Austral Ecology (2009) 34, 69–76
Mammal mycophagy and fungal spore dispersal across a steep environmental gradient in eastern Australia KARL VERNES* AND LINDA DUNN Ecosystem Management, University of New England, Armidale NSW 2351, Australia (Email:
[email protected])
Abstract We examined changes in the types of fungi consumed by six species of small mammals across a habitat gradient in north-eastern New South Wales that graded from swamp, to woodland, to open forest and then to rainforest. All mammals ate some fungus, but only bush rats (Rattus fuscipes) regularly did so, and their diet included most of the fungal taxa that we identified across all mammals in the study.The composition of bush rat diet changed significantly with each change in habitat from woodland, to forest, to rainforest. In particular, there was a significant difference in the diets of rats caught either side of the open forest-rainforest ecotone, which marks the change in fungal community from one dominated by ectomycorrhizal fungi, to a community dominated by arbuscular mycorrhizal fungi. Movement patterns of bush rats living around the open forest-rainforest ecotone suggest that they transport fungal spores between these contrasting fungal communities.Therefore, bush rats have the potential, by way of spore dispersal, to influence the structure of vegetation communities. Key words: ecosystem dynamics, ectomycorrhizal fungi, ecotone, forest health, truffle.
INTRODUCTION In rainforest, arbuscular mycorrhizal (AM) fungi form mycorrhizal associations with the multitude of tree species that comprise tropical Australian rainforests (Brundrett et al. 1996; Hopkins et al. 1996; Smith & Smith 1997). By comparison, ectomycorrhizal (ECM) fungi are the predominant symbionts on the roots of many trees in sclerophyll forests and woodlands, including those from the genus Eucalyptus (Bougher 1995; Hopkins et al. 1996). Many mycorrhizal fungi lack the mechanisms necessary for independent spore dispersal. Those fungi give rise to hypogeous sporebearing structures commonly known as ‘truffles’ that are a staple food for many native Australian mammals (Claridge & May 1994) that unearth, ingest and subsequently disperse the spores (Blaney 1996; Johnson 1996). The great majority of taxa eaten by mycophagous (‘fungus-eating’) mammals are either known to, or thought to form symbiotic relationships with native trees or shrubs (Claridge & May 1994). Because of their potential to transport fungi relatively long distances through their foraging movements, mycophagous mammals might play an important role in distributing the spores of ectomycorrhizal fungi across habitat boundaries and into early successional habitats (Johnson 1996). This may especially be the case in ecotonal sclerophyll habitats that merge into rainforest, because of the AM/ECM dichotomy. *Corresponding author. Accepted for publication February 2008.
© 2009 The Authors Journal compilation © 2009 Ecological Society of Australia
Along the Great Escarpment of north-eastern New South Wales, wet sclerophyll forests dominated by Eucalyptus spp. abut extensive tracts of rainforest, with a sharply defined ecotone separating the two habitats (Campbell & Clarke 2006). Here, as in other eastern Australian rainforests, the position of the rainforest-open forest ecotone depends in part on past fire regimes (see Ash 1988; Unwin 1989; Harrington & Sanderson 1994). Fires that periodically sweep into the region from the dry sclerophyll forests are at times so intense that, under optimal burning conditions, they burn into the less flammable rainforest, killing rainforest trees and clearing the way for the advance of the Eucalyptus-dominated open forest. At other times, runs of wetter years allow the rainforest to encroach upon the open forest (Unwin 1989). The natural separation of symbiotic AM and ECM fungi at the dynamic rainforest-open forest edge means that in order for the establishment of either rainforest or Eucalyptus-dominated forest, a dispersal mechanism is required whereby spores of the appropriate fungi are transported back and forth across the ecotone, so that trees can establish and one or the other forest type is able to dominate the transition zone. Mammals are the likely spore dispersers in this system, similar to the role they have been shown to play elsewhere such as transportation of spores to sterile soil at a glacier forefront from nearby forest (Cazares & Trappe 1994) and in moving spores to drained beaver meadows where no mycorrhizal fungi previously existed (Terwilliger & Pastor 1999). doi:10.1111/j.1442-9993.2008.01883.x
70
K. VERNES AND L. DUNN
Some mammal species only consume a limited number of fungal species and travel only short distances between consumption of fungi and defecation of fungal spores, so it is unlikely that they are important spore dispersers. Other mammals consume a diverse range of species, do so regularly and travel great distances in a night of foraging, making them important to spore dispersal. For example, in northeastern Queensland, the northern bettong (Bettongia tropica) consumes at least 35 hypogeous taxa (Vernes et al. 2001), has a very large home range (Vernes & Pope 2001) and makes extensive nightly foraging movements (Vernes & Haydon 2001), all of which suggests it plays an important role in plant-fungal processes. Recently, work by Vernes (2009) and Vernes and Trappe (2007) has shown that several mammals that live near the rainforest edge, and that move between sclerophyll woodland and rainforest, consume a great diversity of hypogeous fungi. Opportunistic collection at a range of discrete sites in northeastern Queensland by Reddell et al. (1997) indicated that ECM fungal spores were more common in mammal diets sampled in sclerophyll forest compared with rainforest, whereas spores from AM fungi were more frequently found in mammal diets sampled from rainforest. In a Panamanian cloud forest, Mangan and Adler (2000) determined that a diversity of rodent species regularly consumed several species of AM fungi. Vernes and Trappe (2007) speculated that mammals that use both habitat types during their foraging movements probably play a key role in transporting spores of both AM and ECM fungi between habitats. Most previous studies of mycophagy in Australia have focused on consumption and dispersal of spores within habitats that support either AM or ECM fungi. This paper examines patterns in fungal consumption and dispersal by mammals across a steep environmental gradient that includes a number of ECM habitat types defined by sharp ecotones, and that is bordered by sedge swamp (dominated by AM fungi) habitat at one end of the gradient, and rainforest (also dominated by AM fungi) at the other. We hypothesize that changes in habitat along the gradient will be reflected in the fungal composition of mammal diets in those habitats, but that spore dispersal of ECM and AM fungi across habitat boundaries by some small mammals will also be apparent.
METHODS Study sites The study was conducted at Gibraltar Range National Park in north-east New South Wales. This location doi:10.1111/j.1442-9993.2008.01883.x
provided an opportunity to study the small mammal community over a continuous ecological gradient spanning (i) peat swamp dominated by sedges (hereafter ‘swamp’), (ii) woodland with a heath shrub layer (hereafter ‘woodland’), (iii) wet sclerophyll forest with a fern dominated ground layer (hereafter ‘open forest’), and (iv) warm temperate rainforest (hereafter ‘rainforest’). Boundaries between each habitat were distinguished by an obvious and sharp ecotone no greater than 25 m wide.Vegetation types at the site are described more fully by Clarke and Myerscough (2006). At each of two sites along the ecotone, paired transects (100 m apart) were arranged so that each transect traversed each of the four habitats and the intervening ecotones, with trapping grids positioned along these transects in the centre of each habitat, as well as on the ecotone between habitats. Trapping grids consisted of a cluster of nine traps in a 3 ¥ 3 grid with 20-m trap spacings, for a total of seven clusters (63 traps) per transect.The distance between the edges of grids along each transect was about 50 m. The two sites were approximately 250 m apart. The study area, including a map showing location of transects and grids, is described more fully by Vernes et al. (2006).
Sampling and analysis of mammal diets Each transect was trapped for three nights per field trip, with sampling undertaken in November 2003, and February, March and April 2004, totalling 756 trap-nights per trip and 3024 trap-nights over the entire study. The four sampling periods were chosen to capture the greatest fungal diversity in diets over as short a time frame as possible. By sampling diets between summer and late autumn, we were sampling the time of the year during which many species of hypogeous fungi in eastern Australian forests and woodlands form fruit-bodies (Claridge et al. 1993a, 2000). Previous studies have shown that sampling diets at this time yield a correspondingly high diversity of fungal spore types (Bennett & Baxter 1989; Claridge et al. 1993b). The fungal diet of four native rodents: the bush rat (Rattus fuscipes), swamp rat (Rattus lutreolus), fawn-footed melomys (Melomys cervinipes) and New Holland mouse (Pseudomys novaehollandiae); and two marsupials: the brown antechinus (Antechinus stuartii) and eastern pygmy possum (Cercartetus nanus), was examined. Animals were individually numbered, and scats were collected from each individual only for the first capture from a sampling period, so numbers presented in this paper reflect samples that were collected and analysed, not total captures made (the latter is covered by Vernes et al. 2006). Scats were preserved in the field in labelled vials containing 70% ethanol. © 2009 The Authors Journal compilation © 2009 Ecological Society of Australia
M A M M A L M Y C O P H AG Y A N D F U N G A L S P O R E D I S P E R S A L
In the laboratory, each scat sample was crushed in its vial using a glass rod, then allowed to settle for 10 min. A sample of fine fraction (suspended particles above the solid matter) was transferred by pipette from the vial and spread across a microscopic slide. Once the slide had dried, it was sealed using glycerol jelly and a cover slip. Initially, a scan of each slide (¥400 magnification) was made to determine whether fungal spores were present. Then for each slide showing fungal spores, a more detailed analysis was made where 20 random fields of view were examined at ¥400 magnification. Presence/absence of each unique taxon of dietary fungus on each slide was recorded. Spores were identified to the lowest taxonomic level possible using available keys (e.g. Castellano et al. 1989), published taxonomic papers, reference material collected at our site and assistance from expert mycologists (see Acknowledgements).
Habitat use by individual bush rats Records of individual bush rats captured during the four sampling trips were used to calculate the area utilized by individuals during the study, and to determine whether the ranges of these animals incorporated different habitat types. Ranges, plotted as minimum convex polygons enclosing all capture locations, were constructed for all individuals that were captured four or more times during the study.
Statistical analyses The broad consumption of fungi by each mammal species in any one habitat is expressed in this paper as the percentage of samples that contain fungi (i.e. percentage occurrence). Dietary differences are expressed in greater detail by calculating the proportion of samples that contained each taxon of fungi. The bush rat consumed the greatest diversity of fungus, was caught along much of the gradient, and was the species of mammal that we captured most often. For these reasons, we focused on this species as our model mycophagous mammal for examining finer changes in fungal consumption across the environmental gradient, particularly to determine whether a change in dietary fungus occurred when the vegetation changed from an ECM (i.e. Eucalyptus) dominated to an AM (i.e. rainforest) dominated community. Bush rats were not caught in the swamp, or swamp-woodland ecotone, and were caught in very low numbers in the woodland, therefore analyses focused on the section of the gradient from the woodland-open forest ecotone, through to rainforest. Observed trends in the numbers of fungal taxa per sample were examined using analysis of variance © 2009 The Authors Journal compilation © 2009 Ecological Society of Australia
71
(anova). The model used site (1 and 2) and position along the gradient (1 = woodland-open forest ecotone, 2 = open forest, 3 = open forest-rainforest ecotone, and 4 = rainforest) as variables. Multivariate analyses available within the PRIMER 6 software package (Clarke & Gorley 2006) were used to examine trends in consumption of each fungal taxa. To determine whether all data gathered for a mammal species could be pooled by habitat (i.e. across sites), we constructed Bray–Curtis similarity matrices of individual samples (the presence-absence of each fungal taxon within each scat), and examined patterns using multidimensional scaling.We then searched for significant patterns in sample composition using Analysis of Similarity (anosim) tests for data coded by transect. The resultant low global R-value (R = 0.027) and correspondingly high P-value (P = 0.18) supported pooling of data. To examine changes in consumption of taxa across the environmental gradient, we constructed a Bray– Curtis similarity matrix using square root-transformed percentage occurrence data for each fungal taxon ¥ habitat combination, for each of the 37 fungal taxa that were identified in the diet of bush rats. We then used Hierarchical Cluster Analysis using group average (Clarke & Gorley 2006) to display patterns in consumption between habitats, and similarity profile permutation tests to look for significant differences between habitat clusters.
RESULTS Capture success across the habitat gradient Most mammal species were sampled in heath woodland, wet sclerophyll forest and rainforest (or their adjacent ecotones), except for the New Holland mouse, samples from which came almost entirely from the swamp-woodland ecotone (Table 1). Most bush rats and fawn-footed melomys were sampled in the wetter forest types (Table 1), while most swamp rat samples came from the swamp, and the swampwoodland ecotone (Table 1). The numbers of dietary samples obtained from each species varied substantially, with particularly low capture numbers for eastern pygmy possums (n samples = 8, from 8 individuals) and swamp rats (n samples = 7, from 5 individuals; Table 1) making detailed analyses and inferences about their diets challenging. Moderate numbers of samples were obtained for fawn-footed melomys (n samples = 35, from 25 individuals), New Holland mice (n samples = 18, from 11 individuals), and the brown antechinus (n samples = 17, from 16 individuals; Table 1). The greatest number of samples was obtained from the bush rat (n samples = 69, from doi:10.1111/j.1442-9993.2008.01883.x
72
K. VERNES AND L. DUNN
Table 1.
Percentage occurrence of fungi in mammal diets in the habitat, grouped by the habitat in which they were captured Percentage occurrence of fungi in diet
Habitat Sedge swamp Swamp-heath ecotone Heath woodland Heath-wet sclerophyll ecotone Wet sclerophyll forest Wet forest-rainforest ecotone Rainforest
Brown antechinus
Eastern pygmy possum
Fawn-footed melomys
New Holland mouse
Bush rat
Swamp rat
–
–
–
–
–
–
–
–
–
– (1) 66.7 (3) 33.3 (6) 13.3 (15) 10.0 (10)
41.2 (17) – (1) –
0.0 (2) 33.3 (3) –
– (1) 0.0 (4) 33.3 (3) 0.0 (4) 0.0 (4) – (1)
33.3 (3) 50.0 (2) – (1) 0.0 (2)
100.0 (2) 100.0 (16) 90.9 (11) 96.2 (26) 71.4 (14)
– – –
– (1) – (1) – –
Numbers in parentheses are the total number of samples analysed for each species, in that habitat. Percentage occurrence of fungi in samples is not presented for samples totalling
