Seedling predation and growth at a rainforest–pasture ecotone, and the value of shoots as seedling analogues

https://doi.org/10.1016/S0378-1127(01)00519-9Get rights and content

Abstract

Variation in the survival and growth of seedlings of six rainforest plant species was examined at a site where rainforest adjoins pasture in subtropical eastern Australia. The site was regularly used by a feeding group of red-necked pademelons Thylogale thetis, a herbivorous macropodid marsupial. Protection from herbivory, using wire mesh cages to exclude the pademelons, significantly increased seedling growth and survival in five of the six plant species. The caged minus uncaged seedling growth rates were used as a measure of the level of growth suppression due to pademelon herbivory, which varied significantly among plant species, was greater in pasture than in forest, and decreased with distance from the forest edge. This pattern of variation among species and positions was highly correlated with that observed in a separate series of experiments that had assessed overnight damage levels to cut shoots resulting from pademelon browsing. Cut shoots should therefore be useful as seedling analogues, offering a potential for better experimental replication than planted seedlings. Both data sets agree in their support of the proposition that pademelon predation on seedlings is likely to suppress differentially the regeneration of certain plant species at rainforest edges.

Introduction

High mortality during the seedling stage may limit the number of recruits in a plant population (Augspurger, 1984), and one component of this mortality is damage due to grazing or browsing by mammals. Predation by mammals has been shown to strongly affect seedling establishment in tropical rainforests (Sork, 1987, Howe, 1990, Osunkoya et al., 1992), and may even cause local extinction of certain plants through species-specific herbivory (Myster and McCarthy, 1989).

Rainforests of eastern Australia were extensively cleared and fragmented during the 19th and early 20th centuries, mainly to provide pasture and agricultural land (Webb, 1966). These land uses, practised throughout the twentieth century, have created habitat mosaics in which forest remnants abut pasture areas of low biomass and complexity, but in which the productivity and palatability of the pasture grasses may be high. In such situations, ecological processes at habitat edges become increasingly important (Murcia, 1995, Holl and Lulow, 1997). Regeneration of rainforest is dependent on seed dispersal into the cleared areas, and this is commonly highest near the forest edge (Gorchov et al., 1993), potentially resulting in progressive forest regrowth from the edge into the pasture. However, progress from the seed dispersal stage to that of actively growing forest plants may be influenced by the activities of vertebrate herbivores.

Forest animal species show diverse responses to the creation of forest–pasture ecotones, varying among species from avoidance to preference (Murcia, 1995). The remnant subtropical rainforests of eastern Australia contain a variety of mammalian herbivores, including pademelons Thylogale spp., wallabies Wallabia spp., possums Trichosurus spp. and bandicoots Perameles spp. Among these, the red-necked pademelon T. thetis (Lesson) is locally abundant at rainforest–pasture ecotones (Johnson, 1980, Johnson, 1983, Wahungu et al., 1999, Wahungu et al., 2001). These small macropodid marsupials (weight up to about 9 kg, head-body length up to about 600 mm; Johnson, 1983) use the rainforest as cover and feed mainly from dusk to dawn in the pasture, up to 100 m from the forest edge (Johnson, 1980). Within this zone, most feeding activity is concentrated within 15 m of the edge, in proximity to shelter from predators (Wahungu et al., 1999, Wahungu et al., 2001).

The spatial pattern of occurrence of small-bodied mammalian herbivores at forest–pasture ecotones has predictable characteristics that are a consequence of decision trade-offs between the use of the forest for cover and the open areas for feeding (e.g. Newman et al., 1988, Lima and Dill, 1990, Cowlishaw, 1997, Wahungu, 1998, Wahungu et al., 2001). The density of food resources in the pasture may become depleted close to the ecotone (where predation risk is low), so that the resource yield increases with distance from the edge (cf. Covich, 1976), but the risk of predation also increases when animals travel from mature forest into exposed areas (Newman et al., 1988, Cassini, 1991). Therefore, the animals may frequently feed within threshold distances that are a trade-off between resource gains and predation risk (Covich, 1976).

Previous research using cut shoots of different plant species and different life forms (Wahungu et al., 1999; Wahungu et al., unpublished) placed at varying positions in relation to rainforest–pasture ecotones showed that: (1) the level of shoot damage due to pademelon herbivory was higher within the pasture than in the forest; (2) in both habitats, the intensity of damage decreased with increasing distance from the forest edge; and (3) the damage level varied greatly among plant species and life forms. Therefore, patterns of selective herbivory by pademelons may determine where seedlings of rainforest plants will successfully establish, and influence the trajectory of secondary succession at a site. However, robust conclusions concerning the implications of these findings for plant recruitment and forest succession cannot be made unless it is also demonstrated that the pattern of predation observed in shoots is similar to that observed in seedlings.

This paper therefore examines whether selective herbivory by pademelons causes differential survival and growth in seedlings of different species, and whether this is modified by distance from a rainforest–pasture ecotone, into either forest or pasture. We report the results of a field experiment that tested the effects of the following factors on seedling survival and growth, at a single site: protection from grazing, plant species, and position in relation to the edge. We also assess the strength of correlation between previously reported levels of short-term browsing damage to cut shoots (Wahungu et al., 1999) and the longer term survival and growth of seedlings measured in the present study.

Section snippets

Study area and site

The study was carried out at an interface between upland rainforest and pasture on the Lamington Plateau (altitude 700–1000 m, 28°14′S, 153°08′E) of southeast Queensland, where the rainforest mainly comprises higher altitude and high rainfall communities of ‘cool’ complex notophyll vine forest (Young and McDonald, 1987). In some areas, the level plateaux have been cleared of rainforest and converted to pasture for cattle grazing. Sharp ecotones where rainforest adjoins pasture occur in such

Variation in growth and survival with caging, species and distance

The number of seedlings surviving was strongly influenced by caging (t=5.6, P<0.001, N=48 pairs of distance–species combinations). Of the caged seedlings, 88% were alive at the end of the 6-month experimental period compared with 61% of the uncaged seedlings. Furthermore, surviving caged seedlings grew twice as many leaves as surviving uncaged seedlings (cagedmean=35.2±2.6S.E., N=48, uncagedmean=18.4±5.3S.E., N=48, t=3.4, P<0.01), and also were much taller at the end of the 6-month period (caged

Factors leading to observed seedling performance patterns

In the present study, the observed caged minus uncaged growth rates of seedlings varied greatly, from highly positive to highly negative. The sign and magnitude of these values indicate the effect of caging. If caging had no effect on seedling growth, there should be a caged minus uncaged growth rate of zero. Values would then be positive if vertebrate herbivores, which were excluded from the cages, suppressed growth outside the cages. Species preferred by the herbivores should show this

Acknowledgements

We are grateful to Wade Hadwen and Peter O’Reilly (Jr.) for their help with field logistics, and thank John Creed for designing and making the exclusion cages. Bill Magnusson, Steve Mackay and Mark Kennard gave helpful advice regarding data analysis. This research was carried out under the Queensland Department of Environment Scientific Services Permit No. E3/001156/97/SAA. The study was done under an Australian Agency for International Development (AusAID) scholarship.

References (36)

  • R.M.A. Gill

    A review of damage by mammals in north temperate forests. 3. Impact on trees and forests

    Forestry

    (1992)
  • D.L. Gorchov et al.

    The role of seed dispersal in the natural regeneration of rainforest after strip-cutting in the Peruvian Amazon

    Vegetatio

    (1993)
  • D.K. Holl et al.

    Effects of species, habitat, and distance from edge on post-dispersal seed predation in a tropical rainforest

    Biotropica

    (1997)
  • Hopkins, M.S., 1975. Species pattern and diversity in the subtropical rainforest. Ph.D. Thesis. University of...
  • H.F. Howe

    Survival and growth of juveniles of Virola surinamensis in Panama: effects of herbivory and canopy closure

    J. Trop. Ecol.

    (1990)
  • F.M. Jaksic et al.

    Why are native herbs in the Chilean matorral more abundant beneath bushes: microclimate or grazing?

    J. Ecol.

    (1980)
  • K.A. Johnson

    Spatial and temporal use of habitat by the red necked pademelon Thylogale thetis

    Aust. Wildl. Res.

    (1980)
  • Johnson, K.A., 1983. Red-necked pademelon. In: Strahan, R. (Ed.), The Australian Museum Complete Book of Australian...
  • Cited by (12)

    • Plant species richness in the Chaco Serrano Woodland from central Argentina: Ecological traits and habitat fragmentation effects

      2006, Biological Conservation
      Citation Excerpt :

      Edge effects could be involved in species–area relationships, since the proportion of edge habitat is inversely related to area (Connor and McCoy, 2001). In such edges, alterations of microclimatic conditions (Kapos et al., 1997; Didham and Lawton, 1999; Mesquita et al., 1999; Sizer and Tanner, 1999) and biological interactions (Bresciano et al., 1999; Bruna, 1999; Arnold and Asquith, 2002; Wahungu et al., 2002; Donoso et al., 2003) can increase plant mortality (Mesquita et al., 1999; Laurance et al., 2000) and regeneration (Sizer and Tanner, 1999), resulting in different patterns of species number and composition from those found in the interior of continuous forests (Harper et al., 2005). After edge formation, dynamic processes involving regeneration and changes in species composition can seal open spaces, reducing the extent of matrix influence into the forest (Didham and Lawton, 1999; Laurance et al., 2002; Harper et al., 2005).

    View all citing articles on Scopus
    1

    Present address: Department of Wildlife Management, Moi University, PO Box 1125, Eldoret, Kenya.

    View full text