INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Ascomycetous fungi are the principal microbial secondary producers in standing decaying salt marsh grasses (40). This conclusion has been reached by using transmission electron microscopy and direct epifluorescence microscopy and by the dynamics of a fungal index sterol, ergosterol (37, 44, 47). Most monitoring of the fungal mass dynamics in standing decaying smooth cordgrass (Spartina alterniflora), a major salt marsh grass of the western Atlantic Ocean (4, 10, 11, 29), has been performed in Georgia salt marshes (40), although the range of smooth cordgrass extends north into Canada and south into Florida (30, 31).

One examination of fungal mass dynamics in standing decaying smooth cordgrass in New Brunswick, Canada, led to the conclusion that at the northern end of the range, the fungal standing crops are smaller than the fungal standing crops in Georgia (51). Newell (33) and Samiaji and Bärlocher (51) hypothesized that smaller fungal standing crops occur in marshes that are farther north on the basis of visual estimates of the density of fungal sexual structures (ascomata). It has been reported that there are substantial differences in microbial grass decomposition activities over north-to-south gradients (e.g., threefold greater maximum microbial incorporation of ¹⁴C-labeled wheat straw carbon at 43°N compared to incorporation at 61°N in Europe [6]) (1, 59). Here we describe our examination of the hypothesis that the level of fungal decomposition activity in smooth cordgrass is lower at higher latitudes. We conducted our examination by obtaining samples of living fungal standing crops and determining rates of fungal production in standing decaying smooth cordgrass in southern and northern temperate marshes in the autumn, when large standing crops of decaying shoots were present (45) before ice could shear away northern shoots (33). We obtained samples over a 3-year period because of the possibility that there could be interyear differences in marsh function (32, 51, 58), and as a subsidiary project, we obtained samples from standing decaying black needlerush (Juncus roemerianus) and black grass (Juncus gerardi). In addition, we examined variations in fungal standing crop size and activity in relation to nitrogen and phosphorus contents of decaying cordgrass in order to look for indications of potential control of fungi by these nutrients (20, 44).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Sites. Samples were obtained from eight sites over a 3-year period (1996 to 1998) (Table 1). The sites from which samples were obtained are as follows: site ME in the Wells National Estuarine Research Reserve in Maine (western part of Drakes Island marshes) (3); site MA2 at the Plum Island Sound Long Term Ecological Research Site in Massachusetts (south side of Rowley River marshes) (14); site MA1 in the Waquoit Bay National Estuarine Research Reserve in Massachusetts (marshes on the eastern border of Sage Lot Pond) (3, 13, 49, 52); site VA1 at the Virginia Coast Reserve & Long Term Ecological Research Site in Virginia (upper Phillips Creek marshes) (7, 28); site VA2 near the Chesapeake Bay-Virginia National Estuarine Research Reserve in Virginia (York River marshes along Chesapeake Bay, about 5 km east-northeast of station LE4.3 [12]); site FL1 at the Florida State University Marine Laboratory in Florida (inlet marsh just to the east of the laboratory behind the beach dunes, on St. George Sound, Gulf of Mexico); site FL2 in the Anastasia State Recreation Area east of St. Augustine, Fla. (marsh at the south end of Salt Run inlet); and site FL3 at the Halifax River inlet at Ormond Beach in Florida (marsh strip on the west side of the river at Riviera Park). Montague and Wiegert (31) have described the Florida marshes previously.

Collecting and handling. At each site, our goal was to collect standing dead shoots of smooth cordgrass (S. alterniflora) in a median-height, representative marsh plot in an area where tidal flood water should have a salinity in the range from 15 to 30 mg liter¹. When samples were collected at a site in more than one year, the same plots were used each year. On a day between 15 November and 1 December, we selected four shoots that bore at least two attached, dead, intact leaf blades (completely brown, not shredded to the extent that they were split lengthwise more than once over more than 10 cm of the length, and not collapsed onto the sediment). Electrical cable ties were placed tightly around the shoots at the base (to retain the decayed sheaths), and the shoots were cut at the sediment level. The severed shoots were air dried in an air-conditioned room for 1 week. The shoots were then express shipped to Sapelo Island, Ga., in plastic bags containing freshly opened desiccant and processed at the University of Georgia Marine Institute.

Fungal standing crop and productivity. The size of the standing crop of living fungi was measured by determining the ergosterol content of blades (15, 16, 19, 38). A high-pressure liquid chromatography assay method was used; this method included preincubation in a [¹⁴C]acetate solution and measurement of the ¹⁴C incorporated into ergosterol (34). The rates of acetate incorporation (fungal membrane synthesis) were used as an index of fungal productivity (16, 18, 19, 56). Pieces (1.5 cm) were cut from the nonligule end of each of the four air-dried blades obtained from a site for each height (upper or mid) on the shoots and were pooled to obtain one 6-cm sample, which was used for a fungal productivity assay, as described by Newell (36) and Newell et al. (44). Adaptation to seawater (salinity, 15 mg liter¹; filtered with a 0.2-µm-pore-size filter) submergence was accomplished by incubation for 3 h with slow agitation at 20°C under 30 µmol of photosynthetically available radiation m² s¹ (36). The final specific activities of the 5 mM acetate solutions used for incubation were 2.5 to 4.4 dpm pmol of acetate¹ (1 Bq = 1 dps), and preparations containing radiolabeled acetate were incubated for 2 to 2.5 h under the conditions used to allow adaptation to submergence. Reactions were terminated by rinsing away the radioisotope and immediately submerging the preparations in ethanol and storing them at 4°C in the dark (36). The liquid chromatographic conditions used were the conditions described by Newell et al. (43); the ergosterol standards used were kept desiccated under nitrogen at 4°C in the dark (38). Ergosterol peaks were collected and scintillation counting was performed as described by Newell (34, 36).

Ascospores in cordgrass. The rate of expulsion of ascospores was measured (with mid blades; two replicates per site) as described by Newell and Wall (41) and Newell and Wasowski (42). To do this, ascospores were captured on a target coverslip placed below a wet blade sample and counted with a dissecting microscope (×100), and species identities were confirmed at a higher magnification (×400). Preparations were incubated for 72 h at 20°C under 30 µmol of photosynthetically available radiation m² s¹ (12 h of light and 12 h of darkness).

Organic matter content. Blade pieces were dried with a microwave oven, weighted, ignited (450°C, 4 h), and reweighed to determine the organic matter content (34).

Carbon, nitrogen, and phosphorus contents of cordgrass. C and N contents of blades were determined by using a combustive autoanalyzer (Perkin-Elmer model 2400 CHN analyzer) (44). The total phosphorus content of blades was determined by using an inorganic nutrient autoanalyzer (Braun-LuebbeAutoAnalyzerII) after persulfate-autoclaving digestion (2, 22). The efficiency of persulfate extraction was determined by using apple and citrus leaf standards obtained from the National Institute of Standards and Technology, U.S. Department of Commerce. C/N and C/P ratios were calculated as mass ratios.

Statistical procedures. An SPSS statistical software package (48) was used for statistical processing of data. Angular or logarithmic transformations were used with data that were in ratio form or that exhibited statistically significant heteroscedasticity (54). The word "significant" is used below to mean statistically significant (probability of type I error, <0.05 [54]). Least-significant-range testing was performed by using the Student-Newman-Keuls method (54). Means ± standard deviations are given below.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Ergosterol content of cordgrass. Analysis of variance (ANOVA) revealed that the ergosterol contents of upper and mid decaying smooth cordgrass blades were not significantly different ( for upper blades, 521 ± 251 µg g of organic matter¹; for mid blades, 622 ± 219 µg g¹), so the data for the blade categories were pooled for further analysis. There was not a significant difference in the ergosterol contents of smooth cordgrass blades when sites were compared (P = 0.88) (Table 2). Overall, the mean content was 572 ± 237 µg g of organic matter¹. The ergosterol contents of blades were not correlated with latitude (r = 0.13; P = 0.50) (Table 2).

1

1

1

Ergosterol synthesis in cordgrass. The rates of acetate incorporation into ergosterol per unit of ergosterol were not significantly different in different types of blades ( for upper blades, 28 ± 18 pmol µg of ergosterol¹ h¹; for mid blades, 21 ± 14 pmol µg¹ h¹). The rates at different sites were highly significantly different, in contrast to ergosterol contents (Table 2), but like the ergosterol contents, the rates of ergosterol synthesis were not correlated with latitude. The mean rates of ergosterol synthesis were lowest at the site with the highest mean ergosterol content (site MA2) (Table 2) and highest at the site with the lowest mean ergosterol content (site FL3); there was a significant negative correlation between synthesis rate and ergosterol content (r = 0.35; P = 0.03).

1

1

1

1

Ascospores. Because ascospore expulsion values were highly variable for site and year, we pooled the data obtained for Florida and for the two sites north of Cape Cod. An ANOVA revealed that the values for individual or pooled sites were not significantly different (P = 0.58), and there was not a significant correlation between ascospore expulsion values and latitude (r = 0.31; P = 0.25). The coefficient of variation for overall mean ascospore expulsion values was more than 100% (overall , 60 ± 91 spores per mm² of abaxial blade surface per 72 h). The mean ascospore expulsion value was highest in 1996 (158 ± 134 spores mm² 72 h¹) (P = 0.08, as determined by ANOVA). There was not an apparent shift in the species composition of the captured ascospores. The same two major ascomycetous blade decomposers and the same two regular but rarer ascomycetous blade decomposers that were previously found on Sapelo Island (40) were found at all of the sites. The major species were Phaeosphaeria spartinicola and Mycosphaerella species 2 of Kohlmeyer and Kohlmeyer (24), and the regular but rarer species were Phaeosphaeria halima and Buergenerula spartinae; however, no P. halima or B. spartinae was found at sites ME and MA2.

Juncus species. Because two species were involved, J. roemerianus at sites VA1 and VA2 and sites south of these sites and J. gerardi at sites MA1, MA2, and ME, the data for sites were pooled before analysis as follows: data for sites MA1, MA2, and ME; data for sites VA1 and VA2; and data for sites FL1, FL2, and FL3. The ergosterol contents of Juncus samples obtained at different sites were not significantly different (overall , 267 ± 132 µg g of organic matter¹; P = 0.50, as determined by ANOVA). The mean rates of acetate incorporation, however, were more than twofold higher for J. gerardi than for J. roemerianus (but the rates at the sites in Virginia and Florida for J. roemerianus were not significantly different) (J. gerardi rate, 81 ± 38 pmol µg of ergosterol¹ h¹; J. roemerianus rate, 31 ± 20 pmol µg¹ h¹). Although the leaf blades at sites FL1, FL2, and FL3 were larger (diameter, 3.4 ± 0.6 mm) than the leaf blades at sites VA1 and VA2 (diameter, 2.4 ± 0.2 mm), the organic matter density was lower (, 0.18 ± 0.03 mg mm³) for the Juncus blades at sites FL1, FL2, and FL3 than for the blades at sites VA1 and VA2 (0.30 ± 0.04 mg mm³) (P < 0.001). The organic matter density of J. roemerianus at sites VA1 and VA2 was equivalent to the organic matter density of J. gerardi sheaths and stems ( for J. gerardi, 0.31 ± 0.02 mg mm³; diameter, 1.1 ± 0.1 mm).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The hypothesis of Newell (33) and Samiaji and Bärlocher (51) that there should be a south-to-north decrease in the living fungal standing crop and, by extension, a decrease in the potential fungal productivity in standing decaying smooth-cordgrass (S. alterniflora) was shown to be false by the findings described here. The hypothesis was based on visually estimated densities of ascomata (fungal sexual structures) in decaying blades from latitudes north of 44°N. We observed no statistically significant decreases in ascospore expulsion rates (which presumably reflected densities of mature ascomata), fungal standing crop sizes, or rates of ergosterol synthesis at 20°C moving north at sites located from 29 to 43°N. Samiaji and Bärlocher (51) observed small fungal standing crops (maximum mean, 125 µg per g of organic matter consisting of smooth cordgrass leaf blades) (cf. Table 2) for Bay of Fundy marshes (45°N), so it is still possible that a sharp decrease in fungal mass accumulation occurs in standing decaying smooth cordgrass blades at the northern end of the range of this marsh grass. A point at which there could have been a change in fungal activity, based on the known marked shifts in biovariables that occur there, was the boundary between the Acadian and Virginian Provinces (Cape Cod, Mass.; 42°N) (3). However, we obtained mean values for ergosterol content and the ergosterol synthesis rate at 20°C in the Acadian Province that were as high as or higher than some of the mean values obtained for Virginia marshes and marshes that are farther south (Table 2), including the area of maximum development of smooth cordgrass salt marshes in Georgia (40). It appears that substantial potential ascomycetous secondary production based on decay of shoots is the rule wherever smooth cordgrass marshes occur. High rates of autumnal fungal production in northern marshes may be the result of adaptation to short periods when the temperatures permit fungal activity (59). We need to determine levels of fungal productivity at field temperatures at a range of latitudes and in different seasons (instead of our standard 20°C temperature with autumn blades) before we will have a clear picture of the latitudinal range of realized levels of fungal productivity.

Primary productivity of smooth cordgrass is generally limited by nitrogen availability interacting with sediment redox stress (10, 29). There is evidence that nitrogen can limit the efficiency of heterotrophic microbial yields (5) and can limit accumulation of living matter by cordgrass fungi (44). Consistent with these findings was the explanation for part (21%) of the variation in our multilatitudinal data set for the ergosterol content of standing decaying cordgrass. The higher the level of available nitrogen, the more fungal growth and the greater the immobilization of nitrogen, which results in a lower C/N ratio for decaying blades. Higher nitrogen contents of decaying blades were not related to higher instantaneous rates of fungal growth (as determined by acetate incorporation), probably because a higher N content was largely a result of previous fungal immobilization. It has been found that adding nitrogen leads to higher levels of cordgrass fungal productivity, but only at a relatively early point in the decay process (44). There was a hint (P = 0.11) in our data that a low C/P ratio rather than a low C/N ratio was weakly (r² = 0.09) related to higher fungal growth rates; we speculate that this may be a sign that phosphorus can occasionally limit the growth of cordgrass ascomycetes when sufficient nitrogen is present later in the standing decay process (20, 57).

Eighteen percent of the variation (r² = 0.18) in the ergosterol content of our cordgrass blade samples was attributable to variation in the organic matter density of the blades, but this relationship disappeared when the variation attributable to differences in the C/N ratio was factored out. Thus, the relationship between high ergosterol content and low organic matter density was partly due to the tendency of leaves with low organic matter densities to have high nitrogen contents. The relationship between high ergosterol content and low organic matter density could be (i) the result of blades that are more decayed (and thus have lower organic matter densities) having accumulated more fungal matter per unit of remaining decaying system matter (but we did not find a significant difference in ergosterol content between our less-decayed upper blades and more decayed mid blades), and/or (ii) the result of the capacity of leaves with lower lignocellulose contents to permit accumulation of higher living fungal contents. The much lower organic matter densities of the decaying Acadian leaves (sites ME and MA2) (Table 2) probably could be attributed to the trend toward the lower structural carbohydrate contents in living material in more northern cordgrass stands (however, the evidence that this occurs is based on belowground structures [17]). The relative flimsiness of the Acadian cordgrass blades was obvious even to the naked eye and was reflected in the greater flexibility observed during handling of the blades, which implied that the lignin content was low. If the ascomycetes had had less need to lyse very complex lignocellulose polymers in the Acadian blades, they might have had more energy to spend in accumulating mass (per unit of leaf mass) and preparing to expel ascospores. This implies that the ascomycetes in the Acadian blades could have immobilized more nitrogen per unit of decaying leaf matter (40) and so could be consistent with the finding that the ergosterol content was related to the C/N ratio of decaying blades. The linkage between low decaying blade C/N ratios and high latitudes may also be related to the greater (about twofold) output of N into the coastal Atlantic Ocean in the northeastern United States than in the southeastern United States (23).

Since neither latitude nor phosphorus content was associated with differences in ergosterol content, much of the variation in ergosterol content remains unexplained. Part of the variation was due to year-to-year variation (the ergosterol content in 1996 was twofold higher than the ergosterol content in the following 2 years at the three sites sampled in all 3 years). The interannual variation in secondary microbial productivity may have some of the same causes as the interannual variation in the productivity of smooth cordgrass that has been observed (32, 51, 58). Additional potential sources of variation include (i) differences in the duration of decay for the leaves used (i.e., leaf blades may senesce, die, decay, and be lost by breakage at different rates at different sites, so that all of the blades obtained at different sites were not at approximately the same point in the standing decay process, as we had intended); (ii) differences in time spans required for different strains of cordgrass ascomyetes to complete digestion of the blades and/or production of ascomata and/or expulsion of ascospores; and (iii) differences in the impacts of invertebrate leaf grazers on cordgrass ascomycetes. Regarding the third possibility, Graça et al. (21) found that leaf-grazing cordgrass marsh invertebrates have distinctly different effects on cordgrass ascomycetes, ranging from no growth enhancement (periwinkles [Littoraria irrorata]) to twofold enhancement (salt marsh coffeebean snails [Melampus bidentatus]), so different shredder invertebrate compositions (perhaps including species not yet recognized as shredders [40]) could clearly result in variations in the fungal contents of decaying leaves.

It is not likely that differences in the ergosterol contents of ascomycete mycelia were major causes of variation in our ergosterol content data, since species shifting was not observed at different sites (see above) and the ergosterol contents of marsh grass ascomycete species (with no history of maintenance on artificial media) have been found to be homogeneous (the average coefficient of variation was 8% when five species belonging to two large ascomycetous taxa and two marsh grass species were used) (36). However, although temperature probably does not substantially influence the mycelial ergosterol content (19), we cannot rule out the possibility that growth temperature could have an effect on the ergosterol content of marsh grass ascomycetes.

Relatively high ergosterol contents were associated with relatively low rates of fungal production. This relationship has been observed previously with smooth cordgrass (44) and with riparian leaves decaying in streams (18, 55). This relationship probably occurs because the amount of living fungal matter increases until any additional increase in the living fungal matter content of decaying leaves is hindered by limited substrate availability or limited access to one or more nutrients, at which point the growth rate decreases and translocation of fungal organic compounds to points where spores are produced becomes the dominant fungal activity (8). This hypothesis is supported by the finding that as the fungal contents of smooth cordgrass plateaued, the amount of CO₂ released per unit of fungal mass decreased (44).

The high level of noise in our data for rates of ascospore expulsion per square centimeter of blade abaxial surface is curious, yet such noise has appeared (coefficients of variation, >100%) each time that these rates have been measured for standing decaying smooth cordgrass (41, 42). The variation reported for individual sites is variation from blade to blade, but patchiness on 25-mm² spore capture targets for individual blades has also been observed (unpublished data). Some variation in the rate of ascospore expulsion would be expected due to differences in (i) the time elapsed since the onset of fungal pervasion of blades, (ii) the time required to produce mature ascomata, and (iii) the degree of grazing by shredder invertebrates (21). Evidence that these explanations are valid was described by Newell and Wasowski (42). However, high coefficients of variation were obtained even for blades that presumably were essentially identical with respect to the potential three causal factors described above (42). Why should the rates be so patchy? Might it be that a high degree of spatiotemporal patchiness in the rates is advantageous because it helps ensure that a steady stream of new ascospores is released for colonization of new substrates (mature leaf blades)?

Recently, researchers have shown that it is likely that in general fungi dominate secondary microbial production in marine and freshwater standing decaying grass shoots (26, 27, 37, 46, 50, 53). Kohlmeyer et al. (25) found a diverse community of ascomycetous decomposers that are especially adapted to standing decaying blades of black needlerush (J. roemerianus). Therefore, it is not surprising that we found that Juncus species had ergosterol contents of hundreds of micrograms per gram of organic decaying material. Despite the fact that two species were involved (black needlerush and black grass [J. gerardi]), there were not significant differences among the means obtained for Juncus sites from 29°N to 43°N. However, the mean ergosterol content of the Juncus species (267 µg g of organic matter¹) was only about one-half the mean ergosterol content of smooth cordgrass (572 µg g¹), implying that the limits for accumulation of living fungal mass are lower in Juncus species. J. roemerianus decays slowly naturally (9); resistance to build-up of fungal mass may be the principal cause of the slow rate of decomposition. The mean rate of acetate incorporation into ergosterol for black needlerush (30 pmol µg of ergosterol¹ h¹) was in the range of the values obtained for smooth cordgrass (Table 2), but the mean rate for black grass was more than twice as high. One potential explanation for this is that the conversion factor (from acetate incorporation to fungal mass production) for the principal species of fungi that decompose black grass is lower than the conversion factors for the fungi that decompose black needlerush and smooth cordgrass (19, 40).