A gradient analysis of communities of macrofungi and slime moulds on decaying beech logs

575

Mycol. Res. 105 (5) : 575–596 (May 2001). Printed in the United Kingdom.

A gradient analysis of communities of macrofungi and slime moulds on decaying beech logs

Jacob HEILMANN-CLAUSEN* Department of Mycology, University of Copenhagen, Øster Farimagsgade 2D, DK-1353 København K, Denmark. E-mail : jhc!kvl.dk Received 3 January 2000 ; accepted 31 August 2000.

The occurrence of fungi and slime moulds on 70 decaying beech logs was surveyed based on the presence\absence of sporocarps. In total 277 species of fungi and 25 of slime moulds were recorded and summarised in a log\species datamatrix. The structure of the datamatrix was analysed using detrended correspondance analysis (DCA). The ecological nature of the gradients expressed by the first three DCA-axes was then investigated by environmental and log related variables. The first and strongest gradient corresponded to changes in the community development during the decay process. The second gradient was complex and corresponded to both decay rate and microclimatic stress. The third, rather weak, gradient was influenced by soil conditions. The gradients are discussed in a context of fungal ecological strategy theories. A model generalised from the community development on the studied logs is proposed.

INTRODUCTION Increased acceptance of the perception that dead wood is a key element in forest ecosystems (e.g. Harmon et al. 1986, Franklin, Shugart & Harmon 1987, Samuelsson, Gustafsson & Ingelo$ g 1994), has evoked an awareness of the need to understand the biological processes associated with it. Dead wood, which many consider crucially important for biodiversity (Samuelsson et al. 1994, Christensen & Emborg 1996, Ohlson et al. 1997), constitutes a source of nutrients that are gradually released during the decomposition process (Harmon et al. 1986, 1994). Although the process of wood decomposition involves several groups of organisms (i.e. bacteria, protozoa, nematodes, arthropods, and insects), under normal conditions fungi are the primary agents of decay (Rayner & Boddy 1988, Boddy 1992). Decay of large woody units, in particular of whole logs, is a slow process, that may take decades or even centuries to complete (Harmon et al. 1986, Rayner & Boddy 1988). The associated decomposer community development is very complex, even when only fungi are considered (Boddy 1992, Renvall 1995). Microclimate, substrate quality (as determined i.e. by host species and decay stage), and forest history are commonly cited as the most important variables determining the fungal community structure in decaying wood (Gilbertson 1980,

* Present address : Department of Forestry, The Royal Veterinary and Agricultural University, Rolighedsvej 23, DK-1958 Frederiksberg C, Denmark.

Rayner & Boddy 1988, Høiland & Bendiksen 1997, Sippola & Renvall 1999). Several studies have dealt with the separate effect of each factor. For instance, Coates & Rayner (1985a-c), Chapela, Boddy & Rayner (1988), Lange (1992) and Renvall (1995) investigated the community changes occurring during the decay process. Butin & Kowalski (1983, 1986), Griffith & Boddy (1990), and Keizer & Arnolds (1990) compared the wood decaying mycota of several different tree species. Boddy (1983), Boddy, Gibbon & Grundy (1985), and Griffith & Boddy (1991a,b) investigated the effects of microclimate, and Høiland & Bendiksen (1997), Lindblad (1998), and Sippola & Renvall (1999) surveyed the impacts of forest history. As yet, very few studies have tried to evaluate the relative importance of these variables in a multivariate context. Currently several multivariate methods are available for ecological studies (Økland 1990, Kenkel & Booth 1992, Legendre & Legendre 1998). Ordination, widely utilised in vegetation ecology during recent years, has yielded new insights into complex vegetation processes and structures (Økland 1990). In fungal ecological research, ordination has yet to gain popularity. Høiland & Bendiksen (1996) and Lindblad (1998) applied Detrended Correspondence Analysis (DCA) in their studies of fungal communities of decaying conifer logs. However, as the main objective of both studies was to investigate the impact of forestry on the wood mycota, less attention was paid to understanding the general community structures and processes. In the present study, detrended correspondence analysis (DCA) has been used to reveal the relationship between

Macrofungi and slime moulds on beech logs fungal species composition of decaying beech logs (Fagus sylvatica L.) and abiotic and biotic variables. Slime moulds were also included in the study, although they are not considered active wood decayers. The main questions addressed by the study are : (1) Which coenoclines are most important in the studied communities ? ; (2) Do these coenoclines reflect underlying ecological gradients ? ; (3) How do the observed coenoclines\ecological gradients relate to current theories of fungal ecological strategies and decay development ?. The study is entirely based on sporocarp inventory. MATERIAL AND METHODS Site description Suserup Skov (19.2 ha, 7–31 m a.s.l., 55m22hN, 11m34hE) is a near-natural temperate, deciduous forest situated on the island of Zealand in eastern Denmark. The climate is cool-temperate and suboceanic with an annual mean temperature of 8.1 mC and an annual mean precipitation of 635 mm (Emborg, Christensen & Heilmann-Clausen 1996). The forest borders open land with scattered woody vegetation to the north and east, and pastureland with a fairly well-developed woody vegetation to the west. To the south, the forest borders Lake Tystrup. The forest is situated on an undulating elevated plateau to the north with some more or less south-facing slopes towards the lake. The soils are generally deep mull or moder-soils, developed on glacial tills in the elevated parts and on lacustrine deposits towards the lake (Vejre & Emborg 1996). The topsoil pH generally varies from 3.9–4.5, but locally values above 7 are reached (Feilberg 1993, Møller 1997). Dominant tree species are beech (Fagus sylvatica), ash (Fraxinus excelsior), pedunculate oak (Quercus robur), and wych elm (Ulmus glabra), which form a complete and dense canopy in most parts of the forest. The well-developed herb-layer is generally dominated by early flowering herbs such as Anemone spp., Mercurialis perennis and Corydalis bulbosa. Suserup Skov, since 1960 kept as a strict forest reserve, has a long history as a non-intervention forest (Fritzbøger & Emborg 1996). Fossile pollen records have shown that beech has been present at the site for at least 3000 yr with a dominant position in the latest 1000 yr (Hannon, Bradshaw & Emborg 2000). Recent investigations have shown that the forest is now close to a dynamic steady state, with equilibrium

576 between regeneration, mature patches, and breakdown (Emborg, Christensen & Heilmann-Clausen 2000). Dating and selection of logs Initially an attempt was made to establish the time of death (log age) of all dead beech trees with a diameter in breast height (dbh) above ca 60 cm. For this purpose a sequence of high resolution aerial photos were examined. All dead beech known to be dead by 1994 were tracked backwards in the photographic sequence until they were identified as standing, living trees. Log age was successfully determined for 130 beech logs, which were then divided into seven age-classes (1960–67, 1967–72, 1972–78, 1978–81, 1981–1985, 1985–90, 1990–93), and four dbh classes (70–89 cm, 90– 109 cm, 110–129 cm,
130 cm). The dissimilarity in the age classes reflects the unequal distribution of available aerial photos. The seven age classes were used as basis for the selection of logs. Ten logs were selected such that all dbh classes were represented equally within each age class. When conflicts arose, logs were selected randomly. In all 70 logs were selected for the study. Registration of log variables In addition to log age (i) and dbh (ii), several variables were measured or recorded for each log (Tables 1–3). The log type (iii) was determined considering the following four categories ; log uprooted with a distinct root plate, log broken at root neck, log broken 2–7 m above ground-level, log broken 8–15 m above ground-level. The number of fractures (iv) and forks (v) on each log, including major branches, diam equal to or larger than 50 cm, were simply counted. The medium decay class (vi), soil humidity (vii), wind exposure (viii), and sun exposure (ix), were subjectively judged according to the parameters given in Tables 1–2. The bark cover (x), moss cover (xi), and the distribution of stromata and superficially black-stained wood caused by Eutypa spinosa (xii) were estimated to the nearest 10 % of the total log surface. The degree of soil contact was estimated to the nearest 10 % of the log length (xiii). The distance to the nearest forest edge (xiv) was measured to the nearest 5 m on a stem position map. The soil type (xv) was classed in accordance with Vejre & Emborg (1996) with the following exceptions : no distinction was made between flat, undulating

Table 1. Characteristics used for the classification of logs in decay classes. Characteristics Class 1 Class 2 Class 3 Class 4 Class 5

Wood hard, a knife (with a thin blade) penetrates only a few mm into the wood, bark intact, twigs (diam 1 cm) intact. Wood rather hard, a knife penetrates less than 1 cm into the wood, bark starting to break up, twigsplost, branches (diam 1–4 cm) intact. Wood distinctly softened, knife penetrates ca. 1–4 cm into the wood, except for parts colonized by certain pyrenomycetes (in particular Eutypa spinosa, Kretzschmaria deusta and Xylaria hypoxylon), bark partly lost, branchesplost, original log circumference intact. Wood strongly decayed, knife penetrates ca. 5–10 cm into the wood, except for parts colonized by certain pyrenomycetes (see above), bark lost in most places, original log circumference disintegrating. Wood very strongly decayed, either to a very soft crumbly substance, or being flaky and fragile with numerous remnants of pseudosclerotial plates, these defining the log surface, knife penetrates in most places more than 10 cm into the wood, original log circumference not or hardly recognizable.

Jacob Heilmann-Clausen

577

Table 2. Characteristics used for the classification of logs in classes of soil humidity, wind exposure and sun exposure. Soil humidity

Wind exposure

Sun exposure

Class 1

Well drained forest soils on hill tops.

Exposed forest at edges and on hills.

Class 2 Class 3 Class 4

Other high, well drained forest soils. Normal, well drained forest soils. Low, pwaterlogged soils.

Class 5

Swamp soils, pinundated during wintertime.

Open forest, without a proper understorey Normal forest, with sparse understorey. Normal forest, with a well developed understorey. Well sheltered places, protected by a dense understorey.

Forest edges and open forest on south facing slopes. Open, sunlit forest. Normal forest on level areas. Normal forest, either with a dense canopy or on north facing slopes. High forest with a dense canopy and on north facing slopes.

Table 3. Log variables measured or estimated for each log. Variable

Data type

Scale

Range

Log age Dbh. Log type Fractures Bole forks Decay stage Soil humidity Wind exposure Sun exposure Bark cover Moss cover Eutypa cover Soil contact Dist. to edge Soil type Plant cover Floristic deviance Plant diversity Decay rate

Ratio Ratio Nominal Ratio Ratio Ordinal Ordinal Ordinal Ordinal Ratio Ratio Ratio Ratio Ratio Nominal Ratio Ratio Ratio Ordinal

Continuous Continuous 4 levels Count Count 5 levels 5 levels 5 levels 5 levels 12 levels 12 levels 12 levels 11 levels Continuous 3 levels 11 levels 12 levels Count 5 levels

Median

Unit

2–31 70–168

15 110

0–5 0–4 1–5 1–5 1–5 1–5 0–100 0–60 0–70 10–100 5–135

2 1 3 4 4 4 10 5 20 90 70

20–100 0–80 1–6 k2-3

90 * 2 0

Years cm Class Number Number Class Class Class Class % % % % m Class % % Number Class

* The variable, floristic deviance, is further subdivided in eight variables, each representing a plant species ; Anemone nemorosa (Ane nem), Anemone ranuculoides (Ane ran), Corydalis bulbosa (Cor bul), Galium odoratum (Gal odo), Lamiastrum galeobdolon (Lam gal), Mercurialis perennis (Mer per), Ranunculus ficaria (Ran fic), Viola riviana\reichenbachiana (Vio riv), and a ninth varible (Other species) with the remaining species (Alium ursinum, Arum maculatum, Corydalis intermedia, Gagea lutea, Pulmonaria obscura, and Sanicula europaea).

and steep soils ; lacustrine sand and lacustrine clay soil types were pooled since both were very sparingly represented among the logs studied. The presence or absence of 14 plant species (i.e. Allium ursinum, Anemone nemorosa, A. ranunculoides, Arum maculatum, Corydalis bulbosa, C. intermedia, Ranunculus ficaria, Gagea spathacea,Galium odoratum, Lamiastrum galeobdolon, Mercurialis perennis, Pulmonaria obscura, Sanicula europaea, and Viola riviana (Including V. reichenbachiana), was noted around each log in May 1995. The relative distribution of each plant species and of vegetation-less areas was estimated to the nearest 10 % of the forest-floor area surrounding the individual log, and the total degree of plant cover (xvi) was evaluated. The variable floristic deviance (xvii) was derived by considering the forest floor area covered by plants other than the generally dominant A. nemorosa, whereas plant diversity (xviii) was obtained by totalling the number of plant species occurring in the vicinity of each log. The decay rate (xix) was estimated by comparing for each log, the log age with the decay stage. The logs were reinventoried in 1999 for a more precise understanding of the log specific decay development. In this re-inventory an additional decay class (class 6) was introduced for almost

completely decomposed logs. Within each age class the average decay class was calculated. Logs with average or close to average decay class scores were given the decay rate score 0. Correspondingly, logs with a deviating decay class score were given decay rate scores ranging from 1 to 3 to denote the strength of deviance and with a plus (j) or a minus (k) as prefix to denote whether the log was decayed more or less than to be expected. Sporocarp inventory Sporocarps were inventoried on all logs on ten separate occasions between May 1994 and December 1995 (1994 : 8–9 May, 29 May, 10 July, 23–24 Aug, 4 Oct. ; 1995 : 24 Feb., 25 Sep., 1 Oct., 15 Oct, 14 and 17 Dec.). The entire log was investigated, including the stump and branches thicker than 10 cm. Sporocarps were identified in situ or collected for later identification. The September and October 1995 inventories included only species determinable in the field. All basidiomycete and slime mould taxa found were included, but non-stromatic pyrenomycetes were excluded from the ascomycete inventories. Due to the project-scale it was considered unrealistic to inventory all species on all logs. In

Macrofungi and slime moulds on beech logs all likelihood therefore, genera producing very inconspicuous sporocarps (e.g. Basidiodendron, Botryobasidium, and Tulasnella within the basidiomycetes ; and Arachnopeziza and Hyaloscypha within the ascomycetes) are under-represented. Additionally, in the final analysis some common species were treated in an uncritical or broad sense so as to reduce microscopical work. Galerina marginata, Scutellinia scutellata, Mollisia cinerea, Athelia epiphylla, and Peziza micropus were all determined using macroscopical characters, although microscopical examination is necessary for a critical determination of these taxa. This uncritical approach was chosen after microscopical examinations of dozens of specimens in each species-group, failed to reveal more than one species in each case. Furthermore the concept of Bisporella citrina used here is broad and includes a potentially undescribed taxon, microscopically differentiated by asci regularly shorter than 100 µm. Voucher specimens were obtained for all species, with a few exceptions (Fomes fomentarius, Phallus impudicus and other very well-defined species) and are kept at the Botanical Museum at the University of Copenhagen (C). A number of specimens lost after determination have been replaced by newly collected voucher material. Nomenclature Basidiomycete nomenclature and species concepts follow Hansen & Knudsen (1992, 1997), except for the Sistotrema brinkmanii group (cf. Hallenberg 1984), Tomentella (cf. Ko! ljalg 1996) and Stypella and Exidia (cf. Roberts 1998). Ascomycete nomenclature and species concepts follow Hansen & Knudsen (2000), except for instances where specialized discomycete literature has been consulted. Slime mould nomenclature follows Nannenga-Bremekamp (1991). Data Analyses Detrended correspondance analysis (DCA) was chosen as the ordination method for several reasons. DCA is easy to access and assumes unimodal species\gradient response curves, in contrast to Principal Component Analysis (PCA) which assumes linear responses. Linear species\gradient responses appear to be rare in most complex communities, and PCA has in general appeared to perform badly on vegetation data (Økland 1990). DCA enables samples (logs) and species to be ordinated simultaneously in the same ordination space following an iterative procedure, which aims to optimise the dispersion of species optima. Sample scores are derived from species scores by weighted averages. The ordination space is defined by a number of ordination axes, which in DCA are scaled in SD units. For more details on DCA see Hill & Gauch (1980) or Økland (1990). The relative powers of DCA axes can be established by comparing eigen values and gradient lengths. The gradient length indicates the amount of β-diversity (species turnover) expressed along the individual axis, whereas the eigenvalue is a measure of the amount of variation extracted by the axis. The eigenvalue is most informative when compared with eigenvalues of other axes in the same analysis. Økland (1999) and Ejrnæs & Bruun (2000) have shown that the commonly

578 used measures of total inertia and explained variation have little significance. Hence these measures are omitted in the present paper. DCA was executed using the program PC-ORD 3.0 (McCune & Mefford 1997) using default options. Downweighting of rare species was avoided in final ordinations, since this tended to weaken correlations with environmental variables. However, species occurring on two or fewer logs were excluded from the ordination. Stability of the found DCA-axes was evaluated by subset ordination as introduced by Laweson (1997) and Ejrnæs & Bruun (2000). Three subsets, derived from the full data set, were analysed : (1) Slime moulds excluded ; (2) Common species (species occurring on 6 logs or fewer excluded) ; and (3) Subordinate species (species occurring on 20 logs or more excluded). The sample scores from each subset ordination were regressed to the sample scores obtained from the full set ordination, in order to evaluate the stability of the full set coenoclines in the three variously diminished data sets. Significance of correlation between the best correlated axes in each subset and the axes of the full data set was tested with Pearson product moment correlation (Zar 1999). As DCA is entirely based on species data, DCA-axes are per se pure coenoclines which are not associated with any kind of environmental information. However, there are several ways in which coenoclines can be interpreted indirectly in environmental terms (Økland 1990). In this study relationships between DCA-axes and environmental variables were basically evaluated by Kendall rank correlation. Rank correlation was chosen in favour of its parametric alternatives, since several of the recorded environmental variables are on an ordinal scale or are not normally distributed. Kendall rank correlation was favoured over Spearman rank correlation, since the latter method is less efficient in the presence of many tied ranks (Legendre & Legendre 1999). Subsequently multiple regression was applied in order to investigate the complex structures of the ecological gradients expressed along the DCA-axes. Backward elimination (see Zar 1999) was used to evaluate which variables should be selected for the multiple regression models. In the case of curvilinear relationships, polynomial regression was applied. The two nominal variables, soil type and log type, were coded as ‘ dummy ’ variables (see Zar 1999). To ease interpretation of DCA-axes in respect to fungal community development, the recorded fungal species were subdivided into six morphogroups (see Table 4). For each group the relative contribution to overall species richness was calculated per log, prior to correlation analysis. All regression and correlation analysis was run using the program JMP for Macintosh (SAS Institute 1997). RESULTS In total 302 taxa, 277 species of fungi and 25 species of slime moulds, were recorded during the study (Table 4). More than one third (124) of the recorded species were found on one or two logs only and were hence omitted from the ordination. Fifteen species were found on 35 logs or more.

Jacob Heilmann-Clausen

579

Table 4. Total species list subdivided in morphological groups. Note that several heterobasidiomycetes are assigned as corticioid fungi, due to their thin, apressed sporocarps.

Abbreviation in DCA-diagrams

campoly diadisc

eutspin hypgela hypcoha hypfrag hyprubi kredeus neccocc

nematro nemches nemserp xylhypo xylpoly

asccyli biscitr

lacbrev lacimpu lacvirg molcine neopura

orbdeli pezmicr polprui scuscut

athepip

byscori cerexce cerreti conpute

cyllaev

gloanal

Species

Number of logs

Collection-numbers of selected voucher specimens preserved at the Botanical Museum, Copenhagen (C.)

PYRENOMYCETES Camarops polysperma C. tubulina Diatrype disciformis D. flavovirens D. stigma Eutypa spinosa Eutypella quaternata Hypocrea gelatinosa Hypomyces aurantius Hypoxylon cohaerens H. fragiforme H. rubiginosum Kretzschmaria deusta Lopadostroma turgidum Nectria coccinea N. episphaeria N. peziza Nemania atropurpurea N. chestersii N. serpens Protocrea cfr. farinosa Xylaria hypoxylon X. polymorpha

9 2 5 1 1 62 2 4 1 3 15 15 45 1 4 1 2 4 9 6 1 42 17

JHC94-350 C C JHC94-015 C JHC94-139 ; JHC94-351 C C JHC94-146 C C C JHC94-009 C C JHC94-159 ; JHC94-019 ; JHC94-140 ; JHC94-352 ; JHC94-010 C C

DISCOMYCETES Arachnopeziza aurata A. variepilosa Ascocoryne cylichnium Bisporella citrina Cystopezizella conorum Dasyscyphella nivea Hyaloscypha fuckelii Lachnum brevipilosum L. impudicum L. virgineum Mollisia cinerea M. ligni Neobulgaria pura Ombrophila sp. Orbilia cfr. alnea O. delicatula O. epipora Peziza micropus Polydesmia pruinosa Scutellinia scutellata Tapesia lividofusca

3 1 8 10 1 1 1 5 22 4 49 1 6 1 2 19 2 26 12 29 1

JHC94-420 JHC94-035 C JHC95-224 ; JHC95-253 ; JHC95-254 JHC94-359 JHC95-202 JHC94-036 ; JHC95-210 JHC95-252 JHC94-069 ; JHC95-201 JHC94-360 JHC94-432 JHC95-206 C JHC94-421 JHC94-430 ; JHC94-150 JHC94-101b ; JHC94-423 JHC94-037 ; JHC94-102 ; JHC94-103 C JHC95-221 C JHC95-209

CORTICIOID FUNGI Athelia epiphylla Athelopsis glaucina Basidiodendron caesiocinereum Byssomerulius corium Ceriporia excelsa C. reticulata Coniophora arida C. puteana Cristinia gallica C. helvetica Cylindrobasidium laeve Exidiopsis effusa Gloeocystidiellum luridum Gloeohypochnicium analogum

37 1 1 3 8 38 2 17 2 2 7 1 1 3

JHC95-248 ; JHC95-249 JHC94-057 JHC95-207 JHC95-240 JHC94-198 ; JHC94-344 ; JHC95-250 C JHC95-195 C JHC94-183 JHC94-081 JHC95-241 JHC95-217 JHC94-047 JHC94-021 ; JHC95-255

JHC94-144

JHC94-160 JHC94-062 ; JHC94-356 JHC94-143 ; JHC94-162 JHC94-354 ; JHC94-357

Macrofungi and slime moulds on beech logs

580

Table 4 (cont.)

Abbreviation in DCA-diagrams glolact

hypprae hyppube hyprose hypseti hypargu hyppara hypradu hypsuba

hyppunc

pencine peninca penlyci

phllivi phlradi phlrufa phltrem radconf scorimo sisbrin siscoro

sissern steochr stehirs sterugo stygril stysubh sublong thafusi

tomsubl trecoha trefari trehyme

trestev tultoma tulviol

Species

Number of logs

Collection-numbers of selected voucher specimens preserved at the Botanical Museum, Copenhagen (C.)

Gloiothele lactescens Hyphoderma argillaceum H. medioburiense H. pallidum H. praetermissum H. puberum H. roseocremeum H. aff. roseocremeum H. setigerum Hyphodontia arguta H. aspera H. paradoxa H. radula H. sambuci H. subalutacea Hypochnicium eichleri H polonense H. punctulatum Lindtneria cfr. flava Mycoacia uda Peniophora cinerea P. incarnata P. lycii P. nuda P. quercina Phanerochaete sordida P. tuberculata P. velutina Phlebia livida P. radiata P. rufa P. tremellosa Phlebiella allantospora Radulomyces confluens Scopuloides rimosa Sistotrema biggsiae S. brinkmannii S. coroniferum S. oblongisporum S. octosporum S. sernanderi Sistotremastrum niveocremeum Steccherinum ochraceum Stereum hirsutum S. rugosum Stypella griletti S. subhyalinum Subulicystidium longisporum Thanatephorus fusisporus Tomentella ferruginea T. radiosa T. sublilacina Trechispora cohaerens T. farinacea T. hymenocystis T. microspora T. praefocata T. stevensoni Tulasnella eichleriana T. tomaculum T. violea Xenasma pulverulentum

19 2 1 2 6 17 4 2 9 6 2 6 16 1 3 1 2 3 1 1 3 8 3 1 1 1 1 1 10 13 2 6 1 20 4 1 20 3 1 2 5 2 16 21 7 4 4 12 3 1 1 5 9 7 4 1 1 7 1 3 4 1

C JHC94-074 ; JHC94-111 ; JHC95-192 JHC94-025 JHC94-024 ; JHC94-123 JHC94-115 ; JHC95-189 ; JHC94-075 JHC95-214 ; JHC95-215 JHC94-026 ; JHC94-132 ; JHC95-178 JHC95-179 C JHC94-122 ; JHC95-233 JHC95-216 JHC94-022 ; JHC94-076 JHC95-185 JHC94-027 JHC94-029 ; JHC94-113 JHC95-213 C JHC95-232 C JHC95-247 JHC95-186 C JHC94-056 C JHC95-204 JHC95-187 ; JHC95-226 JHC94-422 JHC95-229 C JHC95-184 JHC94-030 ; JHC95-219 ; JHC94-085b JHC94-058 JHC95-235 ; JHC95-237 ; JHC95-181 ; JHC95-182 JHC95-188 JHC94-125 JHC95-180 ; JHC95-239 JHC95-234 C C JHC95-243 JHC94-334 ; JHC95-205 JHC94-033 ; JHC94-336 JHC94-128 JHC94-106 JHC94-105 JHC94-196 JHC94-031 ; JHC94-341 JHC94-089 ; JHC95-244 ; JHC94-088 ; JHC94-109 JHC94-095-227 JHC95-203 JHC95-177 JHC94-092 ; JHC94-096 JHC95-208 JHC94-034 ; JHC94-036 ; C JHC94-117

JHC94-130

JHC95-242

JHC95-228

JHC95-220 ;

JHC95-238

JHC95-245 ;

JHC94-335

Jacob Heilmann-Clausen

581

Table 4 (cont.)

Abbreviation in DCA-diagrams

bjeadus cergilv datmoll Fomfome fompini fusferr ganlips inonodu iscresi mergiga physang phyvitr polbrum polcili polsqua poltube polvari

trahirs travers

armgall bolreti clidiat clinebu cliphyl clihobs colbuty

conbrun contene copdome

copmica

galunic gymjuno hemcucc hencand hydsuba hypfasc kuemuta

Collection-numbers of selected voucher specimens preserved at the Botanical Museum, Copenhagen (C.)

Species

Number of logs

POLYPORES Antrodiella hoehnelii A. semisupina Bjerkandera adusta B. fumosa Ceriporiopsis gilvescens Datronia mollis Fomes fomentarius Fomitopsis pinicola Fuscoporia ferrea Ganoderma lipsiensis G. feifferi Inonotus nodulosus Ischnoderma resinosum Meripilus giganteus Physisporinus sanguinolentus P. vitreus Polyporus badius P. brumalis P. ciliatus P. squamosus P. tuberaster P. varius Skeletocutis nivea Trametes gibbosa T. hirsuta T. versicolor Tyromyces chioneus Tyromyces wynnei

2 2 17 1 9 6 38 6 5 22 1 6 14 18 16 14 1 3 3 3 3 13 2 2 7 20 1 1

C C C JHC95-197 JHC95-251 C

AGARICS and CYPHELLOIDS Armillaria gallica A. mellea Bolbitius reticulatus Clitocybe diatreta C. nebularis C. phyllophila Clitopilus hobsonii Collybia butyracea C. fusipes C. peronata C. tuberosa Conocybe brunnea C. sordida C. tenera Coprinus domesticus C. echinosporus C. lagopus C. micaceus C. xanthothrix Cystolepiota adulterina C. hetieri Entoloma dichroum Flammulaster muricatus Galerina nana G. unicolor Gymnopilus junonius Hemimycena cucullata Henningsomyces candidus Hydropus subalpinus Hypholoma fasciculare Inocybe petiginosa Kuehneromyces mutabilis Laccaria laccata

44 1 7 21 9 3 10 8 1 1 2 3 1 3 3 1 1 36 1 1 1 1 1 1 24 3 3 5 3 35 1 15 1

C C C JHC94-166 C C JHC94-424 C C

C JHC95-194 C C C C C C JHC94-199 C C C C C C JHC94-349 C C C JHC95-093 C

C JHC94-183 JHC94-185 JHC94-184 ; JHC95-147 C JHC94-054 C C JHC94-426 C JHC94-187 JHC94-361 JHC94-370 ; JHC95-092 JHC94-181 C C JHC94-167 C C C JHC94-186 C C

Macrofungi and slime moulds on beech logs

582

Table 4 (cont.)

Abbreviation in DCA-diagrams lacsubd lepaspe lepcris lepfulv

lepflacc macrhac maralli marrotu martorq megplat

mycamic myccroc mycdios mycgale mycgalo mychaem mychiem mycpeli mycpura mycrena mycrose mycspei myctint ompepic oudmuci pansero phoauri phosqua Pleoste plucerv pluchry

pluinqu plunanu pluphle plusali plusemi pluthom pluumbr

psaobtu psapilu

Species

Number of logs

Lactarius subdulcis Lepiota aspera L. boertmannii L. cristata L. fulvella L. jacobii L. ochraceofulva Lepista flaccida L. nuda Macrolepiota rhacodes Marasmius alliaceus M. cohaerens M. rotula M. torquescens M. wynnei Megacollybia platyphylla Melanophyllum haematospermum Melanotus horizontalis Merismodes anomalus Micromphale brassicolens Mycena amicta M. crocata M. diosma M. erubescens M. galericulata M. galopus M. haematopus M. hiemalis M. pelianthina M. polygramma M. pura M. renata M. rosea M. speirea M. tintinabulum Omphalina epichysium Oudemansiella mucida Panellus serotinus Pholiota aurivellus P. squarrosa Pleurotus dryinus P. ostreatus Pluteus cervinus P. chrysophaeus P. ephebeus P. godeyii P. inquilinus P. nanus P. phlebophorus P. podospileus P. salicinus P. semibulbosus P. thomsonii P. umbrosus Psathyrella candolleana P. conopilus P. cortinarioides P. fusca P. lacrymabunda P. obtusata P. piluliformis P. populina P. pygmaea

4 6 1 7 4 1 2 14 1 12 40 1 27 9 1 22 2 1 2 1 3 35 6 2 28 4 50 3 5 2 17 11 4 4 7 8 14 8 6 2 1 10 37 11 2 1 4 5 22 1 8 4 3 5 1 1 1 1 1 19 8 1 1

Collection-numbers of selected voucher specimens preserved at the Botanical Museum, Copenhagen (C.) C C JHC95-261 C C C C C C C C C C C C C C JHC95-191 JHC95-198 C JHC94-053 C C C C C C C C C C C C C C C C C JHC95-155 C C C JHC94-362 JHC94-176 JHC94-174 ; JHC95-094 JHC94-178 ; JHC94-363 JHC94-050 JHC94-175 C JHC94-171 ; JHC94-172 ; JHC95-154 JHC95-153 C JHC94-098 JHC94-188 JHC94-191 JHC94-192 JHC94-425 JHC95-148 ; JHC95-212 JHC94-365 JHC94-193

Jacob Heilmann-Clausen

583

Table 4 (cont.)

Abbreviation in DCA-diagrams psarost

psecyat

strcyan

xerradi

calcorn exiglan exinucl hercora lycperl lycpyri mutcani phaimpu phlfagi ramstri schcomm tremese

arccine arcdenu cerfruc

criargi entlyco fulsept lamarcy lycepid

phycine steaxif stefusc stetyph

triscab trivari

Species

Number of logs

Collection-numbers of selected voucher specimens preserved at the Botanical Museum, Copenhagen (C.)

P. rostellata P. spadiceogrisea P. tephrophylla Pseudoclitocybe cyathiformis Resupinatus trichotis Ripartites tricholoma Simocybe centunculus Stropharia cyanea Tricholoma lascivum Tubaria furfuracea Volvariella hypopithys Xerula radicata

4 1 1 9 1 1 1 3 1 2 1 10

JHC94-194 ; JHC94-195 JHC94-364 C C C C JHC94-368 C C C JHC95-086 C

OTHER BASIDIOMYCETES Auricularia auricula-judae Calocera cornea Cacryomyces stillatus Exidia glandulosa E. nucleata Hericium coralloides Lycoperdon echinatum L. perlatum L. pyriforme Mutinus caninus Phallus impudicus Phleogena faginea Ramaria stricta Schizophyllum commune Tremella mesenterica

1 6 1 4 3 4 1 13 36 3 5 11 15 5 7

C

SLIME MOULDS Arcyria affinis A. cinerea A. denudata A. incarnata Ceratiomyxa fruticulosa Comatricha alta C. nigra Cribraria argillacea Enteridium lycoperdon Fuligo septica Lamproderma arcyrioides Lycogala epidendrum s.l. Metatrichia floriformis M. vesparium Physarum cinereum Stemonitis axifera S. fusca Stemonitopsis hyperopta S. typhina Symphytocarpus amaurochaetoides Trichia contorta T. persimilis T. scabra T. varia Tubifera ferruginosa

1 3 4 2 24 1 2 5 18 19 3 35 1 2 1 7 6 1 6 2 1 1 10 29 1

JHC95-260 C C JHC94-099 C JHC94-002 JHC94-100 JHC94-082 JHC94-361 C JHC95-258 ; JHC95-259 C C JHC95-257 JHC94-094 C C JHC94-090 C JHC94-101 JHC95-256 C JHC94-429 JHC94-427 ; JHC94-428 C

JHC95-200 JHC95-199 C C C C C

C C C C

A ‘ C ’ in the voucher specimen column indicates that the species is preserved at C from the locality, but not from the present investigation.

The number of fungal species per log averaged 30.6, with the highest numbers recorded on logs in intermediate age classes (Fig. 1). On the single-most species-rich log, 78 species were recorded. Logs 5–22 yr old were found to produce significantly more species than both older and younger logs (ANOVA on ranks ; Newman–Keuls test, P 0.05).

The highest number of slime mould species, however, was recorded on slightly older logs (Fig. 2), with 10–22 yr old logs found to be most species-rich (ANOVA on ranks ; Newman– Keuls test P 0.05). Table 5 shows correlation coefficients between the investigated variables. Basically three groups of intercorrelated

Macrofungi and slime moulds on beech logs

584 variables are distinguishable. The first group consists of log size and outline variables (dbh, fractures and bole forks), the second decay related variables (log age, decay class, bark cover and soil contact), whereas the third, complex set, is constituted by variables related to microclimatic conditions (distance to edge, sun exposure, wind exposure, soil humidity, and moss cover). Ordination

Fig. 1. The average number of fungal species per log in the seven age classes. The bars represent 95 % confidence limits. Number of logs l 10 in each class, except for logs 1–4 years old, which are only represented by nine logs.

After a visual inspection of the initial DCA scattergrams, one obvious outlier log was omitted from further analysis. Outliers are potentially interesting, but may cause serious distortions overruling the general patterns of the ordination. The excluded log was newly dead, but had extensively decayed parts. The fact that the recorded mycota were associated with both initial and late stage decay may explain the log’s aberrant performance. Table 6 summarises the properties both of the full data set (after exclusion of the outlier), and the three subsets. General properties of the DCA-axes

Fig. 2. The average number of slime mould species per log in the seven age classes. The bars represent 95 % confidence limits. Number of logs l 10 in each class, except for logs 1–4 years old, which are only represented by nine logs.

The DCA-analysis resulted in three interpretable axes, of which the first have a considerably higher eigenvalue (0.42) and hence more explanatory power compared to the two subsequent axes (eigenvalues 0.15 and 0.13). The gradient length scores are more steadily decreasing from DCA1 (3.4 SD units) to DCA2 (and 2.5 SD units) and DCA3 (1.8 SD units), indicating that despite the low eigenvalue, DCA2 is still associated with a distinct β-diversity gradient. Scatterplots of DCA1–3 in the full data set regressed against the best correlated axes in the subset ordinations (Fig. 3) support the distinct hierarchic ranking of the ordination axes, with DCA1 being consistently expressed in all subsets. Also, DCA2 reproduced fairly well in all subsets, supporting that this axis express an important gradient in the communities studied. In contrast, DCA3 is expressed only in the subset obtained by excluding slime moulds, indicating that this axis lacks stability. It must seem confusing that the rankings of DCA axes in the subsets differ from the rankings in the full data set (e.g. DCA2 in the full data set is best correlated with DCA3 in the common species subset). It is, however, a

Table 5. Kendall rank correaltions between variables.

Dbh. Bole forks Fractures Log age Decay stage Bark cover Soil-contact Dist. to edge Wind exposure Sun exposure Soil-humidity Moss cover Eutypa cover Plant cover Plant diversity Floristic deviance Decay rate

**** P

Dbh.

Bole forks

Fractures Log age

Decay stage

Bark cover

1 0n41**** 0n35*** 0n03 0n12 k0n03 0n16 0n22** 0n19* 0n12 0n07 0n24* k0n04 0n01 k0n12 k0n06 0n07

1 0n46**** k0n18 k0n17 0n15 k0n01 0n06 0n04 0n04 k0n01 0n17 0n99 k0n07 k0n01 k0n07 0n06

1 k0n02 0n06 0n01 0n19 0n06 0n12 0n14 0n12 0n32** 0n13 k0n03 k0n14 k0n07 0n21*

1 k0n81**** 0n50**** k0n02 k0n09 0n25* 0n16 0n19 0n11 k0n06 0n21* k0n12 0n03

1 k0n43**** 1 0n06 0n07 1 0n12 0n02 0n51**** 1 k0n19 0n16 0n13 0n31** 1 k0n09 0n20* k0n09 0n20* 0n14 k0n19 0n25* 0n12 0n35*** 0n26* k0n18 k0n08 k0n03 k0n11 0n05 0n02 0n03 0n21* 0n19 0n05 0n1 k0n15 k0n23* k0n01 k0n11 0n04 k0n08 0n22* 0n1 k0n04 0n06 0n28** 0n15 0n13 0n09

0n0001 ; *** P l 0n0001k

1 0n80**** k0n77**** 0n38**** k0n07 k0n17 0n19* 0n08 0n18 0n21* 0 k0n08 k0n02 k0n21*

0n001 ; ** P l 0n001k

Soilcontact

Dist. to edge

0n01 ; * P l 0n01k

Wind Sun SoilMoss exposure exposure humidity cover

0n05.

Eutypa cover

Plant cover

Plant diversity

Floristic Decay deviance rate

1 0n34** 1 k0n13 0 1 k0n01 k0n02 0n05 1 0n08 0n08 k0n01 0n08 1 0n09 0n14 0n01 0n08 0n63**** 1 0n07 k0n07 k0n27** k0n07 k0n29** 0n21*

1

Jacob Heilmann-Clausen

585

Table 6. Properties of the full data set and of the three investigated subsets.

Number of samples (logs) Number of species Number of non-zero data items

Full set

Slime moulds excluded

Common species

Subordinate species

69 179 2041

69 164 1957

69 96 1794

69 146 1082

Fig. 3. Evaluation of DCA coenoclines by ordinations of subsets. The three columns represent scatterplots of the three DCA-axes of the full set against the best correlated DCA-axes in the subset ordinations. Coefficients of the Pearson product moment correlation are indicated within the plots. Axes are scaled in SD units.

Macrofungi and slime moulds on beech logs subset a

586

subset b

Fig. 4. Distribution of logs in the DCA1\DCA2 ordination space. The dotted line shows the cut-level dividing subset a from subset b.

by comparing their DCA2\log variable correlation quotients with corresponding quotients obtained for the full data set (Table 7). In subset b only one log variable (sun exposure) is significantly correlated with DCA2. In subset a, on the contrary, DCA2 correlates significantly with all variables correlated with DCA2 in the full data set. This indicates that the tongue represents a true structure in the data set and that DCA2 expresses an ecological gradient only expressed at the lower end of DCA1 (Økland 1990, ‘ type c. tongue ’). Accordingly, logs with high DCA1 scores are likely to represent noise in the interpretation of DCA2. A separate ordination of the subsets, as suggested in similar cases by Peet (1980) and Økland (1990), was undertaken but did not increase the informative power, compared to the separate analyses of subsets of the full ordination. This is probably a consequence of the lowered amount of informative data in the diminished data sets. Interpretation of the DCA-axes

Table 7. Kendall rank correlations between DCA2 and log variables in subset a and subset b compared to the full data set.

Log age Decay rate Wind exposure Dist. to dge Fractures Other plants Bole forks Decay stage Bark cover Ane nem Sun exposure Floristic deviance Plantdiv

Subset a (n l 41)

Subset b (n l 28)

Full set (n l 69)

0n5156**** k0n4976**** k0n4661*** k0n3862*** k0n3768** 0n3447** k0n3059* 0n3004* k0n2915* k0n2749* k0n2649* 0n2613* 0n2548*

k0n0770 k0n0193 k0n2320 k0n1853 0n1144 0n0176 k0n0156 k0n1210 0n1089 k0n1490 k0n3046* 0n0958 0n1902

0n2346** k0n3544*** k0n3647**** k0n3134*** k0n2162* 0n1963 k0n2337* 0n1413 k0n1467 k0n2150* k0n2290* 0n1830* 0n2348**

Only variables with significant correlation coefficients in a set are shown. **** P 0n0001 ; *** P l 0n0001k 0n001 ; ** P l 0n001k 0n01 ; * P l 0n01k 0n05.

common feature in DCA, that subordinate axes change ranking or becomes intermixed if the data set is diminished. A plot of DCA1 against DCA2 in the full data set (Fig. 4) reveals a distinct uneven distribution of logs in the ordination space. In the right half of the ordination diagram, the logs cluster closely around a horizontal line, while they are much more widely scattered along DCA2 in the left half. Such a sample distribution is not uncommon in DCA and has been denoted a ‘ tongue ’ (e.g. Økland 1990). In many cases tongues represent true data set structures, but in cases of aberrant data sets, they may be artefacts generated by the DCA-algorithm (Økland 1990). The nature of the DCA tongue was evaluated by dividing the samples (logs) in to two subsets based on their positions in the DCA-diagram, as suggested by Økland (1990), followed by a careful examination of both subsets. Two subsets were constructed by referring logs with DCA1 scores below 1.25 SD units to subset a, whereas subset b contains logs with DCA1 scores above 1.25 SD units. The subsets were examined

In Fig. 5 various key variables are plotted in the DCA1\DCA2 ordination space and Table 8 shows correlation coefficients between DCA axes and variables, both for the full data set and subset a. DCA1 correlates strongly with all decay-associated variables and can readily be interpreted as reflecting a decay process gradient. Below 1.25 SD units (subset a) the correlation to log age, bark cover and soil contact ceases. Decay stage is, however, still distinctly expressed. DCA2 reflects a combined microclimatic stress and decay rate gradient. Both in the full data set and in subset a the axis correlates significantly with several variables, most notably wind exposure, distance to edge and decay rate. In subset a, however, log age is the most highly correlated variable. DCA3 is associated with a turnover in several floristic variables (plant cover, plant diversity and floristic deviation), probably relating to a soil gradient. More specifically, the frequency of Anemone ranunculoides and Ranunculus ficaria subsp. bulbifera is positively correlated with the axis. The strong intercorrelation among several variables in the material (Table 5) obscures the complex environmental gradient structures behind the axes, and prevents a quantitative interpretation of these, since correlation quotients are in no way additive. Multiple regression was applied in an attempt to overcome these shortcomings. Properties of the resulting models are shown in Table 9 and Fig. 6. The DCA1 model supports the interpretation of this axis as representing a very distinct decay process gradient. The model is dominated by a single variable, log age, which accounts for up to 84 % of the sample variation. Inclusion of other decay related variables did not strengthen the model, but a weak improvement was obtained by including distance to edge and soil contact. The DCA2 (subset a) model reveals a complex gradient structure relating both to microclimatic stress and decay rate. The model contains four variables, all of which contribute significantly to explain the sample variation. This model has, compared to the DCA1 model, a lower determination coefficient. Attempts to model for the full DCA3 data set was not very

Jacob Heilmann-Clausen

587

Fig. 5. Biplots showing the log specific scores in the DCA1\DCA2 ordination space with respect to various key variables. Axes are scaled in SD units.

successful. Models were not very powerful (r# about 0.3), or appeared more or less spurious by the inclusion of variables which independently did not correlate with the axis. Therefore, a model based on subset a is presented. The DCA3a model has an even lower determination coefficient than the DCA2a model. Plant diversity is the most strongly expressed variable in the model, but soil contact and the ‘ dummy variable ’ soil type also contributes. With respect to the dummy variable, the model implies that logs with high DCA3 scores tend to be situated on richer soil types (e.g. clay tills and lacustrine deposits), whereas logs on sandy tills tend to have low DCA3 scores. Distribution of species along DCA-axes Fig. 7 shows the distribution of species in the DCA1\DCA2 ordination space. The pattern is obviously difficult to overview, but some help for the interpretation may be gained from the correlation quotients given in Table 10 and from the accompanying overlay plots (Fig. 8).

The species richness of slime moulds is negatively correlated with DCA1, and, except for the corticioid fungi, a distinct shift in relative importance is evident in all fungal morphogroups. Polypores and pyrenomycetes are very strongly positively correlated with the axis, whereas agarics are negatively correlated (Table 10). Weaker, but still significant, correlations are found for discomycetes and ‘ other basidiomycetes ’. DCA2 shows a strongly negative correlation to both fungal and slime mould species richness. Only within the corticioid fungi is this tendency not evident, with the relative contribution of this group to the total fungal species richness strongly increasing along the axis. It is noteworthy that the correlation between DCA2 and overall species richness is stronger than the correlation between DCA2 and any environmental variable (Table 8). In other words, DCA2 represents a species richness gradient in addition to describing an ecological gradient. While DCA3 correlates positively with the relative importance of discomycetes, no other evident trends are expressed along the axis.

Macrofungi and slime moulds on beech logs

588

Table 8. Kendall rank correlations between DCA axes and log variables in the full data set and subset a respectively.

DCA1 DCA2 DCA3 Decay stage Log age Bark cover Soil contact Moss cover Sun exposure Soil humidity Decay rate Wind exposure Dist. to edge Fractures Other plants Bole forks Ane nem Ran fic Ane ran Plant diversity Floristic deviance Plant cover Mer per

DCA1

DCA1a

DCA2

DCA2a

DCA3

DCA3a

1n00 k0n08 0n00 k0n68**** k0n66**** 0n65**** k0n41**** k0n23* k0n22* k0n20* k0n01 0n00 k0n00 k0n06 k0n08 0n14 k0n00 k0n17 0n03 0n07 0n04 k0n06 0n06

1n00 k0n17 k0n04 k0n33** k0n27* 0n30* k0n27* k0n04 0n10 k0n36** k0n17 0n02 0n01 0n08 k0n33** 0n20 0n10 k0n23 k0n01 0n01 k0n03 k0n06 k0n05

k0n08 1n00 k0n04 0n14 0n23* k0n15 k0n09 k0n12 k0n23* k0n07 k0n35*** k0n36**** k0n31*** k0n22* 0n20* k0n23* k0n22* 0n14 0n13 0n23* 0n18* k0n03 0n19

k0n17 1n00 k0n10 0n30* 0n52**** k0n29* k0n05 k0n12 k0n26* k0n04 k0n50**** k0n45*** k0n39*** k0n38*** 0n34** k0n31* k0n27* 0n12 0n16 0n25* 0n26* k0n03 0n16

0n00 k0n04 1n00 k0n11 k0n10 0n05 k0n02 0n12 0n02 0n09 0n01 0n05 k0n07 0n12 0n11 k0n09 0n00 0n37*** 0n28** 0n26** 0n24** 0n23* 0n10

k0n04 k0n04 1n00 k0n24 k0n17 0n12 0n07 0n22 k0n04 0n16 k0n04 0n14 0n01 0n12 0n16 k0n07 k0n04 0n43*** 0n41*** 0n42*** 0n38** 0n25* 0n25*

0n0001 ; *** P l 0n0001k

Only variables with significant correlation quotients in a data set are shown. **** P * P l 0n01k 0n05.

0n001 ; ** P l 0n001k

0n01 ;

Table 9. ANOVA’s comparing successively more complicated multible regression models predicting DCA1, DCA2a, and DCA3a.

Terms DCA1 log age (polynomial, 2. degree) jedge (polynomial, 2. degree) jsoil contact DCA2a decay rate jdecay stage (polynomial, 2. degree) jwind exposure jfloristic deviance DCA3a plant diversity jsoil type jsoil contact

Res. Df.

Sum of squares

67 65 64

362726 15924 7529

0n0001 0n0001 0n003

0n83 0n87 0n89

68 66 65 64

44335 26267 12564 10428

0n0019 0n0113 0n0014 0n0046

0n33 0n53 0n62 0n70

68 67 66

22815 13641 5442

0n0001 0n0048 0n032

0n28 0n44 0n51

Prob. (F)

R#

All values except R# and residual degrees of freedom (Res. Df) refer to the added variable.

DISCUSSION Several authors (e.g. Shigo 1967, Rayner & Todd 1979, Rayner & Boddy 1988) have questioned the relevance of studies of wood decaying fungal communities based solely on sporocarp inventory. The question has been raised as to whether sporocarp distribution corresponds to the mycelial growth and activity in the wood. However, in the present study sporocarp inventory was successfully combined with ordination to gain new insight into fungal community structure and development in decaying wood. This may be linked with the fact that logs and other coarse woody debris generally appears as well-spaced, temporal units on the forest floor, and that the associated fungi consequently are dependent on an effective dispersal. Although cord forming and rhizomorphic basidiomycetes

have efficient vegetative dispersal (e.g. Boddy 1993), the production of spores from sporocarps appear to be the most important dispersal agent for most known wood decay fungi (cf. Rayner & Boddy 1988). Additionally, it must be emphasised that although sporocarp inventories are necessarily indirect in their approach to fungal mycelia, they have the advantage of allowing a fast and extensive recording of fungal structures, generally determinable to species. Inventories on the mycelial level, in contrast, are still linked with serious problems with respect to identification and are extremely time consuming if applied in extensive studies. Interpretation of DCA axes The clear differences in eigenvalues make it clear that DCA1 expresses considerably more sample variation than DCA2 and

Jacob Heilmann-Clausen

589

Fig. 6. Prediction profiles showing the roles of individual variables in the multiple regression models obtained for DCA1–3. The scaling of the x-axes relates to the specific variables included. The y-axes are scaled in SD units.

DCA3. Furthermore, the axis ’ high stability to data set modifications (Fig. 3) and very clear model structure (Table 9, Fig. 6) suggests that the decay process gradient revealed by DCA1 has a dominant role in determining the community structure of the studied logs. This finding is hardly surprising, as previous researchers have consistently reported a distinct species turnover during wood decay, both in sporocarp based inventories (e.g. Kreisel 1961, Lange 1992, Willig & Schlegte 1995, Andersson 1997) and in fungal cultures obtained from wood samples (e.g. Coates & Rayner 1985a-c, Chapela & Boddy 1988c, Chapela et al. 1988 ; see also Ka$ a$ rik 1975). By combining sporocarp inventories with DCA, Renvall (1995), Høiland & Bendiksen (1996) and Lindblad (1997) were able to demonstrate that a decay gradient was the dominant factor structuring communities of decomposer fungi on conifer logs. The somewhat weaker gradient expressed along DCA2 is in this context more interesting. The gradient is complex, as reflected by the axis\variable correlations (Table 8) and by the DCA2a regression model (Table 9), which include several almost equally contributing variables. Several other combinations of variables (not shown) actually produced almost equally strong models for the axis. Accordingly, the model under discussion should not be taken too literally with respect

to individual variables, but rather should be understood to imply that several groups of variables should be considered in interpreting DCA2 ecologically. Decay rate, which has the strongest individual determination coefficient in the DCA2a regression model (Table 9), represent one of two ruling parameters of the gradient. Another variable associated with this parameter is log age, which has the most significant individual correlation coefficient to DCA2a (Table 8). Log age and decay rate are naturally somewhat autocorrelated, since slowly decaying logs last longer than fast decaying logs, and obviously log age is partly substituted by decay age in the DCA2a regression model. However, the clear effect of decay stage in the model implies that this substitution is not perfect. The other obvious key parameter behind the DCA2 gradient is related to the amount of externally induced microclimatic fluctuation. This parameter can be found in the DCA2a regression model represented by wind exposure, but individually also the variable distance to edge and sun exposure are significantly correlated with DCA2a (Table 8). Most wood decay fungi have optima at intermediate moisture levels, but considerable differences do exist between different fungal species with respect both to their optima and tolerances (Boddy 1983, Boddy et al. 1985, Chapela & Boddy 1988b,

Macrofungi and slime moulds on beech logs

590

Fig. 7. Distribution of fungal species in the DCA1\DCA2 ordination space. Names of species are abbreviated according to Table 4. Some species have been moved slightly to give place for all names. Axes are scaled in SD unitsi100.

Griffith & Boddy 1991b). Unfortunately, the effects of a fluctuating microclimate as expressed along DCA2 has never been investigated with respect to fungal communities in decaying wood, but must be considered a distinct stress factor, implying an altered community structure and a decreased decay rate. The role of the last variable in the model, floristic deviation, is more difficult to interpret. The variable might reflect a longterm response to certain microclimatic or soil related conditions that favours plant species other than the dominant Anemone nemorosa. DCA3 seems to add little to the description of the overall community development of the studied logs. The model of the axis (Table 9) is not very powerful, and its low eigenvalue and gradient length scores suggest that the axis may partly represent noise or polynomial distortions (Økland 1990). This interpretation is supported by the poor recovery of the axis in the subset ordinations (Fig. 3). Because the axis seems to respond mainly to a soil gradient, it may be somewhat surprising to find it so weakly expressed, for soil conditions

have previously been found to influence the occurrence and fitness of several wood decaying fungi (Eriksson et al. 1973–88, Rayner & Boddy 1988, Kuyper & de Vries 1990, Abdalla & Boddy 1996). Probably the gradient would be more clearly expressed if there were more logs distributed across a wider spectrum of soil types than were present in this study. The increasing percentage of discomycetes along DCA3 (Table 10) is intriguing but hardly significant, in view of unreliability of the gradient. Omitting DCA3 from consideration, DCA1 and DCA2 together convey a good description of the community development and structure of the logs studied. The overall developmental trend of the community is represented by DCA1, with DCA2 representing deviations to this trend that are related to the log decay rate and the amount of microclimatic stress influencing the individual log. Distribution of species groups along DCA axes From Table 10 it is evident that significant changes in the

Jacob Heilmann-Clausen

591

Fig. 8. Biplots showing the log-specific species richness and the relative share of various fungal morpho-groups in the DCA ordination diagrams. Axes are scaled in SD units.

community structure occur along DCA1 and 2. However, in order to evaluate how these changes are related to the ecological gradients relating to these axes, it is necessary to investigate the distribution of single species and ecologically defined species groups in the ordination space.

DCA1 Among the first species to obtain optima along the DCA1 decay process gradient, two different invasion strategies seem to be represented. Some species (e.g. Nectria coccinea, Hypoxylon

Macrofungi and slime moulds on beech logs

592

Table 10. Kendall rank correlations between various species groups and DCA axes in the full dataset and in subset a. DCA1 Species number of Fungi Slime moulds Share of Pyrenomycetes Discomycetes Agarics Corticioid fungi Polypores Other basidiomycetes

k0n00 k0n24** 0n29*** k0n19* k0n40**** 0n11 0n37**** 0n24**

DCA1a

0n20 0n18 0n11 k0n13 0n09 k0n16 0n22* k0n14

DCA2

DCA2a

DCA3

DCA3a

k0n39**** k0n40****

k0n47**** k0n58****

k0n02 k0n06

0n04 k0n03

0n05 k0n07 k0n11 0n37**** k0n10 k0n20*

0n08 k0n07 k0n20 0n43**** k0n10 k0n30**

0n08 0n30*** k0n10 k0n19 0n02 k0n024

0n04 0n29** k0n09 k0n28 0n05 k0n00

The two first rows considers raw species numbers, the others relative shares. **** P * P l 0n01k 0n05.

fragiforme) have been reported as endophytic, latent invaders in healthy beech branches (Chapela & Boddy 1988a,b). Inonotus nodulosus, Diatrype disciformis, Neobulgaria pura, and Exidia glandulosa might also be considered to belong here. They are specific or selective for beech and all species typically produce extensive fructifications on very recently dead logs, both features highly characteristic of endophytic decay fungi (Rayner & Boddy 1988, Boddy 1994). Experimental work has shown that latent invaders are activated very soon after the wood water content is lowered (e.g. induced by tree death). The early optima of several known or suggested latent invaders, reported here supports this finding and indicate that the latent invasion strategy can secure not only an early infection of dead wood but also an early sporulation. Other species (e.g. Cylindrobasidium evolvens, Schizophyllum commune, Peniophora spp.) may function as unspecialised opportunists, sensu Rayner & Boddy (1988). They typically sporulate on small spots on the log surface and are probably either confined to wood unoccupied by latent invaders or able to replace these under the influence of severe microclimatic stress, e.g. in sun-exposed wood. Slightly later optima are obtained by a number of polypores, notably Bjerkandera adusta, Datronia mollis, Polyporus brumalis, Trametes hirsuta, and T. versicolor. Together with the corticioid Phlebia radiata, P. tremellosa, and Stereum hirsutum, these appear to form a well-defined ecological group (cf. ‘ class 1 ’ in Coates & Rayner (1985c)). The species often co-occur, although T. hirsuta, does show a strong preference for sun exposed wood. All are white rotters and several have been found to cause a very fast decay, either experimentally (Worrall, Anagnost & Zabel 1997) or under field conditions (Chapela et al. 1988, Willig & Schlegte 1995). Studies on mycelial interactions have shown that members of the group typically exhibit an intermediate combative ability. They are characteristically able to replace unspecialised opportunists and latent invaders, yet are themselves replaced a few years after establishment, e.g. by cord-forming basidiomycetes (Boddy & Rayner 1983, Coates & Rayner 1985c, Chapela & Boddy 1988c, Chapela et al. 1988). In line with this, the basidiomycetes that form mycelial cords (i.e. Ramaria stricta, Clitocybe diatreta, Hypholoma fasciculare, Lycoperdon perlatum, L. pyriforme, Megacollybia

0n0001 ; *** P l 0n0001k

0n001 ; ** P l 0n001k

0n01 ;

platyphylla, Mutinus caninus, Phallus impudicus, and Stropharia cyanea), all have their optima in the ‘ older ’ half of the ordination diagram. Cord formers are generally regarded as among the most combative wood decomposers (cf. Coates & Rayner 1985c, Boddy 1993), and under field conditions they have been found to cause a rapid decay (Chapela et al. 1988). Surprisingly, this contrasts with in vitro experiments, in which several cord formers have been shown to have little or no ability to decay fresh wood (Dix & Cairney 1985, Tanesaka, Masuda & Kinugawa 1993, Worrall et al. 1997). The reason behind this paradox may be that some cord formers depend on a pre-conditioning of the wood by earlier colonisers, before they can themselves act as decayers, or that they are unable to initiate decay unless they, as in nature, can draw on a large network of resources. Several agarics, belonging to Pluteus, Psathyrella, and Mycena, show even later optima than those recorded for most cord-formers. Unfortunately, as these genera have been included in very few experimental studies, little is known about their wood degrading abilities and strategies. However, Chapela et al. (1988) found that Psathyrella piluliformis (syn. P. hydrophila) was able to replace several cord formers, and it is possible that strong combative ability may play an important role for the group in general. In this regard, it is noteworthy that isolation of Pluteus species is difficult and depends on special nutritive media (Banerjee & Sundberg 1993, Banerjee 1994). This may explain why that genus has not been reported in studies based on isolation procedures (e.g. Chapela et al. 1988) and suggests that members of the genus are highly dependent of a special nutritive environment, e.g. imposed by fungi occurring earlier in the succession. Considering only the taxa discussed above, it would seem that DCA1 represents a gradient in combative ability. However, it is possible that the role of substrate modification is equally important, especially during the late phases of decay, as suggested previously by Renvall (1995), and Holmer, Renvall & Stenlid (1997). DCA2 Along DCA2 there is an intricate relation between community structure and decay rate. It is possible that both parameters are determined solely by the external microclimatic regime, but the strong independent expression of decay rate in the model

Jacob Heilmann-Clausen suggests that this is not the case. As the decay rate has been reported to vary considerably among species of wood decaying fungi (e.g. Worrall et al. 1997), it seems likely that this is influenced both by the external microclimatic regime and by the community-structure of the individual log. In beech wood several pyrenomycetes, including Xylaria hypoxylon and Nemania serpens (syn. Hypoxylon serpens), have been found to be especially slow decayers, both experimentally (Boddy, Owens & Chapela 1989) and in the field (Chapela et al. 1988). Chapela et al. (1988) found that X. hypoxylon was consistently associated with very dry decay columns sealed by pseudosclerotial plates. Apparently, that species is able to actively decrease the water content of occupied wood, probably as a defence against invasion of more combative fungi. The slow decay rate of the species seems to be a direct result of this strategy (Boddy et al. 1989, Boddy 1992). Similar strategies therefore appear to be distributed particuarly in pyrenomycetes, of which several have been found to be welladapted to water stress (Griffith & Boddy 1991b, Boddy 1992 ; see also Boddy, Bardsley & Gibbon 1987). In the present study, although X. hypoxylon was recorded very frequently (Table 4), the sporocarps were primarily confined to small patches. On the other hand, Eutypa spinosa and Kretzschmaria deusta (syn. Ustulina deusta) appeared to occupy large volumes of wood in several logs, estimated from very extensive fructifications. Both species have been shown to cause a rot characterised by well-developed pseudosclerotial plates (Pearce 1991, Hendry, Lonsdale & Boddy 1998) and are likely to use a strategy of active wood desiccation. In fact decay rate is negatively correlated with the relative cover of E. spinosa (Table 5), but this correlation is not evident with respect to DCA2 (Table 8). Unfortunately, the relative cover of K. deusta was not estimated, but the species ’ rather high score on DCA2 (Fig. 7) may help explain the decreasing decay rate expressed along the axis. This in turn raises the speculation that the shift in community structure along DCA2 is caused not only by external microclimatic stress but also by endogenous microclimatic stress induced by pyrenomycetes, and K. deusta in particular. The increasing relative share of corticioid fungi (Table 10, Fig. 8) along DCA2 may appear surprising as several, rare corticioid fungi are supposed to be highly dependent on a humid forest environment (Eriksson et al. 1973–88, Larsson 1997). In general, however, the group seems to exhibit good adaptations to environments marked by microclimatic stress. Adaptations include : (1) resistance to prolonged desiccation, e.g. by clamydospore formation (Boddy 1992) ; (2) resistant sporocarps reviving functional only during humid periods (Eriksson et al. 1973–88, Nun4 ez & Ryvarden 1993, Nun4 ez 1996) ; (3) psychrophily, allowing growth and sporulation at low temperatures under snow or during moderately cold, humid winterperiods (Gilbertson 1973, Larsen, Jurgensen & Harwey 1981, Griffith & Boddy 1991b) ; and (4) cryptic sporulation under logs and large branches where close soil contact creates a more stable, humid microenvironment (Høiland & Bendiksen 1996). Cryptic sporulation appears to be used by several species with high scores on DCA2 (e.g. Gloiothele lactescens, Hyphoderma praetermissum, Scopoloides rimosa, Stypella grilletii,

593 Trechispora stevensonii). Psychrophily appear to be the strategy of others species (e.g. Brevicellicium olivascens, Hyphodontia radula, Radulomyces confluens, Phlebia livida, Sistotrema sernanderi, Tulasnella violea). In both groups sporocarps adapted to repeated cycles of desiccation and rewetting seem to be an additional feature, but this adaptation seems to be especially prominent among species with low DCA1 scores (notably Cylindrobasidium evolvens, Peniophora cinerea, P. incarnata, and P. lycii). These are all confined to the exposed bark and outermost sapwood of newly dead trees, an environment that offers little protection from the desiccating forces of wind and sun. Schizophyllum commune is also obviously well-adapted to this habitat. Thus it would appear that the role of microclimatic stress is not restricted to DCA2, but should be considered also with respect to DCA1, although probably only during the initial phase of community development. The role of slime moulds The species richness of slime moulds is negatively correlated to both DCA1 and DCA2 (Table 10 ; Fig. 8), which implies that the group has a strong preference for well-decayed, moist wood. This is not surprising since water availability has been stated to be of crucial importance for the group (Ing 1994). Interestingly, the three species with highest DCA2 scores, i.e. Enteridium lycoperdon, Fuligo septica, and Lycoperdon epidendrum, all form large aethalia, in contrast to all other species included in the ordination. The significance of this finding remains unclear. Relatively little is known about the activity of slime moulds in decaying wood, but their role is most likely entirely indirect (Rayner & Boddy 1988, Ing 1994). Several species are able to digest fungi in artificial culture (Howard & Curie 1932) and digestion of fungal mycelia and sporocarps are also recorded in nature (Elliott & Elliott 1920, Ing 1994). Another important food resource of slime moulds is bacteria (Madelin 1984), which could also explain the late peak of the group since bacteria are generally more abundant in well-decayed wood (Rayner & Boddy 1988). Exclusion of slime moulds from the ordination resulted in a less distinct expression of DCA2 (Fig. 3), and in less significant correlations with environmental variables (results not shown). This implies that slime moulds might serve as good indicators of the environmental characteristics of decaying wood, with a rich occurrence highly indicative of a humid wood environment. The frequency and abundance of slime moulds may also serve to indicate the biotic characteristics of wood, particularly in view of a high sporocarp production on some logs. Overall community development Fig. 9 presents a hypothesised community development derived from the DCA1 versus DCA2 ordination diagram of the studied logs (Fig. 4). This proposed community development diagram is inspired by the diagram of fungal community development in decaying wood by Rayner & Webber (1984), but with several modifications. According to the diagram, community development is initiated by a disturbance event (tree-death) leading to the activation of

Macrofungi and slime moulds on beech logs

594

Fig. 9. Diagram of the macrofungal community development on the studied logs, based on the plot of DCA1 against DCA2 (Fig. 4), but laterally reversed. The arrows indicate ecological conditions supposed to determine the community development. The text within the community triangle indicate ecological adoptions essential at various (mainly extreme) stages of community development.

already established latent invaders and the new establishment of unspecialised opportunists. As decay proceeds, combat is generally intensified leading to a community dominated by highly combative species. Depending on external microclimatic conditions and internal community structure, the microclimatic stress regime may increase, leading to a community dominated by stress tolerant fungi. Alternatively, in the absence of microclimatic stress, the importance of substrate modification is increasing leading to a community dominated by more or less combative late-stage specialists. The overall diagram outline, with the broadest, most differentiated part being expressed at late stages of decay, implies that as decay proceeds logs become increasingly diversified with respect to community structure. Renvall (1995) and Høiland & Bendiksen (1996) have previously reported a similar trend during the decay of conifer logs, which Renvall (1995) suggested might reflect an increase in the number of microhabitats present in well-decayed wood. In the present case, however, diversification is linked to an increasing importance of microclimatic stress. This can be explained by extensively decayed wood having a much coarser and more open pore-structure than fresh wood, and hence being more susceptible to fluctuations in external humidity. An additional explanation might be that differences in wood humidity caused by fungal community structure (e.g. dominance of certain pyrenomycetes), becomes increasingly evident as decay proceeds. With respect to overall community development also, the log-age prediction profile in the DCA1 regression model (Fig.

6) deserves comment. The curvilinear profile shows that the biggest axis shift and hence the fastest community development occurs during the first half of the decay period. Thus the overall community development on the logs studied can be summarised as being fast and predictable in the early phases of decay, but becoming increasingly slow and diversified as decay proceeds.

A C K N O W L E D G E M E N TS I wish to thank Thomas Læssøe and two anonymous referees for highly useful comments and improvements to the manuscript. Morten Christensen is thanked for fruitful discussions and for enjoyable hours spent in the field, and Rasmus Ejrnæs for introducing me to the world of ordination. Henrik F. Gøtzsche, Nils Hallenberg, Seppo Huhtinen, Henning Knudsen, Ewald Langer, Thomas Læssøe, Leif O$ rstadius, Ain Raitviı3 r, Peter Roberts, and Jan Vesterholt are thanked for help in determining various critical specimens. The Jakob E. Lange Foundation is thanked for funds used to acquire aerial photos. Finally, I wish to thank the Sorø Akademi Foundation for permitting me to work in Suserup Skov.

REFERENCES Abdalla, S. H. M. & Boddy, L. (1996) Effect of soil and litter type on outgrowth patterns of mycelial systems of Phanerochaete velutina. FEMS Microbiology Ecology 20 : 195–204. Anderson, H. (1997) [Fungi on 10 stumps of Fagus sylvatica – all of the same age – in the O$ lper Holz in Braunschweig]. Zeitschrift fuW r Mykologie 63 : 51–62. [In German.]

Jacob Heilmann-Clausen Banerjee, P. (1994) Successful isolation and growth of tissue cultures of Pluteus species. Mycologist 8 : 132–135. Banerjee, P. & Sundberg, W. J. (1993) Preliminary observations on germination of Pluteus basidiospores. Mycologia 85 : 811–813. Boddy, L. (1983) Effect of temperature and water potential on growth rate of wood-rotting basidiomycetes. Transactions of the British Mycological Society 80 : 141–149. Boddy, L. (1992) Development and function of fungal communities in decomposing wood. In The Fungal Community. Its organization and role in the ecosystem (G. C. Carroll & D. T. Wicklow, ed) : 749–782. 2nd edn. MarcelDekker, New York. Boddy, L. (1993) Saprotrophic cord-forming fungi : warfare strategies and other ecological aspects. Mycological Research 97 : 641–655. Boddy, L. (1994) Latent decay fungi : the hidden foe ? Arboricultural Journal 18 : 113–145. Boddy, L., Bardsley, D. W. & Gibbon, O. M. (1987) Fungal communities in attached ash branches. New Phytologist 107 : 143–154. Boddy, L., Gibbon, O. M. & Grundy, M. A. (1985) Ecology of Daldinia concentrica : effect of abiotic variables on mycelial extensions and interspecific interactions. Transactions of the British Mycological Society 85 : 201–211. Boddy, L., Owens, E. M. & Chapela, I. H. (1989) Small scale variation in decay rate within logs one year after felling : effects of fungal community structure and moisture content. FEMS Microbiology Ecology 62 : 173–184. Boddy, L. & Rayner, A. D. M. (1983) Ecological roles of basidiomycetes forming decay communities in attached oak branches. New Phytologist 93 : 77–88. Butin, H. & Kowalski, T. (1983) [The natural pruning of branches and their biological preconditioning. I. The fungal flora of beech (Fagus sylvatica).] European Journal of Forest Pathology 13 : 322–334. [In German.] Butin, H. & Kowalski, T. (1986) [The natural pruning of branches and their biological conditions. III. The fungal flora of common maple, grey alder, silver beam, hornbeam and common ash.] European Journal of Forest Pathology 16 : 129–138. [In German.] Chapela, I. H. & Boddy, L. (1988a) Fungal colonization of attached beech branches I. Early stages of development of fungal communities. New Phytologist 110 : 39–45. Chapela, I. H. & Boddy, L. (1988b) Fungal colonization of attached beech branches II. Spatial and temporal organization of communities arising from latent invaders in bark and functional sapwood, under different moisture regimes. New Phytologist 110 : 47–57. Chapela, I. H. & Boddy, L. (1988c) The fate of early fungal colonizers in beech branches decomposing on the forest floor. FEMS Microbiology Ecology 53 : 273–284. Chapela, I. H., Boddy, L. & Rayner, A. D. M. (1988) Structure and development of fungal communities in beech logs four and a half year after felling. FEMS Microbiology Ecology 53 : 59–70. Christensen, M. & Emborg, J. (1996) Biodiversity in natural versus managed forests in Denmark. Forest Ecology and Management 85 : 47–51. Coates, D. & Rayner, A. D. M. (1985a) Fungal population and community development in cut beech logs, I. Establishment via the aerial cut surface. New Phytologist 101 : 153–171. Coates, D. & Rayner, A. D. M. (1985b) Fungal population and community development in cut beech logs, II. Establishment via the buried cut surface. New Phytologist 101 : 173–181. Coates, D. & Rayner, A. D. M. (1985c) Fungal population and community development in cut beech logs, III : Spatial dynamics, interactions and strategies. New Phytologist 101 : 183–198. Dix, N. J. & Cairney, J. W. G. (1985) Wood rotting by Phallus impudicus. Transactions of the British Mycological Society. 85 : 514–520. Ejrnæs, R. & Bruun, H. H. (2000) A gradient analysis of temperate grassland on well-drained soil. Journal of Vegetation Science 11 : 573–584. Elliot, W. T. & Elliot, J. E. (1920) The sequence of Fungi and Mycetozoa. Journal of Botany 58 : 273–274. Emborg, J., Christensen, M. & Heilmann-Clausen, J. (1996) The structure of Suserup Skov, a near-natural temperate deciduous forest in Denmark. Forest & Landscape Research 1 : 311–333. Emborg, J., Christensen, M. & Heilmann-Clausen, J. (2000) The structural dynamics of Suserup Skov, a near-natural temperate deciduous forest in Denmark. Forest Ecology and Management 126 : 173–189.

595 Eriksson, J., Hjortstam, K., Larsson, K.-H. & Ryvarden, L. (1973–88) The Corticiaceae of North Europe. 8 vols. Fungiflora, Oslo. Feilberg, J. (1993) [Forest floor vegetation in the forests of Sorø Akademi.] Flora og Fauna 99 : 23–39. [In Danish.] Franklin, J. F., Shugart, H. H. & Harmon, M. E. (1987) Tree death as an ecological process. BioScience 37 : 550–556. Fritzbøger, B. & Emborg, J. (1996) Landscape history of the deciduous forest Suserup Skov, Denmark, before 1925. Forest & Landscape Research 1 : 291–309. Gilbertson, R. L. (1973) Notes on some corticoid fungi associated with snowbanks in Southern Arizona. Persoonia 7 : 171–182. Gilbertson, R. L. (1980) Wood-rotting fungi of North America. Mycologia 72 : 1–49. Griffith, G. S. & Boddy, L. (1990) Fungal decomposition of attached angiosperm twigs. I. Decay community development in ash, beech and oak. New Phytologist 116 : 407–415. Griffith, G. S. & Boddy, L. (1991a) Fungal decomposition of attached angiosperm twigs. II. Moisture relations of twigs of ash (Fraxinus excelsior L.). New Phytologist 117 : 251–257. Griffith, G. S. & Boddy, L. (1991b) Fungal decomposition of attached angiosperm twigs. III. Effect of water potential and temperature on fungal growth, survival and decay of wood. New Phytologist 117 : 259–269. Hallenberg, N. (1984) A taxonomic analysis of the Sistotrema brinkmannii complex (Corticiaceae, Basidiomycetes). Mycotaxon 21 : 389–411. Hannon, G. E., Bradshaw, R. & Emborg, J. (2000) 6000 years of forest dynamics in Suserup Skov, a semi-natural Danish woodland. Global Ecology and Biodiversity Letters 9 : 101–114. Hansen, L. & Knudsen, H. (eds) (1992) Nordic Macromycetes. Vol. 2. Polyporales, Boletales, Agaricales, Russulales. Nordsvamp, Copenhagen. Hansen, L. & Knudsen, H. (eds) (1997) Nordic Macromycetes. Vol. 3. Heterobasidioid, Aphyllophoroid and Gastromycetoid Basidiomycetes. Nordsvamp, Copenhagen. Hansen, L. & Knudsen, H. (eds) (2000) Nordic Macromycetes. Vol. 1. Ascomycetes. Nordsvamp, Copenhagen. Harmon, M. E., Franklin, J. F., Swanson, F. J., Sollins, P, Gregory, S. V., Lattin, J. D., Anderson, N. H., Cline, S. P., Aumen, N. G., Sedell, J. R., Lienkaemper, G. W., Cromack, K., Jr. and Cummins, K. W. (1986) Role of coarse woody debris in temperate ecosystems. Advances in Ecological Research 15 : 133–302. Harmon, M. E., Sexton, J., Caldwell, B. A. & Carpenter, S. E. (1994) Fungal sporocarp mediated losses of Ca, Fe, K, Mg, Mn, N, P and Zn from conifer logs in the early stages of decomposition. Canadian Journal of Forest Research 24 : 1883–1893. Hendry, S. J., Lonsdale, D. & Boddy, L. (1998) Strip-cankering of beech (Fagus sylvatica) : pathology and distribution of symptomatic trees. New Phytologist 140 : 549–565. Hill, M. O. & Gauch, H. G. (1980) Detrended correspondance analysis : an improved ordination tecnique. Vegetatio 42 : 47–58. Høiland, K. & Bendiksen, E. (1996) Biodiversity of wood-inhabiting fungi in a boreal coniferous forest in Sør-Trøndelag County, Central Norway. Nordic Journal of Botany 16, 643–659. Holmer, L., Renvall, P. & Stenlid, J. (1997) Selective replacement between species of wood-rotting basidiomycetes, a laboratory study. Mycological Research 101 : 714–720. Howard, F. L. & Currie, M. E. (1932) Parasitism of myxomycete plasmodia on fungous mycelia. Journal of the Arnold Arboretum 13 : 438–447. Ing, B. (1994) The phytosociology of myxomycetes. New Phytologist 126 : 175–201. Ka$ a$ rik, A. (1975) Succession of microorganisms during wood decay. In Biological Transformation of Wood by Microorganisms (W. Liese, ed.) : 39–51. Springer Verlag, Berlin. Keizer, P. J. & Arnolds, E. (1990) Mycocoenology of marshy forest and scrubs. I. Host range of wood-decomposing Aphyllophorales and Heterobasidiomycetes. Wageningen Agricultural University Papers 90(6) : 77–91. Kenkel, N. C. & Booth, T. (1992) Multivariate analysis in fungal ecology. In The Fungal Community. Its Organization and Role in the Ecosystem (G. C. Carroll & D. T. Wicklow, eds) : 209–227. 2nd edn. Marcel-Dekker, New York.

Macrofungi and slime moulds on beech logs Ko! ljalg, U. (1996) Tomentella (Basidiomycota) and Related Genera in Temperate Eurasia. [Synopsis Fungorum Vol. 6.] Fungiflora, Oslo. Kreisel, H. (1961) [The development of fungal associations on stumps of Fagus in clearcuts in northern Germany.] Feddes Repertorium : Beiheft 139 : 227–232. [In German.] Kuyper, T. W. & De Vries, B. W. L. (1990) Effects of fertilization on the mycoflora of a pine forest. Wageningen Agricultural University Papers 90(6) : 102–111. Lange, M. (1992) Sequence of macromycetes on decaying beech logs. Persoonia 14 : 449–456. Larsen, M. J., Jurgensen, M. F. & Harvey, A. E. (1981) Athelia epiphylla associated with colonization of subalpine fir foliage under psychrophilic conditions. Mycologia 73 : 1195–1201. Larsson, K.-H. (ed.) (1997) [Redlisted fungi in Sweden, species protocol.] ArtData Banken, Uppsala. [In Swedish.] Laweson, J. (1997) Effects of species partition on explanatory variables in direct gradient analysis – a case study from Senegal. Journal of Vegetation Science 8 : 409–414. Legendre, P. & Legendre, L. (1998) Numerical ecology, second English edition. Developments in Environmental Modelling 20, 1–853. Lindblad, I. (1998) Wood-inhabiting fungi on fallen logs of Norway spruce : relations to forest management and substrate quality. Nordic Journal of Botany 18 : 243–255. Madelin, M. F. (1984) Myxomycetes, microorganisms and animals : a model of diversity in animal interactions. In Invertebrate-Microbial Interactions (J. M. Anderson, A. D. M. Rayner & D. W. H. Walton, eds) : Cambridge University Press, Cambridge, UK. McCune, B. & Mefford, M. J. (1997) PC-ORD. Multivariate Analysis of Ecological Data, version 3.0. MjM Software Design, Gleneden Beach, OR. Møller, P. F. (1997) [Biodiversity in Danish natural forests. A comparison between unmanaged and managed woodlands in East Denmark]. [Danmarks og Grønlands Geologiske Undersøgelse. Rapport 1997\41.] Copenhagen. [In Danish.] Nannenga-Bremekamp, N. E. (1991) A Guide to Temperate Myxomycetes. Biopress, Bristol. Nun4 ez, M. (1996) Hanging in the air : a tough skin for a tough life. Mycologist 10 : 15–17 Nun4 ez, M. & Ryvarden, L. (1993) Basidiomycetes in the canopy and on twigs at ground level : a comparison. In Aspects of Tropical Mycology (S. Isaac, J. C. Frankland, R. Watling & A. J. S. Whalley, eds) : Cambridge University Press, Cambridge, UK. Ohlson, M., So$ derstro$ m, L., Ho$ rnberg, G., Zackrisson, O. & Hermansson, J. (1997) Habitat qualities versus long-term continuity as determinants of biodiversity in boreal old-growth swamp forests. Biological Conservation 81 : 221–231.

596 Økland, R. H. (1990) Vegetation ecology : theory, methods and applications with reference to Fennoscandia. Sommerfeltia, Supplement 1 : 5–233. Økland, R. H. (1999) On the variation explained by ordination and constrained ordination axes. Journal of Vegetation Science 10 : 131–136. Peet, R. K. (1980) Ordination as a tool for analyzing complex data sets. Vegetatio 42 : 171–174. Pearce, R. B. (1991) Reaction zone relics and the dynamics of fungal spread in the xylem of woody angiosperms. Physiological and Molecular Plant Pathology 39 : 41–55. Rayner, A. D. M. & Boddy, L. (1988) Fungal Decomposition of Wood ; its biology and ecology. John Wiley & Sons, Chichester. Rayner, A. D. M & Todd, N. K. (1979) Population structure in wooddecomposing basidiomycetes. In Decomposer Basidiomycetes : their biology and ecology. (J. C. Frankland, J. N. Hedger & M. J. Swift, eds) : 109–128. Cambridge University Press, Cambridge, UK. Rayner, A. D. M. & Webber, J. F. (1984) Interspecific mycelial interactions – an overview. In The Ecology and Physiology of the Fungal Mycelium (D. H. Jennings & A. D. M. Rayner, eds) : 383–417. Cambridge University Press, Cambridge, UK. Roberts, P. (1998) A revision of the genera Heterochaetella, Myxarium, Protodontia and Stypella (Heterobasidiomycetes). Mycotaxon 69 : 209–48. Renvall, P. (1995) Community structure and dynamics of wood-rotting basidiomycetes on decomposing conifer trunks in northern Finland. Karstenia 35 : 1–51. Samuelsson, J., Gustafsson, L. & Ingelo$ g, T. (1994) Dying and Dead Trees. A review of their importance for biodiversity. Swedish Threatened Species Unit, Uppsala. SAS Institute (1997) JMP2 Statistics Made Visual. SAS Institute, Cary, NC. Shigo, A. (1967) Succession of organisms in discoloration and decay of wood. International Review of Forest Research 2 : 237–299. Sippola, A.-L. & Renvall, P. (1999) Wood-decomposing fungi and seed-tree cutting : a 40-year perspective. Forest Ecology and Management 115 : 183–201. Tanesaka, E., Masuda, H. & Kinugawa, K. (1993) Wood degrading ability of basidiomycetes that are wood decomposers, litter decomposers, or mycorrhizal symbionts. Mycologia 85, 347–355. Vejre, H. & Emborg, J. (1996) Interactions between vegetation and soil in a near-natural temperate deciduous forest. Forest & Landscape Research 1 : 335–347. Willig, J. & Schlegte, G. B. (1995) [Fungal succession on wood after windthrow in a natural Beech-forest reserve]. Allgemeine Forst Zeitschrift 15/1995 : 814–818. [In German.] Worrall, J., Anagnost, S. E. & Zabel, R. A. (1997) Comparison of wood decay among diverse lignicolous fungi. Mycologia 89 : 199–219. Zar, J. H. (1999) Biostatistical Analysis, fourth edition. Prentice Hall, London. Corresponding editor : S. A. Redhead