFE, and 54 BURN sites by direct ofa series ofplanned bums designed to study harvests (see Jurik, Briggs and Gates, 1985). successional trends. Canopy trees were six to Three 1.5 m x 1.5 m plots on each site were nine m tall in 1984. The Upper Grapevine harvested by sequentially erecting sections of (UPGR) site supported a successional forest scaffolding and collecting all leaves of arbo 60 to 70 years of age. Average ht of trees in rescent, woody species in each 0.5 m or 1.0 m the overstory was approximately 15 m. The ht increment. For each species, leaf area (of Grapevine Trail (GRTR) site was 80 m east one side ofthe leaf) and dry mass were deter ofthe UPGR site and had a canopy similar to mined for 50 leaves (lumped) from each can that at UPGR. In 1984, bigtooth aspen (Pop opy level; total leafarea ofeach species in each ulus grandidentata Michx.) dominatedthecan level was calculated from the total dry leafmass opy at all sites. Red maple (Acer rubrum L.) and the mass/area ratio (i.e., the SLW) deter and red oak (Quercus rubra L.) were also im mined from the 50-leaf sample. For compar portant canopy species, with their abundances ison, SLW was also individually determined varying somewhat among sites. Paper birch on five leaves from each canopy level. Harvest (Betula papyri/era Marsh.) and trembling as plots were arbitrarily located in areas of"uni pen (Populus tremuloides Michx.) were occa form" canopy, except for the LOFE-3 plot, sionallypresenton someofthe sites. TheGRTR which was near a dense clump of red maple site supported a greater number ofspecies, es sprouts in a clone ofbigtooth aspen. The plots pecially in the understory, than did the other were harvested during the last week ofAugust sites. At GRTR, sugar maple (Acer saccharum 1984 and the first two weeks of September Marsh.) and red oak were the most common 1984, before any leaffall occurred. woody understory species, with beech (Fagus Gas exchange ofsingle, attached leaves was grandifolia Ehrh.), striped maple (Acer penn monitored using a mobile laboratory at the sylvanicum L.), basswood (TWa americana L.), UPGR and GRTR field sites. Leaves from 0.5 shadbush (Amelanchier sp.), and several other to 1.5 m above ground, on seedlings and small species also present. saplings, were sampled at GRTR. At UPGR, All deciduous tree species on the study sites a tower made of scaffolding material allowed produced only a single set of leaves, during a use of leaves growing in the upper part ofthe relatively short period in late spring. Aspen canopy. The gas exchange measurement sys has an indeterminatepattern ofleafproduction tem was an open-circulation system with an on many sites over its entire range ofhabitats. infrared CO 2 gas analyzer, as described in Ju Although young aspen sprouts on recently rik, Weber and Gates (1984). Measurements cleared areas at the Biological Station may ex were made in ambient CO 2 and O2 levels with hibit continuous production ofleaves through saturating light on mature leaves during the out the summer, older aspen stems produce period late June to early September during only a single set ofleaves, possibly because of 1981-1983. Data are largely from experimen the relatively low fertility of the soils in this tal runs, made at all times of day, in which region. This study thus dealt with leaves all light and temperature were controlled; some produced in late spring and all dying in au data from daily-time courses ofgas exchange, tumn. Leafexpansion ofall species except big in ambient light and temperature, are also in tooth aspen typically was completed before cluded. Measurements of temperature re

June, whereas bigtooth aspen leaf expansion sponse of light-saturated CO 2 exchange rate often was not completed until the second or (CER) suggested that the optimal temperature third week ofJune. for all species was between 20 and 25 C, with August, 1986] JURIK- LEAF WEIGHT IN HARDWOOD TREES 1085

60 ,20 r-- - - - ""'T- - ...... ----.--_--,

'DO

~ 50 ..E -----. <, ~ 60 s> '-' 40 Vl

3= .0 SUN --l R~d oak (f) 30 1981 60 20 MAY JUNE SEPT Fig. 1. Changes in Specific Leaf Weight (SLW) of red oak leaves in the understory at Grapevine Trail during ~ and immediately after leafexpansion. Leafexpansion was Vl 20 completed in the first week of June. Each datum is for a different leaf. SHADE MAY JUN JUL AUG SEP OCT Fig. 2. Seasonal patterns of Specific Leaf Weight (SLW) of leaves at Upper Grapevine. a) Leaves growing little or no change in temperature optimum in full sun at the top ofthe canopy or, for sugar maple and during the growing season (Jurik, unpubl. data). beech, on isolated trees in clearings. b) Leaves growing in Data presented here for CER were from mea shade in the understory or, for paper birch and bigtooth surements between 20 and 25 C. aspen, at the lowest point they existed in the canopy. Key: o red oak, • paper birch, i:l bigtooth aspen, + red maple, RESULTS- A study ofdaily variation in SLW o beech, x sugar maple. ofleaves at the top ofthe UPGR forest canopy on 30 August 1984, a sunny, mild day, showed that red oak had a larger range ofmean values was a decline in SLW during leaf senescence (84-98 g/m-) than either bigtooth aspen (76 in late SeptemberandOctober. The "SHADE" 85 g/m-) or paper birch (90-95 g/m-). In all leaves ofbigtooth aspen and paper birch were three species, the highest values of SLW oc from the lowermost branches ofcanopy trees, curred in the afternoon, butthe variation among since these species were not present in the leaves at anyone time ofday was large enough understory; these leaves thus may have been that there were no statistically significant dif in slightly higher light than the other species. ferences among the mean values for anyone For sun-grown leaves, red oak and paper birch species (Student-Newman-Keuls multiple had the highest SLW; sugar maple and beech range test, P > 0.05). Leaves in the understory had the lowest SLW, while bigtooth aspen and were not studied because experience showed red maple were intermediate (Fig. 2a). For that variability in sampling was as large as the shade-grown leaves, paper birch and bigtooth probable maximum diurnal effects, as judged aspen had the highest SLW; sugar maple and from the samples at the top of the canopy. beech again had the lowest SLW, while red oak Specific LeafWeight changed greatly during and red maple were intermediate (Fig. 2b). The and just after leaf expansion, as illustrated in species thus can adjust their leafanatomy, and Fig. 1 for red oak in the understory. Specific hence SLW, to different growth light levels. The LeafWeight declined from leafappearance un species also appear to maintain approximately til near the completion ofleafexpansion in the the same relative ranking in regard to the re first week ofJune. SLW then increased for sev sponse ofSLW to light in both sun and shade. eral wk after expansion had ceased. Seasonal patterns of SLW for seven under After the initial changes in SLW attributable story species at GRTR were very similar in to leaf growth and maturation, SLW varied 1980 and 1981 (Fig. 3). The species expanded relatively little over most ofthe growing season their leaves at different times, so the initial (Fig. 2, 3). Each datum is the mean ofthree to decrease in SLW due to leafexpansion (see Fig. five leaves. Leaves growing in full sun and in 1) was not seen consistently in 1980, when the understory at the UPGR site had relatively sampling was not begun as early as in 1981. little change in SLW from June to mid-Sep Red maple, round-leaf dogwood (Cornus ru tember (Fig. 2). (Sampling ofthese leaves start gosa Lam.), and Maianthemum canadense ed near the completion of leaf expansion and Desf., a perennial herb, consistently had higher hence does not show the early changes in SLW SLW than beech, sugar maple, striped maple, during leafexpansion.) For most species, there and maple-leafviburnum (Viburnum acerifol- 1086 AMERICAN JOURNAL OF BOTANY [Vol. 73

45 Higher SLW

~ red oak N ~ 35 -2J 63 15 10 7 ~ 25 VI 1 shadbush red maple b 75 7.3 7.7 10 65

striped maple ~ .uga, maPle] 55 N l' ~ 13.6 -2J 45 beech ~ 11.5 2 VI } white ash 35 Lower SLW Fig. 4. Relative ranking ofspecies according to Specific 25 LeafWeight (SLW). The "distance" between species is the mean difference in SLW (g/m-) of leaves growing in the 15-l--~-~--.,...---.---~-----l same microenvironment (i.e., within 35 em ofeach other). MAY JULY SEPT Fig. 3. Seasonal patterns of Specific Leaf Weight (SLW) ofleaves in the understory at Grapevine Trail. a) maple (10) and sugar maple-beech (3.6) dis 1980, b) 1981. Key: • sugar maple (seedlings), 0 sugar tances was 13.6 (Fig. 4). maple (small saplings), 0 striped maple, • red maple, t::. The harvest plots at the RIFE, LOFE, and beech, + maple-leaf viburnum, x round-leaf dogwood, o Maianthemum canadense. 54 BURN sites revealed patterns ofleafchar acteristics through the canopy. Mean SLW of the individual samples was an excellent pre ium L.). All species had relatively constant SLW dictor ofthe "true" SLW measured by the 50 from mid-June to mid-September, after which leaf samples. For all species and sites com SLW generally declined during leafsenescence. bined, there was very nearly a 1:1 relationship The above ranking ofunderstory species by between the individualmeans (SLWind ) and the SLW incorporates variability in microenvi bulk samples (SLWbulk), as indicated by the ronment as well as variation over time. Larger regression equation SLWbulk = 2.70 + scale variation in habitatscouldalso affect such 0.962*SLWind (N = 91; slope> 0, P < 0.01; comparisons; for example, the values shown r2 = 0.93). For individual species and plots, in Fig. 2 for red oak in the understory were only four of20 regressions ofthe 50-leafsam from a location about 100 m from the site ples versus the individual leafmeanswere non where the data ofFig. 3, for other species, were significant, and these eitherwere based on very collected. Does red oak really have higher SLW few samples (N < 5) or had a very limited than most other understory species, or was it range of values of SLW. These results suggest by chance or necessity growing in an area with that individual determination of SLW of five more light? To eliminate the effect ofmicroen or more leaves is probably sufficient to give a vironment, pairs or triplets of leaves growing reasonable picture of differences in SLW at within 35 em of each other at GRTR were different levels in a canopy. compared. Figure 4 shows the mean difference The average and range of the coefficient of in SLW between any two species, based on the variation (standard deviation divided by the differences in SLW of several sets of paired mean, in percent) for the several canopy levels leaves. Red oak consistently had the highest in each harvest plot, based on the five indi SLW, while white ash (Fraxinus americana L.) vidual determinations of SLW for each level hadthelowest. Thedirection ofdifferences was in the canopy, were similaramongspecies. Over consistent among all species combinations, and all plots, the mean, range, and number ofsam the "distance" among species (i.e., the differ ples were: bigtooth aspen, 8.1, 2.7-19.5, 49; ence in SLW) was remarkably consistent when red oak, 8.6, 3.8-18.3, 19; red maple, 10.3, measured via direct versus indirect compari 3.8-24.4, 24; all (five) species combined, 8.8, sons. For example, the distance from red oak 2.7-24.4, 100. There were no apparent trends to beech was 15; the sum ofthe red oak-sugar with position in the canopy (data not shown), August, 1986] JURIK- LEAF WEIGHT IN HARDWOOD TREES 1087 • 95 .•~----.---....,...--.,....----.--....,...-""""""'1 95 Trembling aspen 90 Bigtooth aspen ---.. • HIFE-l ---.. 85 o 54 BURN-l N 85 ~ • 54 BURN-3 -€. 80 ~ • ~ 75 75 • ~ 70 0 • • U1 65 • • • 65 o o 0 • a 60 d o • 55 --J-.,.~""""""""""''''''''''''''''''''''''"'T"''....,...... ,...~...,..,..,....,.....,...,,;•• 55 o 0.5 1 1.5 2 2.5 3 o 0.5 1 1.5 2 2.5 3 LAI Above Leaf LAI Above Leaf • 95 " 0 70 Bigtooth aspen Red oak 85 6 o HIFE-l o LOFE-2 ---.. 65 0 ---.. N 75 ~ • HIFE-2 60 o • HIFE-3 60 6 -€. Ol 6 HIFE-3 o 6 54 BURN-2 '-" ~ 65 55 06 • • • o ~ o ~ 55 U1 50 o U1 o • • 45 45 • b o e •• 35 -h-.,..,.....,...... ~...,...... ,...... ,....~..,....~...,.....~...,.....~...-l 0.5 1.5 2 2.5 3 3.5 o 0.5 1.5 2 2.5 3 3.5 4 LAI Above Leaf LAI Above Leaf 90 6 80 Bigtooth aspen + + Red maple 85 • ---.. o LOFE-1 ---.. 70 N 80 -€. • • LOFE-2 ~ + Ol 75 0 + '-" 0 6 LOFE-3 Ol 60 '-" • 06 ~ 70 o HIFE-1 ~ 50 • o e 0 U1 65 U1 • HIFE-2 • • 6 .~. .6 40 6 LOFE-3 60 • c o + 54 BURN-2 o 0 """"'~""""''''''''''''''''r"'""'~..,...~..o;- ...... ,~~..,...~.-1 55 ...... 30 +,-""""T"'"T"""""~""""~""""~"""'''''''''T"'"T''''''~''''''''~...-l o 0.5 1.5 2 2.5 3 3.5 o 0.5 1.5 2 2.5 3 3.5 4 4.5 LAI Above Leaf LAI Above Leaf

Fig. 5. Specific Leaf Weight (SLW) of leaves in each harvest interval versus total LAI above the midpoint of the harvest interval, by species and site. Mean SLW was determined from the total area and total dry mass of 50 leaves. a) Trembling aspen (HIFE), b) bigtooth aspen (HIFE), c) bigtooth aspen (LOFE), d) bigtooth aspen (54 BURN), e) red oak, f) red maple. indicating that there is about the same relative the combined plot regressions were all statis degree ofvariation in SLW ofleaves through tically highly significant (Table 1). Regressions out the canopy. using the log ofSLW versus LAI gave slightly In Fig. 5, SLW is plotted as a function of better fits than the EXP model for red maple total LAI above the midpoint ofeach harvest and red oak at RIFE, but were otherwise worse interval. Specific Leaf Weight declined with (data not shown). Similarly, a linear model of increasing cumulative LAI, for all species and SLW versus LAI gave the best fits for red maple sites. Regressions of SLW against the recip and for red oak at RIFE but gave poorer fits rocal ofthe exponential ofLAI ("EXP" model), than the EXP model in the other cases. At with k = 1.0 (the extinction coefficient), were RIFE, red oak and red maple were limited to statistically significant for almost all of the the lower part ofthe canopy, so that a full range samples (Table 1). Regressions using the EXP oflight environments and potential SLWs was model with k = 0.732, an extinction coefficient not present; thelower portion ofan exponential for light found in deciduous forests in Ten curve could easily be perceived as linear, in nessee by Baldocchi et al. (1984), gave essen such cases. tially the same results (data not shown). Higher soil fertility resulted in lower SLW, Regressions using the EXP model and com for any given cumulative LAI (Fig. 5; Table bining data from all harvest plots on a site 1). Bigtooth aspen on the LOFE and 54 BURN typically had goodness-of-fits (,-2) lower than (lower-fertility) sites had SLW about 15 g/m those of regressions for single plots, although higher than bigtooth aspen on the RIFE (higher- 1088 AMERICAN JOURNAL OF BOTANY [Vol. 73

16...----.------..,--...,------.----,r-----o 5,----,------..,---,...... ----, • 14 a d ,C/) • -'c/) 4.5 7 12 • • 7 E E 4 lO ••• ...... • o • i 8 • ~3.5 • ••• • ..... 3- 6 • Red oak 3 • ffi 4 Sugar maple u 201---_---t---+--_+--...... 2. 5 ~---t---__+_-- ~-_____4 20 40 60 120 20 25 30 3S SLW SLW

Fig. 6. CO2 Exchange Rate (CER) versus Specific LeafWeight (SLW) ofleavesat the UpperGrapevine and Grapevine Trail sites. CER was measured at light-saturation, optimum temperature (20-25 C), and ambient CO 2 ; each datum is a measurement on a different leaf. Data are from 1981, 1982, and 1983. The lines are linear regressions of CER versus SLW (see Table 2). a) Red oak, b) red maple, c) bigtooth aspen, d) sugar maple, e) beech, f) basswood. fertility) site; red maple and red oak were 5 were only rarely represented at the top of the and 9 g/m? higher, respectively, on the LOFE canopy and hence there are no high-SLW sam and 54 BURN sites than on the RIFE site. For ples for these species. Bigtooth aspen was ap any given cumulative LAI at RIFE, bigtooth parently limitedby its intolerance ofshade and aspen, red maple, redoak, and tremblingaspen did not produce SLWs lower than 50 g/m- on had successively higher SLW. At the lower these sites. Regressions ofCER on SLW, using fertility sites, the species had the same relative a linear model, had highly statistically signif order (with no trembling aspen present), al icant positive slopes for all species but sugar though absolute values of SLW differed. maple (Table 2). Red oak had the widest range Maximum, light-saturated CO 2 exchange rate ofSLW and CER and clearly exhibited a linear (CER) at optimum temperature in ambient CO 2 relationship of CER to SLW (Fig. 6a). Red generally increased with SLW (Fig. 6). Data are maple (Fig. 6b) had a pattern very similar to all from the UPGR and GRTR sites, with sim that ofred oak, although the range ofSLW was ilar soil fertility. Due to the successional nature more limited. Bigtooth aspen had some in ofthe forest, sugar maple, beech, andbasswood crease in CERwith SLW, although CERseemed August, 1986] JURIK- LEAF WEIGHT IN HARDWOOD TREES 1089

TABLE 1. Regressions ofSpecific LeafWeight (g/m') ver- TABLE 2. RegressionsofCO2 Exchange Rate (CER, umol sus cumulative LeafArea Index (LAI) above a given m! s:') versus Specific LeafWeight (SLW, g/m'l for levelon the harvestplots. The regressionequation was: individual leaves from the UPGR and GRTR sites, SLW = a + b"Exp(-k"LAI), with k = 1.0. The sis- measured over the years 1981-1983. The regression nificance column is for the test ofb = 0 (.., P < 0.05; equation was: CER = a + b"SLW. The significance ...., P < 0.01; ns, P > 0.05) column is for the test ofb = 0 (.., P < 0.05; ...., P < 0.01; ns, P > 0.05). The second equation for sugar Sig- maple excludes the datum at SLW = 26 g/m' Species/site a b n nif. " Bigtooth aspen Sig- Species a b n nif. HIFE-l 45.414 21.217 6 * 0.81 " HIFE-2 42.767 41.843 4 ** 0.99 Bigtooth aspen 5.49 0.123 40 ** 0.33 HIFE-3 44.316 19.073 6 ** 0.93 Red maple 0.44 0.109 16 ** 0.73 LOFE-l 53.355 31.238 8 ** 0.90 Red oak 0.17 0.112 40 ** 0.81 LOFE-2 60.117 18.073 6 ** 0.96 Beech 1.45 0.081 16 ** 0.53 LOFE-3 60.660 29.937 5 ** 0.95 Sugar maple 2.68 0.032 16 ns 0.13 54 BURN-l 55.746 17.553 6 ** 0.95 Sugar maple (exc pt) 2.05 0.051 15 ** 0.46 54 BURN-3 61.638 30.125 8 ** 0.85 Basswood 0.46 0.177 7 ** 0.93 HIFE 1-3 45.22 22.269 16 ** 0.76 LOFE 1-3 58.241 25.246 19 ** 0.81 54 BURN 1,3 60.25 19.868 14 ** 0.51 understory leaves incorporate variation due to Red oak different growth environments. This variation HIFE-3 44.113 45.332 6 * 0.66 makes it difficult to rank species by photosyn LOFE-2 62.269 37.838 6 ** 0.95 54 BURN-2 60.898 36.402 4 ** 0.99 thetic capacity, since one does notknowwheth er a given photosynthetic capacity is due to Red maple innate differences among species or due to en HIFE-l 21.392 476.48 3 ns 0.98 vironmental effects. Species could be growing HIFE-2 42.333 47.936 7 ** 0.93 only in microenvironments that suit their LOFE-3 56.659 113.42 5 ns 0.56 54 BURN-2 57.102 40.469 8 ** 0.90 growth requirements; subtle differences among HIFE 1-2 39.651 53.19 10 ** 0.87 such environments might not be apparent and Trembling aspen might not be averaged out over many mea surements of different leaves growing in mi HIFE-l 53.726 50.653 5 ** 0.99 croenvironments that were classified as the same by a human observer. The ranking of species by SLW presented in Fig. 4 and the to plateau at SLWs above 70 g/m- (Fig. 6c). relationships ofCER to SLW presented in Fig. Bigtooth aspen had higher CER for a given 6 may be combined to reduce much of the SLW than red maple and red oak. Sugar maple uncertainty in ranking species by photosyn and beech in the understory (Fig. 6d, e) had thetic capacity. For SLW of sugar maple ar much smaller ranges of SLW than the above bitrarily chosen to be 27 g/m-, SLWs of red species. In beech, CER increased with SLW oak, red maple, and beech derived from Fig. (Fig. 6e).Ifone probable outlierfor sugarmaple 4 are 39, 27, and 23 g/m-, respectively. This (at 26 g/m-, Fig. 6d) is excluded, sugar maple represents a hypothetical situation in which all also shows a statistically significant increase in species are growing in the same microenvi CER with SLW (Table 2). Differences among roment. For each species, photosynthetic ca and within plants in both long-term effects on pacity (as indicated by light-saturated CER) is CER, such as nutrient supply, and shorter-term calculated from Fig. 6. CO 2 exchange rate per effects on CER, such as mid-afternoon water unit leafarea is highest in red oak (4.54 JLmol stress, undoubtedly contributed to the varia m ? S-I) and successively lower in red maple tion seen in all species. The data for basswood (4.15), sugar maple (3.43), and beech (3.31). (Fig. 6f) show that much ofthe variation is due The species exhibit the reverse pattern for CER to such differences among plants. The bass per unit leaf mass, with beech highest (144 wood data are from a single sapling in the nmol g-I S-I) and sugar maple (127), red maple understory and hence show the effect ofstan (122), and red oak (116) successively lower. dardizing factors such as nutrient supply and light environment. Although the range ofSLW DISCUSSION- Temporal changes in SLW of was very limited, there was a very close rela leaves at any point in the canopy can influence tionship between CER and SLW (Table 2). interpretations ofthe relationship between SLW Since it is nearly impossible to find leaves and other leaf characteristics such as photo of four or more species growing in the same synthetic rate. Although there was little change microenvironment, i.e., within a few ern of in SLW over the course of a day in the tree each other, most comparisons of CER of leaves studied here, the seasonal changes in 1090 AMERICAN JOURNAL OF BOTANY [Vo1. 73

SLW were large and must be considered in tosynthetic capacity. Light primarily appears comparisons of SLW. The largest changes in to influence anatomical development, with SLW occurred during and immediately after higherlight levels leading to thickerleaves with leaf expansion. Coyne and Van Cleve (1977) greater development of the mesophyll (Nobel also found a pattern ofrapid increase in SLW et a1., 1975; Nobel, 1977; Chabot et a1., 1979; (of trembling aspen) during early June, fol Bjorkman, 1981; Juriketa1., 1982). This great lowed by a moregradual increase overthe sum er mesophyll development means greater SLW mer, although their measurements apparently and, often, higher photosynthesis per unit leaf did not start early enough to show a decline in area due to more photosynthesizing cells per SLW very early in leaf growth. Similarly, Le unit leafarea. Leaves apparently integrate the wandowska and Jarvis (1977) found SLW of various light levels received over a day and spruce needles to increase gradually over the adjust their anatomy in response to the total course ofa year. Changes in SLW early in leaf amount oflight received (Nobel, 1976; Chabot growth are presumably due to changes in rel et a1., 1979). Lower growth temperatures lead ative rates of cell division and cell expansion to increased SLW in alfalfa (Ku and Hunt, (Steer, 1971). Decreases in SLW during leaf 1973). Nutrient supply has its largest effect on senescence in autumn are due primarily to re photosynthesis at an intracellular level (Long translocation ofsugars and nutrients (Ostman streth and Nobel, 1980; Jurik et a1., 1982). and Weaver, 1982). The relative constancy of Decreased nutrient availability often leads to SLW from late June to early September found increased SLW, as has been noted in controlled here, especially in understory leaves, reflects studies ofa variety ofspecies (Oxman, Good both relative constancy of environment and man and Cooper, 1977; Longstreth and Nobel, constancy of leaf developmental stage. Since 1980; Jurik et a1., 1982), although decreases in photosynthetic capacity is also constant during SLW have also been found (Yoshida and Co this period but varies during leaf expansion ronel, 1976). Low-nutrient leaves often have and leaf senescence (Jurik, 1986), studies of smaller, more densely packed cells and may the relationship ofSLW to photosynthetic ca accumulate more starch, because photosyn pacity of leaves will be least confounded by thesis outruns the nutrient-limited conversion temporal effects ifperformed in the period late ofphotosynthetic products into more complex June to early September, as was done here. growth substrates (see Juriket a1., 1982). Leaves Even after temporal changes in SLW have on the lower-fertility sites of this study had been accounted for, comparisons of SLW relatively higher SLW, in accord with many among species still must distinguish genetic controlled experimental studies, although and environmental effects on SLW. There is Coyne and van Cleve (1977) found an increase undoubtedly a genetic influence on SLW, as in SLW with fertilization of aspen stands in shown by comparisons among and within the field. Water stress can have direct effects species grown under the same conditions on turgor pressure and hence cell expansion; (Barnes et a1., 1969; McGee, Schmierbach and SLW can be increased by water stress during Bazzaz, 1981; Ledig and Korbobo, 1983), al leaf expansion because decreased cell expan though the magnitudes of genetic differences sion results in approximately the same mass in SLW are generally much less thandifferences being spreadovera smallerleafarea (see Hsiao, due to environmental effects. For a given ge 1973; Smith and Nobel, 1978). notype, SLW can be influenced by a variety of Investigations of spatial variation in SLW environmental conditions, with leaves being within forest canopies typically have found that most sensitive to conditions prevailing during SLW decreases down through the canopy leaf expansion (Jurik, Chabot and Chabot, (Coyne and van Cleve, 1977; Lewandowska 1979). Leaves in shady environments typically and Jarvis, 1977; Schulze, Fuchs and Fuchs, have lower SLW than leaves grown in sunny 1977; van Elsacker and Impens, 1984). This conditions; lower SLW represents a comple trend in SLW apparently occurs in response to mentofleafcharacteristicsincludingdecreased light, since other environmental factors have leaf thickness, decreased palisade cell devel much smaller gradients through the canopy, as opment, decreased RubP carboxylase, in do factors within a tree such as nutrient supply creased chlorophyll per unit leaf mass, lower and water supply. However, there are surpris light-saturation point and maximum rate of ingly few simultaneous measurements ofSLW photosynthesis, decreased respiration rate, and and the light environment in the field. Del Rio so on (Boardman, 1977; Chabot and Chabot, and Berg (1979) found a negatively linear re 1977). lationship of SLW to the log of light received Recent investigations have started to reveal at a given point in the canopy. Although no the modes of operation of various environ direct measurements oflight were made here, mental factors in determining SLWand pho- the known exponential decline in light with August, 1986] JURIK- LEAF WEIGHT IN HARDWOOD TREES 1091 increasing LAI (Anderson, 1964; Ross, 1981) SLW; one must account for as many environ and the negative exponential relationship of mental effects as possible. The consistency of SLW to cumulative LAI found in this study the results ofthe several different studies here imply a strong relationship between light and suggests that it is possible to account for much SLW. The similarity ofpatterns ofSLW versus ofthe environmental variation, so that genetic LAI on sites ofdifferent fertility also indicates differences among species can be compared. A a strong influence oflight level on SLW, since second problem is interpreting different SLWs absolute values varied somewhat according to in evolutionary terms; what is the significance site fertility butthe exponential decline in SLW of SLW? How does SLW relate to the plant's with cumulative LAI was the predominant fea carbon balance, ability to survive, etc.? Com ture of the patterns. Absolute values of SLW puter simulations ofleafcarbon balances have for a given LAI and site are also likely to be shown that leaves with low-light, tolerant char very dependent on how well the assumption acteristics represented by low SLW do indeed ofa uniform canopy (implicit in the exponen have greater carbon gain per unit investment tial equation) is met. in leaf material in shady environments than Some investigators have found a positive would leaves with high-light characteristics correlation between SLWand photosynthetic "grown" in shady environments (Jurik, 1980; capacity per unit leaf area (e.g., Pearce et al., Jurik and Chabot, 1986); such improvement 1969; Dornhoffand Shibles, 1970, 1976), while in carbon gain presumably confers an adaptive others have found no correlation (Ku and Hunt, advantage. Thus, the ranking ofspecies by SLW 1973; see Bjorkman, 1981). Since environ in a given environment does give an idea of mental and genetic factors may have effects on the relative "adaptedness" of a given species leafanatomy and physiological characteristics to shade. The comparisons ofSLW in this study that are not always translated into proportion show that species do indeed have different SLW ate effects on photosynthesis, it is not surpris and photosynthetic capacity for a given mi ing that a positive relationship between SLW croenvironment. Bigtooth aspen, definitely an and leafphotosynthetic capacity has not been early successional species, had the highest pho consistently found, especially in comparisons tosynthetic capacity, even though its highest among species. Specie's may have more or less values ofSLW were not as high as those ofred dissimilar evolutionary histories that have re oak. Otherwise, SLW was a good indicator of sulted in particular combinations ofleafanat photosynthetic capacity and ofrelative ability omy and intracellular photosynthetic capacity, to adapt to shade, with red oak, red maple, so that different photosynthetic capacities may sugar maple, and beech having successively be achieved by leaves of the same SLW in lower SLW. This order ofspecies corresponds different species, and vice versa. The species to that traditionally assigned for successional studied here did show fairly clear relationships status and shade tolerance. between SLW and photosynthetic capacity per unitleafarea. The excellent correlation ofSLW LITERATURE CITED and photosynthetic capacity found in bass ANDERSON, M.e. 1964. Light relations ofterrestrial plant wood emphasizes the effectsin the other species communities and their measurement. BioI. Rev. 39: ofgenetic variation and variation in environ 425-486. mental factors other than light, since such ef BALDOCCHI, D. D., D. R. MATT, B. A. HUTCHISON, AND fects were minimized in the basswood data. R. T. McMILLEN. 1984. Solar radiation within an The relationships ofLAI, light, and SLW sug oak-hickory forest: an evaluation of the extinction gest that SLW was largely influenced by light coefficients for several radiation components during fully-leafed and leafless periods. Agric. For. Meteorol. and can be predicted from LA!. However, the 32: 307-322. larger question ofwhy particular levels ofpho BARNES, D. K., R. B. PEARCE, G. E. CARLSON, R. H. HART, tosynthetic capacity versus SLW have evolved AND C. H. HANSON. 1969. Specific leaf weight dif in different species cannot yet be answered. ferences in alfalfa associated with variety and plant One traditional classification of species is age. Crop Sci. 9: 421-423. BAZZAZ, F. A. 1979. The physiological ecology of plant based on their tolerance or intolerance ofshade, succession. Ann. Rev. Ecol. Syst. 10: 351-371. i.e., their ability to survive and grow under BJORKMAN, O. 1981. Responses to different quantum neath a canopy. Species appearing early in a flux densities. In O. L. Lange, P. S. Nobel, C. B. forest successional sequence are typically in Osmond, and H. Ziegler [eds.], Physiological plant tolerant of shade, while late-successional or ecology I: Responses to the physical environment, pp. "climax" species are tolerant ofshade (Bazzaz, 57-107. Ene. plant physiology, Vol. 12A. Springer 1979). Can species be ranked by SLW relative Verlag, New York. 625 pp. BOARDMAN, N. K. 1977. Comparative photosynthesis of to successional status and shade tolerance? The sun and shade plants. Ann. Rev. PI. Physiol. 28: 355 first problem encountered in answering this 377. question is getting a reasonable comparison of CHABOT, B. F., AND J. F. CHABOT. 1977. Effects oflight 1092 AMERICAN JOURNAL OF BOTANY [Vol. 73

and temperature on leafanatomy and photosynthesis LEWANDOWSKA, M., AND P. G. JARVIS. 1977. Changes in in Fragaria vesca.Oecologia 26: 363-377. chlorophyll and carotenoid content, specific leafarea ---, T. W. JURIK, AND J. F. CHABOT. 1979. Influence and dry weight fraction in Sitka spruce, in response of instantaneous and integrated light flux density on to shading and season. New Phytol. 79: 247-256. leaf anatomy and photosynthesis. Amer. J. Bot. 66: LoNGSTRETH, D. J., AND P. S. NOBEL. 1980. Nutrient 940-945. influences on leafphotosynthesis: Effects ofnitrogen, COYNE, P. I., AND K. VAN CLEVE. 1977. Fertilizer induced phosphorus and potassium for Gossypium hirsutum morphological and chemical responses of a quaking L. PI. Physiol. 65: 541-543. aspen stand in interior Alaska. Forest Sci. 23: 92-102. McGEE, A. B., M. R. SCHMIERBACH, AND F. A. BAZZAZ. DEL RIO, E., AND A. BERG. 1979. Specific leaf area of 1981. Photosynthesis and growth in populations of Douglas-fir reproduction as affected by light and needle Populus deltoides from contrasting habitats. Amer. age. Forest Sci. 25: 183-186. MidI. Natur. 105: 305-311. DORNHOFF, G. M., AND R. SHIBLES. 1970. Varietal dif NOBEL, P. S. 1976. Photosynthetic rates of sun versus ferences in net photosynthesis ofsoybean leaves. Crop shade leaves of Hyptis emoryi Torr. PI. Physiol. 58: Sci. 10: 42-45. 218-223. ---, AND ---. 1976. Leaf morphology and anat ---. 1977. Internal leaf area and cellular CO 2 resis omyin relation to CO 2-exchangerate ofsoybean leaves. tance: photosynthetic implications ofvariations with Crop Sci. 16: 377-381. growth conditions and plant species. Physiol. Plant. HSIAO, T. C. 1973. Plant responses to water stress. Ann. 40: 137-144. Rev. PI. Physiol. 24: 519-570. ---, L. J. ZARAGOZA, AND W. K. SMITH. 1975. Re JURIK, T. W. 1980. Physiology, growth, and life history lationbetween mesophyllsurfacearea, photosynthetic characteristics of Fragaria virginiana Duchesne and rate, and illumination level during development for F. vesca L. (Rosaceae). Ph.D. dissertation. Cornell leavesofPlectra nth us parviflorus Henckel. PI. Physiol. University. 439 pp. 55: 1067-1070. ---. 1986. Seasonal patterns of leaf photosynthetic OSTMAN, N. L., AND G. T. WEAVER. 1982. Autumnal capacity in successional northern hardwood tree nutrient transfers by retranslocation, leaching, and lit species. Amer. J. Bot. 73: 131-138. ter fall in a chestnut oak forest in southern Illinois. ---,J. F. CHABOT, AND B. F. CHABOT. 1979. Ontogeny Can. J. Forest Res. 12: 40-51. ofphotosyntheticperformancein Fragaria virginiana. OXMAN, A. M., P. S. GOODMAN, AND J. P. COOPER. 1977. PI. Physiol. 63: 542-547. The effects ofnitrogen, phosphorus and potassium on --,--, AND --. 1982. Effects of light and rates of growth and photosynthesis of wheat. Photo

nutrients on leafsize, CO 2 exchange, and anatomy in synthetica II: 66-75. wild strawberry iFragaria virginianat. PI. Physiol. 70: PEARCE, R. B., G. E. CARLSON, D. K. BARNES, R. H. HART, 1044-1048. AND C. H. HANSON. 1969. Specific leaf weight and --, J. A. WEBER, AND D. M. GATES. 1984. Short photosynthesis in alfalfa. Crop Sci. 9: 423-426.

term effects ofCO 2 on gas exchange of leaves of big Ross, J. 1981. The radiation regime and architecture of tooth aspen (Populus grandidentata) in the field. PI. plant stands. W. Junk, The Hague. Physiol. 75: 1022-1026. SCHULZE, E.-D., M. FUCHS, AND M. I. FUCHS. 1977. Spa ---, G. M. BRIGGS, AND D. M. GATES. 1985. A com cial distribution of photosynthetic capacity and per parison offour methods for determining leafarea in formance in a mountain spruce forest of northern dex in successional hardwood forests. Can. J. Forest Germany. III. The significance ofthe evergreen habit. Res. 15: 1154-1158. Oecologia 30: 239-248. ---, AND B. F. CHABOT. 1986. Leaf dynamics and SMITH, W. K., AND P. S. NOBEL. 1978. Influence of ir profitability in wild strawberries. Oecologia 69: 286 radiation, soil water potential, and leaf temperature 304. on leafmorphology ofa desert broadleaf, Encelia far KOERPER, G. J., AND C. J. RICHARDSON. 1980. Biomass inosa Gray (Compositae). Amer. J. Bot. 65: 429-432. and net annual primary production regressions for STEER, B. T. 1971. The dynamics of leaf growth and Populusgrandidentata on three sites in northern lower photosyntheticcapacityin CapsicumjrutescensL. Ann. Michigan. Can. J. For. Res. 10: 92-101. Bot. 35: 1003-1015. KU, S. B., AND L. A. HUNT. 1973. Effects oftemperature VAN ELSACKER, P., AND I. IMPENS. 1984. Photosynthetic on the morphologyandphotosyntheticactivityofnewly performance ofpoplar leaves at different levels in the matured leaves of alfalfa. Canad. J. Bot. 51: 1907 canopy. In C. Sybesma [ed.], Advances in photosyn 1916. thesis research, Vol IV, pp. 129-132. LEDIG, F. T., AND D. R. KORBOBO. 1983. Adaptation of YOSHIDA, S., AND V. CORONEL. 1976. Nitrogen nutrition, sugar maple populations along altitudinal gradients: leaf resistance, and photosynthetic rate of the rice photosynthesis, respiration, and specific leaf weight. plant. Soil Sci. Plant. Nutr. 22: 207-211. Amer. J. Bot. 70: 256-265.