Checklist of the Ants of Fiji (Hymenoptera: Formicidae)

Get Complete Project Material File(s) Now! »

Canopy sampling

The foliage of epiphytes, ferns, palms and trees (hereafter referred to as ‘trees’) were sampled by beating. Foliage was brushed/tapped with a 2 m long wooden stick five times to dislodge ants onto a white calico collecting sheet (110 x 75 cm). Sampled foliage was approximately 2 – 4 m off the ground, and is hereafter referred to as ‘canopy sampling’ (sensu Moffett 2000). Although, canopy sampling traditionally refers to the upper areas of trees, Moffett’s (2000) review of canopy sampling terminology advocates that any sampling above the level of the ground should be classified as canopy. There are no identification guides to the native plant species in Fiji or Colo-i-suva Forest Park. Specimens of foliage were taken to park staff to obtain identification, and the publications of Smith (1979) and Watling (2005) used for diagnostics and information. Canopy samples were classified into five broad categories: mahogany (Swietenia macrophylla), palms (Metroxylon vitiense, Pinanga coronata), tree ferns (Cyathea hornei), native tree species (mixture of Calophyllum vitiense, Palaquium spp., Endiandra spp., Canarium vitiense and Garcinia nr vitiensis), and epiphytes (unidentified species). Epiphytes were sampled because several ant species from Fiji are known to inhabit them (J. Wetterer pers. comm.).

Baiting experiment

Within the park, five transects (150 m) were marked, starting 1m from the main edge (road, or walking track) and running into the forest interior, along a north-south axis. Ten stations were located along the 150 m transects, 15 m apart. At each station, three microhabitats were examined; under the litter, on top of litter, and on canopy vegetation approximately 2 m off the ground (see earlier comment on the definition of canopy, Moffett 2000). In order to distinguish the two types of canopy sampling used in this study (i.e. beating and baiting), I refer to canopy baiting as the ‘shrub layer’. Within each microhabitat, three types of baits were used; cotton wool soaked in a saturated sucrose solution (sugar bait), cotton wool soaked in soy cooking oil (oil bait), and tuna (Sealord™ chunky style tuna in spring water, tuna bait). Each cotton wool ball was approximately 5 ml in volume, with an exposed surface area of 4 – 7cm2 . Approximately 2 g of tuna was used in each vial. Fresh baits were placed into a 25 ml plastic vial (25 mm diameter). For the shrub microhabitat, the vial was tied with wire to vegetation approximately 2 m off the ground. The baits within each microhabitat were placed in a triangular array, equidistant from each other with 30 cm spacing and between microhabitats (within a station) there was at least 1 m spacing. After one hour, vials were collected, capped and returned to the laboratory.

Abundance, composition and niche overlap at baits

To examine the abundance of ants at different microhabitat and food resources, each resource state was summed across all stations within a transect. For example, the number of ants caught from the resource state of ‘oil-shrub’ was summed across all stations within a transect. A two-factor ANOVA was used to examine microhabitat and bait on abundance (log transformed data) using SPSS v12.0.2 software. Transects were used as a covariate with Bonferroni post-hoc tests. Differences in the composition of ant species from different microhabitat and food resources were determined by nonmetric multidimensional scaling in PRIMER, using a Bray-Curtis similarity matrix on presenceabsence data from 10 runs (Clarke & Warwick 2005). Pairwise tests between microhabitats and food types were examined using Analysis of Similarities (ANOSIM) with 999 permutations.

Abundance, composition and niche overlap at baits

Abundance data were skewed, but log transformed data were normally distributed (one sample Kolmogorov-Smirnov Z = 0.921, p = 0.365). There was a significant difference in the abundance of ants between microhabitats (two factor ANOVA, F = 26.53, d.f. = 2, p < 0.001, Figure 4.2), but abundance was not significantly different between bait types (F = 1.16, d.f. = 2, p < 0.323). Bonferroni post-hoc tests showed that both litter microhabitats had significantly more ants than shrub (p < 0.05). There was no effect of transect as a covariate (F = 0.005, p < 0.942). However, there was some evidence for an interaction between microhabitat and bait (F = 2.45, d.f. = 4, p < 0.065), with a higher abundance of ants collected from tuna on shrub, than from other baits (oil and sugar) on shrub. The abundance of the most common invasive species, P. vaga, was significantly lower in the shrub microhabitat than either on the top or under the litter (F = 10.76, d.f. = 2, p < 0.001).

READ The multidimensionality and dynamic nature of relevance

Sampling

At each site, the foliage of ten plants was brushed/tapped with a wooden stick five times to dislodge ants onto a white calico collecting sheet (110 x 75 cm). Ants were collected with an aspirator and placed into 75% ethanol. Plants were haphazardly chosen for beating each month. At each site 12 pitfall traps were set in a 6×2 grid with 5 m spacing. Each trap consisted of a 100 mm deep plastic cup with a diameter of 105 mm containing 100 ml of a 75% ethanol to mono-propylene glycol mix (70/30), sunk vertically in the ground. A lid was secured a few centimetres above the trap to minimise debris entering the trap. Traps were left open for 7 days per month and pooled into a ‘site sample’. Sampling was undertaken once a month for twelve months; 23 February 2005 – 2 March, 18 – 25 March, 7 – 14 April, 12 – 19 May, 16 – 23 June, 14 – 21 July, 18 – 25 August, 15 – 22 September, 20 – 27 October, 17 – 24 November, 13 – 20 December, 17 – 24 January 2006.

Contents :

Abstract
Acknowledgements
Contents
List of Tables
List of Figures
List of Appendices
Chapter One: General Introduction
- 1.1 Biological invasions
  - 1.1.1 Overview
  - 1.1.2 Definitions
- 1.2 Invasive ants
  - 1.2.1 Attributes of invasive ants
  - 1.2.2 Susceptibility of communities to invasive ants
  - 1.2.3 Biological invasions: making the link
  - 1.2.4 The distribution of invasive ant species
- 1.3 Aims and layout of thesis
- 1.4 References
Chapter Two: Checklist of the Ants of Fiji (Hymenoptera: Formicidae)
- Abstract
- 2.1 Introduction
- 2.2 Methods
  - 2.2.1 Invasive ants in the Pacific
  - 2.2.2 The Islands of Fiji
  - 2.2.3 Species records from Fiji
- 2.3 Results and Discussion
  - 2.3.1 Invasive ants in the Pacific
  - 2.3.2 The ant species of Fiji
  - 2.3.3 Biogeographical origins and diversification of Fijian endemic species
  - 2.3.4 Published records
  - 2.3.5 Conclusions
- 2.4 References
Chapter Three: Coexistence, Habitat Patterns and the Assembly of Ant Communities in the Yasawa Islands, Fiji
- Abstract
- 3.1 Introduction
- 3.2 Methods
  - 3.2.1 The Yasawa Islands
  - 3.2.2 Sampling
  - 3.2.3 Food baiting experiment
  - 3.2.4 Specimen curation
  - 3.2.5 Statistical analyses
- 3.3 Results
  - 3.3.1 Faunal composition
  - 3.3.2 Coexistence in litter communities
  - 3.4 Discussion
  - 3.4.1 Conclusions
- 3.5 References
Chapter Four: Ecological Partitioning and Invasive Species in a Tropical Rainforest Community
- Abstract
- 4.1 Introduction
- 4.2 Methods
  - 4.2.1 Study site
  - 4.2.2 Canopy sampling
  - 4.2.3 Litter sampling
  - 4.2.4 Baiting experiment
  - 4.2.5 Specimen identification
  - 4.2.5 Statistical analyses
- 4.3 Results
  - 4.3.1 Diversity
  - 4.3.2 Body size
  - 4.3.3 Abundance, composition and niche overlap at baits
- 4.4 Discussion
- 4.5 References
Chapter Five: The Role of Habitat and Competition in Shaping Ant Communities in New Zealand
- Abstract
- 5.1 Introduction
- 5.2 Methods
  - 5.2.1 Habitat partitioning
  - 5.2.2 Competitive interactions
  - 5.2.3 Identification and curation
- 5.3 Results
Chapter Six: Transferability of Distribution Models for Two Invasive Ant Species
Chapter Seven: Modeling the Potential Geographic Distribution of Invasive Ant Species in New Zealand
Chapter Eight: The Diversity and Origin of Exotic Ants Arriving in New Zealand via Human-Mediated Dispersal
Chapter Nine: General Discussion