Historical and Contemporary Fish Dispersal in Australia's Northern Rivers

Author Mills, Courtenay Elyse

Published 2012

Thesis Type Thesis (PhD Doctorate)

School Griffith School of Environment

DOI https://doi.org/10.25904/1912/728

Downloaded from http://hdl.handle.net/10072/366825

Griffith Research Online https://research-repository.griffith.edu.au

Historical and contemporary fish dispersal in Australia's northern rivers

Courtenay Elyse Mills

Bachelor of Arts in Environmental Management and Policy Bachelor of Science with Honours in Australian Environmental Studies

Griffith School of Environment with the Australian Rivers Institute Griffith University

Submitted in fulfilment of the requirements of the degree of Doctor of Philosophy

October 2011

Synopsis

Riverine environments present specific barriers and challenges to dispersing fish, depending on their dispersal capability. While some barriers are obvious and finite (for example a waterfall), the subtle influences of dendritic river network structure can also affect the dispersal, and therefore the genetic structure, of freshwater fish within river catchments. In addition to these natural barriers, artificial barriers like weirs and dams can have significant effects on the flow of genes between populations in rivers.

In order to achieve the effective management and conservation of populations there is a need to understand dispersal and how the movement of individuals impacts populations at a variety of scales. Studies of dispersal are important for the perspective they bring to the movement and migration of individuals and the subsequent flow of genes – both temporally and geographically. It is this spread of genes that largely determines the genetic diversity of populations, which is a quality necessary for the stability and persistence of individuals and populations.

This study aimed to understand how fish disperse through rivers, and the factors that have limited, prevented or aided dispersal in past populations. Specifically, the impacts of the structure and form of river catchments on dispersal were investigated in populations of chequered rainbowfish (Melanotaenia splendida inornata), western rainbowfish (Melanotaenia australis) and sooty grunter (Hephaestus fuliginosus) in the

Mitchell and Daly Rivers in tropical Australia. It was predicted that Melanotaenia spp. would be highly influenced by barriers in these tropical catchments. If this prediction holds true, patterns of gene flow would correspond to the environmental characteristics of distance, elevation and waterway classification. Mitochondrial and microsatellite

Synopsis 2 markers were used to quantify gene flow in the Mitchell and Daly Rivers. Both M. australis and M. s. inornata were strongly structured within the catchments, and restrictions to dispersal were determined by distance (that is isolation by distance

(IBD)), confirming the limited dispersal capabilities for these species. Elevation was also a factor inhibiting population connectivity for M. s. inornata. In addition to these patterns, both the position of a population in the catchment and the type of waterway

(for example isolated waterholes or large perennial rivers) also affected the connectivity within M. australis and M. s. inornata. Overall, there was also evidence for a source – sink metapopulation dynamic in both the Daly and Mitchell Rivers.

For H. fuliginosus, which is considered to be a comparatively strong disperser, it was predicted that there would be relatively unrestricted patterns of gene flow within the study catchments – that is, there would be little influence from distance, elevation or waterway classification. This expectation was confirmed by the low levels of genetic structure detected in both catchments which suggests that this species was able to overcome barriers, although a weak IBD signature was detected in the Daly River. H. fuliginosus is highly dependant on riffles as nursery areas for offspring, and so it was expected that increased riffle access would result in increased connectivity – a relationship that was not supported by the data. It was also predicted that increased riffle habitat would support a larger population, resulting in increased genetic diversity.

Surprisingly, the reverse was found. Evidence for a population expansion in the Daly

River, dated at 120 200 – 262 300kya, suggested that these conclusions of patterns of gene flow must be accepted with caution as gene flow – genetic drift equilibrium may have not yet established.

Synopsis 3

One of the limitations of genetic markers of dispersal is that they integrate generational information, and may not reveal the movements made by individuals within their lifetimes. To overcome this constraint, a dual marker approach was adopted which combined genetic markers with stable isotope markers to determine dispersal.

Specifically, this study aimed to investigate how wet season flooding facilitates dispersal in the headwaters of the Mitchell River. Given that the question of interest spanned a single wet season (a period of several months) the application of ecological markers like stable isotopes was critical. M. s. inornata were collected from four sites in the upper Mitchell River catchment during the dry season and wet season. Two potential types of individual movement – short migratory events (found through the use of stable isotopes) and contemporary dispersal (found through the use of microsatellite markers) were revealed. Gene flow was very high between the sampled sites, effectively making all individuals from the same population. Due to these high rates of dispersal, the genetic markers were not able to detect any specific migrants or a general increase in movement during the wet season. In contrast, the use of stable isotopes revealed migration of fish during flooding, with individuals moving to floodplains and backwaters over the wet season when they could gain access to high quality yet temporally available food resources. The use of stable isotope and genetic markers in this study enabled an ecologically relevant understanding of patterns of movement in this species over time.

It is essential to have a thorough knowledge of genetic variation within and between populations when implementing management strategies. Understanding the manner in which historical river structure can influence population genetic structure is an important part of managing river functionality and health, particularly for river systems under growing development pressure (like those examined in this study). By

Synopsis 4 determining the landscape features that have affected dispersal, these influences can be extrapolated and applied to contemporary populations – thereby informing how we may expect populations to react and be influenced by future artificial and natural environmental and landscape changes.

Synopsis 5

Statement of Originality

This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.

Courtenay Elyse Mills 28th October 2011

Statement of Originality 6

Table of Contents

Synopsis ...... 2 Statement of Originality ...... 6 Table of Contents ...... 7 List of Figures ...... 10 List of Tables ...... 15 Acknowledgements ...... 19 1.0 Introduction ...... 21 1.1 Dispersal and gene flow in aquatic environments ...... 21 1.2 Models of gene flow in aquatic environments ...... 29 1.3 Applying genetic markers in studies of dispersal ...... 34 1.3.1 Mitochondrial DNA ...... 36 1.3.2 Microsatellites ...... 37 1.4 Genetic legacies in tropical rivers ...... 38 1.5 Using stable isotopes to measure dispersal ...... 42 1.6 Combining methods to investigate dispersal ...... 44 1.7 Australia’s tropical rivers ...... 45 1.8 Australia’s tropical freshwater fish ...... 47 1.8.1 The chequered and western rainbowfish (Melanotaeniidae: Castelnau 1875) ...... 47 1.8.2 The sooty grunter (Terapontidae: Macleay 1883) ...... 48 1.9 Aims, objectives and thesis outline ...... 48 2.0 Field and Laboratory Methods ...... 51 2.1 Study design ...... 51 2.2 Field sampling ...... 52 2.2.1 General fish collection methods ...... 54 2.2.2 Collection of samples for SIA ...... 57 2.2.3 Study species ...... 58 2.3 Laboratory methods ...... 60 2.3.1 DNA extraction ...... 60 2.3.2 Mitochondrial and microsatellite primers ...... 61 2.3.3 Reaction and cycling conditions ...... 63 2.3.3.1 Control region and ATPase 6 and 8 ...... 63 2.3.3.2 Microsatellites ...... 64 2.3.4 mtDNA sequencing ...... 65 2.3.5 Stable isotope preparation ...... 66 2.4 Statistical analyses of data ...... 67 2.4.1 Genetic diversity ...... 67 2.4.2 Analyses of molecular variance ...... 68 2.4.3 Testing for isolation by distance ...... 69 2.4.4 Testing for neutrality ...... 69 2.4.5 Assignment testing ...... 72 2.4.6 Pairwise regressions ...... 73 2.4.7 Stable isotope analysis ...... 75 3.0 The influence of the structure and form of dendritic systems on gene flow in rainbowfish populations of tropical Australia ...... 78 3.1 Introduction ...... 78 3.2 Detailed Methods ...... 82 3.2.1 Classification of waterways...... 82 3.2.2 Isolation of waterways ...... 82 3.2.3 Testing landscape influences ...... 83

Table of Contents 7

3.3 Results ...... 84 3.3.1 Intrapopulation genetic diversity ...... 84 3.3.2 Tests of neutrality ...... 95 3.3.3 Partitioning of genetic variation ...... 99 3.3.4 Isolation indices ...... 108 3.3.5 Mantel tests ...... 115 3.3.6 Regression analyses ...... 116 3.4 Discussion ...... 124 3.4.1 Genetic diversity ...... 124 3.4.2 Population expansions and selection ...... 125 3.4.3 Population structure ...... 126 3.4.4 Influence of the landscape on connectivity ...... 129 3.4.5 The influence of elevation and distance on isolation ...... 131 3.5 Conclusion ...... 133 4.0 Does the landscape affect strong dispersers? Investigating riverine influences on gene flow in sooty grunters (Hephaestus fuliginosus)...... 134 4.1 Introduction ...... 134 4.2 Detailed Methods ...... 140 4.2.1 Calculations of Tau ...... 140 4.2.2 Classification of waterways...... 140 4.2.3 Isolation of waterways ...... 141 4.2.4 Testing landscape influences ...... 141 4.2.5 Identification of riffles ...... 142 4.3 Results ...... 144 4.3.1 Intrapopulation genetic diversity ...... 144 4.3.2 Tests of neutrality ...... 150 4.3.3 Partitioning of genetic variation ...... 153 4.3.4 Isolation indices ...... 155 4.3.5 Mantel tests ...... 157 4.3.6 Regression analyses ...... 157 4.4 Discussion ...... 167 4.4.1 Genetic diversity ...... 167 4.4.2 Population expansion in the Daly River ...... 167 4.4.3 Genetic structure and dispersal capability ...... 169 4.4.4 Influence of the landscape on connectivity ...... 172 4.5 Conclusion ...... 176 5.0 The rainbow connection: using stable isotopes and population genetics to evaluate the importance of floods to rainbowfish ...... 179 5.1 Introduction ...... 179 5.2 Detailed Methods ...... 185 5.2.1 Site selection ...... 185 5.2.2 AMOVA design ...... 185 5.2.3 Assignment indices ...... 186 5.2.4 Power testing ...... 186 5.3 Results ...... 188 5.3.1 Data collection ...... 188 5.3.2 Microsatellite analyses ...... 189 5.3.3 Stable isotope analyses ...... 193 5.4 Discussion ...... 198 5.4.1 Genetic differentiation ...... 198 5.4.2 Isotopic differentiation ...... 199 5.5 Conclusion ...... 203

Table of Contents 8

6.0 Conclusion ...... 205 6.1 Patterns of gene flow in tropical catchments...... 205 6.2 Migration in tropical catchments ...... 207 6.3 Management implications for Melanotaenia spp...... 208 6.4 Management implications for H. fuliginosus ...... 210 6.5 Management implications for fish in tropical Australia ...... 211 References...... 212 7.0 Appendix ...... 279

Table of Contents 9

List of Figures

Figure 1.1 Model of IBD along a river (blue arrow indicates direction of flow). Each circle represents a population which exchanges genes with the adjacent populations. Colours represent the gradient of genetic differentiation that has led to the isolation of the red population (upstream) from the green population (downstream)...... 31

Figure 1.2 The Stream Hierarchy Model (SHM), in this example showing how the number of haplotypes (represented by individual colours) contained within a population (represented by circles) increases with the number of stream branches, but also that gene flow within a branch is more likely than between branches...... 33

Figure 1.3 The Death Valley Model (DVM), showing three isolated populations between which no gene flow occurs (migration (m) = 0). Haplotypes are represented by individual colours, showing that each population has become highly divergent...... 33

Figure 1.4 The Headwater Model (HM), showing that dispersal is more likely to occur across land barriers between headwater sites, than via the stream channel. Each population is represented by a circle, with colours indicating the haplotype of a given population...... 34

Figure 1.5 Metapopulation models, showing a) the classic model where cycles of extinction and recolonisation occur between neighbouring populations; and b) the mainland – island model, where the source population (mainland) ‘supplies’ the islands. In this example, circles represent isolated aquatic habitats within the landscape. Arrows show the direction of dispersal...... 35

Figure 2.1 The Daly River catchment (NT) and the Mitchell River catchment (QLD) in Australia...... 53

Figure 2.2 Two of the sampled sites showing variation in flow, where a and b show baseline flow for Bushy Creek and Cooktown Crossing, respectively; and c and d show monsoonal flow for Bushy Creek and Cooktown Crossing, respectively...... 54

Figure 2.3 a) Saltwater crocodile (Crocodylus porosus) at a sampling site (Mt Nancar) in the Daly River; b) inspecting the seine net after use; c) front view of the electrodes; and d) electrofishing in Kingfish Lagoon in the Mitchell River catchment...... 55

Figure 2.4 TRaCK researchers Tim Jardine (R) and Richard Hunt (L) electrofishing a small stream in the Mitchell River (photograph courtesy of Dominic Valdez)...... 56

Figure 2.5 M. s. inornata individual recovering after capture...... 57

Figure 2.6 Melanotaenia australis from the Daly River main channel (left); Melanotaenia splendida inornata from the headwaters of the Mitchell River (right).59

Figure 2.7 Example of an M. s. inornata male with the pronounced deep body...... 60

Figure 2.8 H. fuliginosus from the upper Mitchell River...... 60

List of Figures 10

Figure 2.9 Relationships between landscape variables (in this example – geographic distance) and genetic distances. Individual regressions show the predicted patterns for the different balances of genetic drift and gene flow...... 76

Figure 3.1 The Daly River catchment, showing the 17 sites from which M. australis was collected...... 85

Figure 3.2 The Mitchell River catchment, showing the 16 sites from which M. s. inornata was collected...... 86

Figure 3.3 The Daly River catchment, showing the geographic distribution of M. australis haplotypes. In the top diagram, circles represent each sampled population, with individual colours representing the proportions of haplotypes found within them. In the bottom diagram, circles represent each individual haplotype (frequency corresponds to circle size), with the individual colour corresponding to the haplotypes represented in mapped populations. Small black circles indicate haplotypes that were not sampled, or have become extinct through genetic drift...... 87

Figure 3.4 The Mitchell River catchment, showing the geographic distribution of M. s. inornata haplotypes. In the top diagram, circles represent each sampled population, with individual colours representing the proportions of haplotypes found within them. In the bottom diagram, circles represent each individual haplotype (frequency corresponds to circle size), with the individual colour corresponding to the haplotypes represented in mapped populations. Small black circles indicate haplotypes that were not sampled, or have become extinct through genetic drift...... 88

Figure 3.5 Mismatch distributions modelled for M. australis, showing a) the expected and observed pairwise differences for a population of historically constant size; b) the expected and observed pairwise differences for an expanding population...... 98

Figure 3.6 Mismatch distributions modelled for M. s. inornata, showing a) the expected and observed pairwise differences for a population of historically constant size; b) the expected and observed pairwise differences for an expanding population...... 99

Figure 3.7 Magnitude of ΔK for each estimation of K (1 – 20) for M. australis. ... 102

Figure 3.8 Proportions of membership of M. australis individuals to each of the three groups determined by STRUCTURE v2.3.3...... 103

Figure 3.9 Magnitude of ΔK for each estimation of K (1 – 20) for M. s. inornata. 104

Figure 3.10 Proportions of membership of M. s. inornata individuals to each of the three groups determined by STRUCTURE v2.3.3...... 104

Figure 3.11 Isolation indices for all sampled populations of M. australis based on a) mean FST for ATPase 6 and 8, b) mean ΦST for ATPase 6 and 8, and c) mean FST for microsatellites. Populations are classed as being downstream (blue), midreach (light green) or upstream (peach) and show 95% confidence intervals...... 113

Figure 3.12 Isolation indices for all sampled populations of M. s. inornata based on a) mean FST for ATPase 6 and 8, b) mean ΦST for ATPase 6 and 8, and c) mean FST for

List of Figures 11 microsatellites. Populations are classed as being downstream (blue), midreach (light green) or upstream (peach) and show 95% confidence intervals...... 114

Figure 3.13 Isolation indices for all sampled populations of M. australis based on a) mean FST for ATPase 6 and 8, b) mean ΦST for ATPase 6 and 8, and c) mean FST for microsatellites. Populations are classed as being in the main channel (green), a tributary (orange) or a waterhole (teal) and show 95% confidence intervals...... 115

Figure 3.14 Isolation indices for all sampled populations of M. s. inornata based on a) mean FST for ATPase 6 and 8, b) mean ΦST for ATPase 6 and 8, and c) mean FST for microsatellites. Populations are classed as being in the main channel (green), a tributary (orange) or a waterhole (teal) and show 95% confidence intervals...... 116

Figure 3.15 Mean regression residuals and 95% confidence intervals for decomposed regression analysis. Example taken from correlations with distance for M. australis...... 117

Figure 3.16 Model of best fit for the relationship between FST and stream distance for six populations of M. australis in the Daly River, showing a strong positive correlation (R2 = 0.839, P = 0.000)...... 118

Figure 3.17 Model A of best fit for the relationship between FST and stream distance for eight populations of M. s. inornata in the Mitchell River, showing a strong positive correlation (R2 = 0.887, P = 0.000)...... 118

Figure 3.18 Model A of best fit for the relationship between FST and stream distance for seven populations of M. s. inornata in the Mitchell River, showing a strong positive correlation (R2 = 0.926, P = 0.000)...... 119

Figure 3.19 Model C of best fit for the relationship between FST and elevation for all populations of M. australis in the Daly River, showing a weak, non-significant positive correlation (R2 = 0.029, P = 0.614)...... 119

Figure 3.20 Model D of best fit for the relationship between FST and elevation for 10 populations of M. australis in the Daly River, showing a weak, non-significant positive correlation (R2 = 0.070, P = 0.459)...... 120

Figure 3.21 Model E of best fit for the relationship between isolation and elevation for 10 populations of M. s. inornata in the Mitchell River, showing a strong positive correlation (R2 = 0.836, P = 0.000)...... 120

Figure 3.22 Model F of best fit for the relationship between isolation and elevation for 11 populations of M. s. inornata in the Mitchell River, showing a strong positive correlation (R2 = 0.625, P = 0.004)...... 121

Figure 3.23 Decomposed pairwise regression between FST and distance for M. australis in the Daly River. ‘True outliers’ are represented by dashed lines...... 123

Figure 3.24 Decomposed pairwise regression between FST and distance for M. s. inornata in the Mitchell River. ‘True outliers’ are represented by dashed lines. ... 123

List of Figures 12

Figure 4.1 Example of the elevation profile used to identify potential riffles in the river beds between sites DS to DB. Rectangles highlight the identified riffles...... 144

Figure 4.2 The Daly River catchment, showing the 23 sites from which H. fuliginosus was collected...... 145

Figure 4.3 The Mitchell River catchment, showing the 16 sites from which H. fuliginosus was collected...... 146

Figure 4.4 The Daly River catchment, showing the geographic distribution of H. fuliginosus haplotypes. In the top diagram, circles represent each sampled population, with individual colours representing the proportions of haplotypes found within them. In the bottom diagram, circles represent each individual haplotype (frequency corresponds to circle size), with the individual colour corresponding to the haplotypes represented in mapped populations. Small black circles indicate haplotypes that were not sampled, or have become extinct through genetic drift...... 148

Figure 4.5 The Mitchell River catchment, showing the geographic distribution of H. fuliginosus haplotypes. In the top diagram, circles represent each sampled population, with individual colours representing the proportions of haplotypes found within them. In the bottom diagram, circles represent each individual haplotype (frequency corresponds to circle size), with the individual colour corresponding to the haplotypes represented in mapped populations. Small black circles indicate haplotypes that were not sampled, or have become extinct through genetic drift...... 149

Figure 4.6 Mismatch distributions modelled for H. fuliginosus in the Daly River, showing a) the expected and observed pairwise differences for a population of historically constant size; b) the expected and observed pairwise differences for an expanding population...... 153

Figure 4.7 Isolation indices for all sampled populations of H. fuliginosus from the Daly River based on a) mean FST for control region and b) mean ΦST for control region. Populations are classed as being downstream (blue), midreach (light green) or upstream (peach)...... 157

Figure 4.8 Isolation indices for all sampled populations of H. fuliginosus from the Daly River based on a) mean FST for control region and b) mean ΦST for control region. Populations are classed as being in the main channel (green) or a tributary (orange)...... 157

Figure 4.9 Mean regression residuals and 95% confidence intervals for decomposed pairwise regression analysis. Example taken from correlations with distance for Daly River H. fuliginosus. Here, site DR was identified as the first outlier i.e. the greatest deviant from 0...... 159

Figure 4.10 Model A of best fit for the relationship between FST and stream distance for H. fuliginosus in the Daly River, showing a positive correlation (R2 = 0.155, P = 0.003). Four outlier populations were removed...... 159

Figure 4.11 Model B of best fit for the relationship between FST and stream distance for H. fuliginosus in the Daly River, showing a positive correlation (R2 = 0.100, P = 0.005). Two outlier populations were removed...... 160

List of Figures 13

Figure 4.12 Model C of best fit for the relationship between FST and stream distance for H. fuliginosus in the Daly River, showing a positive correlation (R2 = 0.129, P < 0.001). Outlier population DR was removed...... 160

Figure 4.13 Model D of best fit for the relationship between FST and stream distance for H. fuliginosus in the Daly River, showing a weak positive correlation (R2 = 0.090, P = 0.014). Three outlier populations were removed...... 161

Figure 4.14 Model E of best fit for the relationship between isolation indices (mean FST) and elevation for all population of H. fuliginosus in the Daly River, showing no significant correlation (R2 = 0.002, P = 0.884)...... 163

Figure 4.15 Model F of best fit for the relationship between FST and PRUs for all populations of H. fuliginosus in the Daly River, showing no significant correlation (R2 = 0.004, P = 0.622)...... 164

Figure 4.16 Model G of best fit for the relationship between haplotype diversity and PRUs for all populations of H. fuliginosus in the Daly River, showing a weak, but non- significant negative correlation (R2 = 0.174, P = 0.122)...... 164

Figure 4.17 Model G of best fit for the relationship between haplotype diversity and PRUs for all populations of H. fuliginosus in the Daly River, showing a significant negative correlation (R2 = 0.331, P = 0.031)...... 165

Figure 4.18 Decomposed pairwise regression between FST and distance for H. fuliginosus in the Daly River. ‘True outliers’ are represented by dashed lines. .... 167

Figure 5.1 Sites sampled for M. s. inornata and primary sources of organic matter in the upper Mitchell River catchment...... 185

Figure 5.2 δ13C and δ15N for M. s. inornata at all sites in the upper Mitchell catchment. Each data point represents a single individual...... 194

Figure 5.3 δ13C and δ15N for M. s. inornata at all sites in the upper Mitchell catchment, with all individuals categorised as being captured in either the wet season (in blue), or the dry season (in red). Each data point represents a single individual...... 196

Figure 7.1 δ13C and δ15N for all samples at Bushy Creek in the upper Mitchell catchment, with all samples represented for each sampled season...... 298

Figure 7.2 δ13C and δ15N for all samples at Cooktown Crossing in the upper Mitchell catchment, with all samples represented for each sampled season...... 299

Figure 7.3 δ13C and δ15N for all samples at the McLeod River in the upper Mitchell catchment, with all samples represented for each sampled season...... 300

Figure 7.4 δ13C and δ15N for all samples at Rifle Creek in the upper Mitchell catchment, with all samples represented for each sampled season...... 301

Figure 7.5 Initial investigation into the variability of the marker ATPase 6 and 8 in H. fuliginosus. Black dots denote haplotypes that have not been sampled, or have become extinct from the population. Circle size corresponds to sample size...... 302

List of Figures 14

List of Tables

Table 2.1 Control region primer sequences used to amplify H. fuliginosus...... 61

Table 2.2 ATPase 6 and 8 primer sequences used to amplify M. australis and M. s. inornata...... 62

Table 2.3 Microsatellite primer sequences used to characterise genetic structure in M. australis and M. s. inornata. * denotes loci isolated by Zhu et al. (1998), and ^ denotes loci isolated by Young et al. (2009)...... 62

Table 2.4 Fluorescent tag sequences used in microsatellite reactions...... 64

Table 2.5 Annealing temperatures, fluorescent tags and grouping of microsatellite loci for PCR cycling...... 64

Table 2.6 Classification of patterns based on the regression intercepts and slopes.75

Table 3.1 Gene diversity indices for M. australis in the Daly River. Deviations from HWE are denoted by *, after sequential Bonferroni correction. Monomorphic loci are denoted by -. No tests were performed for populations where n < 3...... 88

Table 3.2 Gene diversity indices for M. s. inornata in the Mitchell River. Deviations from HWE are denoted by *, after sequential Bonferroni correction. Monomorphic loci are denoted by -. No tests were performed for populations where n < 3...... 90

Table 3.3 Tests of neutrality for populations of M. australis (Daly River) and M. s. inornata (Mitchell River). * denotes significant values...... 96

Table 3.4 Goodness of fit test (raggedness) for simulated population expansions in M. australis and M. s. inornata. * denotes significant values where α = 0.05...... 99

Table 3.5 AMOVA calculations for M. australis. * denotes significant fixation indices. ° denotes % variation within populations...... 100

Table 3.6 AMOVA calculations for M. s. inornata. * denotes significant fixation indices. ° denotes % variation within populations...... 100

Table 3.7 Assignment of M. australis individuals to populations using GENECLASS2 v2.0. Putative populations from which individuals could not be assigned back to any of the represented populations and denoted by + (for one individual) and * (for two individuals). ^ denotes lower catchment groups, ° denotes midcatchment groups and ˉ denotes upper catchment groups...... 106

Table 3.8 Assignment of M. s. inornata individuals to populations using GENECLASS2 v2.0. ^ denotes lower fork groups, ° denotes midcatchment groups and ˉ denotes upper fork groups...... 107

Table 3.9 ANOVA results for comparisons of isolation indices for each of the waterway classifications in M. australis and M. s. inornata. Significant values are denoted by *, where α = 0.05...... 109

List of Tables 15

Table 3.10 Post-hoc Scheffé multiple comparison tests between means for catchment positions. * denotes significant values where α = 0.05...... 110

Table 3.11 Post-hoc Scheffé multiple comparison tests between means for waterway types. * denotes significant values where α = 0.05...... 111

Table 3.12 Spearman’s rank correlations between distance from river mouth and average fixation index (isolation index) for M. australis and M. s. inornata (α = 0.05)...... 112

Table 3.13 Mantel test correlation coefficients for M. australis and M. s. inornata (α = 0.05). * denotes significant values...... 115

Table 3.14 Fit of models for distance and elevation showing all populations included, then with outliers sequentially removed. n is number of populations included in each model. Best models are shaded in grey...... 121

Table 3.15 Intercepts, slopes and regression coefficients for all tested populations of M. australis and M. s. inornata. ‘True outliers’ are shaded grey...... 123

Table 4.1 Gene diversity indices and sampling regime for H. fuliginosus in the Daly River...... 149

Table 4.2 Gene diversity indices and sampling regime for H. fuliginosus in the Mitchell River...... 150

Table 4.3 Tests of neutrality for populations of H. fuliginosus in the Daly River and Mitchell River. * denotes significant values...... 151

Table 4.4 Goodness of fit test (raggedness) for simulated population expansions in H. fuliginosus in the Daly River...... 153

Table 4.5 AMOVA calculations for H. fuliginosus in the Daly River. * denotes significant fixation indices. ° denotes % variation within populations...... 154

Table 4.6 AMOVA calculations for H. fuliginosus in the Mitchell River. * denotes significant fixation indices...... 154

Table 4.7 ANOVA results for comparisons of isolation indices for each of the waterway classifications for Daly River H. fuliginosus. No significant relationships were found (α = 0.05)...... 155

Table 4.8 Spearman’s rank correlations between distance from river mouth and average fixation index (isolation index) for H. fuliginosus in the Daly River (α = 0.05)...... 155

Table 4.9 Mantel test correlation coefficients for H. fuliginosus in the Daly and Mitchell catchments (α = 0.05)...... 157

Table 4.10 Fit of models for distance and elevation showing all populations included, then with outliers sequentially removed. n is number of populations included in each model. Best models are shaded in grey...... 161

List of Tables 16

Table 4.11 Fit of models for the PRU influence on FST and haplotype diversity – showing all populations included, then with outliers sequentially removed. n is number of populations included in each model. Best models are shaded in grey...... 165

Table 4.12 Intercepts, slopes and regression coefficients for all tested populations of H. fuliginosus. ‘True outliers’ are shaded grey...... 166

Table 5.1 Gene diversity indices for M. s. inornata collected for microsatellite and stable isotope analyses in the upper Mitchell catchment. Deviations from HWE are denoted by *, after sequential Bonferroni correction...... 189

Table 5.2 AMOVA calculations for M. s. inornata. No fixation indices were significant (α = 0.05). ° denotes % variation within populations...... 191

Table 5.3 mAI and vAI for each group sampled before and during the wet season. All values were non-significant (α = 0.05)...... 192

Table 5.4 δ13C range (CR) and δ15N range (NR) for all individuals, dry season individuals and wet season individuals...... 194

Table 5.5 P-values for the Euclidean distance between centroids (MD) for dry and wet season M. s. inornata individuals (α = 0.05). * denotes significant values...... 194

Table 5.6 P-values for tests of eccentricity (E) for each site, compared across seasons (α = 0.05)...... 195

Table 5.7 P-values for Euclidean distances from the mean sample centroid (CD), and individual’s nearest neighbour (NND) (α = 0.05)...... 196

Table 7.1 Latitude and longitude for sites at which M. australis was sampled...... 278

Table 7.2 Latitude and longitude for sites at which M. s. inornata was sampled. . 279

Table 7.3 Latitude and longitude for sites at which H. fuliginosus was sampled in the Daly River...... 280

Table 7.4 Latitude and longitude for sites at which H. fuliginosus was sampled in the Mitchell River...... 281

Table 7.5 Latitude and longitude for sites at which M. s. inornata and primary sources of organic material was sampled...... 281

Table 7.6 Haplotype distributions for M. australis in the Daly River. Haplotype frequency and sample sizes shown for all sites...... 282

Table 7.7 Haplotype distributions for M. s. inornata in the Mitchell River. Haplotype frequency and sample sizes shown for all sites...... 283

Table 7.8 Mantel test distance matrix for populations of M. australis in the Daly River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims...... 284

List of Tables 17

Table 7.9 Mantel test distance matrix for populations of M. s. inornata in the Mitchell River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims...... 285

Table 7.10 Average allelic diversity estimates for microsatellite loci in populations of M. australis in the Daly River, and M. s. inornata in the Mitchell River...... 286

Table 7.11 Haplotype distributions for H. fuliginosus in the Daly River. Haplotype frequency and sample sizes shown for all sites...... 287

Table 7.12 Haplotype distributions for H. fuliginosus in the Mitchell River. Haplotype frequency and sample sizes shown for all sites...... 288

Table 7.13 Mantel test distance matrix for populations of H. fuliginosus in the Daly River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims...... 289

Table 7.14 Mantel test distance matrix for populations of H. fuliginosus in the Mitchell River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims...... 291

Table 7.15 Standard lengths (mm) and life stage for M. s. inornata individuals analysed for both genetic and isotopic variation in the upper Mitchell River catchment. .... 292

Table 7.16 Two-tailed t-test statistics between M. s. inornata adult and juvenile isotope signatures (C13 and N15) at each sampling time. Significant values are denoted by *, where α = 0.05...... 296

List of Tables 18

Acknowledgements

My supervisors: Professor Jane Hughes has invested a lot into not only helping me with my PhD, but also in teaching me the wily ways of molecular ecology. She has spent so much time reading, and reading, and reading my thesis. Thank you for all your efforts Jane.

Doctor Wade Hadwen has really provided so much support for me (although he thinks he didn’t). He constantly encouraged me the whole way through my PhD. Wade has the knack of turning my back-to-front, circular, and flamboyant sentences into something that actually makes sense! Thank you Wade for helping me on my way.

Doctor Dan Schmidt has been the “go-to” person with any major malfunction or ignorant question. His enthusiasm in the lab is very infectious. I also appreciate his practical advice in the event of any disaster – “go eat some icecream”, or “you need some chocolate before you go any further”.

Doctor Brad Pusey was the voice of reason in many of the results. His critical thinking and probing really helped me to evaluate my data, and make my science well-rounded.

The traditional owners: I want to sincerely acknowledge and thank all the traditional owners, both past and present, of the Mitchell and Daly catchments. In particular, I would like to thank the people of Kowanyama, the Kowanyama rangers, the Wagiman people and rangers, the Western Gugu Yalanji, Gugu Mini, Koko Mullarichee, Kuku Djungan, Mbarbarum and Wokomin Aboriginal Tribes, and the Mitchell River Catchment Traditional Custodian Advisory Group. I am especially indebted to the delightful Marceil Lawrence, Ruth Link, Graham Brady, Rodney Riley, Raferty Riley and Poncie Kernoth. Without their help, guidance and knowledge, I could not have succeeded – either being lost or eaten by a crocodile.

The hunter-gatherers: Many people assisted me in the field, or went out of their way to clip fish for me on their own projects. I am deeply grateful (and also somewhat apologetic!) to Tim Jardine, Dominic Valdez, Richard Hunt, Mark Kennard, Pete Kyne, Dan Warfe, Ian Dixon, Brad Pusey, Kate Masci, Sylvain Arene, Colton Perna, Poncie Kernoth, Virgilio Hermoso, Dave Sternberg, Jimmy Fawcett, Ben Cook, Peter Graham and Greg McDonald. I am sure there were many others that were on fishing expeditions. Thank you for risking life, limb and sanity, and helping me to retain mine! In the lab, I had Tim Jardine, the Geeks (particularly Joel Huey, Kathryn Real and Arlene Butwell), Rene Diocares and Vanessa Fry assisted me with everything – from putting out small fires (on more than one occasion) to calming me when I mixed up samples. Ben Stuart-Koster was invaluable for his statistical analyses and interpretations of my SIA data set. I also would like to thank Mark Kennard and John Spence for their suggestions on my data set.

For being on my side: I am very glad to have had the help of Michael Douglas, Ruth O’Connor, Kathy Carter, Deslie Smith and Petney Dickson when trying to deal with the multitude of administrative, financial, and “really hard forms to fill out” problems. Thank you all for your support.

Acknowledgements 19

The numerous landholders, station owners and managers who graciously allowed me to sample on their property. Also the Department of Environment and Resource Management, the Australian Rivers Institute, Tropical Rivers and Coastal Knowledge and Griffith University for resources and financial support.

Also to Alisha Steward and Jon Marshall, for letting me lurk around their house when I needed a change of scenery. Nissan for making great Patrols with Cruskit’s box shaped centre consoles, Huggies for making excellent baby wipes that clean everything on fieldtrips (the bulldust out of the car, dishes, clothes and me!).

And finally….

I want to thank my family for letting me still be a university student, and learning how to say “molecular ecologist”. Thanks to Nan and Grandad for feeding me every Sunday.

And thank you to my very lovely Dr. Matthew Baddock who loves deserts, dust and inanimate rocks, but still tries to show interest when I talk about animals and rivers.

Acknowledgements 20

1.0 Introduction

1.1 Dispersal and gene flow in aquatic environments

The movements of individuals, in both space and time, can have a variety of immediate and ongoing consequences for populations and communities in both ecological and evolutionary dimensions (Nathan et al. 2003). Whether permanent or migratory, short or long distance; these movements can all have significant repercussions for population dynamics (Koenig et al. 1996; Nathan et al. 2003). Immigrants can alter ecosystem interactions such as resource competition and species co-existence; but also evolutionary directions with the introduction of new genes (Pusey & Wolf 1996; Barber et al. 2002; Laikre et al. 2010). Consequently, it is imperative to understand the mechanisms, patterns and pathways of dispersal and the spatial and temporal context of dispersal in order to appropriately and effectively manage populations.

The term ‘dispersal’ is often used interchangeably with ‘migration’, but for the purposes of this thesis ‘dispersal’ is defined as a permanent movement away from the natal population; ‘migration’ differs in that an individual may return to its source on a regular basis (Nathan et al. 2003; Song et al. 2005). When an individual disperses and is successfully recruited into a new population, potentially influential genetic additions to the existing gene pool may occur through reproduction (Allendorf 1983; Petit &

Excoffier 2009). As a result, dispersal can be identified as the mechanism of gene flow

(Slatkin 1985).

Investigating patterns and rates of gene flow can provide insights into population dynamics through rates of migration or recruitment and the delineations of populations

Chapter 1.0 ∙ Introduction 21 and metapopulations – all of which are important considerations in conservation management (Beebee & Rowe 2008). The characterisation of gene flow can be of particular significance when predicting rates of recolonisation in the event of a population extinction (Trakhtenbrot et al. 2005; Hughes 2007). If a population experiences no gene flow, this indicates that there are no immigrants to the population and therefore recolonisation is unlikely in the event of an extinction (Slatkin 1985;

Milligan et al. 1994; Hughes 2007; Lester et al. 2007). Alternatively, populations experiencing high levels of immigration and subsequent gene flow are likely to re- establish quickly (Barber et al. 2002).

Dispersal and subsequent gene flow can be highly influential in determining gene pools

(Allendorf 1983; Slatkin 1987; Hughes 2007). The exchange or directional flow of genes into populations homogenises existing gene frequencies and inhibits the development of differentiation (Mayr 1954; Slatkin 1985; Dlugosch & Hays 2008).

Alternatively, populations that experience little immigration and gene flow become differentiated and ultimately diverge due to the effects of genetic drift and/or natural selection (Nunney & Campbell 1993; Schluter 2009). Consequently, certain population characteristics can be inferred by determining the gene frequencies within and between them, and determining the factors that influence them (Slatkin 1987; Avise 2009).

Dispersal studies are also useful in predicting the impacts of changes to the natural environment on resident populations (Slatkin 1987; Lenormand 2002). For example, in riverine systems higher levels of hydrological connectivity often support higher rates of dispersal, and thus the flow of genes is sufficient to have a homogenising effect on allele frequencies (Slatkin 1985; Amoros & Bornette 2002; Burridge et al. 2008). The opposite can be observed in riverine systems that experience low, or temporally restricted hydrological connectivity (Crispo et al. 2006).

Chapter 1.0 ∙ Introduction 22

Riverine landscapes also undergo geomorphic changes that can impact dispersal and gene flow on both temporal and spatial scales (Amoros & Bornette 2002). For example, an oxbow in a river may ‘evolve’ and be transformed into a straighter channel; alternatively, over the course of a catchment, the river substrate may change from cobble to sand. The structural components of a river system are constantly evolving due to natural processes such as river capture, flow direction, flow regime (including climate variability and change) and weathering (Waters et al. 2001; Huey et al. 2008). At any given time, a river is composed of a variety of structural elements which can greatly differ between downstream and upstream river reaches, providing a series of landscape changes and barriers (Castric et al. 2001; Fausch et al. 2002; Wilkinson et al. 2006).

These structural characteristics may aid or restrict gene flow, depending on the dispersal requirements and capabilities of a given species (Emerson & Hewitt 2005; Crispo et al.

2006). Features such as waterfalls, ephemeral waterholes, floodplains, rapids, riffles and deep channels can have complex ramifications for gene flow in aquatic biota (Crispo et al. 2006; Huey et al. 2006; Markert et al. 2010; Sheldon et al. 2010). For instance, a large, pelagic fish can easily travel through deep channels, but may find a riffle to be an obstacle to dispersal (Fausch et al. 2002). Alternatively, a small fish inhabiting rocky creek beds may move without difficulty over riffles, but may be unsuited to swimming through deep channels (Schaeffer 2001; Lowe et al. 2006). Such features can ultimately act as ‘filters’ in determining the connectivity of populations, depending on the species.

When riverine features or characteristics isolate populations, the reduced gene flow can lead to changes (via selection, mutation and genetic drift) in the existing gene pool

(Allendorf 1983; Sheldon et al. 2010). Consequently, evolutionary trajectories are less influenced by outside gene flow and distinct gene frequencies evolve (Lenormand

Chapter 1.0 ∙ Introduction 23

2002). Providing these populations are not constantly being subjected to bottleneck events (which would reduce genetic diversity), this reduction in gene flow can support the development of genetic diversity whereby natural selection and genetic drift lead to genetic differentiation (Luikart et al. 1998; Huey et al. 2006). Alternatively, the hydrological regime of some riverine habitats can support population connectivity.

Flooding events that connect rivers and floodplains laterally and longitudinally can assist dispersal and homogenise gene frequencies (Bayley 1995; Hughes & Hillyer

2003; Hänfling et al. 2004). In these instances, individuals have greater opportunities to disperse freely and promote high levels of gene flow between populations – thus eradicating any genetic differentiation between populations as barriers to dispersal intermittently diminish (Amoros & Bornette 2002; Chaput-Bardy et al. 2009).

In rivers and floodplains that experience highly seasonal or episodic flow patterns, many aquatic species (particularly fish) are often restricted to refugial pools, often for years, until connectivity is once again established (Cook et al. 2002; Magoulick &

Kobza 2003; Arthington et al. 2005; Huey et al. 2008). In some instances, local refugia dry completely and population sizes can be greatly reduced, creating population bottlenecks and a significant loss of genetic diversity (Chapman & Kramer 1991;

Vrijenhoek 1998; Douglas et al. 2003; Matthews & Marsh-Matthews 2003). These factors highlight the importance of hydrological connectivity for the persistence of populations and their genetic diversity.

Over time, reduced connectivity via landscape changes can lead to deepening impacts on the genetic structure of a population. For example, in a study of galaxids (Galaxias divergens), a historical drainage reversal resulted in divergence and cladogenesis of populations after connectivity was altered (Waters et al. 2006). It is these kinds of

Chapter 1.0 ∙ Introduction 24 environmental changes that can influence the physical connectivity of a river system and determine patterns of gene flow within a particular species. It is not only natural changes in hydrological connectivity that can influence patterns of dispersal and gene flow. Artificial changes to flow regimes may also have severe ramifications for the distribution of individuals and their alleles (Ormerod 2003; Beneteau et al. 2009).

Critically, the impacts of manmade barriers to dispersal tend to vary between species.

For aquatic taxa that rely heavily on hydrological connectivity for long-distance dispersal and migration events, barriers to dispersal such as weirs and dams will greatly alter the flow of genes within and between populations (Yamamoto et al. 2004;

Trakhtenbrot et al. 2005; Guy et al. 2008). For sessile or sedentary species, the introduction of a barrier may have a negligible effect on gene flow (Avise 1995).

Damming river systems can confine populations to one section within the catchment, thus the natural patterns of gene flow can be severely disrupted and potentially result in the loss of genetic diversity (Kingsford 2000; Jager et al. 2001; Couvet 2002). Such effects are of particular concern for smaller, fragmented populations, where the influences of natural selection and genetic drift can quickly eradicate genetic diversity

(Lande 1988; Neraas & Spruell 2001; Neville et al. 2009). In many North American species, including trout, bull trout, white sturgeon and white-spotted charr, population declines and reduced genetic diversity have been observed following the establishment of a dam (see Deiner et al. 2007; Neraas & Spruell 2001; Brown et al. 1992; Jager et al.

2001; Yamamoto et al. 2004).

Additionally, the artificial augmentation of hydrological flow both within and between rivers has a role to play in determining the genetic structure of a population. Sudden, abnormal releases of large volumes of water can wash individuals downstream,

Chapter 1.0 ∙ Introduction 25 sometimes greatly exceeding the natural extent of dispersal distances (Wishart & Davies

2003). Also, the artificial transfer of water between previously unconnected basins can disturb population delineations (Davies et al. 1992; Hughes et al. 1996; Snaddon et al.

1998). Potential problems arise if the displaced individuals recruit and reproduce into a new population in a new location, thus introducing the risks associated with outbreeding depression and alien introgression (Davies et al. 1992; DeMarais et al. 1992; Snaddon et al. 1998; Moran 2002; Le Roux & Wieczorek 2009).

Hybridisation between two populations can result in fertile offspring, which can then backcross into one or both parent populations – thus spreading foreign alleles in a process called introgression (Anderson & Stebbins 1954; Dowling & Secor 1997;

Currat et al. 2008). Alien introgression can eradicate the indigenous genes of the resident population; thereby eroding the genetic diversity of the species as a whole, introducing the risk of outbreeding depression, and also erasing the evolutionary history of the population (Smith 1992; Rhymer & Simberloff 1996; Neraas & Spruell 2001;

Yamamoto et al. 2004; Laikre et al. 2010). Fish are particularly susceptible to this problem due to 1) the hydrological extremes that man-made structures can impose on natural waterways (as discussed previously); 2) the common practice of sporting-fish introductions to rivers and lakes; 3) fish farms – from which individuals are regularly released (both accidently and intentionally) into the natural environment; 4) the use of fish as biological controls; 5) the intentional release of unwanted aquarium fish into waterways; and 6) the release of baitfish (Rhymer & Simberloff 1996; Laikre et al.

2010). In addition, (and to further exacerbate the problem) fish have high rates of hybridisation relative to other groups, which may indicate that they have weaker pre- and postzygotic isolating mechanisms (Schwenk et al. 2008).

Chapter 1.0 ∙ Introduction 26

It must also be emphasised that introgression is not a phenomenon associated only with artificial introductions – it can also be observed in natural populations (Dowling &

Secor 1997; Nevado et al. 2009). However, there are numerous records of population

(even through to species) extinctions in fish as a result of introgression from non- natives. Pupfish (Cyprinodon bovinus), westslope cutthroat trout (Onchorynchus clarki lewisi), Owens tui chub (Siphateles bicolor snyderi) and Spanish brown trout (Salmo trutta) are just a few examples of species that have undergone considerable introgression (and are under threat from extinction) from introduced species (see

Echelle & Echelle 1997; Rubidge et al. 2001; Chen et al. 2007; Almodóvar et al. 2001).

Such studies highlight the detrimental impacts of habitat alteration on genetic diversity and the need to investigate the specific potential impacts of artificial changes to environmental flows.

Overall, it is important to understand the conservation implications of population structure and genetic diversity, and to recognise the impacts that changes to gene flow patterns can have on maintaining population viability (Frankel 1974; Moritz 1999; Huey et al. 2006). The loss of genetic variability is a slow process, generally occurring over hundreds or thousands of generations (Amos & Balmford 2001). The rate of this loss is largely dependent on effective population size (Ne), whereby the smaller the population, the smaller the capacity to contain all genetic variants (Nunney & Campbell 1993).

These variants (whether beneficial, detrimental or neutral) are all equally likely to be eliminated from the gene pool, due to the impartial nature of genetic drift (Kimura

1968; Willi et al. 2007). Populations are often composed of numerous discrete gene pools, and it is this genetic variability which can aid species when adapting to changes in their environments (Lacy 1988). When a population undergoes a severe reduction in numbers for several generations, much of the genetic variability can be lost by chance

Chapter 1.0 ∙ Introduction 27 through genetic drift (Amos & Balmford 2001; Dlugosch & Parker 2008). The loss of genetic diversity then leads to reduced evolutionary potential or evolvability (Frankel

1974; Fernández & Caballero 2001).

Further to this are issues surrounding heterozygosity within a population. As many genetic disorders are expressed as recessive homozygous traits, it is imperative for heterozygous combinations to remain in a population as these potentially increase individual fitness (Hedrick & Miller 1992; Grueber et al. 2008). The fixation of alleles is also a major concern in instances where genetic drift or selective sweeps ensure certain alleles are highly represented or the only potential combination (Amos &

Harwood 1998; Brodie 2007). It is this fixation of either beneficial or detrimental alleles that reduces genetic variability and the ability to respond to environmental changes

(Montgomery et al. 2000; Reed & Frankham 2003). Subsequently, the loss of genetic variation can increase the extinction risk for many species.

It is a common misconception that population size is always indicative of population health and persistence (Lacy 1987; Frankham 1996; Brodie 2007). Less tangible measures, such as genetic diversity, are often simply overlooked or else considered inferior to more traditional approaches of assessing populations (e.g. observational studies) (Frankel 1974; Soule 1985; Kinnison et al. 2007). This is of particular concern where a population may consist of a great number of individuals, but the effective population size (Ne) is small, or the genetic diversity contained within it is very low.

Consequently, the illusion is that the population is a large one, and that the threat of decline is unlikely (Frankel 1974; Luikart et al. 2010). This facade poses a danger to species and populations when the population is considered to be healthy, and the loss of part of the population is considered acceptable and will have a negligible effect (Haig

Chapter 1.0 ∙ Introduction 28

1998). These risks are particularly problematic in conservation settings; firstly, there can be substantial difficulty in convincing managers and policy makers of the importance of maintaining genetic diversity (Avise 1989; Amos & Balmford 2001;

Brodie 2007) and secondly, the comparative expense of molecular studies means that not all species and populations will be granted the luxury of having their DNA scrutinised.

To effectively and appropriately manage species in a conservation strategy, clear demarcations must be made between population boundaries (Avise 1989; Palsbøll et al.

2007). By identifying and targeting populations or groups (rather than the species as a whole) that are significant to the continuation of the species, more realistic and achievable goals can be set for long-term conservation. Prioritisation is especially important when resources (such as funding, time or appropriate habitat) limit the scale of conservation (Dizon et al. 1992; Moritz 1994). Examinations of dispersal and gene flow can reveal the processes that determine the genetic viability of populations (Nathan

2005). However, they can also allow generalised patterns to be identified which further the understanding of migration, dispersal distances, recruitment and colonisation.

1.2 Models of gene flow in aquatic environments

Like many ecological functions, dispersal and gene flow can often follow generic patterns across taxa or environments. Several models have been used to describe and predict the arrangement of genetic structure that may be observed in real populations.

Isolation by distance (IBD) is a generalised model that is readily applicable to many populations in many environments (Slatkin 1993; Rousset 1997; Björklund et al. 2010).

First developed by Sewell Wright in 1943, it describes the effect that geographic

Chapter 1.0 ∙ Introduction 29 distance has on determining the isolation of populations. Adjacent populations are expected to be more genetically similar than those further away – effectively the geographic distance corresponds with genetic distance (Figure 1.1) (Wright 1943;

Slatkin 1985; Segelbacher et al. 2010).

Figure 1.1 Model of IBD along a river (blue arrow indicates direction of flow). Each circle represents a population which exchanges genes with the adjacent populations.

Colours represent the gradient of genetic differentiation that has led to the isolation of the red population (upstream) from the green population (downstream).

IBD can only be established under certain conditions. Firstly, the dispersal distance of the populations must be less than the total range (Slatkin 1993; Anderson et al. 2010).

So for populations with a strong, unrestricted dispersal capability, IBD would not be expected, as the dispersal distances mean that gene flow occurs between even the most distant populations. Secondly, IBD is only established when the diverging effects of genetic drift and the homogenising effects of gene flow are in equilibrium (Wright

1943; Slatkin & Maddison 1990; Robledo-Arnuncio & Rousset 2010). High gene flow will counteract divergence and mask any IBD effect (Slatkin 1993). Alternatively, genetic drift in highly isolated populations can inflate the population differentiation that would be expected under the IBD model (Rousset et al. 1997). Thirdly, populations should only be separated by distance alone, that is, the distribution should be continuous

(or if the population is physically subdivided, gene flow should be able to occur

Chapter 1.0 ∙ Introduction 30 between populations) (Hardy & Vekemans 1999; Guillot et al. 2009). This condition is particularly crucial when considering apparent IBD in a population – superficially, IBD can be an explanation for population differentiation (Cushman & Landguth 2010).

Consideration must also be given to the landscape, and how other features may also be interacting to isolate populations (Murphy et al. 2008). For example, populations of a freshwater fish are found to fit the IBD model. However, a closer examination of the landscape shows that a series of small waterfalls may actually be the barrier that has resulted in population differentiation, not distance itself. Ultimately, it is important to consider the reasons why IBD may or may not be observed (Storfer et al. 2007; Murphy et al. 2008).

Two models that are perhaps more specific to aquatic populations are the Stream

Hierarchy Model (SHM) and the Death Valley Model (DVM), developed by Meffe &

Vrijenhoek (1988). The SHM predicts that gene flow is more likely to occur between populations within a tributary than between neighbouring tributaries, resulting in the partitioning of genetic variation among basins (Figure 1.2) (Meffe & Vrijenhoek 1988).

Under this model, the genetic structure of populations within a catchment will be hierarchical, and ultimately, the more dendritic the catchment, the greater the genetic structure. The DVM describes the genetic consequences for highly isolated populations that experience no gene flow (Meffe & Vrijenhoek 1988). Under such conditions, the isolated populations will become increasingly divergent over time as the processes of mutation, genetic drift and selection all act to alter the original gene frequencies (Figure

1.3).

Chapter 1.0 ∙ Introduction 31

Finn et al. (2007) developed the Headwater Model (HM) to address the patterns of gene flow in headwater specialist species. The HM predicts that dispersal in headwater adapted species is more likely to occur overland between headwater sites, rather than

Chapter 1.0 ∙ Introduction 32 via the stream channel (Figure 1.4). However, this model will only fit species that have some terrestrial or aerial dispersal capability, as found in giant water bugs (Abedus herberti) and caddisflies (Chaetopterygopsis maclachlani) (Finn et al. 2007; Lehrian et al. 2010).

Figure 1.4 The Headwater Model (HM), showing that dispersal is more likely to occur across land barriers between headwater sites than via the stream channel. Each population is represented by a circle, with colours indicating the haplotype of a given population.

Metapopulation models are also commonly used to reflect the way in which sub-divided populations in a fragmented landscape behave as a “population of populations” (Levins

1969). There are two variants of this model. First is the classic (also known as the island) metapopulation model which was originally developed by Levins (1969) to best predict and manage pest invasions in crops (Hanski & Gilpin 1991). The metapopulation model describes the way in which physically isolated subpopulations undergo cycles of local extinction (which is more likely in smaller subpopulations)

Chapter 1.0 ∙ Introduction 33 followed by recolonisation from other subpopulations (Figure 1.5a) (Hanski 1999;

Fullerton et al. 2010). Second is the mainland – island metapopulation (source – sink) model (Harrison 1991; Kawecki & Ebert 2004). The mainland – island metapopulation model is more specific to situations where there may be a region that is consistently inhabited by a self-sustaining population (the source), from which individuals continually recolonise less favourable habitat (Figure 1.5b) (Hanski 1998; Woodford &

McIntosh 2010).

a) b)

Figure 1.5 Metapopulation models showing a) the classic model where cycles of extinction and recolonisation occur between neighbouring populations; and b) the mainland – island model where the source population (mainland) ‘supplies’ the islands. In this example, circles represent isolated aquatic habitats within the landscape. Arrows show the direction of dispersal.

1.3 Applying genetic markers in studies of dispersal

Molecular markers are parts of a genome which are investigated separately as loci

(Parker et al. 1998; Le Roux & Wieczorek 2009). In studies of molecular ecology, it is important to find variation in these loci where different alleles (or variants of a similar

Chapter 1.0 ∙ Introduction 34 sequence) can be found (Sunnucks 2000; Climer et al. 2009). Diploid organisms have two alleles at each locus – with an allele inherited from each parent. If both of the alleles in a single individual are identical, then that individual is homozygous.

Alternatively, if the alleles are different then the individual is heterozygous at that locus

(Clark 1990). A single locus can have varying levels of this polymorphism, with many different combinations of alternative alleles (Beebee & Rowe 2008). These high levels of polymorphism are widely used in population studies where there is a need to make a distinction between individuals, for example in studies of parentage (Ellegren 1992;

Primmer et al. 1995; Ellegren & Sheldon 2008). However, in population studies a more conservative level of polymorphism is suitable, as it is the proportion of variation shared among populations that is measured (Hey & Kliman 1993; Anne 2006).

Molecular markers and genetic variation are very useful in answering questions about dispersal patterns, especially when other methods may be difficult to apply (Nathan

2001; Schwartz et al. 2007). For instance, mark and recapture studies often have low rates of success as recapturing marked animals can be problematic (Nathan et al. 2003).

Low recapture rates result in low sample sizes, which subsequently reduce the statistical power of analyses (Koenig et al. 1996; Luikart et al. 2010). The use of a genetic marker means that the need for recapture is eliminated as an individual is sampled once for its

DNA (an unchanging measure over the period of the animal’s lifetime), resulting in data which can then be applied to a range of hypotheses (Koenig et al. 1996; Nathan 2001;

Berry et al. 2004). Careful consideration must be given to the study in question – particularly the temporal and spatial scales that will be involved to determine the appropriate application of a genetic marker, and specifically which marker is the most suited to the research questions (Manel et al. 2005). To this end, the following section

Chapter 1.0 ∙ Introduction 35 will outline two genetic markers that were relevant to this study, and discuss their strengths and weaknesses in population genetics research.

1.3.1 Mitochondrial DNA

Mitochondrial DNA (mtDNA) is widely used in population studies to reveal detailed levels of variation within and between populations (Parker et al. 1998; Pemberton

2008). As a uniparentally inherited genome, mtDNA consists of a single haploid chromosome passed on from the maternal line (Avise 1986; Avise 2009). Consequently, the effective population size (Ne) is approximately a quarter that of nuclear DNA

(Slatkin & Hudson 1991). The loss of allelic diversity is inversely related to Ne and so the effects of genetic drift are more pronounced in mtDNA (Lacy 1987; Zhang &

Hewitt 2003). Accordingly, mtDNA is more sensitive to limited gene flow than nuclear

DNA and changes are captured on a relatively short time scale – making it appropriate for inferring patterns of gene flow, structure and diversity within a species (Nathan et al.

2003; Zhang & Hewitt 2003; Hickerson et al. 2010).

mtDNA does not recombine, and so it preserves intraspecific historical mutations

(Mossman & Waser 1999; Balloux 2010). Hence, the order in which these mutations occurred can be reconstructed and the associations represented in networks of genealogical relationships (Excoffier et al. 1992; Templeton 2005). By overlaying these historical relationships on a temporal or spatial geographic plane, links can be made between genetic mutations and geographic distributions – a field known as phylogeography (Avise et al. 1987).

Chapter 1.0 ∙ Introduction 36

However, mitochondrial markers have some limitations in their application. Being a maternally inherited molecule, mtDNA is not effective in reflecting dispersal patterns of both males and females (Avise et al. 1987; Avise 2009). Also, mtDNA is a single locus, thus allowing the potential for biased results under the influence of selection and stochastic processes (Zhang & Hewitt 2003; Hurst & Jiggins 2005). mtDNA reflects the history of only that genome, resulting in data which may not provide an accurate representation of the history and connectivity of populations (Zhang & Hewitt 1996;

Galtier et al. 2009). The use of additional independent markers can reduce this potential bias (Strobeck et al. 1976; Balloux 2010).

1.3.2 Microsatellites

Microsatellites are tandem repeats of one to five nucleotides within the genome of an individual (Jarne & Lagoda 1996; Geffen et al. 2007). The majority of these loci are in non-coding regions of the genome, and are therefore commonly considered to be neutral

(Ellegren 2004). Microsatellites are co-dominant and often reveal high levels of polymorphism because of their comparatively high rates of mutation (Slatkin 1995;

Goldstein & Pollock 1997; Väli et al. 2008). Consequently, they are a valuable resource when revealing differences between individuals – a requirement in fine-scale population studies where recent restrictions to gene flow are expected (Bruford & Wayne 1993;

Selkoe & Toonen 2006). Furthermore, the geographic distributions of alleles can be used to infer patterns of small to large scale dispersal and gene flow between populations (Queller et al. 1993; Anderson et al. 2010). Unlike mtDNA, microsatellites are bi-parentally inherited, and so can be particularly useful when trying to investigate patterns of dispersal in both males and females (Jarne & Lagoda 1996; Balloux &

Lugon-Moulin 2002).

Chapter 1.0 ∙ Introduction 37

While microsatellites are a useful tool for population genetics, they too have some limitations. Homoplasy (where the alleles are the same size, but are not derived from the same lineage) can be problematic as it will obscure the true allelic diversity of the population (Estoup et al. 1995; Garza & Freimer 1996; Estoup et al. 2002). Null alleles can also be problematic if mutations in the flanking regions prevent the primers from annealing correctly during the amplification process (Chapuis & Estoup 2007). These annealing problems then create false homozygosity, as only one allele may be amplified at a given locus (Van Oosterhout et al. 2004; Chapuis & Estoup 2007). If the null allele frequency is high within a population, this can erroneously result in perceived heterozygote deficiencies. Null alleles impact on the perceived population structure, as excess homozygosity tends to overestimate the isolation of a population (Dewoody et al. 2006). Additionally, microsatellites in functional regions of the genome can be linked and are potentially under selection (Selkoe & Toonen 2006). However, as with all molecular markers, the use of additional loci should highlight any abnormalities in the data set, and reduce bias attributed to a single marker type.

1.4 Genetic legacies in tropical rivers

Studies of dispersal using genetic tools can be highly informative when trying to characterise general patterns peculiar to certain ecosystems, species or regions.

Typically, large tropical rivers have been found to support high genetic diversity, but low genetic structure in fish populations. For instance, in the Amazon River, tambaqui

(Colossoma macropomum) and pirarucu (Arapaima gigas) populations were found to be genetically diverse, but with very low population structure (Hrbek et al. 2005; Santos et al. 2007). In the Mekong River, catfish (Pangasianodon hypophthalmus and Pangasius

Chapter 1.0 ∙ Introduction 38 bocourti) were also found to have very high levels of genetic diversity with low (but significant) genetic structure (So et al. 2006(a); So et al. 2006(b)). Prochilodus sp. in

South American rivers also followed this trend of high levels of widely distributed alleles (Revaldaves et al. 1997; Sivasundar et al. 2001). These patterns are generally attributable to the naturally high, constant hydrological connectivity of these tropical river systems, and especially monsoonal flooding (Hrbek et al. 2005; Raeymaekers et al. 2008).

However, tropical fish species have been found to have significant genetic structure as a result of geomorphological changes creating barriers to gene flow and dispersal.

Hurwood et al. (2007) found past restrictions in gene flow in physically connected cyprinid (Henicorhynchus lobatus) populations living in the Mekong and Mun (a tributary) Rivers. While this cyprinid is known to be a highly vagile species, the influences of drainage rearrangement had strongly determined the historical connectivity and genetic structure of the population. A similar pattern was found in cardinal tetras (Paracheirodon axelrodi) where river vicariance has led to highly structured populations in the Rio Negro basin in northern South America (Cooke et al.

2009). Alternatively, genetic studies have also shown currently isolated populations to have had historical connectivity prior to landscape changes. African catfish (Clarias gariepinus) across eastern Africa are one such example where rifting and volcanic activity reduced gene flow between once connected populations (Giddelo et al. 2002).

In southeast Asia, changes in sea levels have largely determined the connectivity and gene flow between populations of the river catfish (Hemibagrus nemurus) (Dodson et al. 1995). In this example, sea levels in the Pleistocene were low enough that land bridges and rivers connected what are now isolated island populations in Java, Sumatra and Borneo, resulting in gene flow from and to the mainland. It is clear that the

Chapter 1.0 ∙ Introduction 39 geomorphology of tropical rivers is influential in isolating and integrating populations, according to the changes that occur over time.

It is unclear if these two broad categories (high diversity with low structure, and genetic structure in accordance with historical geomorphic change) are applicable to tropical aquatic populations worldwide. In Australia, it may be that tropical rivers are atypical environments with atypical genetic influences; alternatively, the genetic connectivity in tropical rivers may follow the patterns found in other continents. The taxa of Australian tropical rivers are poorly understood in terms of the relationship between their genetic lineages and their geographic distributions. However of the studies that have been done, many suggest that aquatic species residing in tropical rivers develop strong genetic structuring, and that many parameters can affect the genetic differentiation of populations. For instance, purple-spotted gudgeons (Mogurnda adspersa) show strong genetic differentiation in the headwaters of the Tully River (QLD), where the Tully

Falls, a significant barrier to upstream dispersal, separates highland and lowland populations (Hurwood & Hughes 1998). However lowland populations showed high levels of contemporary connectivity, possibly facilitated by seasonal flooding (Hurwood

& Hughes 1998). Fly-specked hardyhead (Craterocephalus stercusmuscarum) also show high levels of genetic differentiation where populations are separated by waterfalls (McGlashan & Hughes 2000). Populations on either side of these barriers are genetically subdivided, which is indicative of the way in which historical geomorphological changes have restricted dispersal and gene flow (McGlashan &

Hughes 2000). In addition to geomorphologically driven separations of populations,

Huey et al. (2008) found that for Hyrtl’s tandan (Neosilurus hyrtlii), flow regimes had significant impacts on the levels of genetic diversity. In that study, populations in the rivers of the Gulf of Carpentaria had high levels of genetic diversity due to predictable

Chapter 1.0 ∙ Introduction 40 flow and stable population sizes, whereas populations in the south (in the Lake Eyre

Basin) had lower genetic diversity which was thought to result from population bottlenecks caused by hydrological extremes. These studies highlight the complex relationships between and within river systems and their resident populations, and emphasises the need to understand the processes that facilitate genetic structure in

Australian rivers.

Further to this is the necessity of investigating the little understood interactions between populations within river systems at a finer level. Few population studies have focused on identifying fine-scale gene flow at more local levels in Australian rivers; overall, fish dispersal within catchments is poorly understood. Hurwood & Hughes (1998) found that the purple-spotted gudgeon (Mogurnda adspersa) was subjected to significant restrictions to gene flow within the rivers in the Atherton Tableland (north Queensland) due to drainage rearrangement and waterfalls (as mentioned previously). The Pacific blue eye (Psuedomugil signifer) was also identified as having significant genetic structure within catchments in north Queensland, resulting from the combined effects of

IBD and metapopulation dynamics (McGlashan et al. 2001). Alternatively, Hughes &

Hillyer (2006) showed that dispersal of Australian smelt (Retropinna semoni) and bony bream (Nematolosa erebi) in inland Australia was unrestricted. These types of studies are crucial in determining the environmental and landscape features that may restrict dispersal in natural populations, and assist in making further inferences and predictions on how artificial changes to the landscape might affect dispersal patterns.

Chapter 1.0 ∙ Introduction 41

1.5 Using stable isotopes to measure dispersal

The use of stable isotope markers as indicators of dispersal is a relatively recent application of a technology that can be used to identify resident and immigrant individuals within a population (Hadwen & Arthington 2007), and also population boundaries (West et al. 2006). In ecological studies the most commonly used elements are carbon (C) and nitrogen (N) and both exist naturally in more than one isotopic form, each with differing nuclear masses. Stable isotope approaches use the proportions of these isotopes in organic matter to infer patterns in feeding and movement (Peterson &

Fry 1987; Rubenstein & Hobson 2004; Wunder & Norris 2008). The ratios of these isotopes can be analysed using a mass spectrometer, as the different nuclear weights can be distinguished (Peterson & Fry 1987; Hobson et al. 2010).The different isotopes of C and N are typically referred to as being ‘light’ or ‘heavy’, a reference to their nuclear masses. The ratios between light and heavy isotopes may vary between regions, for example the ratio of 13C (heavy) to 12C (light) may vary between individuals, sites and/or regions (Hobson 2005). Natural materials (e.g. muscle tissue) are composed of C and N derived from the environment through dietary uptake (Phillips & Gregg 2003). In certain situations, land use can also be influential in determining the stable isotope signature of N (also referred to as δ15N) via the addition of 15N-enriched chemicals, fertilisers and human and agricultural wastes (Vitousek et al. 1997). Consequently, a stable isotope analysis of δ13C and δ15N can identify the ecological roles of individuals within a food web as δ13C signatures reflect those of diet, and δ15N reflects trophic position as it fractionates (becomes heavier) with each trophic level (DeNiro & Eppstein

1978; West et al. 2006; Barnes et al. 2008).

Chapter 1.0 ∙ Introduction 42

In addition to being a useful tracer of feeding relationships, stable isotope ratios of C and N can also be used to measure dispersal. However, it is critical that the source of an individual’s diet can be traced back to a specific environment, and hence its home territory (Webster et al. 2002; Nathan et al. 2003; Hobson et al. 2010). This requirement may only be met when there is adequate variation and discrimination between regions, as the isotopic signature of some food sources, like terrestrial vegetation, can remain constant over a large geographic scale (Hobson 1999). This is not to say that each region must be unique – isotopic signatures can be effectively used in answering questions about dispersal provided they are adequately differentiated for the scale of interest (Inger & Bearhop 2008). Where differences in stable isotope signatures do occur between sites of interest, this technique can make strong delineations when distinguishing between resident individuals and new migrants

(Cunjak et al. 2005; Hadwen & Arthington 2007). Therefore, a migrant to a population should be isotopically distinguishable from residents, provided that 1) the environment from which it came is isotopically different to the recruitment site, and 2) the migrant has only recently arrived at the recruitment site, as the dietary uptake in the new environment will change isotopic signatures over the course of a few months (Peterson

& Fry 1987; Hobson 1999). If the ratios of stable isotopes vary from site to site, these migrants can potentially be traced back to their home region (Hobson 1999; Møller et al. 2006). In the same way, residents can easily be identified due to the similarities between their isotopic signatures and their habitat isotopic signatures (Rubenstein &

Hobson 2004). Using stable isotope techniques in this way, it is therefore possible to identify population boundaries and quantify dispersal distances (Nathan et al. 2003;

Hobson 2005; Hobson et al. 2010).

Chapter 1.0 ∙ Introduction 43

1.6 Combining methods to investigate dispersal

When attempting to characterise and evaluate patterns of dispersal it is desirable to utilise more than one technique, so as to clarify and strengthen the conclusions that can be made (Bossart & Pashley Prowell 1998; Nathan 2001; Nathan et al. 2003; Cunjak et al. 2005). By integrating genetic methods, which investigate historical dispersal, and stable isotopes, which investigate contemporary dispersal that has occurred within months of the dispersal event, dispersal can be investigated at different time scales

(Chetkiewicz et al. 2006; Lowe et al. 2006). Combining genetic analyses with stable isotope analyses in dispersal studies has been little explored, but several studies have successfully quantified dispersal using this dual approach. Boulet et al. (2006) used genetic markers to identify three geographic breeding regions in the northern yellow warbler (Dendroica petechia; aestiva group), and to differentiate the specific migratory patterns of each of these groups. The use of a stable isotope marker (deuterium) revealed some connectivity between these groups during migration. Another migratory bird, Wilson’s warbler (Wilsonia pusilla), showed high levels of gene flow using genetic markers, while stable isotopes (hydrogen) showed the exact migratory (and hence gene flow) pathways which supported the genetic findings (Clegg et al. 2003).

Similarly, Cook et al. (2007(b)) discovered populations of shrimp (Paratya australiensis) to be genetically homogeneous, indicating high levels of dispersal and gene flow, with the stable isotopes (carbon and nitrogen) demonstrating that this gene flow was a result of many small-scale dispersal events between refugial pools.

Conversely, in southern pygmy perch (Nannoperca australis), genetic markers revealed separate streams to support independent, isolated populations with varying levels of connectivity within streams highlighted by both the genetic markers and the stable isotopes (carbon and nitrogen) (Cook et al. 2007 (a)). The stable isotope data revealed

Chapter 1.0 ∙ Introduction 44 that dispersal in the southern pygmy perch was highly influenced by seasonal changes in hydrological connectivity. Whilst studies of either stable isotope signatures or population genetics can be highly effective to examine dispersal when used independently, their application together can shed important light on differences or similarities between contemporary and historical processes.

1.7 Australia’s tropical rivers

The tropical rivers of northern Australia are characterised by their highly seasonal rainfall from November through to May, interspersed by long periods of low precipitation (Hamilton & Gehrke 2005). Up to 90% of annual rainfall in these regions occurs between the months of October to April (Blanch et al. 2005). These extremes in flow result in periods of flooding and high hydrological connectivity, but also periods of drought where entire rivers can be reduced to a few refugial pools (Perna & Pearson

2008). Even so, 70% of Australia’s runoff is accounted for in tropical catchments

(Hamilton & Gehrke 2005).

The Mitchell and Daly Rivers are major tropical drainage systems of Australia, and are sizeable systems, covering 70 000km2 and 53 000km2 respectively (Erskine et al. 2003;

Russell et al. 2003; Blanch et al. 2005) (Figure 2.1). Additionally, they experience the highest levels of discharge for rivers in Queensland and the Northern Territory, placing them among the most significant Australian rivers in terms of flow. Both these catchments are under threat due to the impacts of water extraction, introduced species and agricultural practices (Erskine et al. 2003; Hart 2004; Blanch et al. 2005; Petheram et al. 2008).

Chapter 1.0 ∙ Introduction 45

The Daly and Mitchell River catchments mostly consist of flat plateaus which are dominated by savannah and seasonally dry forests (Hamilton & Gehrke 2005; Petheram et al. 2008). However, these river systems have quite different geomorphologies. The

Mitchell River headwaters originate from tablelands reaching 1000m above sea level in the east of the catchment (Forsyth & Nott 2003). Further west, the landscape abruptly changes to flat savannah grasslands that gradually fall in elevation. The catchment itself is long, stretching from the Gulf of Carpentaria almost to Cairns on the east coast.

However, the catchment is fairly consistent in width, with relatively few hierarchical levels of branching within tributaries.

The Daly River is characterised by floodplains, sandstone escarpments, braided channels, gorges and waterfalls (Blanch et al. 2005; Wasson et al. 2010). The middle reaches of the catchments are of low relief, falling to 20m below sea level (Blanch et al.

2005). Areas of higher relief in the rolling landscape rise approximately 40m above sea level (Nott 1994). The vast flows during the wet season play a strong restructuring role, with the shape of the Daly River channels changing from year to year (Blanch et al.

2005). Unlike most other northern rivers, flow in the Daly is perennial (Erskine et al.

2003). During the dry season, limestone aquifers feed the river thus producing waters high in magnesium and calcium, but low in phosphorus and nitrogen (Blanch et al.

2005; Erskine et al. 2003). Consequently, this river is particularly sensitive to the introduction of phosphorus and nitrogen in the form of agricultural fertilisers (Hart

2004; Blanch et al. 2005). The Daly River catchment differs from the Mitchell in that progressive levels of branching initiate from the river mouth, giving it a triangular drainage formation.

Chapter 1.0 ∙ Introduction 46

1.8 Australia’s tropical freshwater fish

1.8.1 The chequered and western rainbowfish (Melanotaeniidae: Castelnau 1875)

Melanotaenids are freshwater fish native to Papua New Guinea and Australia (Pusey et al. 2004). The Melanotaenia genus exhibits a broad range of bright colour variation that has generated the collective name of the genus – the rainbowfish. Melanotaenia splendida (Peters 1866) (eastern rainbowfish) is a small (rarely exceeding 120mm) omnivorous species of the family Melanotaeniidae which is native to Papua New

Guinea and Australia (Pusey et al. 2004). In Australia, a subspecies, Melanotaenia splendida inornata (Castelnau 1875) (chequered rainbowfish) (Figure 2.6), is also widely distributed throughout the Gulf of Carpentaria drainages of Queensland and the

Northern Territory (Pusey et al. 2004). A second species, Melanotaenia australis

(Castelnau 1875) (western rainbowfish) (Figure 2.6) is morphologically and behaviourally very similar to M. s. inornata (Pusey et al. 2004). For both species, the majority of reproduction occurs during the wet season, although spawning can occur year-round. Fish can mature when still at a small size, and breeding occurs in their first year (Pusey et al. 2004). Individuals generally live for a period of two years in the wild.

The distribution of M. australis differs from that of M. s. inornata, as they are found in the Pilbara (WA) through to the Adelaide River (NT) (Allen et al. 2002; Phillips et al.

2009). Given the similarities between these two species and the differences in distribution, M. australis and M. s. inornata were considered to be comparable species in which to assess genetic structure in the Daly and Mitchell Rivers, respectively.

Chapter 1.0 ∙ Introduction 47

1.8.2 The sooty grunter (Terapontidae: Macleay 1883)

Hephaestus fuliginosus (Macleay 1883) (sooty grunter) (Figure 2.8) of the family

Terapontidae are relatively large, heavy bodied and long-lived fish growing to approximately 40cm (Allen et al. 2002). This species ranges in colour from pale gold through to charcoal grey, although they darken when under stress. Juveniles are easily distinguishable by a dark spot on the anal fin. Diet is varied, and encompasses small insects and zooplankton through to frogs, birds and native fruits (Davis et al. 2010). H. fuliginosus is considered to be abundant in the more coastal drainages from the Victoria

River (NT) through to the eastern side of the Cape York Peninsula, and is also found in

Papua New Guinea (Pusey et al. 2004). Breeding coincides with the onset of the wet season, although there are variations in timing across their range in northern Australia

(Pusey et al. 2004).

1.9 Aims, objectives and thesis outline

This study aims to investigate dispersal and patterns of gene flow in Australian fish populations within catchments. As discussed previously (see Section 1.4), there is a need for fine scale studies of dispersal in Australian fish, particularly when predicting and managing the impacts of alterations to hydrological connectivity and the flow-on effects of such change. Specifically, this research focusses on two rainbowfish species

(M. s. inornata and M. australis) and one grunter species (H. fuliginosus) within the

Daly and Mitchell Rivers. These catchments were chosen to align with the aims of the

Tropical Research and Coastal Knowledge (TRaCK) program and its focus on the Daly,

Mitchell and Fitzroy Rivers. For logistical purposes, (given the budget and time frame) the Daly and Mitchell Rivers were selected as the catchments for this research.

Chapter 1.0 ∙ Introduction 48

In using genetic markers to characterise dispersal in populations of the three fish species selected, this study will enhance the limited knowledge of dispersal in tropical freshwater fish within a discrete river system. In addition, by integrating some of the genetic methods commonly used to investigate dispersal with stable isotope analyses, it is hoped this research will reveal patterns of dispersal of M. s. inornata over the wet season from historical through to contemporary times, thus identifying the connectivity within the upper Mitchell River catchment. Specifically, several questions will be addressed:

1. How does the landscape influence the connectivity of fish populations within

catchments?

2. Are there any historical geomorphic or environmental changes that have led to

contemporary population structure within each of the study catchments?

3. What are the small scale (in both time and space) movements of rainbowfish

within the Mitchell River catchment? How are these affected by flooding?

4. What implications do these findings have for the conservation and management

of fish populations in tropical Australia?

This research will investigate population level genetic structure in Melanotaenia spp. and Hephaestus sp. to characterise gene flow in the study areas. Investigation of the genetic differentiation (in mitochondrial and microsatellite markers) found in contemporary populations will provide a better understanding of the influences of geomorphic and environmental changes on the genetic structure of riverine organisms.

Furthermore, a genetic analysis of the structure of populations in these rivers will provide a basis for conservation strategies in these regions.

Chapter 1.0 ∙ Introduction 49

This thesis consists of seven chapters in total. Chapter 2.0 describes the general methods used in the research. Specifically, it provides information on the study species and the study regions. Also, detailed technical information is also presented, detailing the sample collection methods and the general laboratory and analytical procedures used.

Chapter 3.0 is an investigation of the population genetics of Melanotaenia spp. in the

Daly and Mitchell catchments, and aims to investigate how the landscape has contributed to genetic structure in these species. Chapter 4.0 focuses on H. fuliginosus in the Daly and Mitchell catchments, and again looks at landscape genetics and population connectivity. Chapter 5.0 is the final data chapter, and combines genetic markers with stable isotope markers to determine the dispersal patterns of M. s. inornata over the wet season in the headwaters of the Mitchell River. Chapter 6.0 summarises the main findings of Chapters 3.0, 4.0 and 5.0, and presents the overall conclusions and implications for fish in Australia’s tropical rivers. Chapter 7.0 is an appendix, and comprises of relevant additional supplementary material.

Chapter 1.0 ∙ Introduction 50

2.0 Field and Laboratory Methods

2.1 Study design

Three freshwater species were the subjects of this study: Melanotaenia splendida inornata, Melanotaenia australis and Hephaestus fuliginosus. All three species are widely distributed across many of the catchments of northern Australia. However, for the purposes of this study (which was to concentrate on dispersal within catchments), only two catchments were sampled – the Mitchell and Daly River catchments (Figure

2.1). H. fuliginosus was captured in both catchments; however M. australis and M. s. inornata were only sampled in the Daly and Mitchell catchments respectively, according to their ranges.

This project called for extensive sampling in both catchments so that hierarchical analyses (e.g. analysis of molecular variance (AMOVA)) could be performed.

Consequently, genetic variation within and between populations could be compared at different scales. Samples collected for the TRaCK program and the North Australian

Freshwater Fish (NAFF) program were used for this project. Samples were collected over a number of years, and by a number of people (please see Acknowledgements). In the Daly River, individuals were collected from June 2007 through to November 2009.

For the Mitchell River, collections were started in July 2006 through to March 2010.

Chapter 2.0 ∙ Field and Laboratory Methods 51

120°E 130°E 140°E 150°E

10°S Daly River

Mitchell River

20°S NT QLD

30°S

40°S

0 500 1,000 km

Figure 2.1 The Daly River catchment (NT) and the Mitchell River catchment (QLD) in Australia.

2.2 Field sampling

Several additional field trips were made for the specific purpose of completing the collections for this research. One trip was made to the Daly River in June 2009. Three trips were made to the Mitchell catchment: 1) June 2008 which covered much of the whole catchment; 2) October 2009 for a specific collection at four sites in the headwaters to evaluate the population structure and also levels of dispersal (using SIA) between populations prior to seasonal flooding; 3) March 2010 as a return trip to evaluate the impact flooding events had on dispersal due to seasonal connectivity

(Figure 2.2) (refer to Tables 7.1 – 7.5 for all detailed site locations).

Chapter 2.0 ∙ Field and Laboratory Methods 52

In total, fish were collected from 28 sites in the Daly River, and 29 sites in the Mitchell

River. Fish were collected under a Griffith University ethics permit (ENV/09/07/AEC),

Queensland Fisheries Permit (#89212), Northern Territory Fisheries Permit (# S17-2122

& S17/2666), Northern Territory Parks Permit (#29870), a Northern Land Council permit and with approval from the Mitchell River Traditional Custodian Advisory

Group (MRTCAG).

Chapter 2.0 ∙ Field and Laboratory Methods 53

2.2.1 General fish collection methods

Fish were collected using three methods: backpack electrofishing, boat-mounted electrofishing and seine netting (Figure 2.3b, c and d). Each site varied in terms of its depth, flow, submerged debris, and also the potential presence of saltwater crocodiles

(Crocodylus porosus) (Figure 2.3a). An appropriate technique was selected for each site according to these conditions. For instance, in deeper waters where the effectiveness of seining or backpack electrofishing is greatly reduced, boat-mounted electrofishing was employed. In situations where water was shallow and crocodiles were unlikely, then seining or backpack electrofishing was used to capture fish.

a) b)

c) d)

Figure 2.3 a) Saltwater crocodile (Crocodylus porosus) at a sampling site (Mt

Nancar) in the Daly River; b) inspecting the seine net after use; c) front view of the electrodes; and d) electrofishing in Kingfish Lagoon in the Mitchell River catchment.

Chapter 2.0 ∙ Field and Laboratory Methods 54

The dimensions of the seine net were as follows: 15m long, 2m high (with a 2m long bag) with a 9mm mesh size. The net was pulled perpendicularly to the shoreline by two operators (one on the bank, and one in the water) for a short distance. Depending on the site and abundance of fish, this was typically for a distance of approximately 15 – 20m, at which point the net was arced back towards the bank, thereby enclosing fish in the net. Seine nets were particularly successful when targeting rainbowfish. Voltage suitable gloves and waders were worn to protect the operators from electrical shock.

Shocks were deployed at intervals along suitable habitat, working upstream so as to minimise the disturbance to the target species. Incapacitated fish were captured using small hand nets (Figure 2.4).

Figure 2.4 TRaCK researchers Tim Jardine (R) and Richard Hunt (L) electrofishing a small stream in the Mitchell River (photograph courtesy of Dominic Valdez).

Chapter 2.0 ∙ Field and Laboratory Methods 55

Once captured, fish were either 1) returned immediately to the point of capture if the individual was a non-target species, or a target species in excess of the catch limit, or 2) placed in a recovery bucket if a target species within the catch limit (Figure 2.5).

Figure 2.5 M. s. inornata individual recovering after capture.

At each site, up to 30 individuals of each species were collected. Sampling continued until the desired number had been captured, or until it was clear that all (or the majority) of the fish in the sampled region had been caught. Individuals ranged broadly in size and maturity (see Table 7.15), and thus represented larvae (for Melanotaenia spp. only) through to sexually-mature adults. A small fin clip (~3mm length) was taken from either the anal fin or caudal fin. Some individuals had inadequate fin tissue present (for example smaller juveniles, or adults that had lost fins through natural causes). In these instances the individual was immediately euthanised and bagged as a whole specimen.

The tissue was then directly preserved in 90% ethanol, or stored at -20°C in a portable freezer to preserve the DNA and proteins for genetic analyses. Once sufficiently recovered, the fish were released at the point of capture (with the exception of those individuals required whole, as discussed previously. Also see below).

Individuals collected for the combined application of genetic analyses and stable isotope analyses were required whole. Muscle tissue was required as it is the best reflection of

Chapter 2.0 ∙ Field and Laboratory Methods 56 recent changes in an individual’s diet (Hobson & Clark 1992). For these collections, fish were euthanised in an ice slurry, then frozen at -20°C in a portable freezer to slow the cycling of C and N within the sample. All samples collected for this project were stored at -80°C once transported back to the laboratory.

2.2.2 Collection of samples for SIA

To complement the analysis of signatures in the fish tissue, collections were also made of potential C and N sources to the sites. Samples consisted of biofilm (where possible as flooding/scouring events made this difficult), filamentous algae (if present), coarse particulate organic matter (CPOM), fine particulate organic matter (FPOM), seston and riparian vegetation. Biofilm was scrubbed from rocks using a small brush. Filamentous algae was pulled from its anchoring surface, then rinsed in flowing river water to remove any debris that could contaminate the sample. CPOM was defined as organic matter greater than 1mm, whereas FPOM was defined as organic matter particles less than 1mm, but greater than 250μm. These collections were made using a set of two graded sieves. Seston tows were made with a mesh size of 250μm, and deployed in flowing water for a period of approximately 15 minutes. Riparian vegetation was collected from the dominant species at each of the sites. All samples were stored at -

20°C, until required for processing.

Chapter 2.0 ∙ Field and Laboratory Methods 57

2.2.3 Study species

Figure 2.6 Melanotaenia australis from the Daly River main channel (left);

Melanotaenia splendida inornata from the headwaters of the Mitchell River (right).

M. s. inornata and M. australis are abundant schooling species and are often found in greater numbers under the protection of submerged debris or macrophytes. Allen et al.

(2002) suggest that they preferentially inhabit small streams. However Pusey et al.

(2004) state that the distribution of these subspecies covers a wide range of habitats: small streams to large rivers, and also wetlands and floodplain lagoons. Personal in- field observations agree with the latter. Pusey et al. (2004) also point out that reviewed studies of this species found a greater abundance in slower flowing waters, possibly to avoid the physical stresses of fast flows.

The laterally-compressed body is pale with yellow, olive, and/or blue tones; scales generally have an opalescent sheen. Fins of these fish are usually brightly coloured from red through to yellow, or sometimes completely transparent and colourless. In the case of M. s. inornata, fins are patterned with red and yellow ‘chequers’ – thus giving them their common name. Mature individuals often have dark margins on both the anal and dorsal fins. In addition the rear section of the anal fin can become elongated and pointed

Chapter 2.0 ∙ Field and Laboratory Methods 58

(Pusey et al. 2004; Young et al. 2011).The dorsal fin is separated into two. These rainbowfish species exhibit sexual dimorphism, where males tend to be deeper-bodied and more brightly coloured, which is favoured by their female counterparts (Figure 2.7)

(Young et al. 2010). It is notoriously difficult to distinguish between species of the

Melanotaeniidae, and the lack of morphological differences between M. australis and

M. s. inornata makes them no exception (Pusey et al. 2004).

Figure 2.7 Example of an M. s. inornata male with the pronounced deep body.

Figure 2.8 H. fuliginosus from the upper Mitchell River.

H. fuliginosus differs from rainbowfish in that they form loose aggregations, rather than schools, and prefer rocky creek beds in flowing waters with little aquatic vegetation

(Pusey et al. 2004). This species is a desirable catch for recreational fishing, and has been stocked in many impoundments and rivers on the east coast of northern

Chapter 2.0 ∙ Field and Laboratory Methods 59

Queensland (Olden et al. 2008; Chan et al. 2010). Stocking has led to patchy distributions of populations in regions where it does not naturally occur. As a top predator in these systems, such translocations are concerning for both the genetic and ecological impacts on local communities (Pusey et al. 2004).

2.3 Laboratory methods

2.3.1 DNA extraction

Total genomic DNA was extracted from all samples using a variation of the cetyl trimethyl ammonium bromide (CTAB) phenol–chloroform method described in Doyle

& Doyle (1987). A small piece of tissue was added to 700μL of 2xCTAB buffer (50ml

1M Tris-HCl pH 8.0, 175ml 4M NaCl, 20ml 0.5M EDTA, 10g hexadecyltrimethylammonium, 500ml ddH2O) and 5μL of Proteinase K (20mg/mL) in a

1.5mL tube. The tube was then vortexed, and incubated at 50°C for <12 hours on a

Thermoline dry block incubator to activate the Proteinase K. After the sample was digested, 300μL of both phenol and chloroform-isoamyl (24:1) were added, and then the sample was mixed for 15 minutes (using a Clements CEN96181 turn wheel). The sample was then centrifuged (using an Eppendorf 5424 centrifuge) at 13 500rpm for three minutes, after which the supernatant was transferred to a new 1.5mL tube. The addition of phenol and chloroform-isoamyl, and the steps of mixing and centrifuging were then repeated.

The sample was given a final wash with 600μL of chloroform-isoamyl (again mixed for

15 minutes and centrifuged for three minutes), and the resulting supernatant added to

600μL of cold (-20°C) isopropanol and stored at -80°C for one hour to aid DNA

Chapter 2.0 ∙ Field and Laboratory Methods 60 precipitation. Samples were thawed, then centrifuged for 35 minutes at 13 500rpm to form a DNA pellet. The isopropanol was aspirated using a glass pipette and vacuum flask. The pellet was washed in 1000μL of 70% ethanol, centrifuged for 35 minutes at

13 500rpm, aspirated and dried in a vacuum chamber. Samples were suspended in

100μL of ddH2O and stored at 4°C until required for amplification.

2.3.2 Mitochondrial and microsatellite primers

Mitochondrial ATPase 6 and 8 had previously proved to be an appropriate marker for assessing intraspecific genetic structure within Melanotaenids, and was accordingly used for M. australis and M. s. inornata using ATP8.2L and COIII.2H (S. McCafferty, unpublished data) and another set of internally developed primers to improve sequence quality in problematic individuals (Table 2.2). However, in H. fuliginosus only two haplotypes were found at this marker in each catchment (Figure 7.5). Alternatively, mitochondrial control region was found to have higher levels of variation. Primers developed by Palumbi et al. 1991 (PRO-L and MT-H) (Table 2.1) were successful in amplifying H. fuliginosus individuals, and showed sufficient variability for population comparisons. All attempts to characterise a microsatellite markers for H. fuliginosus were unsuccessful, and a marker library was not produced.

Table 2.1 Control region primer sequences used to amplify H. fuliginosus.

PRO-L F: 5’- CTACCTCCAACTCCCAAAGC -3’ MT-H R: 5’- CCTGAAGTAGGAACCAGCTG -3’

Chapter 2.0 ∙ Field and Laboratory Methods 61

Table 2.2 ATPase 6 and 8 primer sequences used to amplify M. australis and M. s. inornata.

ATP8.2 F: 5’- AAAGCRTYRGCCTTTTAAGC -3’ COIII.2 R: 5’- GTTAGTGGTCAKGGGCTTGGRTC -3’ ATP8 RB F F: 5’- GCCCTTGCTCTDASCCTTCC -3’ COIII RB R R: 5’- TGCATGTGCTTGATGTGCCA -3’

Nine microsatellite loci with appropriate levels of variation were taken from published primers for Melanotaenia splendida in Zhu et al. (1998) and M. australis in Young et al. (2009) (see Table 2.3).

Table 2.3 Microsatellite primer sequences used to characterise genetic structure in

M. australis and M. s. inornata. * denotes loci isolated by Zhu et al. (1998), and ^ denotes loci isolated by Young et al. (2009).

MS22* F: 5’- GAGCGGTAAATGTCTGAGAGC -3’ R: 5’- CCGTTATTTGGGAACTGCTGG -3’ MS37* F: 5’- GCATGGGAGAGGTGGGCTG -3’ R: 5’- ACCTCTGTCCAGATGTCACTG -3’ MS40* F: 5’- GACACCATGGTCATCACCTGAG -3’ R: 5’- CCAGGTACTTGAACTTTCAAACC -3’ Ma03^ F: 5’- CAGCACCACCTTGTAAATCTCA -3’ R: 5’- TCCCTTTGAGGTCAATGCTC -3’ Ma04^ F: 5’- CCAAGTTCATCAAAAGAGGATT -3’ R: 5’- GGCTTCACTCCAGCTAACAGA -3’ Ma06^ F: 5’- TCTCCCTTATGAGCCCCTTT -3’ R: 5’- GTGCCGAGTTGTCAGCTAAA -3’ Ma09^ F: 5’- AGGTTACCAACAAATTTGTCCAC -3’ R: 5’- GAAAATCTTTTGGAGCCAAGG -3’ Ma11^ F: 5’- CCAATCAGAAGGCACGTT -3’ R: 5’- GGAGCATGACTGTTGAAGTATCTG -3’ Ma12^ F: 5’- TTCACTTCCTTTCATCATTTTACG -3’ R: 5’- TCGACCTCCATACAAAAGACC -3’

Chapter 2.0 ∙ Field and Laboratory Methods 62

2.3.3 Reaction and cycling conditions

2.3.3.1 Control region and ATPase 6 and 8

The amplification of mitochondrial DNA was performed in 10μL reactions using CTAB

DNA extractions (method previously described in 2.3.1). PCRs amplified approximately 850 base pairs of sequence data from the rainbowfish genome at ATPase

6 and 8, and 600 base pairs of sequence data from the grunter genome at control region.

For ATPase 6 and 8, amplification occurred in 10µL reactions of primers ATP8.2 and

COIII.2 (0.2µL of each), ddH2O (7.28µL), extracted DNA (0.5µL) and Fisher Biotec’s

10 x reaction buffer (1.0µL), MgCl2 [25mM] (0.6µL), dNTPs [10mM] (0.2µL), and

Thermus aquaticus DNA polymerase (Taq) (Fisher Biotec) [5.5 units/µL] (0.02µL).

For control region, amplification occurred in 10µL reactions of primers MT-H and

PRO-L (0.2µL of each), ddH2O (7.1µL), extracted DNA (1.0µL) and Astral Scientific’s

10 x reaction buffer (1.0µL), MgCl2 [25mM] (0.6µL), dNTPs [10mM] (0.2µL), and

Thermus aquaticus DNA polymerase (Taq) (Astral Scientific) [5.5 units/µL] (0.02µL).

These mixtures were then subjected to the same thermal program to amplify the target region: 94C for four minutes, then 35 cycles of: 94C for 30 seconds, 51C for 30 seconds, and 72C for one minute; then 72C for five minutes. Samples were then held at 4C.

Chapter 2.0 ∙ Field and Laboratory Methods 63

2.3.3.2 Microsatellites

The amplification of microsatellites was done using either a multiplex reaction with a total volume of 7.5 µL or a simplex reaction with a total volume of 10.0µL (Table 2.5).

Primer sequences were tailed with one of four fluorescent tags, allowing dye detection for each locus within multiplexed reactions (Table 2.4) (see Real et al. 2009 for detailed technical description). The reactions used total genomic DNA from CTAB extractions

(see 2.3.1).

Table 2.4 Fluorescent tag sequences used in microsatellite reactions.

Fluorescent Tag Tail Sequence NED 5’- CGGCCAGTGAATTGGATTTA -3’ PET 5’- TAGAGTCGACCTGCAGGCAT -3’ VIC 5’- GCTGCAAGGCGATTAAGTTG -3’ FAM 5’- CTCTTCGCTATTACGCCAGC -3’

Table 2.5 Annealing temperatures, fluorescent tags and grouping of microsatellite loci for PCR cycling.

Group and Annealing Primer Set Fluorescent Tag Temperature (TA) MS40 NED

Group A Part 1 - TA 57C MS37 PET Ma04 VIC

Group A Part 2 - TA 58C Ma11 NED Ma12 NED

Group B Part 1 - TA 58C MS22 PET Ma03 FAM Ma09 NED Group B Part 2 - TA 58C Ma06 FAM

Chapter 2.0 ∙ Field and Laboratory Methods 64

For multiplexed reactions, amplification occurred in 7.5µL reactions of primers.

Reactions contained forward primer (0.02µL of each), reverse primer (0.08µL of each), fluorescent tag (0.08µL of each for the associated primer), extracted DNA (0.8µL),

Astral Scientific’s 10 x reaction buffer (1.0µL), MgCl2 [25mM] (0.6µL), BSA (1.5µL), dNTPs [10mM] (0.2µL), and Thermus aquaticus DNA polymerase (Taq) (Astral

Scientific) [5.5 units/µL] (0.15µL), with ddH2O making up the reaction to 7.5µL.

Simplex reactions (for Ma11), contained forward primer (0.07µL of each), reverse primer (0.28µL of each), tag (0.28µL of each), ddH2O (4.9µL), extracted DNA (0.5µL),

Astral Scientific’s 10 x reaction buffer (0.7µL), MgCl2 [25mM] (0.42µL), dNTPs

[10mM] (0.14µL), and Thermus aquaticus DNA polymerase (Taq) (Astral Scientific)

[5.5 units/µL] (0.2µL). All mixtures were then subjected to thermal cycling conditions of: 94C for four minutes, then 35 cycles of: 94C for 30 seconds, TA (see Table 2.5 for specific values) for 30 seconds, and 72C for one minute; then 72C for 20 minutes.

Samples were then held at 4C. For each group, parts 1 and 2 were combined for each individual sample, then run on an ABI 3130 Genetic Analyser.

2.3.4 mtDNA sequencing

To determine success of the PCR, 2.5µL of PCR product was processed through an agarose gel (1%) with an electric current (80V and 400amps for 30 minutes, using a

Bio-Rad Power Pac 300). A DNA size marker (Lambda DNA for control region;

ECO911 (BstEII) for ATPase 6 and 8) denoted the length of the amplified fragment when observed under UV light (Bio-Rad). For positive samples, the remaining PCR product was purified with exo-sap (Fermentas exonuclease (20u/µL) and shrimp alkaline phosphatase (SAP) (1u/µL)). 1.0µL of exonuclease and 0.25µL of SAP was

Chapter 2.0 ∙ Field and Laboratory Methods 65 added to the PCR product, then subjected to a thermal program (37C for 35 minutes,

80C for 20 minutes, final hold at 4C).

Samples were then sequenced using the appropriate forward primer (either ATP8.2 or

PRO-L) (1.0µL), exosap product (1.0µL), ddH20 (5.75µL), 5X sequencing reaction buffer (2.0µL) and 3.1 Big Dye Terminator mix (0.25µL). The samples then underwent a final cycling program as follows: 96C for one minute; then 30 cycles of: 96C for 10 seconds, 50C for five seconds, 60C for four minutes. Samples were stored at 4C.

Approximately 20% of samples were additionally sequenced in reverse using the same procedures, to ensure accuracy and sequencing consistency.

Each reaction was then washed with 100% ethanol (60µL), ddH20 (10µL) and EDTA

[125mM] (5µL), then incubated at room temperature for 15 minutes. Samples were then centrifuged for 40 minutes at 13 500rpm, and the supernatant aspirated leaving the

DNA pellet to be rinsed in 70% ethanol (60µL). The centrifuge step was repeated (13

500rpm for 30 minutes), and the supernatant again aspirated. The DNA pellet was then dried in a vacuum chamber, and then sequenced using the Applied Biosystems 3130xl

Genetic Analyser in the Griffith University DNA Sequencing Facility.

2.3.5 Stable isotope preparation

All collected samples were rinsed in de-ionised water to ensure any contaminating debris was removed. For larger fish, a portion of the dorsal muscle above the lateral line was filleted, and the skin removed in order to provide a pure muscle sample. For smaller fish (<18mm), attempts to obtain pure muscle were very difficult. Consequently, for these individuals, the spinal column was removed, but the sample contained skin tissue.

Chapter 2.0 ∙ Field and Laboratory Methods 66

All samples were then placed in aluminium trays and dried at 60°C in a Thermoline

Scientific Dehydrating Oven for >48 hours. Dried fish tissue was finely ground in a small ceramic mortar and pestle. Other dried samples were placed in an aluminium canister containing a steel ball, and agitated for 180 seconds (at 30 beats/second) in a

Retsch MM200 puck and ring mill grinder. An appropriate mass for each sample was placed in a Sercon tin capsule, and moulded into a small tablet. A Eurovector EA 3000 elemental analyser was used to combust each sample, and the resultant gases were analysed in a GV Isoprime mass spectrometer (GV Instruments) for C and N. The calibrating standards used for analysis were atmospheric N for δ15N and ANU sucrose for δ13C. High lipid content in tissue can cause deviations in δ13C values. The ratios of

C:N were <4 for 97.3% of individuals, indicating low lipid content for the majority of individuals (Logan et al. 2008; Jardine et al. 2011). Consequently, lipid corrections not necessary to correct the data set.

2.4 Statistical analyses of data

2.4.1 Genetic diversity

Sequence data was aligned and edited using the program Sequencher 4.1(Gene Codes

Corporation 2000), resulting in sequences for H. fuliginosus at control region, and M. australis and M. s. inornata at ATPase 6 and 8. TCS v1.21 (Clement et al. 2000) was used to estimate the genetic relationships between haplotypes and construct parsimonious networks. Arlequin v3.5.1.2 (Excoffier et al. 2005) was used to quantify the genetic and nucleotide diversity of the mtDNA data sets.

Chapter 2.0 ∙ Field and Laboratory Methods 67

GeneMapper v4.0 (Applied Biosystems) was used to visualise and score microsatellite genotypes. Scored loci for each population were examined for null alleles, scoring error and allelic dropout using MICRO-CHECKER v2.2.3 (Van Oosterhaut et al. 2004).

MICRO-CHECKER v2.2.3 calculates the expected heterozygote and homozygote allele size difference frequencies for each given population, and uses the principles of Hardy-

Weinberg Equilibrium (HWE) to identify deviations from these expectations.

2.4.2 Analyses of molecular variance

Because of the manner in which gene frequencies can change due to genetic drift and mutation, the isolation of a population can establish a divergence from the original population. Wright’s FST (1949) describes this divergence by identifying the total heterozygosity (HT) of the whole population, and the heterozygosity within populations

(HS). FST denotes the proportion of the heterozygosity between populations, compared to the total heterozygosity.

This statistic lends itself to hierarchical levels of analysis when the sampling design is constructed in a tiered fashion, allowing the scale to widen and focus accordingly. An

Analysis of Molecular Variance (AMOVA) is designed to group the genetic structure of a population among groups and sites, but also within groups and sites (Excoffier et al.

1992). Three fixation indices comparable to Wright’s FST can then identify each hierarchal level as FCT (among groups), FSC (among sites within groups) and FST

(among all sites i.e. overall).

Using Arlequin v3.5.1.2 (Excoffier et al. 2005), AMOVAs were calculated to partition the total genetic variation at three levels – among catchments (FCT), among sites within

Chapter 2.0 ∙ Field and Laboratory Methods 68

catchments (FSC) and within sites (FST). An additional statistic, ΦST, was also used to incorporate genetic distance and genetic divergence to evaluate population structure

(Excoffier et al. 1992). ΦST was calculated using the pairwise differences between haplotypes in Arlequin v3.5.1.2 (Excoffier et al. 1992; Excoffier et al. 2005). AMOVAs were calculated separately for the Daly and Mitchell Rivers, with the highest level being subcatchments.

2.4.3 Testing for isolation by distance

Isolation by distance can be detected using Mantel tests, which reveal correlations between genetic distance and geographic distance (Wright 1943; Slatkin & Maddison

1990; Bohonak 1999). To test for associations between genetic distance and geographic distance, Arlequin v3.5.1.2 (Excoffier et al. 2005) was also used to perform Mantel tests. Here, genetic distance matrices using Slatkin’s Linearised FST (FST = 1/(1-FST)

(Slatkin 1995) for mitochondrial markers were correlated with geographic distance matrices for each genus. Distance was determined in two ways within each catchment:

1) distance via the base flow waterways, and 2) distance as the straight path between sites (as a simulated flooding event). Mantel tests were calculated separately for each of these conditions so as to evaluate the influence of floodplains on dispersal behaviour.

2.4.4 Testing for neutrality

Many of the theoretical and statistical classifications that are made about genetic data sets assume neutrality from selection (Beebee & Rowe 2008). Consequently, it is important to clarify the influences of selection and population size changes on a given data set. A star-like phylogeny can be indicative of a population expansion, and is

Chapter 2.0 ∙ Field and Laboratory Methods 69 characterised by one or two very common internal haplotypes, from which a number of closely related, low frequency haplotypes are generated (Slatkin & Hudson 1991).

During a bottleneck event, low frequency haplotypes have a higher chance of being lost from the population, which generally results in a network of several common haplotypes that may be connected by extinct haplotypes (Aris-Brosou & Excoffier

1996; Harpending et al. 1998).

Tajima’s D is commonly used to test for neutrality, and is able to distinguish between both positive selection (where a mutation is advantageous, and so individuals without this mutation are selected against), and balancing selection (where inflated levels of diversity are maintained) (Tajima 1989). In this test, θ is calculated by comparing the total pairwise differences to segregating sites. In situations where positive selection is occurring, the mutant will become fixed in the population in time, at which point synonymous sites will produce highly similar, low frequency haplotypes (Bamshad &

Wooding 2003; Nielsen 2005). Consequently, a Tajima’s test will result in a negative value, as the number of segregating sites exceeds the pairwise differences (Tajima 1989;

Kreitman 2000). In contrast to this, a number of haplotypes are maintained under balancing selection, which results in positive D values (Bamshad & Wooding 2003;

Kreitman & Di Rienzo 2004). Tajima’s D, while effective, cannot differentiate between selection and population size changes (Lin et al. 2011). Fu’s FS is another test for neutrality that is potentially more sensitive to the signatures of a population expansion as it is designed to detect the high number of low frequency mutations that are expected in recent demographic expansions (Fu 1997; Peck & Congdon 2004). Fu & Li (1993) also developed the F* and D* tests, which detect departures from neutral models of mutation by determining high levels of rare polymorphism caused by selection.

Chapter 2.0 ∙ Field and Laboratory Methods 70

By combining tests, a clearer representation of the factors affecting the observed deviations from neutrality can be reached (Rozas & Rozas 1999). To test whether the observed mutational differences were generated by a change in population size or selection, four statistics were used: Tajima’s D, Fu’s FS, Fu and Li’s F* and D* and raggedness indices (r) (to ascertain whether the mutations were a product of a stable or expanding population (Harpending et al. 1993). Typically, a demographic expansion is characterised by significant Tajima’s D and Fu’s FS values, and non-significant Fu and

Li’s F* and D* (Joseph et al. 2002). Alternatively, significant D* and F* statistics and non-significant Fu’s FS are indicative of selection (Peck & Congdon 2004).

To determine whether populations were in genetic drift – mutation equilibrium, the program DnaSP v5.1 (Librado & Rozas 2009) was used to calculate several tests of neutrality. Expected distributions of pairwise differences between sequences were modelled under sudden expansion and constant population size scenarios and compared to observed pairwise distributions. Raggedness values range from 0 to 1, with smaller values indicating a population expansion, and larger values indicating a population of constant size (Harpending et al. 1993). For a population that has recently expanded, mismatch distributions will have a single, smooth modal peak which is indicative of the accumulation of mutations that is expected in an expansion (Lavery et al. 1996; Dawson et al. 2001). For a population of constant size that is in equilibrium between genetic drift and mutation, distributions will be multimodal, and ragged (Rogers & Harpending

1992; Harpending et al. 1998; Cassone & Boulding 2006). Raggedness (r) (Harpending et al. 1993) was calculated for each species to determine the form of the distribution.

Chapter 2.0 ∙ Field and Laboratory Methods 71

2.4.5 Assignment testing

Assignment tests can be used to measure dispersal by determining the natal populations of individuals (Berry et al. 2004). Like other statistical methods commonly used in population genetics, assignment tests group individuals based on their genetic differentiation. However, assignment testing can be used to eliminate the artificial boundaries and delineations imposed by the researcher. In molecular ecology a population is usually defined as being a sampled site which is usually chosen based on the distance from other sites (Goudge 1955; Waples & Gaggiotti 2006). Sites are chosen in the hope that each of the sampled sites is sufficiently isolated from each other so that the individuals living within them have developed genetic differentiation, and that this genetic structure is captured. Assignment testing considers the population as a whole, and assesses the contained genetic diversity to determine the probability that a given genotype belongs to a given population (Excoffier & Heckel 2006). Furthermore, assignment testing can also incorporate (or discount) known information on putative populations (Beebee & Rowe 2008).

By determining K, which describes the estimated number of populations based on the genetic structure of the data set, a more realistic grouping of sites into populations can then be made (Pritchard et al. 2000). Once K is known, individuals can be grouped into the perceived ‘real’ populations (Evanno et al. 2005). Alternatively, individuals that cannot be assigned to a sampled population (i.e. the genotype is not a product of the sampled populations) are considered to be migrants (Rannala and Mountain 1997;

Paetkau et al. 2004). In certain situations, admixture can be observed, where an individual has a mixed ancestry from two or more populations (Waples & Gaggiotti

Chapter 2.0 ∙ Field and Laboratory Methods 72

2006). Admixture in itself is a useful tool for inferring the connectivity (or isolation) between populations.

CREATE v1.33 (Coombs et al. 2008) was used to convert raw data into the required input files for statistical programs. Two statistical packages, GENECLASS2 v2.0 (Piry et al. 2004) and STRUCTURE v2.3.3 (Pritchard et al. 2000), were used to analyse the genotype data set and assess patterns of gene flow using assignment testing (Manel et al. 2005; Falush et al. 2007). For STRUCTURE v2.3.3 (Pritchard et al. 2000), simulations were run with a 50 000 burnin, a run length of 1 000 000 Markov Chain

Monte Carlo repetitions, and five iterations for each estimate of K (1 through to 20). To obtain the best estimate of K, the method outlined in Evanno et al. (2005) was used. For this approach, ΔK is based on the change between each consecutive log probability of K.

The mode of the distribution of ΔK identifies the number of groups within the data set

(K), and the height of this mode reflects the strength of the signal within the data set.

GENECLASS2 v2.0 (Piry et al. 2004) was used to analyse each microsatellite data set and identifiy the most likely population of origin for each individual – assigned using a probability threshold of 95% (Manel et al. 2005; Falush et al. 2007). An admixture model with correlated allele frequencies was used, based on the proximity of the sampled sites to each other, and the probability of gene flow using the Bayesian assignment method of Rannala & Mountain (1997) (Falush et al. 2003; Hubisz et al.

2009).

2.4.6 Pairwise regressions

Decomposed pairwise regressions were used to investigate the relationships between gene flow and landscape variables in both the Daly and Mitchell River catchments. The

Chapter 2.0 ∙ Field and Laboratory Methods 73 decomposed regression analyses followed the method described in Koizumi et al.

(2006), and aimed to remove outlying populations from the data set that were potentially increasing the dispersion of regression residuals. Such populations potentially mask correlations within a data set by reducing the power to detect the influence of landscape factors on genetic isolation (Slatkin 1993; Rousset 1997;

Hutchison & Templeton 1999). Decomposed regression analyses involved two stages in the process. The most outlying populations were sequentially removed from the data set, and were determined as the populations that deviated the most from 0 (provided

95% confidence intervals did not include 0 – e.g. see Figure 3.15, where the population from the Fergusson River (site DFF) was the first removed) and Figure 4.9, where the population from Stray Creek (site DR) was the first removed) (Koizumi et al. 2006).

Secondly, in order to identify whether these populations were ‘true’ outliers, Akaike

Information Criteria (AIC) (Akaike 1974) were used to determine the model that best fitted the data. Populations excluded from the model were confirmed as ‘true outliers’ to the data set. Models with AIC differences (ΔAICC) of 2.0 or less were considered equally likely (Johnson & Omland 2004). Finally, for the pairwise data sets, individual regressions were made for each of the populations (including outliers) to determine the effects of outliers on the overall regression residuals – that is, how outliers influenced the perceived relationships between the landscape and gene flow. For each of the regressions, the characteristics of gene flow and genetic drift were identified for each population, according to the model described by Koizumi et al. (2006) (Figure 2.9).

From identifiable patterns, inferences were made regarding the connectivity of populations. For example, populations conforming to Pattern 1 (where drift is much greater than gene flow) were highly isolated, whereas populations conforming to Pattern

4 were highly connected (as the effects of gene flow are greater than those of genetic drift). These patterns were identified by the significance of intercepts and slopes for

Chapter 2.0 ∙ Field and Laboratory Methods 74 individual population regressions (Table 2.6). Patterns 1 and 2 were identified by significant intercepts, but for Pattern 1 the slope was non-significant. Pattern 3 was identifiable by a non-significant incept and a significant slope; regressions where both the intercept and slope were non-significant were classed as being Pattern 4.

Table 2.6 Classification of patterns based on the regression intercepts and slopes.

Pattern Intercept Slope 1 Significant Non-significant 2 Significant Significant 3 Non-significant Significant 4 Non-significant Non-significant

2.4.7 Stable isotope analysis

In a similar manner to assignment testing, the analysis of a group of stable isotope signatures can differentiate between populations, and identify individuals that do not fit

Chapter 2.0 ∙ Field and Laboratory Methods 75 within the realm of the sampled populations. Each sample of fish tissue, biofilm, filamentous algae, CPOM, FPOM, seston and riparian vegetation produced a value which was expressed as the ratio of heavy to light isotopes: 13C/12C (δ13C) and 15N/14N

(δ15N). Delta (δ) values were calculated according to an international standard in the equation δX = (Rsample/Rstandard - 1) x 1000 (Werner & Brand 2001). These values were then plotted against each other to visualise the distribution of values across the isoscape.

A residual permutation procedure described in Turner et al. (2010) was used to test for significant differences in the stable isotope ratios (δ13C: δ15N) between dry season and wet season fish collections at each of four sites in the Mitchell River headwaters (see

Chapter 5.2.1). Empirical p-values were estimated based on the permutation of residuals from a linear model. This model was used to estimate group centroids and the dispersion around the centroids (Layman et al. 2007; Turner et al. 2010).

Metrics summarising the distributions of individuals in isotopic space were: 1)

Euclidean distance between group centroids (MD), 2) mean Euclidean distance of individuals from the group centroid (CD) (a measure of dispersion), 3) average distance of each individual to its nearest neighbour (a second measure of dispersion) (NND), 4) the distance between the most enriched and the most depleted δ15N samples (NR) and 5) the distance between the most enriched and the most depleted δ13C samples (CR). MD was used to determine whether the position of group centroids changed from dry season to wet season. CD and NND were used to test for migrants in each of the groups – individuals that were outside the distribution for a given group. NR and CR were used to show the spread of trophic positions (and land uses) and carbon sources at each of the sites. A total of 999 permutations of the residual vectors were used for each metric, excluding calculations of NR and CR. The analysis was conducted in the R statistical

Chapter 2.0 ∙ Field and Laboratory Methods 76 environment (R Development Core Team 2009) using a modified permutation code developed by Turner et al. (2010), available as supplementary material (Ecological

Archives E091-157-S1).

Chapter 2.0 ∙ Field and Laboratory Methods 77

3.0 The influence of the structure and form of dendritic systems on

gene flow in rainbowfish populations of tropical Australia

3.1 Introduction

It is widely recognised that the genetic diversity of a population is an essential criterion for population persistence (Frankham 1995; Pertoldi et al. 2007). In preserving and managing this diversity, it is necessary to identify the processes which may positively or negatively alter intraspecific genetic variation (Lacy 1988; Latta 2008). Genetic diversity is the buffer that can allow a species to react to selective pressure (Reed &

Frankham 2003). Effectively, it is the key to ‘evolvability’, and without this genetic diversity the plasticity and resilience of a species is impaired in the event of environmental change (Palstra & Ruzzante 2008; Hoffmann & Willi 2008). Dispersal is the means through which existing genetic diversity may be transferred within and between populations – that is, gene flow (Allendorf 1983; Slatkin 1987). By quantifying the genetic connectivity of a population, further inferences can be made regarding population structure, population boundaries, migration, recruitment, dispersal capability and also historical events (Palstra & Ruzzante 2008; Hoelzel 2010).

Studies of gene flow may also highlight the impact of landscape characteristics on connectivity, by identifying how the genetic structure and diversity of a population has been influenced by geomorphic features in their range (Storfer et al. 2007; Anderson et al. 2010). This relationship is an important one to consider when determining the genetic dynamics of a population, and how the evolution of both the landscape and the populations within it are linked (Holderegger & Wagner 2008). Within a landscape, distance itself can drive isolation between populations. For taxa with a dispersal

Chapter 3.0 ∙ Rainbowfish 78 capability that is less than the geographic range of the population, gene flow may be limited to neighbouring populations, thereby restricting the diffusion of genes across the landscape (Kimura & Weiss 1964; Slatkin 1993; Primmer et al. 2006). Isolation by distance (IBD) describes this ‘transitional’ isolation, where increased distance between populations leads to increased genetic divergence (Wright 1943; Jenkins et al. 2010).

As a result, neighbouring populations will be more similar than those further afield.

Recent advances in landscape genetics have challenged the significance of IBD in contrast to the effects of more tangible physical features (e.g. waterfalls, roads, weirs or mountain ranges). However, many studies have shown that IBD is still an influential component that can contribute to the observed spatial variation, in addition to that of a physical barrier (Kittlein & Gaggiotti 2008; Jenkins et al. 2010). Consequently, it is important to gauge the impact of IBD on population structure so that limits to dispersal distance are accounted for in management plans (Manel et al. 2003; Epps et al. 2007).

While populations in a landscape may be isolated as a result of distance, populations in elevated regions can also experience restrictions to gene flow (Narum et al. 2008;

Morrissey & de Kerckhove 2009). Elevated sites may present a new set of challenges for the establishment of populations. This may be due to the direct effects of elevation, such as changed temperatures and lower atmospheric oxygen concentrations (Altshuler

& Dudley 2006; Dillon et al. 2006). Alternatively, the limiting factor may be the change in vegetation, or the abundance of food resources. For fish, the slope of the landscape contributes greatly to the isolation of populations, and can be especially influential in determining ‘directionally biased’ gene flow (Buisson et al. 2008; Morrissey & de

Kerckhove 2009; Cook et al. 2011). Directionally biased gene flow can be attributable to the way that the slope of the landscape increases the velocity of flow in the

Chapter 3.0 ∙ Rainbowfish 79 waterways, making it more difficult and energetically expensive for aquatic biota to disperse against higher flows (Lowe et al. 2006). For these reasons, populations in elevated regions can be under pressure to adapt to these specific environments, particularly if they are on the ‘edge’ of the elevation limit (Sexton et al. 2009).

In addition to the potential diverging effects of selection are the implications of genetic drift in isolated populations at higher elevations. For populations where elevation restricts dispersal, reduced gene flow allows further differentiation through genetic drift

(Meeuwig et al. 2010). Collectively, this results in populations that can be quite divergent, and highly adapted to high altitudes and local conditions (Garland & Adolph

1991; Crossin et al. 2004). A mountain population may quite literally become an island in a lowland sea. Consequently, populations that are in elevated regions need to be managed carefully, as recolonisation may be very difficult from ill-adapted lowland populations (Hughes et al. 2009).

Within a river catchment, water drains from the highest regions to the lowest regions.

While it might be expected that higher altitude aquatic populations are more isolated, there are further complications from the dendritic structure of the river (Fagan 2002).

Generally, the sequential draining of each tributary results in the main channel of a river being the most hydrologically connected waterway type. Those tributaries attached to the main channel experience less discharge and this pattern continues up the stream hierarchy to the headwaters. Smaller streams with no supplementary input to flow, aside from precipitation, may become disconnected from the rest of the system over periods of reduced precipitation; the main channel, under most conditions, should maintain its linear connectivity. Off-channel waterholes would also be disconnected, particularly if only unusually high flows reconnect them to the river system. As discussed in Chapter

Chapter 3.0 ∙ Rainbowfish 80

1.2, the population connectivity of aquatic taxa can follow that of the Stream Hierarchy

Model (SHM), where populations within tributaries are more genetically connected than populations in neighbouring tributaries (Meffe & Vrijenhoek 1988). So it might be expected that populations within tributaries and waterholes become genetically differentiated from the rest of the catchment (Morrissey & de Kerckhove 2009). It follows that the more tributaries a river has (alternatively, the more structured the river), the greater the genetic structure (Hughes et al. 2009).

The Daly and Mitchell Rivers are two of the largest catchments in northern Australia.

Being in a tropical region, both experience periodic episodes of high monsoonal flows.

In addition, the dendritic structure of these rivers is complex, with many tributaries and branches. Within the catchments, there are regions of higher elevation, rising to more than 400m above sea level. This study aimed to investigate the impact of the landscape on gene flow between populations of M. australis and M. s. inornata in the Daly and

Mitchell Rivers, respectively. Several predictions were made:

1. That connectivity between populations of rainbowfish would be limited by the

catchment landscape – specifically the size of the catchments, elevation and the

hydrological connectivity of waterways.

2. That the connectivity of populations would also be determined by the dendritic

structure of catchments.

Chapter 3.0 ∙ Rainbowfish 81

3.2 Detailed Methods

The following methods give specific detail that is additional to the general methods outlined in Chapter 2.0.

3.2.1 Classification of waterways

Populations were classed as being 1) downstream – that is, <170km from the river mouth, 2) midreach – that is 171-310km from the river mouth, or 3) upstream – that is

>311km from the river mouth. These groupings were made by dividing each catchment

(i.e. the Daly and Mitchell catchments) into three, approximately equal lengths.

Additionally, each population was identified as being 1) a tributary population – that is a population sampled in a connected tributary, 2) a main channel population – that is a population that was sampled within the permanent delineations of the main channel, or

3) a waterhole population – that is a population that was confined to a waterhole that may only experience connectivity during the wet season.

3.2.2 Isolation of waterways

An ‘isolation index’ (calculated as the mean FST between a site and all other sites) was determined for each site within each catchment. First, populations were classified as being either 1) downstream, 2) midreach, or 3) upstream – according to the steps in

Chapter 3.2.1. Second, populations were classified as being either 1) main channel populations, 2) tributary populations, or 3) waterhole populations. Significant differences between these groups were determined using one-way analyses of variance

(ANOVA) and Scheffé’s multiple comparison tests. Spearman’s Rank Correlations

Chapter 3.0 ∙ Rainbowfish 82 were also conducted to test the relationship between distance from river mouth (x) and the fixation value for the given site (y).

3.2.3 Testing landscape influences

A Mantel test was used to evaluate the relationship between genetic and geographic distance, using Slatkin’s linearised FST (where (FST =1/(1−FST)) (Slatkin 1995) for genetic distances, and river pathway between sites for geographic distance (Tables 7.8 and 7.9). A decomposed regression analysis was used to identify any relationship between genetic isolation and two landscape characteristics. These were 1) the elevation of each population, and 2) the stream distance between populations (using pairwise data).

Chapter 3.0 ∙ Rainbowfish 83

3.3 Results

3.3.1 Intrapopulation genetic diversity

Collections of M. australis resulted in 182 individuals from 17 sites in the Daly River

(Table 3.1 and Figure 3.1). For M. s. inornata, 185 individuals from 16 sites in the

Mitchell River were captured (Table 3.2 and Figure 3.2).

Figure 3.1 The Daly River catchment, showing the 17 sites from which M. australis was collected.

Chapter 3.0 ∙ Rainbowfish 84

Figure 3.2 The Mitchell River catchment, showing the 16 sites from which M. s. inornata was collected.

Direct sequencing of a 510bp fragment of the ATPase mitochondrial region yielded significant levels of haplotypic variation for both species of melanotaenids. For M. australis, this resulted in 21 unique haplotypes from 182 individuals (Figure 3.3). For

M. s. inornata, this resulted in 27 unique haplotypes from 185 individuals (Figure 3.4)

(refer to Tables 7.6 and 7.7 for detailed haplotype distributions for M. australis and M. s. inornata). In each species, some haplotypes formed the centre of a star-like pattern

(two arrangements for M. australis – Figure 3.3, and several arrangements for M. s. inornata – Figure 3.4), where several common internal haplotypes were surrounded by recently derived haplotypes. No haplotypes were shared across all of the sampled sites.

Chapter 3.0 ∙ Rainbowfish 85

In the bottom diagram, circles represent each individual haplotype (frequency corresponds to circle size), with the individual colour corresponding to the haplotypes represented in mapped populations. Small black circles indicate haplotypes that were not sampled, or have become extinct through genetic drift.

Chapter 3.0 ∙ Rainbowfish 86

Chapter 3.0 ∙ Rainbowfish 87

Table 3.1 Gene diversity indices for M. australis in the Daly River. Deviations from HWE are denoted by *, after sequential Bonferroni

correction. Monomorphic loci are denoted by -. No tests were performed for populations where n < 3.

ATP8 MS37 Ma04 Site Sample Haplotype Nucleotide Site H H P H H P Code Size Diversity Diversity E O E O Beeboom DA 9 0.750 0.005 0.824 0.889 0.444 0.765 0.889 1.000 Claravale/Sawmill DM 35 0.600 0.004 0.824 0.900 0.933 0.763 0.900 0.599 Middle Creek DB 4 0.500 0.002 0.825 0.899 0.915 0.792 0.750 0.551 Fergusson River DFF 27 0.074 0.000 0.677 0.769 0.445 0.687 0.733 0.421 U/S Centipede Dreaming and Snowdrop Creek DE 5 0.400 0.004 0.643 0.500 0.429 0.650 0.625 0.510 UF Daly River Road DEE 4 0.000 0.000 0.768 0.500 0.055 0.810 0.556 0.201 Flora River DGG 3 0.000 0.000 0.733 0.667 1.000 0.800 0.667 0.601 Fish River DJ 4 0.500 0.004 0.911 0.800 0.489 0.600 0.800 1.000 Big Bird Billabong DJJ 2 1.000 0.006 Umbrawarrah Gorge DKK 8 0.000 0.000 0.808 0.625 0.005 0.692 0.625 0.046 Galloping Jack's DL 26 0.569 0.002 0.809 0.684 0.492 0.718 0.842 0.569 Sandy/Hermit DLL 5 0.400 0.002 0.857 0.714 0.205 0.727 0.667 0.308 Mathison Creek DP 7 0.000 0.000 0.692 0.857 0.261 0.692 0.714 1.000 Stray Creek DR 10 0.822 0.005 0.804 0.778 0.465 0.810 0.778 0.159 Douglas River DS 12 0.318 0.001 0.834 0.923 0.588 0.732 0.923 0.602 Oolloo Crossing (upstream) DT 17 0.838 0.005 0.811 0.900 0.961 0.689 0.921 0.153 Oolloo Crossing DX 4 0.833 0.004 0.929 1.000 1.000 0.643 0.750 1.000

Continued over page…

Chapter 3.0 ∙ Rainbowfish 88

Ma11 Ma12 MS22 Ma06 Site H H P H H P H H P H H P Code E O E O E O E O DA 0.967 0.750 0.013 0.758 0.778 0.640 0.538 0.714 0.511 0.843 0.556 0.052 DM 0.970 0.684 0.001 0.580 0.526 0.002 0.638 0.429 0.069 0.808 0.800 0.768 DB 0.934 0.978 0.509 0.633 0.750 0.701 0.143 0.143 1.000 0.492 0.250 0.077 DFF 0.921 0.948 0.609 0.248 0.200 0.202 0.519 0.182 0.061 0.395 0.200 0.036 DE 0.975 0.875 0.187 0.670 0.286 0.027 - - - 0.739 0.667 0.278 DEE 0.938 0.500 0.000* 0.680 0.818 1.000 0.303 0.333 1.000 0.775 0.750 0.855 DGG - - - 0.333 0.333 1.000 0.281 0.265 0.954 0.600 0.667 1.000 DJ 0.956 0.800 0.240 0.644 0.200 0.016 - - - 0.756 0.400 0.049 DJJ DKK 0.933 0.750 0.156 0.717 0.500 0.082 0.440 0.140 0.021 0.608 0.375 0.166 DL 0.969 0.632 0.004 0.502 0.556 0.384 0.601 0.250 0.011 0.752 0.842 0.995 DLL 0.909 0.333 0.001 0.835 0.857 0.788 0.652 0.500 0.269 0.758 0.429 0.116 DP 0.978 0.800 0.106 0.703 0.714 0.588 0.670 0.714 0.775 0.802 0.571 0.204 DR 0.950 0.875 0.435 0.451 0.333 0.528 0.742 0.593 0.123 0.824 0.778 0.735 DS 0.938 0.667 0.013 0.695 0.692 0.240 0.551 0.583 0.820 0.212 0.077 0.120 DT 0.963 0.700 0.006 0.600 0.500 0.842 0.608 0.500 0.215 0.779 0.900 0.757 DX 0.929 0.750 0.309 0.607 0.500 0.431 0.607 0.750 1.000 0.607 0.250 0.143

Chapter 3.0 ∙ Rainbowfish 89

Table 3.2 Gene diversity indices for M. s. inornata in the Mitchell River. Deviations from HWE are denoted by *, after sequential

Bonferroni correction. Monomorphic loci are denoted by -. No tests were performed for populations where n < 3.

ATP8 MS40 MS37 Ma04 Site Sample Haplotype Nucleotide Site H H P H H P H H P Code Size Diversity Diversity E O E O E O Ten Mile Lagoon MB 8 0.821 0.005 0.833 0.875 0.992 0.890 1.000 1.000 0.813 0.571 0.152 Mitchell River @ Gamboola MC 2 1.000 0.002 Cairo Lagoon ME 19 0.930 0.006 0.858 0.500 0.006 0.429 0.500 1.000 0.346 0.355 1.000 Fishhole Creek MF 1 1.000 0.000 Twelve Mile Lagoon MG 13 0.795 0.005 0.908 0.875 0.291 0.800 0.624 0.067 0.833 0.500 0.332 Koolatah Gauging Station MH 3 0.667 0.005 0.800 0.667 0.601 0.560 0.578 1.000 - - - Mitchell River @ Koolatah MI 1 1.000 0.000 Magnificent Creek MJ 7 0.667 0.004 ------Saltwater Creek MK 25 0.000 0.000 0.814 0.733 0.223 0.769 0.714 0.298 0.358 0.200 0.009 Walsh River @ Nullinga ML 4 0.500 0.001 0.533 0.400 0.335 0.511 0.200 0.110 0.429 0.500 1.000 Palmer River @ Drumduff MM 19 0.889 0.007 0.863 0.444 0.001 0.790 0.765 0.265 0.752 0.588 0.309 Kingfish Lagoon MN 18 0.830 0.005 0.916 1.000 0.861 0.868 0.706 0.307 0.679 0.500 0.033 Dickson's Hole MO 14 0.143 0.001 0.860 0.786 0.528 0.702 0.923 0.247 0.612 0.538 0.104 Tate River @ Ootan MP 11 0.444 0.004 0.658 0.818 0.677 0.091 0.091 1.000 0.663 0.083 0.000* Emu Creek @ Petford MQ 20 0.100 0.000 0.760 0.600 0.096 0.737 0.650 0.094 0.533 0.333 0.154 Mitchell River @ St. George MR 20 0.726 0.003 0.857 0.500 0.008 0.852 1.000 0.961 0.280 0.308 1.000

Continued over page…

Chapter 3.0 ∙ Rainbowfish 90

Ma11 Ma12 MS22 Ma03 Ma09 Ma06 Site H H P H H P H H P H H P H H P H H P Code E O E O E O E O E O E O MB 0.967 0.750 0.033 0.742 0.500 0.238 0.879 0.500 0.061 0.567 0.250 0.021 0.725 0.750 0.168 0.800 0.750 0.163 MC ME 0.957 0.667 0.002 0.533 0.353 0.026 0.800 0.800 1.000 - - - 0.651 0.333 0.005 0.752 0.789 0.728 MF MG 0.978 0.857 0.168 0.685 0.417 0.060 0.711 0.400 0.143 - - - 0.659 0.214 0.000* 0.797 0.923 0.622 MH - - - 0.867 0.333 0.066 0.649 0.648 1.000 0.833 0.500 0.334 0.733 0.333 0.201 0.733 0.333 0.199 MI MJ 0.941 0.799 0.152 0.659 0.286 0.027 0.849 0.512 0.334 - - - 0.567 0.010 0.001 0.848 0.833 0.433 MK 0.944 0.813 0.196 0.590 0.435 0.040 0.733 0.920 0.745 - - - 0.479 0.400 0.421 0.526 0.375 0.042 ML 0.978 0.800 0.102 0.356 0.400 1.000 0.742 0.500 0.076 0.833 0.500 0.334 0.511 0.600 1.000 0.692 0.571 0.133 MM 0.986 0.824 0.002 0.775 0.684 0.091 0.530 0.333 0.514 0.929 0.500 0.030 0.677 0.450 0.001 0.727 0.650 0.662 MN 0.962 0.875 0.180 0.696 0.714 0.726 0.861 0.647 0.076 0.941 0.556 0.001 0.622 0.115 0.016 0.781 0.444 0.008 MO 0.932 0.846 0.044 0.632 0.643 0.916 0.791 0.429 0.000* 0.889 0.333 0.000* 0.325 0.214 0.087 0.696 0.571 0.083 MP 0.918 0.818 0.438 0.498 0.636 1.000 0.571 0.818 0.298 0.683 0.625 0.348 0.159 0.000 0.044 0.312 0.000 0.008 MQ 0.940 0.556 0.000* 0.385 0.308 0.526 0.591 0.600 0.119 0.780 0.286 0.004 0.273 0.198 0.490 0.462 0.368 0.606 MR 0.984 1.000 1.000 0.699 0.750 0.705 0.649 0.615 0.726 0.774 0.600 0.258 0.582 0.444 0.422 0.783 0.750 0.237

Chapter 3.0 ∙ Rainbowfish 91

After sequential Bonferroni correction (Rice 1989), deviations from Hardy Weinberg

Equilibrium (HWE) were detected in 0.98% of tests for M. australis, and 3.47% of tests for M. s. inornata (Table 3.1 and 3.2; see Table 7.10 for allelic diversity estimates) The use of MICRO-CHECKER v2.2.3 (Van Oosterhout et al. 2004) showed these departures to be the result of heterozygote deficiencies in all cases as a result of null alleles. The program FREENA (Chapuis & Estoup 2007) was used to correct the genotype frequencies at all loci for each population, and evaluate the influence of null alleles on FST values. FREENA had little effect on the original values, and so the raw

(uncorrected) data set was used for analysis.

For the microsatellite data set, some deviations from HWE were observed as a result of null alleles. Null alleles themselves are caused when mutations in the primer binding sites result in non-amplification of one or more of the alleles at a given locus (Van

Oosterhout et al. 2004; Chapuis & Estoup 2007). Consequently, apparent homozygotes and homozygote excesses can be generated in the population if the frequency of null alleles is high. Null alleles can be problematic when looking at genetic structure, as the heterozygote deficiencies will artificially inflate the FST values (Van Oosterhout et al.

2006; Falush et al. 2007). However, within the rainbowfish microsatellite data sets only a small percentage of populations at some loci (0.98% of tests for M. australis, and

3.47% of tests for M. s. inornata) deviated from random expectations. This result suggests that the frequency of null alleles was low. In all instances of departures from

HWE, the original chromatograms were reassessed to ensure the scoring was accurate and not influencing HWE (Morin et al. 2009).

Deviations from HWE can also be a product of selection (Van Oosterhout et al. 2004).

Selection in microsatellite loci is considered to be unlikely, as they are generally

Chapter 3.0 ∙ Rainbowfish 92 assumed to be neutral markers (Queller et al. 1993; Jarne & Lagoda 1996; Ellegren

2004). Indirect selection is possible in cases where a locus is linked to a functional gene, which is itself under selection (Maynard Smith & Haigh 1974; Kaplan et al. 1989;

Selkoe & Toonen 2006). The effect of indirect selection is difficult to determine, as selection pressures will vary between sites and may not be operating equally in each of the sampled populations. However, it is also important to note that trimers can be found in coding genes, as mutations do not change the reading frames (Edwards et al. 1991).

Two loci involved in the study were trimers – MS40 and Ma12. For MS40, no departures from HWE were detected. A single population (Palmer River @ Drumduff

(MM)) was out of HWE at Ma12 in M. s. inornata – potentially the result of direct selection.

Another driver behind departures from HWE is the Wahlund effect, which occurs when two or more non-interbreeding populations are combined (Wahlund 1928; Johnson &

Black 1984). The Wahlund effect may be a natural occurrence – for instance when two habitats suddenly become linked; however, it may also be artificial if a ‘population’ consists of separate sub-populations which may then be mistaken for a single stock

(Christiansen 1988; Sugg et al. 1996; Kempf et al. 2010). In such cases, deviations from

HWE would be observed across a number of loci in the population – a finding that was not observed in either of the rainbowfish data sets.

A similar concept to the Wahlund effect is the impact of age structure on a data set. Age structure is an outcome from non-random sampling, where age classes are disproportionately represented in the collection (Palstra & Ruzzante 2010). These separate cohorts may distort the population differentiation, particularly if there are high levels of recruitment from neighbouring populations (Jorde & Ryman 1995). The

Chapter 3.0 ∙ Rainbowfish 93 majority of collections were made using a boat mounted electrofisher. While every effort was made to capture individuals from all size classes, it may be that adults are overrepresented in the data set, as juveniles are more difficult to see from the prow of the boat. Additionally, at many sites the electro-fishing settings were set to shock fish over a broad range of size classes, and so juveniles (<15mm) may not have been stunned. Like the Wahlund effect, departures from HWE would be expected across a number of loci for an age-structured population, but this was not found. While age structure may be affecting the data set, it is difficult to confirm the impact as for this collection, no length measurements were recorded for the sampled individuals.

The patchy recruitment hypothesis is also potentially responsible for the seemingly random deviations from HWE in some populations. This hypothesis describes how unusual genetic diversity and structure are produced when the population is dominated by the offspring of a limited number of matings (Bunn & Hughes 1997). While this concept was originally developed for freshwater invertebrates, this pattern may also be observed in M. australis and M. s. inornata, provided that 1) there are a limited number of breeding individuals in each population, and 2) the population is highly philopatric

(Bunn & Hughes 1997; Schultheis et al. 2008). However, while these species show some philopatry (higher rates of self-assignment identified in the GENECLASS2 v2.0 results), the rates were low for most M. s. inornata populations identified as being out of HWE. M. australis at UF Daly River Road (site DEE) was the only significant deviation from HWE at locus Ma11 in this species. At this site, self-assignment was estimated at 61.54%. However, these philopatric individuals only represented 62.50% of the total individuals assigned back to that population, in which case the effects of patchy recruitment would be negated. It should also be pointed out that philopatry at this site is not optional as UF Daly River Road is a highly disconnected off-channel

Chapter 3.0 ∙ Rainbowfish 94 waterhole. Consequently, patchy recruitment does not seem to be a biologically plausible rationalisation for the departures from HWE.

Allendorf & Phelps (1981) also point out that departures from neutrality can also be introduced to data sets in two ways. First, low sample sizes can inaccurately portray the genetic structure of a population – as by chance, individuals collected may give a poor representation of the true levels of gene flow. In this scenario, Allendorf and Phelps give the example of two pairs of adult fish (all from a panmictic population) which spawn in two different tributaries. If these individuals were sampled, it may be concluded (provided that there is adequate genetic variability) that there are two non- interbreeding populations. Second, the resulting offspring of these matings would be overrepresented in the gene pool due to genetic drift, in a manner similar to patchy recruitment. Again, this would confirm the conclusion of two non-interbreeding populations, as the true biological interpretation is obscured (Waples 1998). Given the seemingly random departures from HWE, the Allendorf – Phelps effect is a plausible explanation in the microsatellite data sets.

3.3.2 Tests of neutrality

For both M. australis and M. s. inornata, neutrality tests showed some significant results (Table 3.3). In M. australis, Fu and Li's D* and F* were both significant.

However Tajima’s D, Fu’s FS and R2 were not. In contrast, for M. s. inornata Fu’s FS was significant, with Tajima’s D, Fu and Li's D* and F* and R2 being non-significant.

For individual populations (where n > 10) neutrality tests were all non-significant.

Chapter 3.0 ∙ Rainbowfish 95

Table 3.3 Tests of neutrality for populations of M. australis (Daly River) and M. s. inornata (Mitchell River). * denotes significant values.

Test Statistic P α

Tajima's D -0.931 0.186 0.050

Fu's FS -6.324 0.069 0.020 Fu and Li's D* -3.386* 0.004 0.020

australis Fu and Li's F* -2.897* 0.013 0.050 M. M. R2 0.056 0.200 0.050

Tajima's D -0.966 0.167 0.050

Fu's FS -11.010* 0.014 0.020 Fu and Li's D* -1.159 0.094 0.020

Fu and Li's F* -1.299 0.097 0.050 M. s. inornata s. M. R2 0.055 0.190 0.050

Pairwise differences for constant and expanding populations were modelled for both species (Figures 3.5 and 3.6). Observed pairwise differences most resembled those of expanding populations for both species (Figures 3.5b and 3.6b). Goodness of fit tests

(raggedness) were non-significant for M. australis, indicating that the mismatch distributions reflect demographic expansions (Table 3.4). However, r was low but significant for M. s. inornata, suggesting that the mismatch distribution was not smooth, and was more representative of a stable population.

Chapter 3.0 ∙ Rainbowfish 96

Chapter 3.0 ∙ Rainbowfish 97

Chapter 3.0 ∙ Rainbowfish 98

Table 3.4 Goodness of fit test (raggedness) for simulated population expansions in M. australis and M. s. inornata. * denotes significant values where α = 0.05.

Raggedness Indices P M. australis 0.149 0.801 M. s. inornata 0.020* 0.028

3.3.3 Partitioning of genetic variation

A global FST showed significant genetic structure for both M. australis (FST mtDNA =

0.406, ΦST mtDNA = 0.532 and FST microsatellites = 0.132) (Table 3.5) and M. s. inornata (FST mtDNA = 0.384, ΦST mtDNA = 0.404 and FST microsatellites = 0.128)

(Table 3.6). Significant genetic structure was detected among subcatchments, within subcatchments and among all populations in all genetic markers for both species

(Tables 3.5 and 3.6). Greater genetic variation was also detected among populations within catchment positions, compared to among catchment positions – a finding that was consistent across all markers for both species. None of the genetic variation was attributable to ‘among waterway types’ with either species. Significant structure was detected among populations within waterway types, and among all populations for both

M. australis and M. s. inornata. Overall, populations within waterway types and catchment positions accounted for much of the variation, indicating that gene flow was still restricted within these classifications.

Chapter 3.0 ∙ Rainbowfish 99

Table 3.5 AMOVA calculations for M. australis. * denotes significant fixation indices. ° denotes % variation within populations.

ATP8 Microsatellites % % % F P Φ P F P Variation ST Variation ST Variation ST Among populations, within - 0.406* 0.000 - 0.532* 0.000 - 0.132* 0.000 the Daly River Among subcatchments in 30.76 0.308* 0.008 39.60 0.396* 0.009 6.92 0.069* 0.000 the Daly River Among populations, within 11.83 0.171* 0.000 15.57 0.258* 0.000 6.78 0.073* 0.000 subcatchments Among all populations 57.40° 0.426* 0.000 44.83° 0.552* 0.000 86.31° 0.137* 0.000

Table 3.6 AMOVA calculations for M. s. inornata. * denotes significant fixation indices. ° denotes % variation within populations.

ATP8 Microsatellites % % % F P Φ P F P Variation ST Variation ST Variation ST Among populations, within - 0.384* 0.000 - 0.404* 0.000 - 0.128* 0.000 the Mitchell River Among subcatchments in 33.05 0.330* 0.003 32.24 0.322* 0.006 6.60 0.070* 0.000 the Mitchell River Among populations, within 6.12 0.091* 0.017 8.85 0.131* 0.006 6.43 0.069* 0.006 subcatchments Among all populations 60.84° 0.392* 0.000 58.90° 0.411* 0.000 86.97° 0.130* 0.000

Chapter 3.0 ∙ Rainbowfish 100

According to the cluster algorithm in STRUCTURE v2.3.3 (Pritchard et al. 2000), K

(the number of populations represented in the genotype data) was estimated to be four for M. australis under the Evanno et al. (2005) method using ΔK. The highest modal value was 342.49 (Figure 3.7). For most individuals, a large proportion of their genes were largely derived from a single source population. However, many individuals had more equal contributions from two, or even all of the defined populations. Overall, this highlighted the genetic structure within M. australis – particularly for sites such as the

Fergusson River (DFF) (Figure 3.8).

Magnitude of K For M. australis

400

350

300

250

ΔKK 200

150

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 K

Figure 3.7 Magnitude of ΔK for each estimation of K (1 – 20) for M. australis.

Chapter 3.0 ∙ Rainbowfish 101

Figure 3.8 Proportions of membership of M. australis individuals to each of the three groups determined by STRUCTURE v2.3.3.

For M. s. inornata, STRUCTURE v2.3.3 identified K as being 3, with the highest modal peak at 53.66, although a K of four was also likely with a modal peak of 51.91 (Figure

3.9). The majority of genes for each individual were from a single population, but in some cases there were more equal genetic contributions from two or more populations.

STRUCTURE v2.3.3 also showed the flow of genes into populations, for example in

Dickson’s Hole (MO), a single individual from the MP/MQ cluster (sites which are some distance east of MO) has dispersed to join the MO population.

Chapter 3.0 ∙ Rainbowfish 102

Magnitude of K For M. s. inornata

ΔK K 30

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 K

Figure 3.9 Magnitude of ΔK for each estimation of K (1 – 20) for M. s. inornata.

Figure 3.10 Proportions of membership of M. s. inornata individuals to each of the three groups determined by STRUCTURE v2.3.3.

In GENECLASS2 v2.0 for M. australis, the majority of individuals were assigned to populations within three regions: 1) DJJ, DEE and DA – assigned 30.36% of individuals, 2) DM, DFF, DR and DL – assigned 39.88% of individuals, and 3) DE – assigned 5.36% of individuals. For M. australis, the assignment of most individuals could be grouped as a lower catchment (sites DJJ, DEE and DA), midcatchment (sites

DM, DFF, DR and DL) or upper catchment (site DE) region (Table 3.7). Eight

Chapter 3.0 ∙ Rainbowfish 103 individuals could not be assigned to any population. This result was not unexpected, as there were significant sampling gaps between some sites. So it is highly likely that the population of origin for these individuals was not represented in the data set. For five sites (DA, DGG, DJ, DLL and DX), no individuals collected were assigned back to that population.

Similarly, M. s. inornata individuals were assigned to populations within three regions:

1) MN, MG and ME – assigned 63.98% of individuals, 2) MR – assigned 6.45% of individuals, and 3) MK – assigned 10.75% of individuals. Overall, there was a midcatchment population (sites MN, MG and ME), an upper fork population (site MR) and a lower fork population (site MK). Again for seven sites (MB, MC, MF, MH, MI,

ML and MO), no individuals were assigned to these regions.

GENECLASS2 v2.0 identified self-assignment for a number of individuals (i.e. the individual was assigned back to the site from which it was collected). Within the Daly

River, there were low to high rates of self-assignment at sites – ranging between 0.00 to

80.00% (Table 3.7). The highest rates of self-assignment were observed at three sites in the Daly River – Upstream Centipede Dreaming and Snowdrop Creek (site DE), UF

Daly River Road (site DEE) and the Fergusson River (site DFF) which was also reflected by the STRUCTURE v2.3.3 analysis (Figure 3.8). For the Mitchell River, less self-assignment was observed, but the values were still low to moderate – ranging between 0.00 to 38.89% (Table 3.8). In this catchment, the highest rates of self- assignment were observed in the Fish River (site MJ), Saltwater Creek (MK) and

Kingfish Lagoon (MN). Another notable feature of the assignment testing was that for some sites, none of the individuals were assigned back to the population from which they were collected (i.e. no self-assignment). Furthermore, for M. australis in the Fish

Chapter 3.0 ∙ Rainbowfish 104

River (site DJ), and M. s. inornata in the Mitchell River @ Gamboola (site MC), Walsh

River (site ML) and Dickson’s Hole (site MO), none of the individuals collected in the entire data set were assigned back to these sites. This result may be a reflection of the unequal sample sizes giving a heavier probability ‘weight’ to some sites over others

(e.g. assignment to site MK may be more likely because of the higher number of individuals).

Chapter 3.0 ∙ Rainbowfish 105

Table 3.7 Assignment of M. australis individuals to populations using GENECLASS2 v2.0. Putative populations from which

individuals could not be assigned back to any of the represented populations and denoted by + (for one individual) and * (for two

individuals). ^ denotes lower catchment groups, ° denotes midcatchment groups and ˉ denotes upper catchment groups.

M. australis Assigned Population % Correctly DA DB DE DEE DFF DGG DJ DJJ DKK DL DLL DM DP DR DS DT DX Assigned DA 2 1 1 1 1 1 1 0.00+ DB 1 2 2 2 1 25.00 DE 5 1 1 1 1 55.56 DEE 1 8 3 1 61.54 DFF 1 12 1 1 80.00+ DGG 1 2 0.00 DJ 1 1 2 1 0.00 DJJ 2 1 1 1 1 1 14.29 DKK 3 3 50.00* +

Putative Population Putative DL 2 2 6 5 2 1 33.33

DLL 1 1 4 0.00+ DM 2 2 7 5 2 1 1 25.00 DP 2 1 1 1 1 1 14.29 +

M. australis M. DR 2 1 2 2 1 25.00 DS 1 1 2 1 2 6 46.15 DT 1 2 3 2 1 1 10.00 DX 3 1 0.00 % Total 11.31^ 1.79 5.36 10.71^ 7.14° 0.60 0.00 8.33^ 1.79 13.69° 2.38 13.10° 1.19 5.95° 4.76 2.98 4.17 Population

Chapter 3.0 ∙ Rainbowfish 106

Table 3.8 Assignment of M. s. inornata individuals to populations using GENECLASS2 v2.0. ^ denotes lower fork groups, ° denotes

midcatchment groups and ˉ denotes upper fork groups.

M. s. inornata Assigned Population % Correctly MB MC ME MF MG MH MI MJ MK ML MM MN MO MP MQ MR Assigned MB 3 1 4 0.00 MC 1 1 0.00 ME 5 1 1 5 8 25.00 MF 1 0.00 MG 2 1 4 7 28.57 MH 2 0.00 MI 1 0.00 MJ 1 1 3 1 2 37.50

Putative Population Putative MK 1 2 2 1 1 1 9 6 1 1 36.00

ML 3 1 2 1 0.00 MM 1 1 3 2 3 10 15.00 MN 3 1 3 1 3 7 38.89 MO 2 3 3 6 0.00

M. s. inornata s. M. MP 4 1 3 3 1 8.30 MQ 1 6 3 3 6 1 5.00 MR 7 1 1 4 1 7.14 % Total 0.50 0.00 14.50° 4.80 16.70° 0.50 3.20 3.20 10.80^ 0.00 4.80 32.80° 0.00 0.50 1.10 6.50ˉ Population

Chapter 3.0 ∙ Rainbowfish 107

3.3.4 Isolation indices

For each isolation index based on mitochondrial DNA and microsatellites, midreach populations were the least isolated and upstream populations were the most isolated, with downstream populations being intermediate between the two (Figure 3.11 and

3.12). Populations within each of the three catchment positions were statistically differentiated from each other, based on mitochondrial DNA but not on microsatellites

(Table 3.9). Scheffé pairwise comparisons showed this difference to be between midreach and upstream populations for both M. australis and M. s. inornata (Table

3.10). Spearman’s rank correlations between distance from river mouth and the fixation value were non-significant in all cases (Table 3.12).

Chapter 3.0 ∙ Rainbowfish 108

Table 3.9 ANOVA results for comparisons of isolation indices for each of the waterway classifications in M. australis and M. s. inornata. Significant values are denoted by *, where α = 0.05.

Species Isolation Index Type Source SS df MS F P

mtDNA FST Between catchment positions 0.149 2 0.075 4.626 0.029*

M. australis mtDNA ΦST Between catchment positions 0.138 2 0.069 4.713 0.027*

msats FST Between catchment positions 0.004 2 0.002 0.300 0.745

mtDNA FST Between catchment positions 0.222 2 0.111 5.953 0.015*

M. s. inornata mtDNA ΦST Between catchment positions 0.304 2 0.152 9.008 0.004*

msats FST Between catchment positions 0.019 2 0.009 3.014 0.084

mtDNA FST Between waterway types 0.135 2 0.068 3.953 0.044*

M. australis mtDNA ΦST Between waterway types 0.125 2 0.062 4.001 0.042*

msats FST Between waterway types 0.017 2 0.008 1.689 0.220

mtDNA FST Between waterway types 0.220 2 0.110 5.822 0.016*

M. s. inornata mtDNA ΦST Between waterway types 0.252 2 0.126 6.013 0.014*

msats FST Between waterway types 0.020 2 0.010 3.347 0.067

Chapter 3.0 ∙ Rainbowfish 109

Table 3.10 Post-hoc Scheffé multiple comparison tests between means for catchment positions. * denotes significant values where α =

0.05.

Species Waterway Classification Marker Pairwise Comparison df MS P

M. australis Catchment position mtDNA FST Midreach/downstream 14 0.016 0.136

M. australis Catchment position mtDNA FST Midreach/upstream 14 0.016 0.048*

M. australis Catchment position mtDNA FST Downstream/upstream 14 0.016 0.999

M. australis Catchment position mtDNA ΦST Midreach/downstream 14 0.015 0.212

M. australis Catchment position mtDNA ΦST Midreach/upstream 14 0.015 0.035*

M. australis Catchment position mtDNA ΦST Downstream/upstream 14 0.015 0.903

M. australis Catchment position msats FST Midreach/downstream 14 0.006 0.938

M. australis Catchment position msats FST Midreach/upstream 14 0.006 0.873

M. australis Catchment position msats FST Downstream/upstream 14 0.006 0.763

M. s. inornata Catchment position mtDNA FST Midreach/downstream 13 0.019 0.154

M. s. inornata Catchment position mtDNA FST Midreach/upstream 13 0.019 0.015*

M. s. inornata Catchment position mtDNA FST Downstream/upstream 13 0.019 0.632

M. s. inornata Catchment position mtDNA ΦST Midreach/downstream 13 0.017 0.154

M. s. inornata Catchment position mtDNA ΦST Midreach/upstream 13 0.017 0.004*

M. s. inornata Catchment position mtDNA ΦST Downstream/upstream 13 0.017 0.257

M. s. inornata Catchment position msats FST Midreach/downstream 13 0.003 0.410

M. s. inornata Catchment position msats FST Midreach/upstream 13 0.003 0.085

M. s. inornata Catchment position msats FST Downstream/upstream 13 0.003 0.727

Chapter 3.0 ∙ Rainbowfish 110

Table 3.11 Post-hoc Scheffé multiple comparison tests between means for waterway types. * denotes significant values where α = 0.05.

Species Waterway Classification Marker Pairwise Comparison df MS P

M. australis Waterway type mtDNA FST Main channel/tributary 14 0.017 0.067

M. australis Waterway type mtDNA FST Main channel/waterhole 14 0.017 0.113

M. australis Waterway type mtDNA FST Tributary/waterhole 14 0.017 0.839

M. australis Waterway type mtDNA ΦST Main channel/tributary 14 0.016 0.054

M. australis Waterway type mtDNA ΦST Main channel/waterhole 14 0.016 0.150

M. australis Waterway type mtDNA ΦST Tributary/waterhole 14 0.016 0.954

M. australis Waterway type msats FST Main channel/tributary 14 0.005 0.236

M. australis Waterway type msats FST Main channel/waterhole 14 0.005 0.871

M. australis Waterway type msats FST Tributary/waterhole 14 0.005 0.750

M. s. inornata Waterway type mtDNA FST Main channel/tributary 13 0.019 0.179

M. s. inornata Waterway type mtDNA FST Main channel/waterhole 13 0.019 0.549

M. s. inornata Waterway type mtDNA FST Tributary/waterhole 13 0.019 0.020*

M. s. inornata Waterway type mtDNA ΦST Main channel/tributary 13 0.021 0.197

M. s. inornata Waterway type mtDNA ΦST Main channel/waterhole 13 0.021 0.480

M. s. inornata Waterway type mtDNA ΦST Tributary/waterhole 13 0.021 0.017*

M. s. inornata Waterway type msats FST Main channel/tributary 13 0.003 0.622

M. s. inornata Waterway type msats FST Main channel/waterhole 13 0.003 0.412

M. s. inornata Waterway type msats FST Tributary/waterhole 13 0.003 0.068

Chapter 3.0 ∙ Rainbowfish 111

Table 3.12 Spearman’s rank correlations between distance from river mouth and average fixation index (isolation index) for M. australis and M. s. inornata (α =

0.05).

M. australis ρ P

mtDNA FST 0.093 0.722

mtDNA ΦST 0.174 0.504

msats FST -0.339 0.184

M. s. inornata ρ P

mtDNA FST 0.291 0.274

mtDNA ΦST 0.450 0.080

msats FST 0.215 0.425

Isolating Influence of Catchment Position (M. australis )

0.6

0.5 Downstream Midreach Upstream 0.4 0 e d 0.3 f g i Fixation Value 0.2 h ZERO

0.1

0 FST Msats FSTjkfldsja mtDNA ΦST mtDNA

Chapter 3.0 ∙ Rainbowfish 112

Isolating Influence of Catchment Position(M. s. inornata )

0.6 Downstream Midreach 0.5 Upstream Series4 Series6 0.4 Series5 Series7 Series8 0.3 Series10 Series9 Series11 Fixation Value 0.2

0.1

FST Msats FST fdsfmtDNA ΦST mtDNA

Figure 3.12 Isolation indices for all sampled populations of M. s. inornata based on a) mean FST for ATPase 6 and 8, b) mean ΦST for ATPase 6 and 8, and c) mean FST for microsatellites. Populations are classed as being downstream (blue), midreach

(light green) or upstream (peach) and show 95% confidence intervals.

When populations were classed as being 1) a main channel population, 2) a tributary population, or 3) a waterhole population (as described in Chapter 3.2.1), there were inconsistent patterns across the catchments (Figures 3.13 and 3.14). For waterhole populations, the pattern of isolation differed between the two catchments. In the Daly

River, the isolation indices for waterholes sat within the range of the tributary populations for values based on mtDNA (Figure 3.13). For the microsatellite data set, the relative rank of waterholes was reduced from the relatively moderate/high isolation, to low/moderate isolation, and was lower than the ‘tributary’ values. In the Mitchell

River, waterhole populations were the most connected sites of the catchment for isolation indices based on both mitochondrial and microsatellite data, and were consistently differentiated from main channel and tributary populations. Tributary populations were generally the most isolated populations, with the exception of values based on microsatellites. Significant differences were determined between waterway

Chapter 3.0 ∙ Rainbowfish 113

types, although microsatellite FST values were again non-significant (Table 3.9). Scheffé analyses showed significant differences between waterholes and tributaries for M. s. inornata only (Table 3.11).

Isolating Influence of Waterway Type (M. australis )

0.7 Main Channel 0.6 Tributary Waterhole Series4 0.5 Series5 Series6 Series7 0.4 Series8 Series9 0.3 Series10

Series11 Fixation Value 0.2

0.1

0 FST Msats FST gdsamtDNA ΦST mtDNA

Chapter 3.0 ∙ Rainbowfish 114

Isolating Influence of Waterway Type (M. s. inornata )

0.6 Main Channel Tributary 0.5 Waterhole Series4 Series5 0.4 Series6 Series7 Series8 0.3 Series9 Series10 Series11 Fixation Value 0.2

0.1

0 FST Msats FSTfdsafd mtDNA ΦST mtDNA

3.3.5 Mantel tests

Mantel tests were significant for M. s. inornata in the Mitchell River, but non- significant for M. australis in the Daly River (Table 3.13). This result suggested that there was a positive correlation between geographic and genetic distance in M. s. inornata only.

Table 3.13 Mantel test correlation coefficients for M. australis and M. s. inornata (α

= 0.05). * denotes significant values.

Mantel Test Correlation Coefficient P M. australis 0.104 0.230 M. s. inornata 0.284* 0.021

Chapter 3.0 ∙ Rainbowfish 115

3.3.6 Regression analyses

Decomposed pairwise regression analyses identified significant, positive correlations

2 between distance and ATPase 6 and 8 FST estimates for both M. australis (R = 0.839, P

= 0.000) (Figure 3.16) and M. s. inornata (two models: R2 = 0.887, P = 0.000 (Model

A); R2 = 0.926, P = 0.000 (Model B)) (Figures 3.17 and 3.18). There was no significant correlation between isolation and elevation for M. australis (R2 = 0.029, P = 0.614

(Model C); R2 = 0.070, P = 0.459 (Model D)) (Figures 3.19 and 3.20); alternatively, a strong correlation was observed for M. s. inornata (R2 = 0.836, P = 0.000 (Model E); R2

= 0.625, P = 0.004 (Model F)) (Figures 3.21 and 3.22).

0.5 0.4 0.3 0.2 0.1 0 -0.1 -0.2

Mean Regression Residuals Regression Mean -0.3 -0.4 -0.5

DA DE DL DP DR DS DT DM DFF DKK DLL Population

Figure 3.15 Mean regression residuals and 95% confidence intervals for decomposed regression analysis. Example taken from correlations with distance for M. australis.

Chapter 3.0 ∙ Rainbowfish 116

DPR - M. australis

0.6

0.5

0.4 R2 = 0.839

0.3 ST F 0.2

0.1

-0.1 0 50 100 150 200 250 300 350 400 450 Distance (km)

DPR - M. s. inornata Model A

0.5

0.4

0.3 R2 = 0.887

ST 0.2 F

0.1

-0.1 0 100 200 300 400 500 600 Distance (km)

Chapter 3.0 ∙ Rainbowfish 117

DPR - M. s. inornata Model B

0.4 0.35 R2 = 0.926 0.3 0.25 0.2

ST 0.15 F 0.1 0.05 0 -0.05 -0.1 0 50 100 150 200 250 300 350 400 Distance (km)

Regression - M. australis Model C

0.6

0.5 R2 = 0.029

0.4

0.3

Isolation Index Isolation 0.2

0.1

0 0 50 100 150 200 250 300 350 400 Elevation (m)

Chapter 3.0 ∙ Rainbowfish 118

Regression - M. australis Model D

0.6

0.5 R2 = 0.070 0.4

0.3

Isolation Index Isolation 0.2

0.1

0 0 50 100 150 200 250 300 350 400 Elevation (m)

Regression - M. s. inornata Model E

0.7

0.6 R2 = 0.836 0.5

0.4

0.3

Isolation Index Isolation 0.2

0.1

0 0 100 200 300 400 500 600 Elevation (m)

Chapter 3.0 ∙ Rainbowfish 119

Regression - M. s. inornata Model F

0.7

0.6 R2 = 0.625 0.5

0.4

0.3

Isolation Index Isolation 0.2

0.1

0 0 100 200 300 400 500 600 Elevation (m)

The best model correlating distance with FST in M. australis included six populations, and identified five populations as outliers (Table 3.14). For M. s. inornata, two models

(Models A and B) were equally likely, and included eight and seven populations respectively. Several populations were identified as being true outliers. Models correlating elevation with isolation included all the M. australis populations (Model C), or highlighted only site DLL as an outlier population (Model D). Similarly, the best isolation-elevation model for M. s. inornata included either all the populations (Model

F), or found site MP as a true outlier (Model E). Individual pairwise regressions for all populations identified all of the patterns described by Koizumi et al. (2006) (Table

3.15).

Chapter 3.0 ∙ Rainbowfish 120

2 Populations Removed n R AICC ΔAICC DFF, DKK, DS, DE, DP 6 0.839 -25.111 0.000 DFF, DKK, DS, DE, DP, DL 5 0.820 -21.225 3.885 DFF, DKK, DS, DE 7 0.716 -20.583 4.528 DFF, DKK, DS, DE, DP, DL, DM 4 0.889 -19.781 5.330 M. australis DFF, DKK, DS 8 0.301 -15.502 9.609 DFF, DKK 9 0.192 -12.547 12.563 DISTANCE DFF 10 0.118 -9.179 15.932 None 11 0.086 -7.116 17.995 MK, MO, MQ 8 0.887 -40.006 0.000 MK, MO, MQ, MJ 7 0.926 -39.043 0.964 M. s. inornata MK, MO 9 0.666 -22.314 17.692 MK 10 0.567 -19.804 59.811 None 11 0.567 -18.451 21.556 None 11 0.029 -42.599 0.000 M. australis DLL 10 0.070 -41.522 1.077 ELEVATION MP 10 0.836 -51.518 0.000 M. s. inornata None 11 0.625 -49.833 1.685

Chapter 3.0 ∙ Rainbowfish 121

M. australis - Distance

0.8

0.6

ST 0.4 F

0.2

0 0 100 200 300 400 500 600 700

-0.2 Distance (km)

Figure 3.23 Decomposed pairwise regressions between FST and distance for M. australis in the Daly River. ‘True outliers’ are represented by dashed lines.

M. s. inornata - Distance

0.8

0.6

ST 0.4 F

0.2

0 0 100 200 300 400 500 600 700

-0.2 Distance (km)

Figure 3.24 Decomposed pairwise regressions between FST and distance for M. s. inornata in the Mitchell River. ‘True outliers’ are represented by dashed lines.

Chapter 3.0 ∙ Rainbowfish 122

Table 3.15 Intercepts, slopes and regression coefficients for all tested populations of M. australis and M. s. inornata. ‘True outliers’ are shaded grey.

Intercept P Slope P Population R2 Pattern (*10-2) (intercept) (*10-2) (slope) DA -0.100 0.979 0.100 0.128 0.593 4 DM -5.400 0.484 0.200 0.037 0.812 3 DL -15.500 0.026 0.200 0.002 0.970 2 DLL -20.000 0.229 0.200 0.028 0.841 3 DR -7.600 0.102 0.100 0.007 0.934 3 DT -0.900 0.859 0.100 0.058 0.750 4 DS 25.000 0.001 29.300 0.178 0.137 1 DP 27.700 0.339 0.100 0.435 0.078 4 DKK 28.500 0.208 0.100 0.234 0.172 4 DE -42.700 0.330 0.200 0.091 0.315 4 DFF 41.900 0.125 0.100 0.508 0.057 4 MB -5.600 0.045 0.100 0.002 0.927 1 ME -4.800 0.013 0.100 0.000 0.970 2 MG -4.600 0.092 0.100 0.002 0.937 3 MM -1.300 0.425 0.100 0.001 0.957 3 MN 0.100 0.949 0.100 0.002 0.933 3 MP -3.600 0.755 0.100 0.059 0.631 4 MR -4.600 0.352 0.100 0.005 0.887 3 MJ -36.000 0.046 0.200 0.001 0.754 2 MK 8.700 0.378 0.200 0.000 0.812 3 MO 3.400 0.685 0.200 0.000 0.806 3 MQ 30.900 0.140 0.100 0.093 0.313 4

Chapter 3.0 ∙ Rainbowfish 123

3.4 Discussion

3.4.1 Genetic diversity

High levels of genetic diversity were found in both M. australis and M. s. inornata, with 21 and 27 unique haplotypes captured within species respectively. This finding was in keeping with the braod pattern of high genetic diversity in tropical rivers (see Sivasundar et al.

2001; So et al. 2006(a)). Divergence between these haplotypes was mostly the result of a single base pair change from a common ancestral haplotype, which resulted in star-like arrangements in some parts of the phylogeny. Importantly, none of the haplotypes for either

M. australis or M. s. inornata were shared across all the sites, which indicates that there is some restriction to gene flow (as was hypothesised), thereby allowing genetic drift to randomly determine the genes conserved within each population.

While overall genetic diversity was high in both catchments, the lower values observed at some sites could be the product of several processes. Firstly, indirect selection can occur, where regions linked to functional genes (that are under direct selection) are also subjected to fixation through genetic-hitchhiking (Ballard & Kreitman 1995; Barton 2000; Galtier et al. 2009). Secondly, population bottlenecks or founder events can result in low genetic diversity, as the small numbers of surviving/establishing individuals carry a limited amount of genetic diversity (Templeton 1980). Further losses of genetic diversity are incurred through genetic drift (Amos & Balmford 2001; Dlugosch & Parker 2008). Lastly, genetic diversity is correlated with Ne, so populations with a small Ne are expected to have correspondingly low genetic diversity (Turner et al. 2006). Neutrality tests identified selection in M. australis and a change in demographic size in M. s. inornata as the potential

Chapter 3.0 ∙ Rainbowfish 124 causes of low genetic diversity. These interpretations were based on the expectation that populations under selection will have significant Fu and Li’s F* and D* values, and non- significant Fu’s FS values – as was observed in M. australis (Fu 1997). The reverse is indicative of a demographic expansion, as was found in M. s. inornata.

3.4.2 Population expansions and selection

For both M. australis and M. s. inornata, parts of the haplotype networks were arranged in star-like patterns, which can be indicative of a recent demographic change, where the population has undergone a sudden expansion (Slatkin & Hudson 1991; Harpending et al.

1998). The frequency of singletons (haplotypes which were observed in only a single individual) is a particularly recognisable feature of a recent population expansion; but over time, singletons should become extinct by chance through genetic drift (Posada & Crandall

2001). The melanotaeniid haplotype networks show a high proportion of singletons and other low frequency haplotypes, which indicates that the number of generations (that is, the opportunities for haplotypes to be removed through genetic drift) since the population expansion has been relatively few, and that the expansion is relatively recent.

Neutrality tests for populations of M. s. inornata in the Mitchell River provided evidence for a population expansion, with the exclusion of R2, which was non-significant. This results was surprising, as the R2 statistic is thought to perform better (compared to other statistics) with small sample sizes (n=10) (Ramos-Onsins & Rozas 2002). Alternatively, for

M. australis significant values of Fu and Li’s F* and D* indicated that this population was

(and is) under selection. Paradoxically, r values indicated the reverse – that M. australis has undergone an expansion, and M. s. inornata has not. However, given that the power of r is

Chapter 3.0 ∙ Rainbowfish 125 not particularly strong, it is likely that this test was too conservative to detect the signature of a population expansion in the mismatch distributions (Ramos-Onsins & Rozas 2002).

Importantly, given that neutrality tests by population were all non-significant, this suggests that the overall deviations from neutrality may simply be attributable to the high levels of genetic structure.

Mitochondrial markers are widely used in population genetic studies, as their lineages of descent are preserved due to the lack of recombination. However, it is a genome that is influenced by selection, particularly due to the high rates of deleterious mutations

(Nachman 1998; Meiklejohn et al. 2007). The impacts of selection can be problematic when attempting to evaluate demographic history, as population size is expected to correspond with genetic diversity in the absence of selection. As a result, the estimates of population size may not be accurately represented if the mitochondrial genome has undergone selection (either positive or negative). In larger populations, selection is expected to be more influential than genetic drift in determining the genetic diversity of the population. Given the abundance of M. australis, it is not surprising that selection seems to be overriding the effects of genetic drift (Allendorf 1983).

3.4.3 Population structure

As predicted, high levels of global genetic structure were observed within both M. australis and M. s. inornata, although FST indices based on the microsatellite loci were weaker, yet still significant (0.132 and 0.128 respectively). The genetic structure was considerably higher than was detected by Phillips et al. (2009) in M. australis in the Kimberley region of

Western Australia (using allozyme variation). The haplotype frequencies at sites showed

Chapter 3.0 ∙ Rainbowfish 126 clear genetic structure across the landscape. This pattern immediately suggests that the contemporary rates of gene flow are restricted, thus contributing to the genetic structure within these populations. While populations within the catchments may be physically connected, especially during the wet season, it is possible that biological or physical barriers are inhibiting dispersal in these species. As predicted by the Stream Hierarchy

Model (SHM) (Meffe & Vrijenhoek 1988), populations within subcatchments had consistently lower fixation indices than those among subcatchments, with the single exception of the FST for microsatellites in M. australis (although the value was still very low – 0.069). These fixation values suggest that dispersing individuals are more likely to stay within a subcatchment, than to move into neighbouring subcatchments – a pattern that was also detected in a previous study of M. australis in Western Australian catchments

(Phillips et al. 2009). Overall, this finding lends support to the prediction that the dendritic structure of a catchment would influence the connectivity of populations.

STRUCTURE v2.3.3 estimated there to be four populations of M. australis in the Daly

River, and three populations of M. s. inornata in the Mitchell River. GENECLASS2 v2.0 results further supported this by showing how the majority of individuals were assigned back to populations within three distinct regions. Additionally, many of the individual population pairwise regressions were catergorised as following Pattern 4 (as described by

Koizumi et al. 2006). In Pattern 4 populations, the effects of gene flow are stronger than that of genetic drift – indicating that these populations are ‘sinks’, and are not differentiated from other populations (the sources). Overall, these results suggest that there is a source- sink situation, where sites (sinks) are being constantly colonised by several ‘source’ populations. Further to this finding was the self-assignment of individuals, which indicated how some individuals were philopatric and recruited back to their population (self-

Chapter 3.0 ∙ Rainbowfish 127 recruitment). However, differences in sample sizes may have contributed to the over- assignment of individuals to certain sites. The higher representation of alleles at sites with higher numbers of individuals would increase the likelihood of assignment back to those populations. While the Paetkau et al. (2004) algorithm incorporates variance in sample sizes into the assignment of individuals, it may still be that biases have led to the over- rejection of individuals to their true populations (Piry et al. 2004).

Interestingly, some individuals were assigned to populations that were a great distance away from their point of collection. For example, an M. australis individual collected at

Beeboom (site DA) was assigned back to the population at Galloping Jack’s (site DL) – a distance of some 230km. Given the strong genetic structure detected for these species, a literal interpretation of occasional long-distance dispersal may not be appropriate.

However, a potential explanation for this apparent discrepancy could be that individuals are dispersing through to new populations; yet these migrants are not reproducing with the resident population and thus, are not contributing to gene flow. Consequently, the genetic structure of the populations is maintained, even through the high dispersal rate of new individuals. So while the physical connectivity between populations may be greater than was hypothesised, the genetic connectivity is certainly restricted.

An alternative explanation is that sex-biased dispersal is occurring, where the females are more philopatric than the males. Sex-biased dispersal can be inferred when the overall population structure (FST) for a uniparentally inherited marker is higher or lower than that of a biparentally inherited marker (Prugnolle & de Meeus 2002). As mtDNA markers are maternally inherited, the Ne is smaller, and therefore more susceptible to genetic drift.

Consequently, it is expected that population differentiation would be higher for a

Chapter 3.0 ∙ Rainbowfish 128 mitochondrial marker, even with equal dispersal of both sexes. However, comparisons between uniparentally and biparentally inherited marker FST values can reveal sex-biased dispersal. For mitochondrial DNA (a maternally inherited genome), a higher FST value

(when compared to the biparentally inherited microsatellites) would indicate that the females are philopatric, with males dispersing (Goudet et al. 2002; Lawson Handley &

Perrin 2007). Male-biased dispersal could account for the lower FST for microsatellites in the melanotaenids, as ATPase 6 and 8 would only reflect female gene flow. This rationale has been applied to many studies where sex of the individuals was not determined as part of the study design. For example, studies of big eye tuna (Thunnus obesus), Patagonian toothfish (Dissostichus eleginoides) and black bream (Acanthopagrus butcherii) have all inferred sex-biased dispersal from the discrepancy between nuclear and mitochondrial markers (Durand et al. 2005; Shaw et al. 2004; Burridge & Versace 2007). Unfortunately, it is not possible to confirm the influence of sex-biased dispersal in this study, as demographic characteristics were not assessed. So while sex-biased dispersal is possible, it is also likely that the discordance between fixation indices can be attributed to the influences of genetic drift on smaller Ne. Additionally, the difference in FST values between the two markers is not adequately differentiated to conclusively infer sex-biased dispersal.

3.4.4 Influence of the landscape on connectivity

Within both the Daly and Mitchell catchments, genetic variation was clearly partitioned across the landscape, as was hypothesised. Further investigations of the landscape influences on gene flow revealed several key factors that had varying levels of impact.

First, dendritic structure was found to influence the dispersal of fish. The use of isolation indices identified patterns based on the position in catchment, where the mitochondrial

Chapter 3.0 ∙ Rainbowfish 129 isolation indices show a trend where the midreach populations were the least isolated, the upstream populations were the most isolated, and the downstream populations intermediate were between the two. As predicted, upstream populations were the most isolated; however, it was the midreach populations that showed the greatest connectivity, not the downstream populations, as was expected. This finding suggests that the periods of high flow in these tropical catchments are not transporting individuals (be it eggs, larvae, juveniles or adults) downstream during the wet season. However, microsatellite loci isolation indices did not show any clear grouping of catchment position categories. As discussed previously, this may be due to sex-biased dispersal, where the females are philopatric, or simply the effects of genetic drift. Overall, these results support the hypothesis that dendritic structure would influence connectivity, but the outcome of this influence was unanticipated.

Contrary to expectation, tributaries were the most isolated waterway types for M. s. inornata, not waterholes. In fact, for M. s. inornata, waterholes were the least isolated habitat type; however, this is not to say that all waterholes in the Mitchell River are extremely well-connected, or that waterholes are more connected than the river itself. It seems likely that the proximity of these specific waterholes to the main channel allows regular connection, and that in this instance, waterholes are ‘behaving’ like main channel environments. For the Daly River, waterholes were generally identified as being more isolated environments. This pattern was particularly true for UF Daly River Road (site

DEE), and this can be attributed to the complete disconnection of this site from the main channel. Generally, tributaries in the Daly River were moderately isolated, and the main channel the most connected for M. australis.

Chapter 3.0 ∙ Rainbowfish 130

The most isolated populations, identified by the isolation index, were those populations with higher self recruitment. For M. australis, populations from UF Daly River Road (site

DEE), Upstream Centipede Dreaming and Snowdrop Creek (site DE) and the Fergusson

River (site DFF) had the highest rates of self-recruitment, and were also some of the most isolated populations. The Fergusson River site in particular was consistently highlighted as the most isolated site, and also had the highest rate of self-recruitment (80%). This site was also the first outlier identified through the regression models, although it was identified as a

‘sink’ population in which IBD was not in effect. In the Mitchell River, the highest rates of self-recruitment were in the Fish River (site MJ), Saltwater Creek (site MK) and the

Mitchell River @ Gamboola (site MN). However, these sites were not particularly disconnected, based on isolation indices. The connectivity was reflected in the lower rates of self-recruitment (the highest being 38.89%). Consequently, it is unlikely that the self- recruitment is a product of the isolation of the sites, but is perhaps more due to these populations being ‘sources’. The decomposed pairwise regressions did highlight three populations (DL, ME and MJ) that followed Pattern 2, which establishes in large populations when there is a barrier to gene flow. It may be that there are barriers preventing dispersal at these sites, but the exact nature of these barriers is unclear.

3.4.5 The influence of elevation and distance on isolation

It was expected that IBD would be a key explanation for the diffusion of alleles across the landscape. Overall, it seems that for M. s. inornata, the landscape in the Mitchell catchment is highly influential in determining patterns of gene flow between populations. By comparison, it appears that population connectivity for M. australis in the Daly River is dependent on distance, rather than elevation, although it was hypothesised that both

Chapter 3.0 ∙ Rainbowfish 131 parameters would be influential. Two populations were found to follow Pattern 1 (Koizumi et al. 2006), in which IBD is not established due to either a bottleneck/founder event, or complete isolation. Given the position of these populations within the catchment, it seems unlikely that they are completely isolated; a bottleneck or founder event may be a more rational explanation for the lack of IBD. While a Mantel test only detected IBD in M. s. inornata, it may be that this evaluation was too conservative to identify any relationship in

M. australis. IBD correlations were also detected in a study of M. australis (using allozyme data) in Western Australia, which supports the expectation and inference of such a relationship in the Daly River (Phillips et al. 2009). Distance and elevation seemed to inhibit dispersal in M. s. inornata in many sites tested in the DPR. Additionally, individual regressions found that for many populations, ‘genetic drift – gene flow equilibrium’ had been established (Pattern 3) (Koizumi et al. 2006). This finding lends further support to the conclusion of IBD, as equilibrium must be reached in order for IBD to establish (Wright

1943; Slatkin & Maddison 1990; Robledo-Arnuncio & Rousset 2010).

Chapter 3.0 ∙ Rainbowfish 132

3.5 Conclusion

This study aimed to determine how distance, elevation and dendritic characteristics impact on the isolation of populations of M. australis in the Daly River, and M. s. inornata in the

Mitchell River. It was predicted that distance, elevation and waterway type would have positive correlations with the isolation of populations. Additionally, it was expected that the catchments’ dendritic strucutre would impact the isolation of populations. Overall, this study found that M. australis and M. s. inornata were affected by the landscape and dendritic formation of the Daly and Mitchell Rivers, respectively. While distance was found to contribute to the genetic structure of both species, it was found that elevation had little influence on determining the genetic structure of M. australis. These findings support the hypotheses that connectivity between populations of rainbowfish is limited by the landscape and dendritic structure of catchments.

This study provides information that needs to be considered in the management of the Daly and Mitchell catchments. Primarily, the source populations in each catchment need to be carefully conserved by ensuring that the habitat at each of these source locations remains intact and undisturbed. This strategy will support the persistence of M. australis in the Daly

River, and M. s. inornata in the Mitchell River. Secondly, isolated populations that are particularly prone to the effects of genetic drift (i.e. Pattern 1 and 2 populations) are at greater risk of extinction. The low level of gene flow to these populations indicates that dispersal to these populations is a slow process, and recolonisation may be unlikely; but also, reductions in genetic diversity ultimately impair the resilience of these populations to perturbation.

Chapter 3.0 ∙ Rainbowfish 133

4.0 Does the landscape affect strong dispersers? Investigating riverine

influences on gene flow in sooty grunters (Hephaestus fuliginosus)

4.1 Introduction

The appropriate management and conservation of riverine environments and populations is a challenging task due to the complexity of freshwater systems (Sparks 1995; Arthington et al. 2010). Often, a ‘blanket’ management strategy is put in place, in the hope that a generalised plan will support as many species as possible (Collares-Pereira & Cowx 2004;

Saunders et al. 2002). Indeed, this is often the only approach that can be taken given the limited financial and staff resources available to catchment managers (Bottrill et al. 2008).

However, this ‘one size fits all’ approach is rarely optimal, and is likely to favour some species over others (Dudgeon 2005; Newson 2010). The different dispersal requirements of fish can greatly complicate the way in which a river should be managed (Nel et al. 2009).

Further to this is the way that artificial and natural changes to riverine systems can impact on species that require access to the ocean/river reaches, or that naturally disperse or migrate large distances within a river system (Bunn & Arthington 2002).

Human modification of natural river systems can have many immediate and ongoing consequences for fish populations (Fausch et al. 2002). While sedentary species (that experience little or no dispersal/gene flow) may not be especially impacted by the fragmentation of a river, species that disperse quite freely can be severely disrupted by artificial barriers (Jager et al. 2001; Nilsson et al. 2005). In managing freshwater fish populations, it is necessary to maintain the natural dispersal/migration regime to counteract

Chapter 4.0 ∙ Sooty Grunter 134 the effects of isolation – namely reduced genetic diversity, inbreeding and potentially population extirpation (Vrijenhoek 1998; Keller & Waller 2002; Knaepkens et al. 2004(a)).

Risk is introduced when a population is sub-divided, particularly if populations naturally experience high rates of gene flow over large distances (Templeton et al. 1990; Landguth et al. 2010). Many studies have demonstrated how the alteration of river connectivity has caused population fragmentation, a decline in genetic diversity and population extinction.

Such examples can be seen in white-spotted charr (Salvelinus leucomaenis) (Yamamoto et al. 2004), silvery minnow (Hybognathus amarus) (Aló & Turner 2005), Macquarie perch

(Macquaria australasica) (Faulks et al. 2011) and three-spined stickleback (Gasterosteus aculeatus) (Raeymaekers et al. 2009). The disturbance of population connectivity is demonstrably a serious issue that has a range of impacts that may be partly determined by dispersal requirements.

While the dispersal requirements of a species is a critical element of population persistence and conservation, it is further complicated by how a species utilises specific habitats (e.g. vegetative cover or substrates), and whether access to such habitats is necessary for the persistence of the population (Heinrichs et al. 2010). While access to either the ocean or rivers is necessary for diadromous species, there are also more subtle habitat requirements that are crucial for solely freshwater species (McDowall 1992; Labbe & Fausch 2000; Bond

& Lake 2003). In some instances, these habitats may be accessed infrequently, and may not be the usual habitat for the species – yet they are critical to survivorship (Mallen-Cooper

1993; Rosenfeld & Hatfield 2006). For instance, many fish species in Australia utilise inundated floodplains to feed, even in systems that flood infrequently and unpredictably

(Humphries et al. 1999; King et al. 2003; Arthington et al. 2005). Another example of specific habitat requirements is the way that common galaxids (Galaxias maculates)

Chapter 4.0 ∙ Sooty Grunter 135 require access to overhanging riparian vegetation to deposit eggs for terrestrial development (Hickford et al. 2010). For salmonids, gravel of a specific size and depth must be found to build spawning redds – again a habitat that is important, but intermittently accessed by the species (Barlaup et al. 2008; Louhi et al. 2008; Palm et al. 2009).

Preserving the availability and accessibility of required habitats should be an important management consideration, especially in the context of life histories. For species where a specific habitat is essential, the preservation of a large section of river may be futile if it does not include critical habitat.

Critical habitat can be highly influential in determining patterns of gene flow and genetic diversity in many aquatic habitat specialists. Micro- and meso-habitat interactions and dependency has been found to have far-reaching genetic implications on both fine and broad scale population connectivity. Faulks et al. (2010(a)) found that genetic diversity in

Macquarie perch (Macquaria australasica) was largely determined by the amount of riffle habitat available to the population, with more riffle habitat supporting a larger population.

The Oxleyan pygmy perch (Nannoperca oxleyana) is another such example where a reliance on a specific habitat (in this case low pH streams with heath vegetation) has led to isolation in suitable regions and a subsequent reduction in genetic diversity (Hughes et al.

1999). In these examples, habitat availability is a clear determinant in the genetic health of specialist species.

The sooty grunter (Hephaestus fuliginosus) is a species that requires two different types of habitat during its life cycle – both riffle habitat and pool habitat (Pusey et al. 2004). For juveniles, the nursery habitat is the numerous, clear, fast-flowing rocky riffles that are common in the tropical rivers of Australia (Stewart-Koster et al. 2011). In fact, in

Chapter 4.0 ∙ Sooty Grunter 136 predictive modelling estimates undertaken by Chan et al. (2010), this association with rocky riffle regions was proven to be essential in providing a nursery habitat for developing fry and juveniles. In contrast, the adults inhabit pools and runs (Pusey et al. 2004). This reliance on two differing habitats makes the management of this species more complex, as adequate access to both riffles and pools must be maintained (Chan et al. 2010; Faulks et al. 2010(a)). These important habitat relationships make this species vulnerable to the impacts of water extraction or environmental change that alters critical habitat availability and the dispersal and connectivity of populations (Preston & Jones 2008; Chan et al. 2010;

Stewart-Koster et al. 2011).

While riffles are essential habitat for sooty grunter, they can also represent both barriers and pathways for dispersal at different life stages. For juveniles, pools form barriers (due to increased risk of predation), and they are restricted to riffles for the first part of their juvenile life stages. For adults, riffles may be barriers to dispersal, as the depth may be insufficient to allow passage over them. The genetic ramifications of this relationship between habitat and dispersal are not understood for this species. It might be expected that more riffle habitat should support a larger population, and ultimately greater genetic diversity within that population (Hartl & Clark 1997; Knaepkens et al. 2004(b); Faulks et al. 2010(a)). Alternatively, greater riffle habitat may restrict the dispersal of adults, and impact on the overall levels of gene flow between populations. However, juveniles may transcend this barrier and be the main drivers of gene flow between populations. Clearly, this relationship between habitat and dispersal is a complex one which may have many different outcomes for gene flow and genetic diversity (Lowe et al. 2006).

Chapter 4.0 ∙ Sooty Grunter 137

The tropical rivers of Australia are currently under consideration for water extraction to support the agricultural industry in these catchments, but also to supplement the domestic water demands in the southern parts of the country (Hart 2004; Blanch 2008). Additionally, the longer term implications of climate change are predicted to alter the flow regime (due to changed timing, constancy and levels of rainfall) in Australia’s tropical rivers, and disturb many of the ecological balances in these areas (Hughes 2003; Hamilton & Gehrke 2005).

For H. fuliginosus, this has two implications in terms of its conservation. First, water extraction or reduced precipitation due to climate change may potentially diminish the suitable riffle habitat in the tropical catchments. By reducing the level and flow conditions of a river system, riffle nursery areas are likely to become inaccessible to spawning adults and their resulting offspring (Stewart-Koster et al. 2011). Chan et al. (2010) expressed concern that the loss of this essential habitat will dramatically reduce the population of H. fuliginosus in regulated rivers. Second, reduced hydrological connectivity may impact on the connectivity of populations of H. fuliginosus in these river systems, thus altering the natural levels of dispersal and gene flow in these regions.

This study investigated the dispersal of H. fuliginosus through the Daly and Mitchell

Rivers, with specific emphasis on determining how the landscape has affected the population genetics. By investigating the dispersal capability of this species using mtDNA, and identifying the landscape characteristics that have led to reduced population connectivity and patterns of gene flow, it was hoped that further inferences, predictions and suggestions could be made regarding the conservation for this species. The following results were expected:

Chapter 4.0 ∙ Sooty Grunter 138

1. The catchment landscape (specifically the dendritic structure, elevation, catchment

size and hydrological connectivity) would not influence connectivity in H.

fuliginosus; however,

2. Riffle habitat would be an important determinant in the isolation/connectivity and

genetic diversity of populations.

Chapter 4.0 ∙ Sooty Grunter 139

4.2 Detailed Methods

The following methods give specific detail that is additional to the general methods outlined in Chapter 2.0.

4.2.1 Calculations of Tau

Where there was evidence of population expansions from neutrality tests, the formula t =

τ/2µ was used to calculate the time elapsed since population expansion, where t is time in generations (for this species t is four years as sexual maturity is reached in the second year, and reproduction continues until death at ~6 years) and µ is the mutation rate, per sequence, per generation (Harpending et al. 1993). The best available estimate of mutation rate was

7.7% to 16.8% per million years, based on divergences in tasselfish (Polynemus sheridani)

(Chenoweth & Hughes 2003).

4.2.2 Classification of waterways

In keeping with the classification system used for the melanotaenids in Chapter 3.0, populations were again classed as being 1) downstream – that is, <170km from the river mouth, 2) midreach – that is 171-310km from the river mouth, or 3) upstream – that is

>311km from the river mouth. These groupings were made by dividing each catchment (i.e. the Daly and Mitchell catchments) into three, approximately equal lengths. Additionally, each population was identified as being 1) a tributary population – that is a population sampled in a connected tributary, 2) a main channel populations – that is a population that was sampled within the permanent delineations of the main channel, or 3) a waterhole

Chapter 4.0 ∙ Sooty Grunter 140 population – that is a population that was confined to a waterhole that may only experience connectivity during the wet season.

4.2.3 Isolation of waterways

Second, populations were classified as being either 1) main channel populations, 2) tributary populations, or 3) waterhole populations. Significant differences between these groups were determined using one-way analyses of variance (ANOVA). Spearman’s Rank

Correlations were used to examine the relationship between distance from river mouth (x) and the fixation value for the given site (y).

4.2.4 Testing landscape influences

A Mantel test was used to evaluate the relationship between genetic and geographic distance, using Slatkin’s linearised FST (where (FST =1/(1−FST)) (Slatkin 1995) as genetic distances, and river pathway between sites as geographic distance (see Tables 7.13 and

7.14). A decomposed regression analysis was used to identify any relationship between genetic isolation and several landscape characteristics. These were 1) the elevation of each population, 2) the distance between populations, and 3) the Potential Riffle Units (PRUs) between populations.

Chapter 4.0 ∙ Sooty Grunter 141

4.2.5 Identification of riffles

Attempts were made to identify the number of riffles between sites using satellite images from Google Earth (see http://www.google.com/earth/index.html). However, this was problematic as 1) the images for the catchment were a patchwork of high flow and low flow events (i.e. photographs from the wet season or dry season), which obscured the presence of riffles, 2) in places where the river channel was narrow, riparian vegetation obscured the view of the waterway, 3) the resolution was sub-optimal in some regions, and 4) satellite imagery would only reveal the white water from larger, more turbulent riffles, thus a visual inspection would not identify rolling riffles with unbroken water. Consequently, an alternative approach to identifying riffles was taken. Using ArcMAP v9.3 (ESRI 2006), a

Digital Elevation Model (DEM) profile was generated for the waterways between each site.

The DEM presented the elevation at intervals along the specified river bed. Potential riffles were identified by measuring the change in elevation between each DEM profile. The minimum slope at which a riffle can form is 0.0099 (Jowett 1993). This slope was verified using known riffle locations in the Daly and Mitchell Rivers. The elevation profile was also graphically represented for all sections of each catchment to ensure that the riffle estimations were plausible, as the data was ‘noisy’ (Figure 4.1). While this method did not include depth data, it was an attempt to identify the regions that may become riffles, or are currently riffles. Consequently, the measurement was a “potential riffle unit” (PRU).

Pairwise measurements of PRUs were calculated for all sites (as the cumulative distance of potential riffle habitat between sites) by determining the length of each DEM measurement and the change in elevation between each consecutive DEM measurement. The slope of the river surface was calculated as: Slope = change in elevation/DEM length. Positive or negative DEM slopes of 0.0099 or greater were classed as PRUs, depending on the model

Chapter 4.0 ∙ Sooty Grunter 142 e.g. Figure 4.1, where negative slopes represented the Douglas River, and positive slopes indicated the mouth of Middle Creek. PRUs were the total length (m) of river sections where the slope was adequate for the formation of a riffle.

Daly River Elevation Profile: Sites DS To DB

75 Convergence point 70 between the two river branches Douglas River (DS) Middle Creek (DB)

65 Elevation(m)

55 0 5000 10000 15000 20000 Distance (m)

Figure 4.1 Example of the elevation profile used to identify potential riffles in the river beds between sites DS to DB. Rectangles highlight the identified riffles.

Chapter 4.0 ∙ Sooty Grunter 143

4.3 Results

4.3.1 Intrapopulation genetic diversity

H. fuliginosus was collected from 23 sites in the Daly River (Figure 4.2 and Table 4.1) and

16 sites in the Mitchell River (Figure 4.3 and Table 4.2). In total, 255 individuals were captured.

Figure 4.2 The Daly River catchment, showing the 23 sites from which H. fuliginosus was collected.

Chapter 4.0 ∙ Sooty Grunter 144

Figure 4.3 The Mitchell River catchment, showing the 16 sites from which H. fuliginosus was collected.

The sequenced mitochondrial control region resulted in a 339bp fragment that showed appropriate levels of variation across both catchments. In the Daly catchment, 32 unique haplotypes from 207 individuals were observed (Figure 4.4). A notably divergent individual was observed in the Daly River, where a singleton from Oolloo Crossing

(Upstream) (site DT) is divergent from the rest of the sampled population by four extinct or unsampled haplotypes. For the Mitchell catchment, 12 unique haplotypes from 48 individuals were revealed (Figure 4.5). Sample sizes in this catchment were limited by the low abundance at the sampled sites. None of these haplotypes were shared between the two catchments (refer to Tables 7.12 and 7.13 for detailed haplotype distributions for H. fuliginosus). The haplotype network for the Daly catchment was indicative of a

Chapter 4.0 ∙ Sooty Grunter 145 demographic expansion, with many low frequency haplotypes surrounding common ancestral haplotypes (Figures 4.4 and 4.5). Gene diversity indices ranged between 0.600 and 1.000 for samples from the Daly River, and 0.000 and 1.000 for samples from the

Mitchell River.

Chapter 4.0 ∙ Sooty Grunter 146

Figure 4.4 The Daly River catchment, showing the geographic distribution of H. fuliginosus haplotypes. In the top diagram, circles represent each sampled population, with individual colours representing the proportions of haplotypes found within them. In the bottom diagram, circles represent each individual haplotype (frequency corresponds Chapter 4.0 ∙ Sooty Grunter 147 to circle size), with the individual colour corresponding to the haplotypes represented in mapped populations. Small black circles indicate haplotypes that were not sampled, or have become extinct through genetic drift.

Chapter 4.0 ∙ Sooty Grunter 148

to circle size), with the individual colour corresponding to the haplotypes represented in

mapped populations. Small black circles indicate haplotypes that were not sampled, or

have become extinct through genetic drift.

Table 4.1 Gene diversity indices and sampling regime for H. fuliginosus in the Daly

River.

Site Sample Nucleotide Haplotype Site Name Code Size Diversity Diversity DB Middle Creek 7 0.007 0.905 DM Claravale/Sawmill 21 0.005 0.867 DQ Flora River (upstream) 9 0.006 0.917 DN Mt Nancar 2 0.009 1.000 DC Mission Hole 6 0.003 0.600 DD Bottom of Centipede Dreaming Gorge 2 0.009 1.000 DE U/S Centipede Dreaming and Snowdrop Creek 4 0.006 1.000 DA Beeboom 26 0.007 0.892 DT Oolloo Crossing (upstream) 12 0.007 0.879 DR Stray Creek 6 0.005 0.733 DP Mathison Creek 13 0.006 0.936 DU Upstream of Manyallaluk 6 0.005 0.600 DJ Fish River 5 0.002 0.600 DFF Fergusson River 27 0.005 0.900 DS Douglas River 5 0.006 1.000 DX Oolloo Crossing 12 0.004 0.803 DL Galloping Jack's 8 0.006 0.857 DH Downstream of Galloping Jack’s 21 0.008 0.919 ID1 ID1 Green Ant Creek 3 0.004 1.000 ID2 ID2 Green Ant Creek 3 0.008 1.000 ID3 Station Creek 4 0.005 0.833 ID4 Edith River 2 0.009 1.000 ID5 ID5 Stray Creek 3 0.006 1.000

Chapter 4.0 ∙ Sooty Grunter 149

Table 4.2 Gene diversity indices and sampling regime for H. fuliginosus in the Mitchell

River.

Site Sample Nucleotide Haplotype Site Name Code Size Diversity Diversity 58 Lynd River @ Torwood 2 0.003 1.000 59 Lynd River @ Rocky Creek 1 0.000 1.000 60 Lynd River/Pinnacle Creek 4 0.003 0.833 61 Mitchell River (10 Mile Creek) 1 0.000 1.000 62 Lynd River (1 Mile Creek) 1 0.000 1.000 63 Mitchell River (Kingfish) 1 0.000 1.000 118 Walsh River (Leadingham Creek Road Crossing) 10 0.006 0.933 MA Bruce Weir 11 0.005 0.800 MC Mitchell River @ Gamboola 2 0.003 1.000 MD Walsh River @ Rookwood 2 0.009 1.000 MJ Magnificent Creek 1 0.000 1.000 MK Saltwater Creek 4 0.003 0.833 ML Walsh River @ Nullinga 1 0.000 1.000 MO Dickson's Hole 1 0.000 1.000 MP Tate River @ Ootan 2 0.000 0.000 MQ Emu Creek @ Petford 4 0.005 0.833

4.3.2 Tests of neutrality

All tests of neutrality were significant for the Daly River population (Table 4.3), indicating

an excess of low frequency haplotypes in the data set. In contrast, all results were non-

significant for the Mitchell River population (Table 4.3). Models of pairwise differences

were generated for the Daly River population for constant and expanding scenarios (Figure

4.6). The observed pairwise differences closely followed that of the modelled expanding

population. Additionally, r was non-significant indicating that the distribution was smooth

– an expectation under expansion (Table 4.4).

Chapter 4.0 ∙ Sooty Grunter 150

Table 4.3 Tests of neutrality for populations of H. fuliginosus in the Daly River and

Mitchell River. * denotes significant values.

Test Statistic P α

Tajima's D -1.787* 0.007 0.050

Fu's FS -22.003* 0.000 0.020 Fu and Li's D* -3.808* 0.002 0.020

Fu and Li's F* -3.581* 0.002 0.050 Daly River Daly

R2 0.030* 0.004 0.050

Tajima's D -0.948 0.192 0.050

Fu's FS -2.837 0.153 0.020 Fu and Li's D* -0.607 0.379 0.020

Fu and Li's F* -0.844 0.222 0.050 MitchellRiver R2 0.073 0.152 0.050

Chapter 4.0 ∙ Sooty Grunter 151

Chapter 4.0 ∙ Sooty Grunter 152

Table 4.4 Goodness of fit test (raggedness) for simulated population expansions in H. fuliginosus in the Daly River.

Raggedness Index P Daly River 0.092 0.880

DnaSP v5.1 calculated τ as 1.712 for the Daly River population (Librado & Rozas 2009).

Time since expansion (t) was estimated to be 30 060 – 65 586 generations, or approximately 120 200 – 262 300 years.

4.3.3 Partitioning of genetic variation

Global FST values were calculated separately for each catchment. In the Daly River, the population was found to have weak, but significant genetic structure (FST = 0.023, P =

0.019; ΦST = 0.020, P = 0.071) (Table 4.5). In the Mitchell River, the population was found to be panmictic with no detectable structure (FST = -0.022, P = 0.625; ΦST = 0.007, P =

0.465) (Table 4.6). Within the Daly River population, there was significant genetic structure among all populations for all calculations. Significant restrictions in gene flow were also found among populations within catchment positions and within waterway types, indicating that dispersal is limited within these environments.

Chapter 4.0 ∙ Sooty Grunter 153

Table 4.5 AMOVA calculations for H. fuliginosus in the Daly River. * denotes significant fixation indices. ° denotes % variation within populations.

Control Region % % F P Φ P Variation ST Variation ST Among populations, within - 0.023* 0.019 - 0.020 0.071 the Daly River Among subcatchments in 0.85 0.008 0.250 1.38 0.014 0.171 the Daly River Among populations, within 1.56 0.016 0.175 0.80 0.008 0.409 subcatchments Among all populations 97.60° 0.024* 0.031 97.83 0.022 0.070

Table 4.6 AMOVA calculations for H. fuliginosus in the Mitchell River. * denotes significant fixation indices.

Control Region % % F P Φ P Variation ST Variation ST Among populations, within - -0.022 0.625 - 0.007 0.465 the Mitchell River

Chapter 4.0 ∙ Sooty Grunter 154

4.3.4 Isolation indices

There was a trend for both mean population FST and ΦST fixation values where downstream populations were the most isolated, midreach populations were the most connected, and upstream populations were intermediate between the two (Figure 4.7). However, this pattern was not statistically significant for either ANOVAs (Table 4.7), or Spearman’s rank correlations (Table 4.8). For populations in main channels or tributaries, isolation indices did not reflect a specific pattern (Figure 4.8). This result was further supported by non- significant ANOVA results showing that the two groups (main channel and tributaries) could not be differentiated (Table 4.7).

Table 4.7 ANOVA results for comparisons of isolation indices for each of the waterway classifications for Daly River H. fuliginosus. No significant relationships were found (α

= 0.05).

Isolation Index Type Source SS df MS F P Between catchment mtDNA F 0.005 2 0.003 0.757 0.482 ST positions Between catchment mtDNA Φ 0.000 2 0.000 0.079 0.924 ST positions Between waterway mtDNA F 0.001 1 0.001 0.349 0.561 ST types Between waterway mtDNA Φ 0.000 1 0.000 0.005 0.946 ST types

Table 4.8 Spearman’s rank correlations between distance from river mouth and average fixation index (isolation index) for H. fuliginosus in the Daly River (α = 0.05).

ρ P

mtDNA FST 0.004 0.984

mtDNA ΦST 0.125 0.568

Chapter 4.0 ∙ Sooty Grunter 155

Isolating Influence of Catchment Position (Daly River)

0.1

0.08 Downstream Midreach 0.06 Upstream Series4 Series6 0.04 Series5 Series7

Fixation Value 0.02

-0.02 FST mtDNA fdsaf ΦST mtDNA

Figure 4.7 Isolation indices for all sampled populations of H. fuliginosus from the Daly

River based on a) mean FST for control region and b) mean ΦST for control region.

Populations are classed as being downstream (blue), midreach (light green) or upstream

(peach).

Isolating Influence of Waterway Type (Daly River)

0.035 0.03 0.025 0.02 Main Channel 0.015 Tributary 0.01 Series3 Series4

0.005 Series5 Fixation Value 0 -0.005 -0.01 -0.015

FST mtDNA huk ΦST mtDNA

Figure 4.8 Isolation indices for all sampled populations of H. fuliginosus from the Daly

River based on a) mean FST for control region and b) mean ΦST for control region.

Populations are classed as being in the main channel (green) or a tributary (orange).

Chapter 4.0 ∙ Sooty Grunter 156

4.3.5 Mantel tests

An initial Mantel test showed no significant relationship between genetic distance and geographic distance for either catchment (Table 4.9).

Table 4.9 Mantel test correlation coefficients for H. fuliginosus in the Daly and Mitchell catchments (α = 0.05).

Mantel Test Correlation Coefficient P Daly catchment 0.147 0.129 Mitchell catchment -0.084 0.711

4.3.6 Regression analyses

Three landscape features (distance, elevation and riffle habitat) were selected which were hypothesised to be the most influential in determining population isolation in the Daly

River. First, to further investigate the lack of IBD, a decomposed pairwise regression was used to identify outliers in the data set, so that the model of best fit could be determined for geographical distance and gene flow. Using Akaike’s Information Criterion (Akaike 1974), four models were identified as being equally likely (Table 4.10). Model A (Figure 4.10) had the largest Akaike weight (-51.274), and included 11 populations in a model with a weak but significant correlation between distance and fixation indices (R2 = 0.155, P =

0.003). Models B, C and D (Figures 4.11, 4.12 and 4.13) were all very similar to Model A – showing a positive, weak correlation between distance and FST values. Overall, these results contrasted with the lack of IBD inferred by the Mantel test, and showed a weak IBD signature in the data.

Chapter 4.0 ∙ Sooty Grunter 157

0.15

0.1

0.05

-0.05

Mean Regression Residuals Regression Mean -0.1

-0.15

DJ DP DS DT DL DB DM DC DFF DA DQ DR DU DX DH Population

Figure 4.9 Mean regression residuals and 95% confidence intervals for decomposed pairwise regression analysis. Example taken from correlations with distance for Daly

River H. fuliginosus. Here, site DR was identified as the first outlier i.e. the greatest deviant from 0.

DPR - Daly River Model A

0.15

0.1 R2 = 0.155

0.05

ST F 0

-0.05

-0.1 0 50 100 150 200 250 300 350 400 Distance (km)

Figure 4.10 Model A of best fit for the relationship between FST and stream distance for

H. fuliginosus in the Daly River, showing a positive correlation (R2 = 0.155, P = 0.003).

Four outlier populations were removed.

Chapter 4.0 ∙ Sooty Grunter 158

DPR - Daly River Model B

0.2

0.15

0.1 R2 = 0.100 0.05

ST 0 F -0.05

-0.1 -0.15

-0.2 0 100 200 300 400 500 600 Distance (km)

Figure 4.11 Model B of best fit for the relationship between FST and stream distance for

H. fuliginosus in the Daly River, showing a positive correlation (R2 = 0.100, P = 0.005).

Two outlier populations were removed.

DPR - Daly River Model C

0.2

0.15 R2 = 0.129 0.1

0.05

ST 0 F -0.05

-0.1 -0.15

-0.2 0 100 200 300 400 500 600 Distance (km)

Figure 4.12 Model C of best fit for the relationship between FST and stream distance for

H. fuliginosus in the Daly River, showing a positive correlation (R2 = 0.129, P < 0.001).

Outlier population DR was removed.

Chapter 4.0 ∙ Sooty Grunter 159

DPR - Daly River Model D

0.2

0.15

0.1 R2 = 0.090

ST 0.05 F

-0.05

-0.1 0 100 200 300 400 500 600 Distance (km)

Figure 4.13 Model D of best fit for the relationship between FST and stream distance for

H. fuliginosus in the Daly River, showing a weak positive correlation (R2 = 0.090, P =

0.014). Three outlier populations were removed.

Chapter 4.0 ∙ Sooty Grunter 160

2 Populations Removed n R AICC ΔAICC DR, DJ, DS, DU 11 0.155 -51.274 0.000 DR, DJ 13 0.100 -50.152 1.122

DR 14 0.129 -49.718 1.556

DR, DJ, DS 12 0.090 -49.284 1.991 DR, DJ, DS, DU, DT 10 0.254 -49.075 2.199

Distance None 15 0.072 -46.002 5.272 DR, DJ, DS, DU, DT, DQ 9 0.172 -45.465 5.809 DR, DJ, DS, DU, DT, DQ, DL, DC 7 0.000 -41.084 10.190 DR, DJ, DS, DU, DT, DQ, DL 8 0.234 -40.543 10.732 None 15 0.002 -93.926 0.000

DS 14 0.001 -89.756 4.171

DS, DR, DJ, DU 11 0.540 -88.111 5.815 DS, DR 13 0.001 -87.136 6.791

DS, DR, DJ 12 0.002 -86.870 7.056 Elevation DS, DR, DJ, DU, DA 10 0.720 -82.941 10.985 DS, DR, DJ, DU, DA, DM 9 0.770 -75.185 18.741

A decomposed regression analysis exploring the influence of elevation on isolation showed no relationship between these measures (R2 = 0.002, P = 0.884) (Table 4.10). For this analysis, the best model included all populations, and did not identify any true outliers

(Figure 4.14).

Chapter 4.0 ∙ Sooty Grunter 161

Regression - Daly River Model E

0.14 0.12 0.1 0.08 R2 = 0.002 0.06 0.04

Isolation Index Isolation 0.02 0 -0.02 -0.04 0 20 40 60 80 100 120 140 160 180 200 Elevation (m)

Investigations of how riffle habitat may facilitate connectivity (using FST values) between populations showed no correlation between these factors. Model F (Figure 4.15) was the most highly weighted (-65.913), and identified no true outliers. This model showed no relationship between connectivity and PRUs (R2 = 0.004, P = 0.622). However, regressions of haplotype diversity with PRUs identified a negative correlation between these parameters, where increasing riffle habitat correlated with decreased haplotype diversity.

Two models were found to be equally likely (Models G (Akaike weight of -62.389) and H

(Akaike weight of -62.004)) (Figures 4.16 and 4.17; Table 4.11).

Chapter 4.0 ∙ Sooty Grunter 162

DPR - Daly River Model F

0.2

0.15

0.1

0.05 R2 = 0.004

ST 0 F -0.05

-0.1 -0.15

-0.2 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Potential Riffle Units (m)

Figure 4.15 Model F of best fit for the relationship between FST and PRUs for all populations of H. fuliginosus in the Daly River, showing no significant correlation (R2 =

0.004, P = 0.622).

Regression - Daly River Model G

1.2

1 R2 = 0.174 0.8

0.6

0.4 Haplotype Diversity Haplotype 0.2

0 0 1000 2000 3000 4000 5000 6000 7000 8000 Potential Riffles Units (m)

Figure 4.16 Model G of best fit for the relationship between haplotype diversity and

PRUs for all populations of H. fuliginosus in the Daly River, showing a weak, but non- significant negative correlation (R2 = 0.174, P = 0.122).

Chapter 4.0 ∙ Sooty Grunter 163

Regression - Daly River Model H

1 0.9 0.8 R2 = 0.331 0.7 0.6 0.5 0.4

0.3 Haplotype Diversity Haplotype 0.2 0.1 0 0 1000 2000 3000 4000 5000 6000 7000 8000 Potential Riffle Units (m)

Figure 4.17 Model G of best fit for the relationship between haplotype diversity and

PRUs for all populations of H. fuliginosus in the Daly River, showing a significant negative correlation (R2 = 0.331, P = 0.031).

Chapter 4.0 ∙ Sooty Grunter 164

2 Populations Removed n R AICC ΔAICC None 15 0.004 -65.913 0.000 DR 14 0.070 -48.794 17.119

DR, DS, DJ 12 0.037 -48.603 17.310 ST

F DR, DS 13 0.114 -48.448 17.465

- DR, DS, DJ, DQ 11 0.030 -46.283 19.630 DR, DS, DJ, DQ, DT 10 0.049 -43.017 22.895 PRUs PRUs DR, DS, DJ, DQ, DT, DFF 9 0.062 -38.342 27.571 DR, DS, DJ, DQ, DT, DFF, DL 8 0.044 -33.465 32.448 DR, DS, DJ, DQ, DT, DFF, DL, DA 7 0.015 -28.679 37.234 None 15 0.174 -62.389 0.000

DS 14 0.331 -62.004 0.385 DS, DFF 13 0.452 -59.509 2.881

DS, DFF, DQ 12 0.488 -55.339 7.051

diversity

- DS, DFF, DQ, DP 11 0.579 -52.929 9.461 DS, DFF, DQ, DP, DB 10 0.713 -51.653 10.736

PRUs PRUs DS, DFF, DQ, DP, DB, DH 9 0.709 -46.554 15.836 DS, DFF, DQ, DP, DB, DH, DC 8 0.754 -43.346 19.043

Individual pairwise regressions of distance and FST values showed outlier populations to follow two patterns described in Koizumi et al. (2006) as: Pattern 2, where drift has more impact than gene flow (site DS); Pattern 4, where the effect of drift is weaker than gene flow (site DR and DJ) (Figure 4.18; Table 4.12). For several populations (including outlier

DU), patterns could not be identified. Non-outlier populations followed either Pattern 2, 3 or Pattern 4 (Table 4.12).

Chapter 4.0 ∙ Sooty Grunter 165

H. fuliginosus - Distance

0.25

0.2

0.15

0.1 ST F 0.05

-0.05

-0.1 0 50 100 150 200 250 300 350 400 450 500 Distance (km)

Figure 4.18 Decomposed pairwise regression between FST and distance for H. fuliginosus in the Daly River. ‘True outliers’ are represented by dashed lines.

Table 4.12 Intercepts, slopes and regression coefficients for all tested populations of H. fuliginosus. ‘True outliers’ are shaded grey.

Intercept P Slope P Population R2 Pattern (*10-2) (intercept) (*10-2) (slope) DB 4.100 0.702 0.000 0.921 0.001 4 DM -0.300 0.887 0.000 0.364 0.104 4 DC 11.300 0.044 0.000 0.486 0.063 - DFF -3.200 0.049 0.000 0.014 0.549 2 DA 2.300 0.261 0.000 0.731 0.016 - DP -2.500 0.264 0.000 0.045 0.414 3 DQ -5.700 0.002 0.000 0.005 0.643 2 DT 2.300 0.246 0.000 0.245 0.165 4 DX 3.300 0.250 0.000 0.977 0.000 4 DH -2.400 0.315 0.000 0.034 0.448 3 DL 2.000 0.732 0.000 0.354 0.108 4 DJ 4.400 0.322 0.000 0.143 0.170 4 DR 0.600 0.931 0.100 0.092 0.218 4 DS -9.300 0.008 0.000 0.021 0.370 2 DU 14.500 0.063 0.000 0.575 0.027 -

Chapter 4.0 ∙ Sooty Grunter 166

4.4 Discussion

4.4.1 Genetic diversity

Similarly to the melanotaenids in Chapter 3.0, high levels of genetic diversity were revealed in both the Daly and Mitchell catchments. Most of these haplotypes were distributed across the entire scale of sampling. Wide haplotype dstributions are indicative of high levels of dispersal and gene flow between the populations – a finding that supported the hypothesis that connectivity between populations would be unrestricted. Furthermore, in the Daly River population, these haplotypes were arranged in a star-like phylogeny, with a common internal haplotype surrounded by many recently generated, low frequency haplotypes. While diversity in some sites was slightly lower, this was mostly the result of lower sample sizes which may not have adequately captured the true spectrum of haplotypes. This was particularly true of the Mitchell River, where sample sizes were small.

In this catchment, sampling effort was high, but it was not possible to catch greater numbers of H. fuliginosus.

4.4.2 Population expansion in the Daly River

The star-like haplotype phylogeny in the Daly River suggests a recent population expansion

(Slatkin & Hudson 1991; Harpending et al. 1998). Singletons and low frequency alleles are considered to be a signature of such an event, as after a population bottleneck, a founder event or a demographic expansion, mutations will generate small numbers of closely- related haplotypes (Avise et al. 1984; Posada & Crandall 2001). Such a pattern was not observed in the Mitchell River population, and the absence of a population expansion was Chapter 4.0 ∙ Sooty Grunter 167 further confirmed by the non-significant results for all the neutrality tests. In contrast, neutrality tests were all significant for H. fuliginosus in the Daly River, which indicated an excess of low frequency sequence variants as is expected under an expansion model.

Additional support was given by the non-significant r value, which indicated a smooth mismatch distribution typical of an expanding population. Population expansions can occur after a founder event, where a group of individuals established in a new environment and the conditions were favourable and conducive to a demographic expansion. Alternatively, the population may have been a small but stable group, which expanded after some change

(for example, warmer temperatures, increased availability of food or increased habitat) that supported an increase in population size (Hewitt 2000).

The dating of this population expansion in the Daly River was estimated at 120 200 – 262

300 years before present. For the northern hemisphere, population expansions in aquatic taxa have been found to coincide with the end of the last glacial maxim – approximately 10

000 to 80 000 years ago (Bennett 2004; Hewitt 2004). Threespine sticklebacks

(Gasterosteus aculeatus), European bullhead (Cottus gobio) and European whitefish

(Coregonus lavaretus) are all examples where glacial retreat has triggered a population expansion (Malhi et al. 2006; Hänfling et al. 2002; Østbye et al. 2005). In general, fish population bottlenecks and expansions tend to be closely linked to Milankovitch cycles, which determine periods of warming and cooling every 100, 41 and 21 thousand years

(Bennett 2004; Hewitt 2004). Lower temperatures can substantially decrease population size; increased temperatures allow those refugial populations to expand. Consequently, population expansions tend to coincide with the onset of warmer periods. For Australia, glacial ages had an indirect effect on the climate, and impacted the aridity of the continent more so than the temperature (Nanson et al. 1992). For fish in Australia, glacial events in Chapter 4.0 ∙ Sooty Grunter 168 the north would have resulted in both a drop in sea level (as globally, water became

‘unavailable’ as snow and ice) and a reduction in fresh surface waters (Unmack 2001). The end of these periods saw sea levels rising and the aridity lessening – leading to an increase in available habitat for fish and a subsequent increase in their populations (Chappell 1991;

Hewitt 2004). Impacts of these changes in Australia can be seen in estuary and Port Jackson perchlets (Ambassis marianus and Ambassis jacksoniensis), eastern rainbowfish

(Melanotaenia splendida splendida) and Oxleyan pygmy perch (Nannoperca oxleyana)

(Mills et al. 2008; Hurwood & Hughes 2001; Knight et al. 2009). The population expansion in H. fuliginosus corresponds very closely with the expansion dates for populations of Ambassis marianus and Ambassis jacksoniensis in the estuarine reaches of eastern Australia (Mills et al. 2008). Given the timing of these expansions for many species, it seems likely that there was an Australia wide climatic change that resulted in riverine fish being able to colonise new reaches and experience population growth. An expansion may also be due to other factors that support population growth (for instance increased food availability); however such variables are difficult to identify. Given the strong influence of climate change (with glacial action being the major influence in the northern hemisphere and sea level change being the main influence in the southern hemisphere) on supporting population expansions in fish the world over, it seems that changes to climatic conditions were the most likely cause for an expansion in H. fuliginosus in the Daly River.

4.4.3 Genetic structure and dispersal capability

In both the Daly and Mitchell catchments, low levels of genetic structure were observed.

Overall, this supported the hypothesis that the dispersal capability of this species would be Chapter 4.0 ∙ Sooty Grunter 169 reflected by low levels of genetic structure within each catchment. To some extent, this was also represented in the absence of Pattern 1 in the decomposed pairwise regressions performed for the Daly River population. Pattern 1 is indicative of the complete isolation of a population – clearly the dispersal capability of this species was sufficient to overcome any

‘complete’ barriers in these catchments (Koizumi et al. 2006). In the Mitchell River, fixation values indicated that gene flow was sufficient to homogenise alleles within the catchment, resulting in a panmictic population. So for the H. fuliginosus in the Mitchell

River, dispersal appears to be unrestricted. Low, but significant levels of genetic structure were identified in the Daly River, indicating that some factor has inhibited movement of this species within this catchment. The finding of genetic structure was contrary to the expectation that dispersal would be unrestricted across the catchment.

It is important to consider whether movement throughout these catchments is the result of active or passive dispersal. A complicating factor is that dispersal may not be undertaken by the adults and more developed juveniles. It is possible, but unlikely that dispersal is a result of passive larval dispersal – a mechanism which often results in low levels of genetic structure (Bohonak 1999; Levin 2006). However, if this is the mechanism of gene flow in these catchments, gene flow would be unidirectional towards the mouth of the catchment, following the flow of the river (Congdon 1995). Unidirectional gene flow was not observed in the data set (at least for the Daly River), as it would be expected that downstream populations would have high levels of genetic diversity, and upstream populations would have very low genetic diversity (Allendorf 1983; Bohonak 1999; Chaput-Bardy et al.

2009). Instead, H. fuliginosus had high levels of genetic diversity at many sites, regardless of whether they were downstream or not. This result indicates that active dispersal (via juveniles and adults) against the flow of the river is distributing alleles to upstream Chapter 4.0 ∙ Sooty Grunter 170 populations. And so if larval dispersal is occurring, it is either not a frequent event, or else gene flow via active dispersal is strong enough to obliterate the genetic signature of unidirectional gene flow (Matschiner et al. 2009). Additionally, most freshwater fish in

Australia have adapted breeding strategies that prevent the passive distribution of eggs and developing fry, as heavy losses would be incurred by offspring being swept to the estuaries or ocean (Growns 2004). Passive larval dispersal is frequently observed in marine species, and contributes to the general pattern of panmixia in these environments (Riginos &

Nachman 2001; Jones et al. 2009). Such dispersal strategies are rarely seen in freshwater fish.

Isolation by distance (IBD) was only mildly influential in the Daly River, and was negligible in the Mitchell River. IBD is only expected in populations with certain dispersal characteristics, and so the presence or absence of IBD it itself can be informative (see

Chapter 1.2 for details). The absence of IBD suggests the dispersal distances are equal to or greater than the sampled range. Given that much of each catchment was sampled, it seems unlikely that IBD would be revealed with further collections – overall, the dispersal distances of H. fuliginosus within catchments are extensive. Additionally, IBD can only be established when there is equilibrium between the diverging effects of genetic drift, and the homogenising effects of gene flow (Wright 1949; Wright 1951). This relationship is fully established in populations where Pattern 3 is operating, which was only observed in sites

DP and DH. In populations following Pattern 2, IBD is in effect, but genetic drift is still greater than gene flow, as was observed in several sites (Koizumi et al. 2006). Due to the recent population expansion detected in the Daly catchment, this suggests that insufficient time has elapsed to allow the population to reach equilibrium. The high levels of gene flow within the catchment are counteracting genetic divergences, thus obscuring the genetic Chapter 4.0 ∙ Sooty Grunter 171 signature of IBD. This counteraction was particularly highlighted by the number of populations that followed Pattern 4 in individual pairwise regressions. Pattern 4 establishes when gene flow is more influential than genetic drift. Koizumi et al. (2006) suggests that this can be indicative of ‘sink’ populations (as was found in Chapter 3.0), but it is more likely that in H. fuliginosus this is just a signature of the high levels of dispersal in general.

Overall, these findings lend partial support to the hypothesis that distance would not impact on the connectivity of populations; however, it is clear that IBD does have some influence on gene flow.

Isolation indices in the Mitchell catchment were quite high for some sites. However, given the overall panmixia in this catchment, it seems that these FST values are over estimated, due to the low sample sizes (of two or less). At these sites, only a single haplotype was observed, which erroneously suggests a lack of gene flow to these populations. Increased sampling sizes at these sites would presumably incorporate more variation, thus lowering the FST values as gene flow is captured. Consequently, these isolation indices did not accurately portray the isolation, and there was insufficient evidence to make any conclusions regarding population isolation in this catchment.

4.4.4 Influence of the landscape on connectivity

As shown in the initial Mantel test investigating the partitioning of genetic variation within the landscape, populations were highly connected within each catchment and did not seem to be influenced by IBD. However, decomposed regression analyses refuted this, and showed that IBD is a factor in determining the genetic structure of populations in the Daly catchment. This discrepancy between analyses may be attributed to the true outliers in the Chapter 4.0 ∙ Sooty Grunter 172 data set which clouded the signature of IBD. These outliers (DJ, DS, DR and DU) were not considered to be particularly isolated, and no other explanation could be found which may explain their deviation from other populations. For example, site DA (Beeboom) is a natural rocky barrier which has been built up to form a substantial road crossing at

Tipperary Station. Downstream are sites DJ, DN and DC, of which only site DJ was an outlier. If this barrier is influencing gene flow downstream, then it would be expected that all three of these sites would be affected.

It was also found that the elevation of sites had little impact on the isolation of populations.

The best model included all populations, and showed that increasing elevation had no effect on isolation. The result supported the hypothesis that the dispersal capability of this species would overcome this barrier. In many species of freshwater fish, elevation has been found to be a contributing factor in limiting the dispersal and connectivity of populations. Narum et al. (2008) found that elevation was negatively correlated with genetic diversity in populations of rainbow trout (Onchorhynchus mykiss), which was the result of reduced connectivity imposed by the elevation barrier. Brook charr (Salvelinus fontinalis) populations are similarly affected by the elevation barrier to dispersal (Castric et al. 2001).

In Australia, elevation has been a determinant in restricting connectivity between populations of northern trout gudgeon (Mogurnda mogurnda) (Cook et al. 2011). The absence of such a correlation in H. fuliginosus is testament to the dispersal capability of this species, and highlights its tolerance to both the changes associated with higher altitudes, and the difficulties of dispersing against fast flow.

Finally, it was predicted that riffle habitat would make a significant contribution in determining the connectivity of populations within the catchment. Contrary to expectations, Chapter 4.0 ∙ Sooty Grunter 173 riffles did not seem to influence gene flow, as no correlation was found between gene flow and the amount of riffle habitat between populations (PRUs). However, it is important to point out that this result does not by any means negate or lessen the importance of riffle habitat to this species (Chan et al. 2010; Stewart-Koster et al. 2011). It simply indicates that riffles are not a determinant of population connectivity. Given the dispersal capability of this species, this result is not surprising, as adult dispersal can easily bridge the distances between suitable riffle habitats (provided that the surface water is of an adequate depth).

However, PRUs did seem to influence the genetic diversity, but in a relationship that was not expected. Increasing riffle habitat led to a reduction in genetic diversity. As riffle habitat did not determine the connectivity of populations, it seems that this reduction in diversity is not the result of restricted gene flow. It is possible that only a few, appropriately sized individuals (i.e. individuals that are sexually mature, but small enough for shallow surface water) are able to traverse the riffles. If this was the case, then only these individuals would be represented by the resulting offspring. This situation would be akin to the patchy recruitment hypothesis. However, in order to determine whether this is truly the case, temporal sampling would be needed to identify whether this relationship remains constant and whether the diversity changes over time (Bunn & Hughes 1997). These questions need to be addressed using a more variable marker, such as microsatellites.

Unfortunately, despite numerous attempts, it was not possible to develop microsatellite markers for H. fuliginosus within the timeframe of this study. Genetic studies always greatly benefit from the use of two (or more genetic markers). The use of an additional marker brings greater clarity and confidence in interpreting results because of the two (or more) independant assessments of gene flow that can be made. Importantly, the Chapter 4.0 ∙ Sooty Grunter 174 introduction of microsatellites may give an alternative view of patterns of gene flow in H. fuliginosus. As microsatellites reflect contemporary dispersal, the differing time frames

(whereby mitochondrial markers are more representative of historical gene flow) can also reveal differing patterns of gene flow. In the case of H. fuliginosus, it is possible that contemporary dispersal is comparatively more restricted than in historical populations. For example, the construction of artificial barriers in the Daly and Mitchell Rivers may be quite significant in impacting connectivity of populations. However, it would only be through the use of variable microsatellite markers with fast mutation rates that this would be detected.

Alternatively, it may be that the microsatellite markers would support the findings presented in this study. Either way, a second marker would be informative in either confirming the interpretation, or else bringing a different perspective in regards to the temporal changes of population connectivity (Karl et al. 2012).

Chapter 4.0 ∙ Sooty Grunter 175

4.5 Conclusion

Overall, this study showed the dispersal capability of H. fuliginosus is far ranging, which has resulted in the very low levels of genetic structure. A recent population expansion

(~283 000kya) in the Daly River has generated high levels of low frequency alleles, which have been distributed right across the catchment due to the high levels of gene flow. The distribution of the genetic diversity across the catchment showed how the dispersal capability of these populations is so great that there was seemingly no influence from any of the tested landscape characteristics – which unexpectedly included riffle habitat. These results highlight the importance of dispersal in this species, and how such gene flow has served in maintaining the genetic diversity within sites.

Careful consideration must be given to the implications of the dispersal requirements for this species. If the proposed water extraction in the Daly River was to occur, changes in riverine connectivity would alter the natural patterns of dispersal observed in this species.

Additionally, adequate access to critical riffle habitat must be maintained in order to support the population size, and the genetic diversity contained within it.

Chapter 4.0 ∙ Sooty Grunter 176

Chapter 4.0 ∙ Sooty Grunter 177

Rainbows are visions, but only illusions, And rainbows have nothing to hide. So we've been told and some choose to believe it I know they're wrong, wait and see. Someday we'll find it, the rainbow connection, The lovers, the dreamers and me.

All of us under its spell, we know that it's probably magic....

Kenny Ascher and Paul Williams (1979) for Kermit The Frog

Chapter 5.0 ∙ The Rainbow Connection 178

5.0 The rainbow connection: using stable isotopes and population genetics

to evaluate the importance of floods to rainbowfish

5.1 Introduction

Studies of dispersal and migration are often based on a single method which can sometimes result in ill-informed, weak or shallow appraisals and conclusions (Nathan 2001; Nathan et al. 2003; Cunjak et al. 2005). Within ecology there is a tendency for researchers to use only a single approach when attempting to quantify movement; sometimes a second method may be useful in elucidating patterns of movement (Bélisle 2005; Lowe et al. 2006).

Furthermore, the true evaluation of the extent of dispersal may be limited by the method applied (Bossart & Pashley Prowell 1998). For instance, when using radio-tracking, only a specific window of time is captured. Two tools to track movements are genetic and stable isotope markers (see Chapter 1.6). However, when used independently they can only reveal dispersal in certain time frames – potentially over many generations in the case of genetic markers, and within the course of an individual’s lifetime with stable isotopes (Sunnucks

2000; Nathan et. al 2003; Rubenstein & Hobson 2004). By overlaying the results of these techniques, a stronger and more holistic evaluation of dispersal can be achieved

(Chetkiewicz et al. 2006; Lowe et al. 2006).

While few studies have used this integrated method, it is an approach that is potentially applicable to studies attempting to quantify dispersal in overlapping temporal scales

(Nathan et al. 2003). Such a study can potentially show the genetic stock (and hence the origin) from which the individual has come – a dispersal event which may have occurred

Chapter 5.0 ∙ The Rainbow Connection 179 from anywhere from the beginning of an individual’s life, to just before the individual was captured (Berry et al. 2004). Alternatively, a stable isotope analysis of carbon and nitrogen can reveal very recent dispersal, and can quantify more specifically how an individual uses its environment, and what movements an individual might make over a specific window of time (Peterson & Fry 1987; Rubenstein & Hobson 2004). By using isotopes of carbon and nitrogen, which have a fast change-over rate, these signatures ‘evolve’ at a pace that can capture the change of environmental signatures (i.e. a dispersal event) within weeks to months (Peterson & Fry 1987; Hobson 1999; Voight et al. 2003).

These combined approaches are particularly relevant when attempting to infer migratory patterns in populations (Webster et al. 2002; Clegg et al. 2003). A genetic marker will not reveal migration when genes are not transferred into a population – the genetic evidence is simply not available and so a dispersal event is ‘silent’ unless that specific individual is captured (Nathan 2005). Moreover, an analysis of stable isotope signatures may only show the recent movements of an individual (West et al. 2006). Stable isotope markers work with genetic markers to determine both dispersal and migration – characteristics which are poorly understood in fish in Australia, particularly tropical Australia (Mallen-Cooper 1993;

Humphries et al. 1999).

The Mitchell River in tropical north Queensland is a major system covering 70 000km2 and has one of the highest discharge rates in Australia (Brooks et al. 2009). It is subjected to extreme flooding during the wet season period of December through to April, at which time the catchment will receive >80% of its annual rainfall. When in flood, the high connectivity between habitats in the Mitchell River has the potential to facilitate movement between aquatic populations. Movement can be categorised as either 1) dispersal – which is a Chapter 5.0 ∙ The Rainbow Connection 180 vectored, permanent movement away from the natal population (Song et al. 2005), or 2) migration – in which an individual may move away for a period of time, but ultimately rejoins the natal population (Gaines & McClenaghan1980; Dingle & Drake 2007).

However, while two populations may be physically connected through the medium of water, this does not necessarily mean that dispersal/migration takes place. There are a variety of conditions and situations that can determine the actual connectivity of populations of aquatic organisms in the Mitchell River.

For fish, there are certain environmental conditions in floodplains that may limit dispersal.

In shallow waters, temperatures can exceed the tolerance limits of some species (Closs &

Lake 1996; Pusey & Arthington 2003). Furthermore, warm shallow waters can become less saturated with oxygen (hypoxic), thereby becoming physically stressful for fish (Bishop

1980; Matthews & Berg 1997; McNeil & Closs 2007). Barriers to dispersal in floodplains are not purely limited to the environmental conditions encountered. When leaving the natal population, individuals may be exposed to a behavioural barrier – avoidance of greater predation pressure (Zollner & Lima 2005; Thuesen et al. 2008). For instance, a fish moving through a deeper channel may be consumed by an actively hunting piscivorous fish (Sih &

Wooster 1994; Heg et al. 2004). Alternatively, moving through shallow regions may put fish within the striking range of raptors and wading birds (Cucherousset et al. 2007).

Ultimately, there are always risks associated with dispersal or migration, and dispersal behaviours will reflect this (Ronce 2007).

While shallow backwaters and floodplains represent stressful environments that may limit dispersal, deeper areas are also challenging environments during periods of elevated flow.

For example, rivers that experience extreme levels of high flow often have very turbulent Chapter 5.0 ∙ The Rainbow Connection 181 flow paths. Fish that are not well suited to coping with such hydrological forces seek refuge in calmer reaches or behind hydraulic shelter, and therefore tend not to disperse during these events (Allouche 2002; Cook & Coughlin 2010). Moving from the relative safety of a home territory may bring fish into contact with high flows. Not only can swift, turbulent water potentially kill or injure fish; high flows may transport an individual to a new environment for which it is ill-adapted (Chapman & Kramer 1991; del Mar Delgado et al.

2010). For example, a freshwater fish could be swept several kilometres downstream where salinity may be outside the tolerance limits of that individual.

While there are hazards associated with moving away from a home site, there are also a great number of benefits which may balance or outweigh the negative effects of dispersal/migration (Cote & Clobert 2010). An individual may experience high intra- and inter-species competition for resources – for example, suitable substrate upon which to attach eggs, or access to patches of protective macrophytes; this can be avoided by dispersing (Bowler & Benton 2005). In some situations, the competition for suitable mates may be very strong, and so it may not be advantageous for an individual to stay in the natal territory – moving to a new environment can present greater opportunities for reproduction

(Johnson & Gaines 1990; Stiver et al. 2006). In many species, young move on to join different populations, which can be regarded as an adaptation to ensure adequate resources for the natal group, and to limit inbreeding (Bowler & Benton 2005). In any case, there are a variety of incentives that may stimulate an individual, or a group of individuals to leave the natal territory, and these are weighted against the costs of dispersing, and the costs of remaining philopatric.

Chapter 5.0 ∙ The Rainbow Connection 182

Floodplains are particularly noteworthy regions where the benefits-balance may favour migrating fish (Bayley 1995; Humphries et al. 1999; King et al. 2003). When the Mitchell

River is in flood, fish are able to take advantage of floodplains as resources for feeding, breeding grounds, nursery grounds, and also as refuges from fast flows associated with flooding (Balcombe et al. 2005; Bunn et al. 2006; Balcombe et al. 2007). They may also act as defensive barriers from larger predatory fish that may be unable to easily navigate through shallower waters (Magoulick 2003). While it is an energetic expense to actively travel to another location, a floodplain may be particularly productive and energetically positive; the energy used for dispersal may be regained (and surpassed) by feeding in this new environment (Balcombe et al. 2005; Bowler & Benton 2005). Furthermore, this surplus energy can then be directed into reproduction (Guderley & Pörtner 2010; Palstra et al. 2010). Also, by being in calmer waters, adults expend less energy in resisting flow, and also deposit their eggs and resulting fry in a nursery environment which will offer some protection for their development (Copp 1989; King 2004; Zeug & Winemiller 2007; Stoll

& Fischer 2011). However, floodplains tend to be intermittently (and sometimes unpredictably) available to fish in many parts of Australia; even for regions where floodplain inundation is predictable, access is still only for a limited time (Bunn et al.

2006). So while floodplains are temporary habitats for fish, they are extremely important for many aspects of fish productivity (Bayley 1995). These factors highlight the importance of hydrological connectivity for the persistence of populations, and the need to understand dispersal in the context of flooding.

Patterns of dispersal and migration of fish during floods are currently poorly understood in

Australia’s tropical rivers. Flooding as a cue for fish dispersal has been well-documented for many of the sub-tropical, arid and temperate regions of Australia (in particular the Lake Chapter 5.0 ∙ The Rainbow Connection 183

Eyre and Murray-Darling Basins) (Humphries et al. 1999; Arthington et al. 2005; Chapman

& Warburton 2006; King et al. 2010). However, the unpredictable, episodic flooding in the central and southern regions of Australia is quite different to the predictable monsoonal flooding of the north (Hamilton & Gehrke 2005; Bunn et al. 2006). Also, many of the fish species of tropical river systems are unique and endemic to this region (Pusey & Kennard

1996; Russell et al. 2003). It is therefore necessary to identify the importance of seasonal flooding for dispersal in tropical freshwater fish, and to determine how this might support the persistence of populations in northern rivers.

This study of M. s. inornata in four sites in the headwaters of the Mitchell River aims to use two markers (genetic and stable isotope) of dispersal to examine how elevated wet season flows influence the dispersal behaviour of this species.

Three predictions were made:

1. That microsatellite analyses and stable isotope analyses would reveal migrant

individuals that were genetically and isotopically differentiated from the resident

population.

2. That these migrant individuals would move in response to increased hydrological

connectivity during the wet season.

3. That M. s. inornata individuals would move to backwaters and floodplains (for

shelter from fast flows, feeding etc.) during the wet season.

Chapter 5.0 ∙ The Rainbow Connection 184

5.2 Detailed Methods

The following methods give specific detail that is additional to the general methods outlined in Chapter 2.0.

5.2.1 Site selection

As discussed in Section 1.5, stable isotopes can only be used as markers of dispersal when sites within a study area are isotopically differentiated (Cunjak et al. 2005; Inger &

Bearhop 2008). Previous studies of food webs in the Mitchell River were used to identify potential sites with distinguishable differences in stable isotope ratios (T. Jardine, unpublished data). Potential sites also needed to be within close proximity to each other, so that dispersal of fish between sites was possible during the wet season. Of these previously characterised sites, four were found to be sufficiently close to each other within the desired smaller geographic scale, and were individually distinguishable using stable isotopes.

5.2.2 AMOVA design

A hierarchical analysis of molecular variance (AMOVA) was used to compare the temporal variation within each of the four sites with the spatial variation among the four sites in

Arlequin v3.5.1.2 (Excoffier et al. 2005). For this analysis, sites were grouped according to season (i.e. wet season individuals were one group, and dry season individuals the other) so as to investigate how the variation was partitioned within and among groups. Additionally, overall FST estimates were made for both groups to determine if there was a temporal change in dispersal and gene flow during this time. Chapter 5.0 ∙ The Rainbow Connection 185

5.2.3 Assignment indices

Assignment indices, first employed by Paetkau et al. (1995), were used as an additional method to verify the overall absence of migrants in each of the populations. An assignment index is determined as being the likelihood that the genotype of an individual across all loci arose from the population in which the individual was sampled (Paetkau et al. 1995; Waser

& Strobeck 1998; Prugnolle & de Meeus 2002). A negative corrected assignment index indicates that the individual has uncommon alleles at a number of loci that has most likely originated from an outside population and is therefore a migrant. To determine the overall change (if any) in rates of migration between seasons, a mean assignment index (mAI) and a variance assignment index (vAI) were calculated across all individuals for both the dry season and during wet season groups in FSTAT v2.9.3.2 (Goudet 2001). A bias between two groups of individuals (which in this case were the dry season group and the wet season group) should result in a significantly lower mAI and vAI for the dispersing population

(Favre et al. 1997; Prugnolle & de Meeus 2002; Estes-Zumpf et al. 2010). Consequently, this analysis was used to detect any changes in rates of dispersal between the two sampling periods.

5.2.4 Power testing

POWSIM v4.0 (Ryman & Palm 2006) was used to estimate the power to identify genetic structure within the sampled populations. This analysis considered the sample sizes, the allele frequency and diversity, number of loci and the effective population size. 1 000 sampling simulations from the existing data were used to investigate homogeneity for each

Chapter 5.0 ∙ The Rainbow Connection 186 of the nine loci. Effective population size was calculated using NeEstimator v1.3 (Ovenden et al. 2007).

Chapter 5.0 ∙ The Rainbow Connection 187

5.3 Results

5.3.1 Data collection

In total, 212 M. s. inornata individuals were captured at the four sites (Figure 5.1) and screened across nine microsatellite loci, and for δ13C and δ15N. At each site, and at each time, other primary sources of organic material (riparian vegetation, biofilm, seston,

CPOM, FPOM and filamentous algae) provided a site stable isotope signature with which to identify the resident fish population (see Figures 7.1 – 7.5 for full dual isotope plots for each site).

Figure 5.1 Sites sampled for M. s. inornata and primary sources of organic matter in the upper Mitchell River catchment.

Chapter 5.0 ∙ The Rainbow Connection 188

5.3.2 Microsatellite analyses

The software package MICRO-CHECKER v2.2.3 (Van Oosterhout et al. 2004) showed populations at most loci to fall within the expectations of Hardy-Weinberg Equilibrium

(HWE). Arlequin v3.5.1.2 was used to investigate deviations from HWE for each population at each locus (Table 5.1). After sequential Bonferroni correction (Rice 1989) deviations from HWE were detected in 5.56% of tests for M. s. inornata. Four populations at three loci (Ma11, MS22 and Ma03) were identified as having heterozygote deficiencies caused by the presence of null alleles. The program FREENA (Chapuis & Estoup 2007) was used to correct the genotype frequencies at all loci for each population, and evaluate the influence of null alleles on FST values. FREENA had little effect on the original values, and so the raw (uncorrected) data set was used for analysis.

Null alleles were identified within the data set at three loci (5.56% of tests). Typically, null alleles create perceived excess homozygosity and deviations from HWE. In this instance, excess homozygosity occurs when heterozygote individuals are incorrectly assigned a homozygote genotype when primer site mutations prevent amplification during PCR (Van

Oosterhout et al. 2004; Chapuis & Estoup 2007). The scoring of alleles at these loci was verified to ensure this was not an artefact of scoring or data entry error. Null alleles were identified in four tests at the loci Ma11 (for CC1 and MR1), MS22 (MR1) and Ma03

(BC2).

Chapter 5.0 ∙ The Rainbow Connection 189

Table 5.1 Gene diversity indices for M. s. inornata collected for microsatellite and stable isotope analyses in the upper

Mitchell catchment. Deviations from HWE are denoted by *, after sequential Bonferroni correction.

MS40 MS37 Site Sample Average Allelic Site H H P H H P Code Size Diversity E O E O Cooktown Crossing (Dry) CC1 30 0.644 0.890 0.767 0.272 0.785 0.767 0.429 Rifle Creek (Dry) RC1 16 0.728 0.823 0.813 0.841 0.752 0.667 0.124 McLeod River (Dry) MR1 30 0.620 0.879 0.733 0.013 0.755 0.800 0.133 Bushy Creek (Dry) BC1 29 0.682 0.832 0.893 0.980 0.777 0.862 0.896 Cooktown Crossing (During wet) CC2 28 0.704 0.834 0.893 0.800 0.801 0.643 0.053 Rifle Creek (During wet) RC2 24 0.722 0.897 0.783 0.217 0.817 0.708 0.470 McLeod River (During wet) MR2 25 0.706 0.859 0.917 0.758 0.793 0.720 0.102 Bushy Creek (During wet) BC2 30 0.640 0.840 0.800 0.444 0.797 0.800 0.904

Continued over page...

Chapter 5.0 ∙ The Rainbow Connection 190

Ma04 Ma11 Ma12 MS22 Site H H P H H P H H P H H P Code E O E O E O E O CC1 0.472 0.467 0.762 0.958 0.769 0.000* 0.584 0.519 0.687 0.713 0.579 0.199 RC1 0.651 0.563 0.473 0.979 0.857 0.057 0.515 0.267 0.005 0.765 0.714 0.023 MR1 0.533 0.500 0.199 0.972 0.786 0.000* 0.468 0.414 0.132 0.776 0.385 0.000* BC1 0.507 0.448 0.072 0.971 0.815 0.021 0.507 0.556 1.000 0.721 0.636 0.312 CC2 0.732 0.786 0.872 0.960 0.826 0.006 0.660 0.667 0.983 0.790 0.609 0.050 RC2 0.526 0.522 0.604 0.962 0.889 0.271 0.522 0.500 0.272 0.758 0.667 0.585 MR2 0.558 0.522 0.040 0.954 0.810 0.004 0.425 0.304 0.110 0.794 0.793 1.000 BC2 0.564 0.600 0.435 0.964 0.840 0.025 0.566 0.655 0.691 0.778 0.519 0.029

Ma03 Ma09 Ma06 Site H H P H H P H H P Code E O E O E O CC1 0.543 0.444 0.051 0.427 0.400 1.000 0.727 0.714 0.899 RC1 0.189 0.200 1.000 0.570 0.667 0.520 0.712 0.688 0.122 MR1 0.448 0.375 0.339 0.494 0.533 0.537 0.758 0.630 0.404 BC1 0.504 0.320 0.016 0.461 0.407 0.664 0.716 0.714 0.402 CC2 0.278 0.174 0.049 0.502 0.429 0.490 0.743 0.714 0.148 RC2 0.297 0.067 0.004 0.503 0.409 0.422 0.766 0.833 0.700 MR2 0.822 0.200 0.003 0.505 0.560 0.805 0.736 0.640 0.200 BC2 0.305 0.074 0.000* 0.472 0.433 0.706 0.641 0.567 0.614

Chapter 5.0 ∙ The Rainbow Connection 191

An AMOVA showed that all the genetic variation was contained within populations, and seasonal differences had no effect on FST estimates (Table 5.2). Additionally, a global FST value was calculated for each season, determining that genetic structure remains constant, panmictic and non-significant across seasons. No significant structure was found at any level of the analysis, indicating panmixia among populations.

Table 5.2 AMOVA calculations for M. s. inornata. No fixation indices were significant

(α = 0.05). ° denotes % variation within populations.

Microsatellites

% Variation FST P Among dry season - 0.000 0.517 populations Among wet season - 0.000 0.507 populations

Among seasons 0.170 0.002 0.213

Within seasons, among 0.000 0.000 0.501 populations

Among all populations 99.820° 0.002 0.420

Assignment indices revealed no bias in dispersal for either the dry season or wet season groups. A bias between two groups would produce lower mAI and vAI for dispersing populations. Although both the mAI (0.065 for the population sampled before the wet season; -0.064 for the population sampled during the wet season; P = 0.785) and the vAI

(12.098 for the population sampled during the dry season; 11.425 for the population sampled during the wet season; P = 0.877) were lower for the wet season individuals, these values were not significantly different from one another (Table 5.3).

Chapter 5.0 ∙ The Rainbow Connection 192

Table 5.3 mAI and vAI for each group sampled before and during the wet season. All values were non-significant (α = 0.05).

Corrected Assignment Indices mAI vAI Dry Season 0.065 12.098 Wet Season -0.064 11.425

A power test (which used the effective population size (Ne) in the analysis) showed that the microsatellite markers used were able to detect FST values of 0.0035 or larger, more than

98.3% of the time. Using NeEstimator v1.3 (Ovenden et al. 2007) the effective population size was calculated to be 140.6 (upper and lower 95% confidence limits were 115.2 and

177.6 respectively).

5.3.3 Stable isotope analyses

Although there was the unavoidable use of two different methods of tissue preparation (i.e. pure muscle for larger adults; tissue inclusive of skin and bones for small juveniles) for the two different size classes of fish (see Chapter 2.3.6), both methods produced isotope signatures within the expected range for each site. Additionally, given the range of size classes, each site was grouped into juveniles (<38mm) and adults (>38mm) according to the specifications described in Pusey et al. (2004) to investigate whether the isotope signatures varied between these age classes (see Table 7.15 for individual standard lengths).

At some sites, significant differences were found between δ13C or δ15N signatures of adults and juveniles – a reflection of the ontogenetic changes in diet and trophic level of this species (see Table 7.16 for full details). All individuals were represented in a dual isotope

Chapter 5.0 ∙ The Rainbow Connection 193 plot (Figure 5.2). This graphical representation revealed some differences between sites.

Most notably, Bushy Creek was particularly enriched in δ15N relative to all other sites

(highest and lowest δ15N value at 8.9 and 4.6 respectively in the total nitrogen range (NR) of 8.9 - 0), and Cooktown Crossing had a comparatively wide distribution of δ13C relative to all other sites (highest and lowest δ13C value at 10.3 and 1.0 respectively in the total carbon range (CR) of 10.3 – 0) (Table 5.4). The McLeod River and Rifle Creek sites were similar in their isotope distributions; however, Rifle Creek was slightly more enriched in

δ15N. In summary, the site differences were suitably differentiated to aid in the identification of migrants.

Figure 5.2 δ13C and δ15N for M. s. inornata at all sites in the upper Mitchell catchment.

Each data point represents a single individual.

Chapter 5.0 ∙ The Rainbow Connection 194

Table 5.4 δ13C range (CR) and δ15N range (NR) for all individuals, dry season individuals and wet season individuals.

CR NR All individuals, all seasons 10.3 8.9 Dry season 7.6 7.8 Wet season 8.7 7.4

Table 5.5 P-values for the Euclidean distance between centroids (MD) for dry and wet season M. s. inornata individuals (α = 0.05). * denotes significant values.

Site P Bushy Creek 0.001* Cooktown Crossing 0.001* McLeod River 0.877 Rifle Creek 0.001*

To examine the fluctuation of isotope signals from dry season to the wet season, each individual was then grouped according to the time of capture (i.e. dry season versus wet season) (Figure 5.3). This grouping identified a trend where sources of δ13C changed between sampling times. For all sites, except the McLeod River, a significant difference was found between sites over the wet season using Euclidean distance between the centroid for each site (MD) (a positive change towards more enriched δ13C signatures) (Table 5.5).

A test of eccentricity was used to identify whether the dispersion of individual signatures from each site changed between seasons. There was no significant difference, indicating that the spread of signatures remained similar in both sampled seasons (Table 5.6).

Overall, the δ13C range (CR) for all individuals was 10.3; δ15N range (NR) for all individuals was 8.9 (Table 5.4). For each season, these values were more specific, however,

Chapter 5.0 ∙ The Rainbow Connection 195 the CR for the wet season individuals did increase from 7.6 to 8.7 whereas the NR did not markedly differ (7.8 to 7.4). This result indicated that the trophic position of M. s. inornata did not change during the wet season, but the sources of dietary carbon did expand.

Table 5.6 P-values for tests of eccentricity (E) for each site, compared across seasons (α

= 0.05).

Site P Bushy Creek 0.225 Cooktown Crossing 0.747 McLeod River 0.255 Rifle Creek 0.905

Chapter 5.0 ∙ The Rainbow Connection 196

From a visual inspection of the distributions of each group, several individuals were identified as being potential migrants (i.e. outliers in the data). Two methods were used to evaluate the dispersion of individuals at each site for each season. Euclidean distance from the sample centroid (CD) and the Euclidean distance from each individual’s nearest neighbour (NND) were calculated to detect changes between the dry and wet seasons. For each of these tests, each individual was found to be within the expected distribution of values (i.e. no migrants were detected), with non-significant results for each site, at each time (Table 5.7).

Table 5.7 P-values for Euclidean distances from the mean sample centroid (CD), and individual’s nearest neighbour (NND) (α = 0.05).

P Site CD NND Bushy Creek 0.237 0.303 Cooktown Crossing 0.131 0.552 McLeod River 0.100 0.590 Rifle Creek 0.703 0.471

Chapter 5.0 ∙ The Rainbow Connection 197

5.4 Discussion

5.4.1 Genetic differentiation

The analysis of molecular variance suggested that all the sampled populations were panmictic. Given the proximity of the sites within a relatively small area of the catchment, it is likely that these populations are freely mixing, and that contemporary gene flow is high. The mAI and vAI did not show any change in rates of dispersal between seasons. So while these results did not reveal specific individual movement that occurred between

October (2009) and March (2010), there is evidence that dispersal is certainly taking place.

Although microsatellites were unable to identify any specific genetic immigrants, these results do not conclude that dispersal during the wet season is not occurring as the high levels of contemporary gene flow and the lack of genetic structure suggest that dispersal/migration is frequent. Consequently, the hypothesis that migrant individuals would be detected by microsatellite analyses was refuted. Genetic analyses were not able to distinguish dispersal events during the selected time frame, and this may be for several reasons. Firstly, movement of individuals during the wet season may be migratory only.

This migration may also happen during the initial flows of the wet season, after which individuals may return quickly. If this is the case, then the sampling time during the wet season may have failed to capture this migratory event. Secondly, the 2009/2010 wet season was abnormally delayed in its onset, experienced less rainfall, and had a shorter duration (personal communication G. Brady, P. Brown, D. Verri). Consequently, there may have been ramifications in providing the cues that fish potentially rely on to initiate dispersal/migration. It is possible that a greatly reduced number of M. s. inornata Chapter 5.0 ∙ The Rainbow Connection 198 individuals chose to move away from their home population for the season of interest which may have produced abnormal results.

And finally, in order for the analyses to detect migrants, the migrants must be genetically identifiable. As the high levels of gene flow between these four populations have resulted in effectively a single population, any actual migrant (that is, a migrant from within the population) would not be identified as a genetic migrant, as it would bear the same genes.

In order for a genetic migrant to be detected, the individual would have to originate from a genetically differentiated population to the one that was sampled. Clearly, the geographic scale of the study was not sufficient to capture two or more genetically distinct populations.

In this sense, this approach was not successful in identifying actual migrants, but it was still able to encapsulate the high levels of dispersal between the sampled sites, and presumably beyond. Using stable isotopes, the nuances of migration were determined; however the genetic markers were more effective in showing the overall patterns of dispersal within this region.

5.4.2 Isotopic differentiation

The δ15N enrichment observed in the site Bushy Creek, is most likely due to the effects of land use in the region. Bushy Creek was the only site located within an agricultural area; specifically, the eastern riparian zone onwards was devoted to grazing cattle, and the western riparian zone onwards was under continual sugarcane cropping. Furthermore, a barramundi aquaculture farm is stationed upstream from this site. It seems that the collective input of fertilisers, chemical sprays, bovine faeces, and waste from the fish farm is likely to be responsible for the enriched δ15N signatures at this site. Many studies have Chapter 5.0 ∙ The Rainbow Connection 199 shown how landuse and fertilisers can enrich nitrogen isotope signatures in streams

(Bedard-Haughn et al. 2003; Deutsch & Voss 2006). Given the design of this study, and the benefits of having isotopically differentiated sites, this δ15N enrichment at Bushy Creek was beneficial. Indeed, Hadwen & Arthington (2007) showed how anthropogenically 15N- enriched sites can be useful in detecting new recruits to fish populations.

It was also observed that Cooktown Crossing had the broadest range of carbon sources.

Cooktown Crossing was the most downstream site (refer to Figure 5.2). Despite differences among sites in the positioning and dispersion of δ15N and δ13C signatures, the absence of differences between the dry and wet seasons suggest that the dispersion of signatures was not influenced by the periods of high flow during the wet season. For all sites except the

McLeod River, there was a positive shift from dry to wet season where the δ13C signatures of fish became more enriched.

Current stable isotope food web research being undertaken in the Mitchell River has identified a specific floodplain epiphyton signature of -22.8 ± 2.6 for δ13C (Jardine et al. in press). Given the shift of fish signatures towards these more enriched δ13C signatures, it is likely that they have migrated onto the floodplains to feed during the flood, and then returned to this region of the Mitchell River catchment. The significant positive shift in observations between seasons (excluding the McLeod River) certainly indicates that there is an increased contribution of floodplain δ13C sources during the wet season. This interpretation was supported by the increased range in diet (CR).

There are two possible interpretations of this result. First, fish could have migrated to the floodplain to feed, thus shifting their stable isotope signatures towards that of the Chapter 5.0 ∙ The Rainbow Connection 200 floodplain. Second, it is possible that the floodplain δ13C sources were transported into the river system at the end of the flood, and are therefore changing the signatures of the fish. In summary, it is uncertain whether it is the fish or the δ13C that is dispersing through the system, but it is clear that floodplain sources of carbon are reflected in the fish stable isotope signatures. This finding has important ramifications for the management of this species. Floodplain environments are evidently an important contributor to the diets of M. s. inornata, and are potentially important habitats as well. If changes in flow regime (e.g. influence of climate change or river regulation) were imposed on the Mitchell River, presumably floodplains would become abnormally disconnected from the main river system. Given the importance of floodplain food sources, it is likely that there will be serious ecological consequences for populations of M. s. inornata in changed conditions.

It is possible that migrants were captured, but were unable to be identified for several different reasons. Firstly, it may have been that the environment from which migrants came was isotopically the same as the new site in which they were collected. While the sites included in the study were isotopically differentiated, it is possible that migrants from outside these sites (i.e. those from the floodplain) carry the same isotopic signatures, and are therefore unable to be linked to a particular source population. Another explanation for the apparent lack of migrants is that dispersal or migration may occur in response to the first rainfalls of the wet season, and not continue after this. If this is the case, then a sufficient amount of time may have elapsed for the signatures to meld to the resident population (i.e. the site signatures). However, the residency time for C and N in this species is unknown, although it is estimated to be between two and eight months for fish tissue

(McCarthy & Waldron 2000; Rubenstein & Hobson 2004). A final interpretation of these results is that dispersal is not occurring at all during the wet season, and that populations Chapter 5.0 ∙ The Rainbow Connection 201 are remaining within the bounds of their territory; however, the levels of gene flow suggest that movement is quite frequent in this species.

Chapter 5.0 ∙ The Rainbow Connection 202

5.5 Conclusion

The use of nine microsatellite loci determined four sampled populations of M .s. inornata in the Mitchell River headwaters to be genetically panmictic, both temporally and spatially.

Also, the use of corrected assignment indices showed that the rate of dispersal did not change between October 2009 and March 2010. These results suggest that small-scale dispersal events between populations of M. s. inornata are frequent. However, panmixia made it difficult to identify actual immigrants to the sampled populations using genetic markers, as the power to detect individual movement was limited by the low differentiation among the populations.

While the stable isotope data suggested that individuals were migrating to the floodplain during the wet season, it was also possible that floodplain sources of 13C and 15N were being transported downstream to the sampled sites. For either interpretation, it is evident that floodplains in the upper Mitchell River catchment are a significant food resource for populations of M. s. inornata, and the connectivity of these floodplains should be maintained. However, more conclusive evidence would come from artificially labelling 15N or 13C to trace the exact pathways of movement. Additionally, sampling at regular intervals throughout the wet season would specify the time frame for dispersal, and ensure that the isotope turnover would not compromise the integrity of the isotope signatures.

By applying both molecular and stable isotope techniques to estimate dispersal, a greater understanding of the ecological influence of seasonal flow and the levels of population connectivity were reached. By integrating these methods, the scope and depth of movement was more fully quantified in a temporal sense. Using microsatellites, this study showed that Chapter 5.0 ∙ The Rainbow Connection 203 effective dispersal (i.e. dispersal that resulted in gene flow) rates were high within the study region. The stable isotope analysis was able to discriminate the migratory events (i.e. movement that may not have resulted in gene flow) that were stimulated by flooding. These findings show that the combined application of genetics and stable isotopes signatures in estimating patterns of movement are an approach worthy of further scrutiny, and could be tested on a broader range of situations and environments.

Chapter 5.0 ∙ The Rainbow Connection 204

6.0 Conclusion

This study investigated how dispersal and connectivity has led to the genetic structure of M. australis, M. s. inornata and H. fuliginosus in tropical Australia. By using genetic markers, the interactions between the environment and dispersal were characterised. The coupling of stable isotope analyses with a microsatellite data set revealed a greater depth of understanding in patterns of movement over the wet season in M. s. inornata.

6.1 Patterns of gene flow in tropical catchments

One of the key findings of this thesis was the different genetic responses to environmental influences between species. For rainbowfish, a small species of lesser dispersal capability, almost all of the tested obstacles to dispersal inhibited gene flow. Alternatively, for H. fuliginosus, a large, capable disperser, gene flow was almost unrestricted by environmental factors. These results highlighted the importance of determining how the environment impacts on gene flow in rivers, but also how the different dispersal capability of species influences movement throughout a catchment. From a management perspective, the differing dispersal capabilities, and therefore gene flow, of these species highlights the fact that river management needs to account for and accommodate these differences, especially with respect to any future changes in the environment and changes to barriers to dispersal

(both natural and artificial).

Chapter 6.0 ∙ Conclusion 205

While this study was by no means a full evaluation of patterns of gene flow in the tropical rivers of Australia, it certainly contributes to the knowledge of aquatic populations in these regions. Small species, including Pacific blueyes (Psuedomugil signifier), purple-spotted gudgeons (Mogurnda adspersa), eastern rainbowfish (Melanotaenia splendida splendida), western rainbowfish (in the Kimberley region) (Melanotaenia australis) and fly-specked hardyhead (Craterocephalus stercusmuscarum) are generally structured according to streams and rivers within catchments, and are also sensitive to physical barriers to dispersal

(see Wong et al. 2004; Hurwood & Hughes 1998; McGlashan et al. 2001; Phillips et al.

2009; McGlashan & Hughes 2000). From previous studies of small, tropical fish, the

Melanotaenia spp. of the Daly and Mitchell Rivers aligned well with the patterns observed in similar species – following that of the metapopulation model or IBD model, rather the

SHM. To date, the low genetic structure observed in H. fuliginosus was relatively rare in the literature. The only comparable example is that of barramundi (Lates calcarifer), a catadromous species, which also exhibits low genetic structure across its range (Shaklee &

Salini 1985). In the current state of knowledge, the sooty grunter is decidedly different to most tropical species, although it was found to be mildly influenced by IBD.

In comparison to other regions in Australia, gene flow in tropical rivers is more complex and variable among species. For instance, in inland river systems, populations tend to have low genetic diversity, but high genetic structure among catchments. In the more southern parts of Australia, such examples can be seen in Hyrtl’s tandan (Neosilurus hyrtlii), the north-west glassfish (Ambassis sp.) and golden perch (Macquaria ambigua) (see Huey et al. 2008; Huey et al. 2011; Faulks et al. 2010(b)). However, while species in tropical

Australia tend to have high levels of structure and diversity, it is clear that further studies in this region are needed to fully characterise patterns of gene flow in resident populations. Chapter 6.0 ∙ Conclusion 206

This study highlighted the complexity of dispersal in tropical riverine fish, and showed that a ‘one size fits all’ approach to management may not be optimal for the incredible variety of species.

Furthermore, this study also highlighted that variation exists between catchments, and that generalisations cannot be made regarding the connectivity of populations within all tropical catchments. For instance, comparisons of FST values between catchments for both

Melanotaenia spp. and H. fuliginosus showed that populations in the Daly catchment were more structured. This may be a reflection of the differences in dendritic structure between the two catchments (where the Daly River shows more hierarchical branching). The influence of other environmental parameters (e.g. elevation) varied between species and catchments, stressing the influence of differing dispersal capabilities within the same environment.

6.2 Migration in tropical catchments

While genetic markers were able to detect dispersal within the Daly and Mitchell catchments, such studies do not necessarily incorporate the intricacies of migration

(Bohonak 1999). The use of stable isotope analyses in a part of the Mitchell River showed that migration to the floodplains during the wet season is a key behaviour for M. s. inornata. However, it is not known whether migration is also carried out by populations in the rest of the catchment, although it is expected. Furthermore, it is unclear whether M. australis and H. fuliginous also migrate during certain times of the year. Migration patterns are important to identify, especially if such requirements as feeding and breeding habitat are highly subsidised by floodplains (Humphries et al. 1999). Chapter 6.0 ∙ Conclusion 207

This is one of the greatest difficulties with genetic markers – they do not reveal migratory events if they do not include reproduction with different populations (Epps et al. 2007;

Broquet & Petit 2009). So while this study was able to determine dispersal and population connectivity through gene flow, these may not truly reflect actual dispersal and migration events. To fully clarify dispersal from historical through to contemporary times, further use should be made of stable isotopes and/or other tracking methods, such as telemetry, mark- recapture or otolith microchemistry (Webster et al. 2002; Nathan et al. 2003; Rubenstein &

Hobson 2004).

6.3 Management implications for Melanotaenia spp.

This study highlighted several specific management implications for the melanotaenids in the Daly and Mitchell Rivers. In most of the ‘outlier’ populations (detected through the

DPR analyses), genetic drift was found to be more influential than gene flow (Pattern 2)

(Koizumi et al. 2006). Consequently, this needs to be taken into consideration for this species management for two reasons. First, the dispersal distance may be inadequate to supplement (or recolonise in the event of extinction) highly isolated populations, and so care must be taken to maintain the connectivity of populations, and ensure habitat/population fragmentation does not exceed the sustainable limits of population persistence (Keyghobadi 2007). Ultimately, reduced gene flow should result in a divergent population due to the effects of drift, mutation and selection (Crandall et al. 2000;

Segelbacher et al. 2010). Secondly, if an isolated population is extirpated, recolonisation will be difficult and a divergent lineage contributing to the genetic diversity of the population will be lost (Blanchet et al. 2010).

Chapter 6.0 ∙ Conclusion 208

By understanding how past and present populations are affected by the landscape, predictions about population consequences can be made with respect to future environmental change (Manel et al. 2003). While landscape genetic studies can illuminate the features that have led to current population structuring, they can also play an important role in conservation genetics for the future (Manel et al. 2003; Allendorf et al. 2010). The use of assignment testing provided evidence that a source-sink dynamic within catchments potentially complicates the appropriate management of both M. australis and M. s. inornata.

Populations composed of a number of sources and sinks are particularly susceptible and sensitive to extirpation, particularly in scenarios where there are few source populations, and many sinks (Dias 1996). The risk is concentrated – a species composed of six self- sustaining populations is comparatively buffered against extinction when compared to six populations, only two of which are sources. In many cases, the loss of a population is not necessarily of great impact to the overall sustainability of the species if the population in question is self-sustaining (Gaggiotti 1996). However, the loss of a source population is of great concern, as it effectively is the additional loss of a number of populations that are

‘fed’ by the source (Crowder et al. 2000; Lowe & Allendorf 2010). It is imperative that the source populations identified in both M. australis and M. s. inornata are recognised in the management of the Daly and Mitchell Rivers to minimise and control the risk associated with source-sink metapopulations.

Chapter 6.0 ∙ Conclusion 209

6.4 Management implications for H. fuliginosus

Due to the strong dispersal capability of this species, it seems that landscape barriers present very little challenge to the movement of H. fuliginosus. The high levels of gene flow are positive for the species in terms of maintaining the genetic diversity contained within the populations (Allendorf 1983; Ficetola & Bonin 2011). Additionally, in the event of population extirpation or decline, recolonisation or recovery would be relatively fast and highly likely (Slatkin 1977; Hughes 2007). Individuals dispersing from surrounding areas would seemingly have very little to encumber their transition into a new site. However, while recolonisation is likely if a population is lost, this should not be seen as a safeguard for inappropriate management strategies.

The proposed water extraction from the tropical catchments of Australia would have genetic implications for dispersal in this species, and would certainly place the whole population under pressure (Hart 2004; Blanch 2008; Chan et al. 2010; Stewart-Koster et al.

2011). Reductions in hydrological connectivity would not support the high levels of dispersal in this species; ultimately, such a practice would result in population isolation

(Hughes et al. 2009). Reduced gene flow would see the erosion of genetic diversity within these separated populations, decreasing the ability of this species to adapt to environmental change (Carvalho 1993; Hoffmann & Willi 2008). Additionally, while H. fuliginosus is a strong disperser, it is important to acknowledge that barriers may be impacting on the dispersal of these populations. However, these barriers may have only recently arisen, which may not yet have had time to impact on the genetic structure of these populations.

Further studies of this species, using microsatellites, would show how the contemporary environment contributes to patterns of dispersal and gene flow in H. fuliginosus Chapter 6.0 ∙ Conclusion 210 populations. Management and conservation strategies should consider this when planning for the sustainability of this species.

6.5 Management implications for fish in tropical Australia

By monitoring the levels of genetic diversity and structure in these species and catchments, the impacts of environmental stressors may be identified (Leary & Allendorf 1989). The loss of genetic diversity can be a lengthy process, and by the time a decline in genetic health is detected, it is rather retrospective (Hoffmann & Hercus 2000). For example, it may take many generations before the impact of a dam or weir is detected through the use of neutral genetic markers (Deiner et al. 2007; Clemento et al. 2009). In order to quickly and effectively identify selection pressures from environmental change, loci that are subjected to selection are perhaps more appropriate for genetic monitoring in the aquatic populations of northern Australia (Hoffmann & Willi 2008; Gebremedhin et al. 2009). The use of stable isotopes or tagging can give an immediate indication of whether this obstruction is a barrier to dispersal within a single generation (Charles et al. 2006).

However, it is important to remember that sustainable levels of gene flow can be achieved with only a few individuals (Couvet 2002) – individuals that may not be detected using other markers of dispersal (Hobson 2005), and so caution is also required when interpreting the results of traditional mark-recapture approaches. For optimal monitoring of aquatic populations in tropical Australia, both genetic markers and other tracking approaches should be employed to determine the immediate and ongoing consequences of both natural and artificial environmental change.

Chapter 6.0 ∙ Conclusion 211

References

Akaike H (1974) A new look at the statistical model identification. IEEE Transactions on

Automatic Control 19, 716-723.

Allen GR, Midgley SH, Allen M (2002) 'Field guide to the freshwater fishes of Australia.'

(Western Australian Museum: Perth).

Allendorf FW & Phelps SR (1981) Use of allelic frequencies to describe population

structure. Canadian Journal of Fisheries and Aquatic Sciences 38, 1507-1514.

Allendorf FW (1983) Isolation, gene flow, and genetic differentiation among populations.

In ‘Genetics and conservation: A reference for managing wild animal and plant

populations’. (Eds CM Schonewald-Cox, SM Chambers, B MacBryde, WL Thomas)

pp. 51-65. (Benjamin Cummings: Menlo Park).

Allendorf FW, Hohenlohe PA, Luikart G (2010) Genomics and the future of conservation

genetics. Nature Reviews Genetics 11, 697-709.

Allouche S (2002) Nature and functions of cover for riverine fish. Bulletin Français de la

Pêche et de la Pisciculture 365-366, 297-324.

References 212

Almodóvar A, Suárez J, Nicola GG, Nuevo M (2001) Genetic introgression between wild

and stocked brown trout in the Douro River basin, Spain. Journal of Fish Biology 59,

68-74.

Aló D & Turner TF (2005) Effects of habitat fragmentation on effective population size in

the endangered Rio Grande silvery minnow. Conservation Biology 19, 1138-1148.

Altshuler DL & Dudley R (2006) Adaptations to life at high elevation: An introduction to

the symposium. Integrative and Comparative Biology 46, 3-4.

Amoros C & Bornette G (2002) Connectivity and biocomplexity in waterbodies of riverine

floodplains. Freshwater Biology 47, 761-776.

Amos W & Harwood J (1998) Factors affecting levels of genetic diversity in natural

populations. Philosophical Transactions of the Royal Society of London Series B-

Biological Sciences 353, 177-86.

Amos W & Balmford A (2001) When does conservation genetics matter? Heredity 87, 257-

265.

Anderson E & Stebbins Jr GL (1954) Hybridization as an evolutionary stimulus. Evolution

8, 378-388.

References 213

Anderson CD, Epperson BK, Fortin MJ, Holderegger R, James PM, Rosenberg MS,

Scribner KT, Spear S (2010) Considering spatial and temporal scale in landscape-

genetic studies of gene flow. Molecular Ecology 19, 3565-3575.

Anne C (2006) Choosing the right molecular genetic markers for studying biodiversity:

from molecular evolution to practical aspects. Genetica 127, 101-120.

Aris-Brosou S & Excoffier L (1996) The impact of population expansion and mutation rate

heterogeneity on DNA sequence polymorphism. Molecular Biology and Evolution 13,

494.

Arthington AH, Balcombe SR, Wilson GA, Thoms MC, Marshall J (2005) Spatial and

temporal variation in fish-assemblage structure in isolated waterholes during the 2001

dry season of an arid-zone floodplain river, Cooper Creek, Australia. Marine and

Freshwater Research 56, 25-35.

Arthington AH, Naiman RJ, McClain ME, Nilsson C (2010) Preserving the biodiversity

and ecological services of rivers: new challenges and research opportunities.

Freshwater Biology 55, 1-16.

Avise JC, Neigel JE, Arnold J (1984) Demographic influences on mitochondrial DNA

lineage survivorship in animal populations. Journal of Molecular Evolution 20, 99-

105.

References 214

Avise JC (1986) Mitochondrial-DNA and the Evolutionary Genetics of Higher Animals.

Philosophical Transactions of the Royal Society of London Series B-Biological

Sciences 312, 325-342.

Avise JC, Arnold J, Ball RM, Bermingham E, Lamb T, Neigel JE, Reeb CA, Saunders NC

(1987) Intraspecific phylogeography: the mitochondrial DNA bridge between

population genetics and systematics. Annual Review of Ecology and Systematics 18,

489-522.

Avise JC (1989) A role for molecular genetics in the recognition and conservation of

endangered species. Trends in Ecology & Evolution 4, 279-281.

Avise JC (1995) Mitochondrial DNA polymorphism and a connection between genetics

and demography of relevance to conservation. Conservation Biology 9, 686-690.

Avise JC (2009) Phylogeography: retrospect and prospect. Journal of Biogeography 36, 3-

15.

Balcombe SR, Bunn SE, McKenzie Smith FJ, Davies PM (2005) Variability of fish diets

between dry and flood periods in an arid zone floodplain river. Journal of Fish

Biology 67, 1552-1567.

Balcombe SR, Bunn SE, Arthington AH, Fawcett JH, McKenzie Smith FJ, Wright A

(2007) Fish larvae, growth and biomass relationships in an Australian arid zone river:

links between floodplains and waterholes. Freshwater Biology 52, 2385-2398. References 215

Ballard JWO & Kreitman M (1995) Is mitochondrial DNA a strictly neutral marker?

Trends in Ecology & Evolution 10, 485-488.

Balloux F & Lugon-Moulin N (2002) The estimation of population differentiation with

microsatellite markers. Molecular Ecology 11, 155-165.

Balloux F (2010) The worm in the fruit of the mitochondrial DNA tree. Heredity 104, 419-

420.

Bamshad M & Wooding SP (2003) Signatures of natural selection in the human genome.

Nature Reviews Genetics 4, 99-111.

Barber PH, Moosa MK, Palumbi SR (2002) Rapid recovery of genetic diversity of

stomatopod populations on Krakatau: temporal and spatial scales of marine larval

dispersal. Proceedings of the Royal Society B: Biological Sciences 269, 1591-1597.

Barlaup BT, Gabrielsen SE, Skoglund H, Wiers T (2008) Addition of spawning gravel—a

means to restore spawning habitat of atlantic salmon (Salmo salar L.), and

anadromous and resident brown trout (Salmo trutta L.) in regulated rivers. River

Research and Applications 24, 543-550.

Barnes C, Jennings S, Polunin NVC, Lancaster JE (2008) The importance of quantifying

inherent variability when interpreting stable isotope field data. Oecologia 155, 227-

235.

References 216

Barton NH (2000) Genetic hitchhiking. Philosophical Transactions of the Royal Society of

London. Series B: Biological Sciences 355, 1553-1562.

Bay L, Caley MJ, Crozier R (2008) Meta-population structure in a coral reef fish

demonstrated by genetic data on patterns of migration, extinction and re-colonisation.

BMC Evolutionary Biology 8, 248.

Bayley PB (1995) Understanding large river floodplain ecosystems. BioScience 45, 153-

158.

Bedard-Haughn A, Van Groenigen JW, Van Kessel C (2003) Tracing 15N through

landscapes: potential uses and precautions. Journal of Hydrology 272, 175-190.

Beebee T & Rowe G (2008) 'An introduction to molecular ecology.' 2nd Edn (Oxford

University Press: New York).

Bélisle M (2005) Measuring landscape connectivity: the challenge of behavioral landscape

ecology. Ecology 86, 1988-1995.

Beneteau CL, Mandrak NE, Heath DD (2009) The effects of river barriers and range

expansion of the population genetic structure and stability in Greenside Darter

(Etheostoma blennioides) populations. Conservation Genetics 10, 477-487.

References 217

Bennett KD (2004) Continuing the debate on the role of Quaternary environmental change

for macroevolution. Philosophical Transactions of the Royal Society of London.

Series B: Biological Sciences 359, 295-303.

Berry O, Tocher MD, Sarre SD (2004) Can assignment tests measure dispersal? Molecular

Ecology 13, 551-561.

Bishop KA (1980) Fish kills in relation to physical and chemical changes in Magela Creek

(East Alligator River system, Northern Territory) at the beginning of the tropical wet

season. Australian Zoologist 20, 485-500.

Björklund M, Bergek S, Ranta E, Kaitala V (2010) The effect of local population dynamics

on patterns of isolation by distance. Ecological Informatics 5, 167-172.

Blanch S, Rea N, Scott G (2005) Aquatic Conservation Values of the Daly River

Catchment, Northern Territory, Australia. WWF-Australia, Charles Darwin

University and the Environment Centre NT, Darwin.

Blanch S (2008) Steps to a sustainable Northern Australia. Ecological Management &

Restoration 9, 110-115.

Blanchet S, Rey O, Etienne R, Lek S, Loot G (2010) Species specific responses to

landscape fragmentation: implications for management strategies. Evolutionary

Applications 3, 291-304.

References 218

Bohonak AJ (1999) Dispersal, Gene Flow, and Population Structure. The Quarterly Review

of Biology 74, 21-45.

Bond NR & Lake PS (2003) Characterizing fish–habitat associations in streams as the first

step in ecological restoration. Austral Ecology 28, 611-621.

Bossart JL & Pashley Prowell D (1998) Genetic estimates of population structure and gene

flow: Limitations, lessons and new directions. Trends in Ecology & Evolution 13,

202-206.

Bottrill MC, Joseph LN, Carwardine J, Bode M, Cook C, Game ET, Grantham H, Kark S,

Linke S, McDonald-Madden E, Pressey RL, Walker S, Wilson KA, Possingham HP

(2008) Is conservation triage just smart decision making? Trends in Ecology &

Evolution 23, 649-654.

Boulet M, Gibbs HL, Hobson KA (2006) Integrated analysis of genetic, stable isotope, and

banding data reveal migratory connectivity and flyways in the northern yellow

warbler (Dendroica petechia; aestiva group). Ornithological Monographs 61, 29-78.

Bowler DE & Benton TG (2005) Causes and consequences of animal dispersal strategies:

relating individual behaviour to spatial dynamics. Biological Reviews 80, 205-225.

Brodie ED (2007) Population size is not genetic quality. Animal Conservation 10, 288-290.

References 219

Brooks AP, Shellberg JG, Knight J, Spencer J (2009) Alluvial gully erosion: an example

from the Mitchell fluvial megafan, Queensland, Australia. Earth Surface Processes

and Landforms 34, 1951-1969.

Broquet T & Petit EJ (2009) Molecular estimation of dispersal for ecology and population

genetics. Annual Review of Ecology, Evolution, and Systematics 40, 193-216.

Brown JR, Beckenbach AT, Smith MJ (1992) Influence of pleistocene glaciations and

human intervention upon mitochondrial DNA diversity in white sturgeon (Acipenser

transmontanus) populations. Canadian Journal of Fisheries and Aquatic Sciences 49,

358-367.

Bruford MW, Wayne RK (1993) Microsatellites and their application to population genetic

studies. Current Opinion in Genetics & Development 3, 939-43.

Buisson L, Blanc L, Grenouillet G (2008) Modelling stream fish species distribution in a

river network: the relative effects of temperature versus physical factors. Ecology of

Freshwater Fish 17, 244-257.

Bunn SE & Hughes JM (1997) Dispersal and recruitment in streams: evidence from genetic

studies. Journal of the North American Benthological Society, 338-346.

Bunn SE & Arthington AH (2002) Basic principles and ecological consequences of altered

flow regimes for aquatic biodiversity. Environmental Management 30, 492-507.

References 220

Bunn SE, Thoms MC, Hamilton SK, Capon SJ (2006) Flow variability in dryland rivers:

boom, bust and the bits in between. River Research and Applications 22, 179-186.

Burridge CP & Versace VL (2007) Population genetic structuring in Acanthopagrus

butcheri (Pisces: Sparidae): Does low gene flow among estuaries apply to both sexes?

Marine Biotechnology 9, 33-44.

Burridge CP, Craw D, Jack DC, King TM, Waters JM (2008) Does fish ecology predict

dispersal across a river drainage divide? Evolution 62, 1484-1499.

Caballero A & Toro MA (2002) Analysis of genetic diversity for the management of

conserved subdivided populations. Conservation Genetics 3, 289-299.

Carvalho GR (1993) Evolutionary aspects of fish distribution: genetic variability and

adaptation. Journal of Fish Biology 43, 53-73.

Cassone BJ & Boulding EG (2006) Genetic structure and phylogeography of the lined

shore crab, Pachygrapsus crassipes, along the northeastern and western Pacific

coasts. Marine Biology 149, 213-226.

Castric V, Bonney F, Bernatchez L (2001) Landscape Structure and Hierarchical Genetic

Diversity in the Brook Charr, Salvelinus fontinalis. Evolution 55, 1016-1028.

Chan TU, Hart BT, Kennard MJ, Pusey BJ, Shenton W, Douglas MM, Valentine E, Patel S

(2010) Bayesian network models for environmental flow decision making in the Daly References 221

River, Northern Territory, Australia. River Research and Applications. DOI:

10.1002/rra.1456.

Chapman LJ & Kramer DL (1991) The consequences of flooding for the dispersal and fate

of poeciliid fish in an intermittent tropical stream. Oecologia 87, 299-306.

Chapman P & Warburton K (2006) Postflood movements and population connectivity in

gambusia (Gambusia holbrooki). Ecology of Freshwater Fish 15, 357-365.

Chappell J (1991) Late Quaternary environmental changes in eastern and central Australia,

and their climatic interpretation. Quaternary Science Reviews 10, 377-390.

Chapuis MP & Estoup A (2007) Microsatellite null alleles and estimation of population

differentiation. Molecular Biology and Evolution 24, 621-631.

Chaput-Bardy A, Fleurant C, Lemaire C, Secondi J (2009) Modelling the effect of in-

stream and overland dispersal on gene flow in river networks. Ecological Modelling

220, 3589-3598.

Charles K, Roussel JM, Lebel JM, Baglinière JL, Ombredane D (2006) Genetic

differentiation between anadromous and freshwater resident brown trout (Salmo trutta

L.): insights obtained from stable isotope analysis. Ecology of Freshwater Fish 15,

255-263.

References 222

Chen Y, Parmenter S, May B (2007) Introgression Between Lahontan and Endangered

Owens Tui Uhubs, and Apparent Discovery of a New Tui Chub in the Owens Valley,

California. Conservation Genetics 8, 221-238.

Chenoweth SF & Hughes JM (2003) Oceanic interchange and nonequilibrium population

structure in the estuarine dependent Indo Pacific tasselfish, Polynemus sheridani.

Molecular Ecology 12, 2387-2397.

Chetkiewicz CLB, St. Clair CC, Boyce MS (2006) Corridors for conservation: integrating

pattern and process. Annual Review of Ecology Evolution and Systematics 37, 317-

342.

Christiansen FB (1988) The Wahlund effect with overlapping generations. The American

Naturalist 131, 149-156.

Clark AG (1990) Inference of haplotypes from PCR-amplified samples of diploid

populations. Molecular Biology and Evolution 7, 111-122.

Clegg SM, Kelly JF, Kimura M, Smith TB (2003) Combining genetic markers and stable

isotopes to reveal population connectivity and migration patterns in a neotropical

migrant, Wilson's warbler (Wilsonia pusilla). Molecular Ecology 12, 819-30.

Clement M, Posada D, Crandall KA (2000) TCS: a computer program to estimate gene

genealogies. Molecular Ecology 9, 1657-1659.

References 223

Clemento AJ, Anderson EC, Boughton D, Girman D, Garza JC (2009) Population genetic

structure and ancestry of Oncorhynchus mykiss populations above and below dams in

south-central California. Conservation Genetics 10, 1321-1336.

Climer S, Jäger G, Templeton AR, Zhang W (2009) How frugal is mother nature with

haplotypes? Bioinformatics 25, 68-74.

Closs GE & Lake PS (1996) Drought, differential mortality and the coexistence of a native

and an introduced fish species in a south east Australian intermittent stream.

Environmental Biology of Fishes 47, 17-26.

Collares-Pereira MJ & Cowx IG (2004) The role of catchment scale environmental

management in freshwater fish conservation. Fisheries Management and Ecology 11,

303-312.

Congdon BC (1995) Unidirectional gene flow and maintenance of genetic diversity in

mosquitofish Gambusia holbrooki (Teleostei: Poeciliidae). Copeia 1, 162-172.

Cook BD, Bunn SE, Hughes JM (2002) Genetic structure and dispersal of Macrobrachium

australiense (Decapoda: Palaemonidae) in western Queensland, Australia. Freshwater

Biology 47, 2098-2112.

Cook BD, Bunn SE, Hughes JM (2007(a)) Molecular genetic and stable isotope signatures

reveal complementary patterns of population connectivity in the regionally vulnerable

southern pygmy perch (Nannoperca australis). Biological Conservation 138, 60-72. References 224

Cook BD, Bunn S, Hughes J (2007(b)) A comparative analysis of population structuring

and genetic diversity in sympatric lineages of freshwater shrimp (Atyidae: Paratya):

concerted or independent responses to hydrographic factors? Freshwater Biology 52,

2156-2171.

Cook BD, Kennard MJ, Real K, Pusey BJ, Hughes JM (2011) Landscape genetic analysis

of the tropical freshwater fish Mogurnda mogurnda (Eleotridae) in a monsoonal river

basin: importance of hydrographic factors and population history. Freshwater Biology

56, 812-827.

Cook CL & Coughlin DJ (2010) Rainbow trout Oncorhynchus mykiss consume less energy

when swimming near obstructions. Journal of Fish Biology 77, 1716-1723.

Cooke GM, Chao NL & Beheregaray LB (2009) Phylogeography of a flooded forest

specialist fish from central Amazonia based on intron DNA: the cardinal tetra

Paracheirodon axelrodi. Freshwater Biology 54, 1216-1232.

Coombs JA, Letcher BH, Nislow KH (2008) CREATE: a software to create input files

from diploid genotypic data for 52 genetic software programs. Molecular Ecology

Resources 8, 578-580.

Copp GH (1989) The habitat diversity and fish reproductive function of floodplain

ecosystems. Environmental Biology of Fishes 26, 1-27.

References 225

Cote J & Clobert J (2010) Risky dispersal: avoiding kin competition despite uncertainty.

Ecology 91, 1485-1493.

Couvet D (2002) Deleterious Effects of Restricted Gene Flow in Fragmented Populations.

Conservation Biology 16, 369-376.

Crispo E, Bentzen P, Reznick DN, Kinnison MT, Hendry AP (2006) The relative influence

of natural selection and geography on gene flow in guppies. Molecular Ecology 15,

49-62.

Crossin GT, Hinch SG, Farrell AP, Higgs DA, Lotto AG, Oakes JD, Healey MC (2004)

Energetics and morphology of sockeye salmon: effects of upriver migratory distance

and elevation. Journal of Fish Biology 65, 788-810.

Crow JF & Aoki K (1984) Group Selection for a Polygenic Behavioral Trait: Estimating

the Degree of Population Subdivision. Proceedings of the National Academy of

Sciences 81, 6073-6077.

Crowder LB, Lyman SJ, Figueira WF, Priddy J (2000) Source-sink population dynamics

and the problem of siting marine reserves. Bulletin of Marine Science 66, 799-820.

Cucherousset J, Carpentier A, Paillisson JM (2007) How do fish exploit temporary waters

throughout a flooding episode? Fisheries Management and Ecology 14, 269-276.

References 226

Cushman SA & Landguth EL (2010) Spurious correlations and inference in landscape

genetics. Molecular Ecology 19, 3592-3602.

Cunjak RA, Roussel JM, Gray MA, Dietrich JP, Cartwright DF, Munkittrick KR, Jardine

TD (2005) Using stable isotope analysis with telemetry or mark-recapture data to

identify fish movement and foraging. Oecologia 144, 636-646.

Currat M, Ruedi M, Petit RJ, Excoffier L (2008) The hidden side of invasions: massive

introgression by local genes. Evolution 62, 1908-1920.

Davies BR, Thoms M, Meador M (1992) An assessment of the ecological impacts of inter

basin water transfers, and their threats to river basin integrity and conservation.

Aquatic conservation: Marine and freshwater ecosystems 2, 325-349.

Davies N, Villablanca FX, Roderick GK (1999) Determining the source of individuals:

multilocus genotyping in nonequilibrium population genetics. Trends in Ecology &

Evolution 14, 17-21.

Davis AM, Pusey BJ, Thorburn DC, Dowe JL, Morgan DL, Burrows D (2010) Riparian

contributions to the diet of terapontid grunters (Pisces: Terapontidae) in wet–dry

tropical rivers. Journal of Fish Biology 76, 862-879.

Dawson MN, Staton JL, Jacobs DK (2001) Phylogeography of the tidewater goby,

Eucyclogobius newberryi (Teleostei, Gobiidae), in coastal California. Evolution 55,

1167-1179. References 227

Deiner K, Garza JC, Coey R, Girman DJ (2007) Population structure and genetic diversity

of trout (Oncorhynchus mykiss) above and below natural and man-made barriers in

the Russian River, California. Conservation Genetics 8, 437-454.

Del Mar Delgado M, Ratikainen II, Kokko H (2010) Inertia: the discrepancy between

individual and common good in dispersal and prospecting behaviour. Biological

Reviews 86, 717-732.

DeMarais BD, Dowling TE, Douglas ME, Minckley WL, Marsh PC (1992) Origin of Gila

seminuda (Teleostei: Cyprinidae) through introgressive hybridization: implications

for evolution and conservation. Proceedings of the National Academy of Sciences 89,

2747.

DeNiro MJ & Epstein S (1978) Influence of diet on the distribution of carbon isotopes in

animals. Geochimica et Cosmochimica Acta 42, 495-506.

Deutsch B & Voss M (2006) Anthropogenic nitrogen input traced by means of δ15N values

in macroalgae: Results from in-situ incubation experiments. Science of the Total

Environment 366, 799-808.

Dewoody J, Nason JD, Hipkins VD (2006) Mitigating scoring errors in microsatellite data

from wild populations. Molecular Ecology Notes 6, 951-957.

Dias PC (1996) Sources and sinks in population biology. Trends in Ecology & Evolution

11, 326-330. References 228

Dillon ME, Frazier MR, Dudley R (2006) Into thin air: physiology and evolution of alpine

insects. Integrative and Comparative Biology 46, 49-61.

Dingle H & Drake VA (2007) What is migration? BioScience 57, 113-121.

Dizon AE, Lockyer C, Perrin WF, Demaster DP, Sisson J (1992) Rethinking the stock

concept: a phylogeographic approach. Conservation Biology 6, 24-36.

Dlugosch KM & Hays CG (2008) Genotypes on the move: some things old and some

things new shape the genetics of colonization during species invasions. Molecular

ecology 17, 4583.

Dlugosch KM & Parker IM (2008) Founding events in species invasions: genetic variation,

adaptive evolution, and the role of multiple introductions. Molecular Ecology 17,

431-449.

Dodson JJ, Colombani F & Ng PKL (1995) Phylogeographic structure in mitochondrial

DNA of a South-east Asian freshwater fish, Hemibagrus nemurus (Siluroidei;

Bagridae) and Pleistocene sea-level changes on the Sunda shelf. Molecular Ecology 4,

331-346.

Douglas MR, Brunner PC, Douglas ME (2003) Drought in an evolutionary context:

molecular variability in Flannelmouth Sucker (Catostomus latipinnis) from the

Colorado River Basin of western North America. Freshwater Biology 48, 1254-1273.

References 229

Dowling TE & Secor CL (1997) The role of hybridization and introgression in the

diversification of animals. Annual Review of Ecology and Systematics 28, 593-619.

Doyle JJ & Doyle JL (1987) CTAB DNA extraction in plants. Phytochemical Bulletin 19,

11–15.

Dudgeon D (2005) River rehabilitation for conservation of fish biodiversity in monsoonal

Asia. Ecology and Society 10, 15-35.

Durand JD, Tine M, Panfili J, Thiaw OT, Laë R (2005) Impact of glaciations and

geographic distance on the genetic structure of a tropical estuarine fish, Ethmalosa

fimbriata (Clupeidae, S. Bowdich, 1825). Molecular Phylogenetics and Evolution 36,

277-287.

Echelle AA & Echelle AF (1997) Genetic Introgression of Endemic Taxa by Non natives:

A Case Study with Leon Springs Pupfish and Sheepshead Minnow. Conservation

Biology 11, 153-161.

Edwards A, Civitello A, Hammond HA, Caskey CT (1991) DNA typing and genetic

mapping with trimeric and tetrameric tandem repeats. American Journal of Human

Genetics 49, 746-756.

Ellegren H (1992) Polymerase-chain-reaction (PCR) analysis of microsatellites: a new

approach to studies of genetic relationships in birds. The Auk 109, 886-895.

References 230

Ellegren H (2004) Microsatellites: Simple sequences with complex evolution. Nature

Reviews Genetics 5, 435-445.

Ellegren H & Sheldon BC (2008) Genetic basis of fitness differences in natural

populations. Nature 452, 169-175.

Emerson BC & Hewitt GM (2005) Phylogeography. Current Biology 15, 367-371.

Epps CW, Wehausen JD, Bleich VC, Torres SG, Brashares JS (2007) Optimizing dispersal

and corridor models using landscape genetics. Journal of Applied Ecology 44, 714-

724.

Erskine W, Begg GW, Jolly P, Georges A, O’Grady A, Eamus D, Rea N, Dostine P,

Townsend S, Padovan A (2003) Recommended environmental water requirements for

the Daly River, Northern Territory, based on ecological, hydrological and biological

principles. Supervising Scientist Report 175, National River Health Program,

Environmental Flows Initiative, Technical Report 4, Darwin.

Estes-Zumpf WA, Rachlow JL, Waits LP, Warheit KI (2010) Dispersal, gene flow, and

population genetic structure in the pygmy rabbit (Brachylagus idahoensis). Journal of

Mammalogy 91, 208-219.

Estoup A, Tailliez C, Cornuet JM, Solignac M (1995) Size homoplasy and mutational

processes of interrupted microsatellites in two bee species, Apis mellifera and Bombus

terrestris (Apidae). Molecular Biology and Evolution 12, 1074-1084. References 231

Estoup A, Jarne P, Cornuet JM (2002) Homoplasy and mutation model at microsatellite

loci and their consequences for population genetics analysis. Molecular Ecology 11,

1591-1604.

Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals

using the software STRUCTURE: a simulation study. Molecular Ecology 14, 2611-

2620.

Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular variance inferred from

metric distances among DNA haplotypes: application to human mitochondrial DNA

restriction data. Genetics 131, 479-491.

Excoffier L, Laval G, Schneider S (2005) Arlequin (version 3.0): An integrated software

package for population genetics data analysis. Evolutionary Bioinformatics 1, 47-50.

Excoffier L & Heckel G (2006) Computer programs for population genetics data analysis: a

survival guide. Nature Reviews Genetics 7, 745-758.

Fagan WF (2002) Connectivity, fragmentation, and extinction risk in dendritic

metapopulations. Ecology 83, 3243-3249.

Falush D, Stephens M, Pritchard JK (2003) Inference of population structure using

multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164,

1567-1587.

References 232

Falush D, Stephens M, Pritchard JK (2007) Inference of population structure using

multilocus genotype data: dominant markers and null alleles. Molecular Ecology

Notes 7, 574-578.

Faulks LK, Gilligan DM, Beheregaray LB (2010(a)) Evolution and maintenance of

divergent lineages in an endangered freshwater fish, Macquaria australasica.

Conservation Genetics 11, 921-934.

Faulks LK, Gilligan DM, Beheregaray LB (2010(b)) Islands of water in a sea of dry land:

hydrological regime predicts genetic diversity and dispersal in a widespread fish from

Australia’s arid zone, the golden perch (Macquaria ambigua). Molecular Ecology 19,

4723-4737.

Faulks LK, Gilligan DM, Beheregaray LB (2011) The role of anthropogenic vs. natural in

stream structures in determining connectivity and genetic diversity in an endangered

freshwater fish, Macquarie perch (Macquaria australasica). Evolutionary

Applications 4, 589-601.

Fausch KD, Torgersen CE, Baxter CV, Li HW (2002) Landscapes to Riverscapes: Bridging

the Gap between Research and Conservation of Stream Fishes. Bioscience 52, 483-

498.

Favre L, Balloux F, Goudet J, Perrin N (1997) Female-biased dispersal in the monogamous

mammal Crocidura russula: evidence from field data and microsatellite patterns.

References 233

Proceedings of the Royal Society of London. Series B: Biological Sciences 264, 127-

132.

Fernández J & Caballero A (2001) A comparison of management strategies for

conservation with regard to population fitness. Conservation Genetics 2, 121-131.

Ficetola GF & Bonin A (2011) Conserving adaptive genetic diversity in dynamic

landscapes. Molecular Ecology 20, 1569-1571.

Finn DS, Blouin MS, Lytle DA (2007) Population genetic structure reveals terrestrial

affinities for a headwater stream insect. Freshwater Biology 52, 1881-1897.

Forsyth A & Nott J (2003) Evolution of drainage patterns on Cape York Peninsula,

northeast Queensland. Australian Journal of Earth Sciences 50, 145-155.

Frankel OH (1974) Genetic Conservation: Our Evolutionary Responsibility. Genetics 78,

53-65.

Frankham R (1995) Conservation Genetics. Annual Reviews in Genetics 29, 305-327.

Frankham R (1996) Relationship of Genetic Variation to Population Size in Wildlife.

Conservation Biology 10, 1500-1508.

Fu YX, Li WH (1993) Statistical tests of neutrality of mutations. Genetics 133, 693-709.

References 234

Fu YX (1997) Statistical Tests of Neutrality of Mutations Against Population Growth,

Hitchhiking and Background Selection. Genetics 147, 915-925.

Fullerton AH, Burnett KM, Steel EA, Flitcroft RL, Pess GR, Feist BE, Torgersen CE,

Miller DJ, Sanderson BL (2010) Hydrological connectivity for riverine fish:

measurement challenges and research opportunities. Freshwater Biology 55, 2215-

2237.

Gaggiotti OE (1996) Population genetic models of source-sink metapopulations.

Theoretical Population Biology 50, 178-208.

Gaines MS & McClenaghan LR (1980) Dispersal in small mammals. Annual Review of

Ecology and Systematics 11, 163-196.

Galtier N, Nabholz B, Glémin S, Hurst GDD (2009) Mitochondrial DNA as a marker of

molecular diversity: a reappraisal. Molecular Ecology 18, 4541-4550.

Garland T & Adolph SC (1991) Physiological differentiation of vertebrate populations.

Annual Review of Ecology and Systematics 22, 193-228.

Garza JC & Freimer NB (1996) Homoplasy for size at microsatellite loci in humans and

chimpanzees. Genome Research 6, 211-217.

References 235

Gebremedhin B, Ficetola GF, Naderi S, Rezaei HR, Maudet C, Rioux D, Luikart G,

Flagstad Ø, Thuiller W, Taberlet P (2009) Frontiers in identifying conservation units:

from neutral markers to adaptive genetic variation. Animal Conservation 12, 107-109.

Geffen E, Luikart G, Waples RS (2007) Impacts of modern molecular genetic techniques

on conservation biology. In ‘Key topics in conservation biology’. (Eds DW

Macdonald & K Service) pp 46-63. (Blackwell Publishing Ltd: Oxford)

Gene Codes Corporation (2000) Sequencher® version 4.1 sequence analysis software, Ann

Arbor, MI USA http://www.genecodes.com

Giddelo CS, Arndt AD & Volckaert FAM (2002) Impact of rifting and hydrography on the

genetic structure of Clarias gariepinus in eastern Africa. Journal of Fish Biology 60,

1252-1266.

Gilpin M (1991) The genetic effective size of a metapopulation. Biological Journal of the

Linnean Society 42, 165-175.

Goldstein DB & Pollock DD (1997) Mutation Processes and Methods of Phylogenetic

Inference. Journal of Heredity 88, 335-342.

Goudet J (2001) FSTAT, a program to estimate and test gene diversities and fixation

indices (version 2.9.3). http://www2.unil.ch/popgen/softwares/fstat.htm

References 236

Goudet J, Perrin N, Waser P (2002) Tests for sex biased dispersal using bi parentally

inherited genetic markers. Molecular Ecology 11, 1103-1114.

Goudge TA (1955) What Is a Population? Philosophy of Science 22, 272-279.

Growns I (2004) A numerical classification of reproductive guilds of the freshwater fishes

of south eastern Australia and their application to river management. Fisheries

Management and Ecology 11, 369-377.

Grueber CE, Wallis GP, Jamieson IG (2008) Heterozygosity-fitness correlations and their

relevance to studies on inbreeding depression in threatened species. Molecular

Ecology 17, 3978.

Guderley H & Pörtner HO (2010) Metabolic power budgeting and adaptive strategies in

zoology: examples from scallops and fish. Canadian Journal of Zoology 88, 753-763.

Guillot G, Leblois R, Coulon A, Frantz AC (2009) Statistical methods in spatial genetics.

Molecular Ecology 18, 4734-4756.

Guy TJ, Gresswell RE, Banks MA (2008) Landscape-scale evaluation of genetic structure

among barrier-isolated populations of coastal cutthroat trout, Oncorhynchus clarkii

clarkii. Canadian Journal of Fisheries and Aquatic Sciences 65, 1749-1762.

References 237

Hadwen WL & Arthington AH (2007) Food webs of two intermittently open estuaries

receiving super 15N-enriched sewage effluent. Estuarine, Coastal and Shelf Science

71, 347-358.

Haig SM (1998) Molecular contributions to conservation. Ecology 79, 413-425.

Hamilton SK & Gehrke PC (2005) Australia's tropical river systems: current scientific

understanding and critical knowledge gaps for sustainable management. Marine and

Freshwater Research 56, 243-252.

Hänfling B, Hellemans B, Volckaert FAM, Carvalho GR (2002) Late glacial history of the

cold adapted freshwater fish Cottus gobio, revealed by microsatellites. Molecular

Ecology 11, 1717-1729.

Hänfling B, Durka W, Brandl R (2004) Impact of habitat fragmentation on genetic

population structure of roach, Rutilus rutilus, in a riparian ecosystem. Conservation

Genetics 5, 247-257.

Hanski I & Gilpin M (1991) Metapopulation dynamics: brief history and conceptual

domain. Biological Journal of the Linnean Society 42, 3-16.

Hanski I (1998) Metapopulation dynamics. Nature 396, 41-49.

Hanski I (1999) Habitat connectivity, habitat continuity, and metapopulations in dynamic

landscapes. Oikos 87, 209-219. References 238

Hardy OJ & Vekemans X (1999) Isolation by distance in a continuous population:

reconciliation between spatial autocorrelation analysis and population genetics

models. Heredity 83, 145-154.

Harpending HC, Sherry ST, Rogers AR, Stoneking M (1993) The Genetic Structure of

Ancient Human Populations. Current Anthropology 34, 483.

Harpending HC, Batzer MA, Gurven M, Jorde LB, Rogers AR, Sherry ST (1998) Genetic

traces of ancient demography. Proceedings of the National Academy of Sciences 95,

1961-1967.

Harrison S (1991) Local extinction in a metapopulation context: an empirical evaluation.

Biological Journal of the Linnean Society 42, 73-88.

Hart BT (2004) Environmental risks associated with new irrigation schemes in Northern

Australia. Ecological Management & Restoration 5, 106-110.

Hartl DL & Clark AG (1997) 'Principles of population genetics.' 4th Edn. (Sinauer

Associates: Sunderland)

Hedrick PW & Miller PS (1992) Conservation Genetics: Techniques and Fundamentals.

Ecological Applications 2, 30-46.

References 239

Heg D, Bachar Z, Brouwer L, Taborsky M (2004) Predation risk is an ecological constraint

for helper dispersal in a cooperatively breeding cichlid. Proceedings of the Royal

Society of London. Series B: Biological Sciences 271, 2367-2374.

Heinrichs JA, Bender DJ, Gummer DL, Schumaker NH (2010) Assessing critical habitat:

Evaluating the relative contribution of habitats to population persistence. Biological

Conservation 143, 2229-2237.

Hewitt G (2000) The genetic legacy of the Quaternary ice ages. Nature 405, 907-913.

Hewitt GM (2004) Genetic consequences of climatic oscillations in the Quaternary.

Philosophical Transactions of the Royal Society B: Biological Sciences 359, 183-195.

Hey J & Kliman RM (1993) Population genetics and phylogenetics of DNA sequence

variation at multiple loci within the Drosophila melanogaster species complex.

Molecular Biology and Evolution 10, 804-822.

Hickerson MJ, Carstens BC, Cavender-Bares J, Crandall KA, Graham CH, Johnson JB,

Rissler L, Victoriano PF, Yoder AD (2010) Phylogeography's past, present, and

future: 10 years after Avise, 2000. Molecular Phylogenetics and Evolution 54, 291-

301.

Hickford MJH, Cagnon M, Schiel DR (2010) Predation, vegetation and habitat-specific

survival of terrestrial eggs of a diadromous fish, Galaxias maculatus (Jenyns, 1842).

Journal of Experimental Marine Biology and Ecology 385, 66-72. References 240

Hobson KA & Clark RG (1992) Assessing avian diets using stable isotopes II: factors

influencing diet-tissue fractionation. Condor, 189-197.

Hobson KA (1999) Tracing origins and migration of wildlife using stable isotopes: a

review. Oecologia 120, 314-326.

Hobson KA (2005) Using stable isotopes to trace long-distance dispersal in birds and other

taxa. Diversity & Distributions 11, 157-164.

Hobson KA, Barnett-Johnson R, Cerling T (2010) Using isoscapes to track animal

migration. In ‘Isoscapes: Understanding movement, pattern, and process on Earth

through isotope mapping’. (Eds J.B. West, GJ Bowen, TE Dawson, KP Tu) pp. 273-

298. (Springer Science: New York).

Hoelzel AR (2010) Looking backwards to look forwards: conservation genetics in a

changing world. Conservation Genetics 11, 655-660.

Hoffmann AA & Hercus MJ (2000) Environmental stress as an evolutionary force.

Bioscience 50, 217-226.

Hoffmann AA & Willi Y (2008) Detecting genetic responses to environmental change.

Nature Reviews Genetics 9, 421-432.

Holderegger R & Wagner HH (2008) Landscape genetics. BioScience 58, 199-207.

References 241

Hrbek T, Farias IP, Crossa M, Sampaio I, Porto JIR & Meyer A (2005) Population genetic

analysis of Arapaima gigas, one of the largest freshwater fishes of the Amazon basin:

implications for its conservation. Animal Conservation 8, 297-308.

Hubisz MJ, Falush D, Stephens M, Pritchard JK (2009) Inferring weak population structure

with the assistance of sample group information. Molecular Ecology Resources 9,

1322-1332.

Huey JA, Hughes JM, Baker AM (2006) Patterns of gene flow in two species of eel-tailed

catfish, Neosilurus hyrtlii and Porochilus argenteus (Siluriformes: Plotosidae), in

western Queensland's dryland rivers. Biological Journal of the Linnean Society 87,

457-467.

Huey JA, Baker AM, Hughes JM (2008) The effect of landscape processes upon gene flow

and genetic diversity in an Australian freshwater fish, Neosilurus hyrtlii. Freshwater

Biology 53, 1393-1408.

Huey JA, Schmidt DJ, Balcombe SR, Marshall JC, Hughes JM (2011) High gene flow and

metapopulation dynamics detected for three species in a dryland river system.

Freshwater Biology 56, 2378-2390.

Hughes JM, Bunn SE, Hurwood DA, Choy S, Pearson RG (1996) Genetic differentiation

among populations of Caridinazebra (Decapoda: Atyidae) in tropical rainforest

streams, northern Australia. Freshwater Biology 36, 289-296.

References 242

Hughes J, Ponniah M, Hurwood D, Chenoweth S, Arthington A (1999) Strong genetic

structuring in a habitat specialist, the Oxleyan Pygmy Perch Nannoperca oxleyana.

Heredity 83, 5-14.

Hughes JM & Hillyer MJ (2003) Patterns of connectivity among populations of Cherax

destructor (Decapoda: Parastacidae) in western Queensland, Australia. Marine and

Freshwater Research 54, 587-596.

Hughes JM & Hillyer MJ (2006) Mitochondrial DNA and allozymes reveal high dispersal

abilities and historical movement across drainage boundaries in two species of

freshwater fishes from inland rivers in Queensland, Australia. Journal of Fish Biology

68, 270-291.

Hughes JM (2007) Constraints on recovery: using molecular methods to study connectivity

of aquatic biota in rivers and streams. Freshwater Biology 52, 616-631.

Hughes JM, Schmidt DJ, Finn DS (2009) Genes in streams: using DNA to understand the

movement of freshwater fauna and their riverine habitat. BioScience 59, 573-583.

Hughes L (2003) Climate change and Australia: trends, projections and impacts. Austral

Ecology 28, 423-443.

Humphries P, King AJ, Koehn JD (1999) Fish, Flows and Flood Plains: Links between

Freshwater Fishes and their Environment in the Murray-Darling River System,

Australia. Environmental Biology of Fishes 56, 129-151.

References 243

Hurst GDD & Jiggins FM (2005) Problems with mitochondrial DNA as a marker in

population, phylogeographic and phylogenetic studies: the effects of inherited

symbionts. Proceedings of the Royal Society B: Biological Sciences 272, 1525-1534.

Hurwood DA & Hughes JM (1998) Phylogeography of the freshwater fish, Mogurnda

adspersa, in streams of northeastern Queensland, Australia: evidence for altered

drainage patterns. Molecular Ecology 7, 1507-17.

Hurwood DA & Hughes JM (2001) Historical interdrainage dispersal of eastern

rainbowfish from the Atherton Tableland, north eastern Australia. Journal of Fish

Biology 58, 1125-1136.

Hurwood DA, Adamson EAS & Mather PB (2008) Evidence for strong genetic structure in

a regionally important, highly vagile cyprinid (Henicorhynchus lobatus) in the

Mekong River Basin. Ecology of Freshwater Fish 17, 273-283.

Hutchison DW & Templeton AR (1999) Correlation of pairwise genetic and geographic

distance measures: inferring the relative influences of gene flow and drift on the

distribution of genetic variability. Evolution 53, 1898-1914.

Inger R & Bearhop S (2008) Applications of stable isotope analyses to avian ecology. Ibis

150, 447-461.

References 244

Jager HI, Chandler JA, Lepla KB, Van Winkle W (2001) A Theoretical Study of River

Fragmentation by Dams and its Effects on White Sturgeon Populations.

Environmental Biology of Fishes 60, 347-361.

Jardine TD, Hunt RJ, Pusey BJ, Bunn SE (2011) A non-lethal sampling method for stable

carbon and nitrogen isotope studies of tropical fishes. Marine and Freshwater

Research 62, 83-90.

Jardine TD, Pusey BJ, Hamilton SK, Pettit NE, Davies PM, Douglas MM, Sinnamon V,

Halliday IA, Bunn SE (In press) Fish mediate high food web connectivity in the lower

reaches of a wet-dry tropical floodplain river. Oecologia. DOI: 10.1007/s00442-011-

2148-0

Jarne P & Lagoda PJL (1996) Microsatellites, from molecules to populations and back.

Trends in Ecology & Evolution 11, 424-429.

Jenkins DG, Carey M, Czerniewska J, Fletcher J, Hether T, Jones A, Knight S, Knox J,

Long T (2010) A meta-analysis of isolation by distance: relic or reference standard for

landscape genetics? Ecography 33, 315-320.

Johnson JB & Omland KS (2004) Model selection in ecology and evolution. Trends in

Ecology & Evolution 19, 101-108.

References 245

Johnson ML, Gaines MS (1990) Evolution of dispersal: theoretical models and empirical

tests using birds and mammals. Annual Review of Ecology and Systematics 21, 449-

480.

Johnson MS & Black R (1984) The Wahlund effect and the geographical scale of variation

in the intertidal limpet Siphonaria sp. Marine Biology 79, 295-302.

Jones GP, Almany GR, Russ GR, Sale PF, Steneck RS, van Oppen MJH, Willis BL (2009)

Larval retention and connectivity among populations of corals and reef fishes: history,

advances and challenges. Coral Reefs 28, 307-325.

Jorde PE & Ryman N (1995) Temporal allele frequency change and estimation of effective

size in populations with overlapping generations. Genetics 139, 1077-1090.

Joseph L, Wilke T, Alpers D (2002) Reconciling genetic expectations from host specificity

with historical population dynamics in an avian brood parasite, Horsfield’s Bronze

Cuckoo Chalcites basalis of Australia. Molecular Ecology 11, 829-837.

Jowett IG (1993) A method for objectively identifying pool, run, and riffle habitats from

physical measurements. New Zealand Journal of Marine and Freshwater Research

27, 241-248.

Kaplan NL, Hudsont RR, Langle CH (1989) The “Hitchhiking Effect” Revisited. Genetics

123, 887-899.

References 246

Kawecki TJ & Ebert D (2004) Conceptual issues in local adaptation. Ecology Letters 7,

1225-1241.

Keller LF & Waller DM (2002) Inbreeding effects in wild populations. Trends in Ecology

& Evolution 17, 230-241.

Kempf F, McCoy KD, De Meeûs T (2010) Wahlund effects and sex-biased dispersal in

Ixodes ricinus, the European vector of Lyme borreliosis: New tools for old data.

Infection, Genetics and Evolution 10, 989-997.

Keyghobadi N (2007) The genetic implications of habitat fragmentation for animals.

Canadian Journal of Zoology 85, 1049-1064.

Kimura M & Weiss GH (1964) The Stepping Stone Model of Population Structure and the

Decrease of Genetic Correlation with Distance. Genetics 49, 561-76.

Kimura M (1968) Evolutionary rate at the molecular level. Nature 217, 624-626.

King AJ, Humphries P, Lake PS (2003) Fish recruitment on floodplains: the roles of

patterns of flooding and life history characteristics. Canadian Journal of Fisheries

and Aquatic Sciences 60, 773-786.

King AJ (2004) Ontogenetic patterns of habitat use by fishes within the main channel of an

Australian floodplain river. Journal of Fish Biology 65, 1582-1603.

References 247

King AJ, Ward KA, O’Connor P, Green D, Tonkin Z, Mahoney J (2010) Adaptive

management of an environmental watering event to enhance native fish spawning and

recruitment. Freshwater Biology 55, 17-31.

Kingsford RT (2000) Ecological impacts of dams, water diversions and river management

on floodplain wetlands in Australia. Austral Ecology 25, 109-127.

Kinnison MT, Hendry AP, Stockwell CA (2007) Contemporary evolution meets

conservation biology II: impediments to integration and application. Ecological

Research 22, 947-954.

Kittlein MJ & Gaggiotti OE (2008) Interactions between environmental factors can hide

isolation by distance patterns: a case study of Ctenomys rionegrensis in Uruguay.

Proceedings of the Royal Society B: Biological Sciences 275, 2633.

Knaepkens G, Verheyen E, Galbusera P, Eens M (2004(a)) The use of genetic tools for the

evaluation of a potential migration barrier for the bullhead. Journal of Fish Biology

64, 1737-1744.

Knaepkens G, Bervoets L, Verheyen E, Eens M (2004(b)) Relationship between population

size and genetic diversity in endangered populations of the European bullhead (Cottus

gobio): implications for conservation. Biological Conservation 115, 403-410.

References 248

Knight JT, Nock CJ, Elphinstone MS, Baverstock PR (2009) Conservation implications of

distinct genetic structuring in the endangered freshwater fish Nannoperca oxleyana

(Percichthyidae). Marine and Freshwater Research 60, 34-44.

Koenig WD, Van Vuren D, Hooge PN (1996) Detectability, philopatry, and the distribution

of dispersal distances in vertebrates. Trends in Ecology & Evolution 11, 514-517.

Koizumi I, Yamamoto S, Maekawa K (2006) Decomposed pairwise regression analysis of

genetic and geographic distances reveals a metapopulation structure of stream

dwelling Dolly Varden charr. Molecular Ecology 15, 3175-3189.

Kreitman M (2000) Methods to detect selection in populations with applications to the

human. Annual Review of Genomics and Human Genetics 1, 539-559.

Kreitman M & Di Rienzo A (2004) Balancing claims for balancing selection. Trends in

Genetics 20, 300-304.

Labbe TR & Fausch KD (2000) Dynamics of intermittent stream habitat regulate

persistence of a threatened fish at multiple scales. Ecological Applications 10, 1774-

1791.

Lacy RC (1987) Loss of genetic diversity from managed populations: Interacting effects of

drift, mutation, immigration, selection, and population subdivision. Conservation

Biology 1, 143-158.

References 249

Lacy RC (1988) A Report on Population Genetics in Conservation. Conservation Biology

2, 245-248.

Laikre L, Schwartz MK, Waples RS, Ryman N (2010) Compromising genetic diversity in

the wild: unmonitored large-scale release of plants and animals. Trends in Ecology &

Evolution 25, 520-529.

Lande R (1988) Genetics and demography in biological conservation. Science 241, 1455.

Landguth EL, Cushman SA, Schwartz MK, McKelvey KS, Murphy M, Luikart G (2010)

Quantifying the lag time to detect barriers in landscape genetics. Molecular Ecology

19, 4179-4191.

Latta RG (2008) Conservation genetics as applied evolution: from genetic pattern to

evolutionary process. Evolutionary Applications 1, 84-94.

Lavery S, Moritz C, Fielder DR (1996) Genetic patterns suggest exponential population

growth in a declining species. Molecular Biology and Evolution 13, 1106-1113.

Lawson Handley LJ & Perrin N (2007) Advances in our understanding of mammalian sex

biased dispersal. Molecular Ecology 16, 1559-1578.

Layman CA, Arrington DA, Montana CG, Post DM (2007) Can stable isotope ratios

provide for community-wide measures of trophic structure? Ecology 88, 42-48.

References 250

Leary RF & Allendorf FW (1989) Fluctuating asymmetry as an indicator of stress:

implications for conservation biology. Trends in Ecology & Evolution 4, 214-217.

Lehrian S, Bálint M, Haase P, Pauls SU (2010) Genetic population structure of an autumn-

emerging caddisfly with inherently low dispersal capacity and insights into its

phylogeography. Journal of the North American Benthological Society 29, 1100-

1118.

Lenormand T (2002) Gene flow and the limits to natural selection. Trends in Ecology &

Evolution 17, 183-189.

Le Roux J & Wieczorek AM (2009) Molecular systematics and population genetics of

biological invasions: towards a better understanding of invasive species management.

Annals of Applied Biology 154, 1-17.

Lester SE, Ruttenberg BI, Gaines SD, Kinlan BP (2007) The relationship between dispersal

ability and geographic range size. Ecology Letters 10, 745-758.

Levin LA (2006) Recent progress in understanding larval dispersal: new directions and

digressions. Integrative and Comparative Biology 46, 282-297.

Levins R (1969) Some demographic and genetic consequences of environmental

heterogeneity for biological control. Bulletin of the ESA 15, 237-240.

References 251

Librado P & Rozas J (2009) DnaSP v5: a software for comprehensive analysis of DNA

polymorphism data. Bioinformatics 25, 1451-1952.

Lin K, Li H, Schlötterer C, Futschik (2011) A Distinguishing Positive Selection From

Neutral Evolution: Boosting the Performance of Summary Statistics. Genetics 187,

229-244.

Logan JM, Jardine TD, Miller TJ, Bunn SE, Cunjak RA, Lutcavage ME (2008) Lipid

corrections in carbon and nitrogen stable isotope analyses: comparison of chemical

extraction and modelling methods. Journal of Animal Ecology 77, 838-846.

Louhi P, Mäki Petäys A, Erkinaro J (2008) Spawning habitat of Atlantic salmon and brown

trout: general criteria and intragravel factors. River Research and Applications 24,

330-339.

Lowe WH, Likens GE, Power ME (2006) Linking Scales in Stream Ecology. BioScience

56, 591-597.

Lowe WH & Allendorf FW (2010) What can genetics tell us about population

connectivity? Molecular Ecology 19, 3038-3051.

Luikart G, Sherwin WB, Steele BM, Allendorf FW (1998) Usefulness of molecular

markers for detecting population bottlenecks via monitoring genetic change.

Molecular Ecology 7, 963-974.

References 252

Luikart G, Ryman N, Tallmon DA, Schwartz MK, Allendorf FW (2010) Estimation of

census and effective population sizes: the increasing usefulness of DNA-based

approaches. Conservation Genetics 11, 355-373.

Magoulick DD & Kobza RM (2003) The role of refugia for fishes during drought: a review

and synthesis. Freshwater Biology 48, 1186-1198.

Magoulick DD (2004) Effects of predation risk on habitat selection by water column fish,

benthic fish and crayfish in stream pools. Hydrobiologia 527, 209-221.

Malhi RS, Rhett G, Bell AM (2006) Mitochondrial DNA evidence of an early Holocene

population expansion of threespine sticklebacks from Scotland. Molecular

Phylogenetics and Evolution 40, 148-154.

Mallen-Cooper M (1993) Habitat changes and declines of freshwater fish in Australia:

What is the evidence and do we need more. Bureau of Resource Sciences

Proceedings, 118-123.

Manel S, Schwartz MK, Luikart G, Taberlet P (2003) Landscape genetics: combining

landscape ecology and population genetics. Trends in Ecology & Evolution 18, 189-

197.

Manel S, Gaggiotti OE, Waples RS (2005) Assignment methods: matching biological

questions with appropriate techniques. Trends in Ecology & Evolution 20, 136-142.

References 253

Markert JA, Schelly RC, Stiassny MLJ (2010) Genetic isolation and morphological

divergence mediated by high-energy rapids in two cichlid genera from the lower

Congo rapids. BMC Evolutionary Biology 10, 1-9.

Matschiner M, Hanel R, Salzburger W (2009) Gene flow by larval dispersal in the

Antarctic notothenioid fish Gobionotothen gibberifrons. Molecular Ecology 18, 2574-

2587.

Matthews KR & Berg NH (1997) Rainbow trout responses to water temperature and

dissolved oxygen stress in two southern California stream pools. Journal of Fish

Biology 50, 50-67.

Matthews WJ & Marsh-Matthews E (2003) Effects of drought on fish across axes of space,

time and ecological complexity. Freshwater Biology 48, 1232-1253.

Maynard Smith J & Haigh J (1974) The hitch-hiking effect of a favourable gene. Genetics

Research 23, 23-35.

Mayr E (1954) Change of genetic environment and evolution. In ‘Evolution As Process’.

(Eds J Huxley, AC Hardy, EB Ford) pp. 157-180. (Allen and Unwin: London).

McCarthy ID & Waldron S (2000) Identifying migratory Salmo trutta using carbon and

nitrogen stable isotope ratios. Rapid Communications in Mass Spectrometry 14, 1325-

1331.

References 254

McDowall RM (1992) Particular problems for the conservation of diadromous fish. Aquatic

Conservation: Marine and Freshwater Ecosystems 2, 351-355.

McGlashan DJ & Hughes JM (2000) Reconciling patterns of genetic variation with stream

structure, earth history and biology in the Australian freshwater fish Craterocephalus

stercusmuscarum (Atherinidae). Molecular Ecology 9, 1737-1751.

McGlashan DJ & Hughes JM (2000) Reconciling patterns of genetic variation with stream

structure, earth history and biology in the Australian freshwater fish Craterocephalus

stercusmuscarum (Atherinidae). Molecular Ecology 9, 1737-1751.

McGlashan DJ, Hughes JM, Bunn SE (2001) Within-drainage population genetic structure

of the freshwater fish Pseudomugil signifer (Pseudomugilidae) in northern Australia.

Canadian Journal of Fisheries and Aquatic Sciences 58, 1842-1852.

McNeil DG & Closs GP (2007) Behavioural responses of a south east Australian floodplain

fish community to gradual hypoxia. Freshwater Biology 52, 412-420.

Meeuwig MH, Guy CS, Kalinowski ST, Fredenberg WA (2010) Landscape influences on

genetic differentiation among bull trout populations in a stream-lake network.

Molecular Ecology 19, 3620-3633.

Meffe GK & Vrijenhoek RC (1988) Conservation genetics in the management of desert

fishes. Conservation Biology 2, 157-169.

References 255

Meiklejohn CD, Montooth KL, Rand DM (2007) Positive and negative selection on the

mitochondrial genome. Trends in Genetics 23, 259-263.

Milligan BG, Leebens-Mack J, Strand AE (1994) Conservation genetics: Beyond the

maintenance of marker diversity. Molecular Ecology 3, 423-435.

Mills CE, Hadwen WL, Hughes JM (2008) Looking through glassfish: marine genetic

structure in an estuarine species. Marine and Freshwater Research 59, 627-637.

Møller AP, Hobson KA, Mousseau TA, Peklo AM (2006) Chernobyl as a population sink

for Barn Swallows: tracking dispersal using stable-isotope profiles. Ecological

Applications 16, 1696-1705.

Montgomery ME, Woodworth LM, Nurthen RK, Gilligan DM, Briscoe DA, Frankham R

(2000) Relationships between population size and loss of genetic diversity:

comparisons of experimental results with theoretical predictions. Conservation

Genetics 1, 33-43.

Moran P (2002) Current conservation genetics: building an ecological approach to the

synthesis of molecular and quantitative genetic methods. Ecology of Freshwater Fish

11, 30-55.

Morin PA, Leduc RG, Archer FI, Martien KK, Huebinger R, Bickham JW, Taylor BL

(2009) Significant deviations from Hardy–Weinberg equilibrium caused by low levels

of microsatellite genotyping errors. Molecular Ecology Resources 9, 498-504. References 256

Moritz C (1994) Defining ‘Evolutionarily Significant Units' for conservation. Trends in

Ecology & Evolution 9, 373-375.

Moritz C (1999) Conservation Units and Translocations: Strategies for Conserving

Evolutionary Processes. Hereditas 130, 217-228.

Morrissey MB & De Kerckhove DT (2009) The maintenance of genetic variation due to

asymmetric gene flow in dendritic metapopulations. The American Naturalist 174,

875-889.

Mossman CA & Waser PM (1999) Genetic detection of sex-biased dispersal. Molecular

Ecology 8, 1063-1067.

Murphy MA, Evans JS, Cushman SA, Storfer A (2008) Representing genetic variation as

continuous surfaces: an approach for identifying spatial dependency in landscape

genetic studies. Ecography 31, 685-697.

Nachman MW (1998) Deleterious mutations in animal mitochondrial DNA. Genetica 102,

61-69.

Nanson GC, Price DM, Short SA (1992) Wetting and drying of Australia over the past 300 ka. Geology 20, 791-794.

References 257

Narum SR, Zendt JS, Graves D, Sharp WR (2008) Influence of landscape on resident and

anadromous life history types of Oncorhynchus mykiss. Canadian Journal of

Fisheries and Aquatic Sciences 65, 1013-1023.

Nathan R (2001) The challenges of studying dispersal. Trends in Ecology & Evolution 16,

481-483.

Nathan R, Perry G, Cronin JT, Strand AE, Cain ML (2003) Methods for estimating long-

distance dispersal. Oikos 103, 261-273.

Nathan R (2005) Long-distance dispersal research: building a network of yellow brick

roads. Diversity & Distributions 11, 125-130.

Nel JL, Roux DJ, Abell R, Ashton PJ, Cowling RM, Higgins JV, Thieme M, Viers JH

(2009) Progress and challenges in freshwater conservation planning. Aquatic

Conservation: Marine and Freshwater Ecosystems 19, 474-485.

Neraas LP & Spruell P (2001) Fragmentation of riverine systems: the genetic effects of

dams on bull trout (Salvelinus confluentus) in the Clark Fork River system. Molecular

Ecology 10, 1153-1164.

Nevado B, Koblmüller S, Sturmbauer C, Snoeks J, Usano Alemany J, Verheyen E (2009)

Complete mitochondrial DNA replacement in a Lake Tanganyika cichlid fish.

Molecular Ecology 18, 4240-4255.

References 258

Neville H, Dunham J, Rosenberger A, Umek J, Nelson B (2009) Influences of wildfire,

habitat size, and connectivity on trout in headwater streams revealed by patterns of

genetic diversity. Transactions of the American Fisheries Society 138, 1314-1327.

Newson M (2010) ‘Catchment Consciousness’–Will Mantra, Metric or Mania Best Protect,

Restore and Manage Habitats? In ‘Salmonid Fisheries: Freshwater Habitat

Management’ (Eds P Kemp) pp. 28-54. (Blackwell Publishing Ltd: Singapore).

Nielsen R (2005) Molecular signatures of natural selection. Annual Review of Genetics 39,

197-218.

Nilsson C, Reidy CA, Dynesius M, Revenga C (2005) Fragmentation and flow regulation

of the world's large river systems. Science 308, 405-408.

Nott J (1994) Long-term landscape evolution in the Darwin region and its implications for

the origin of landsurfaces in the north of the Northern Territory. Australian Journal of

Earth Sciences 41, 407-415.

Nunney L & Campbell KA (1993) Assessing minimum viable population size: demography

meets population genetics. Trends in Ecology & Evolution 8, 234-234.

Olden JD, Kennard MJ, Pusey BJ (2008) Species invasions and the changing biogeography

of Australian freshwater fishes. Global Ecology and Biogeography 17, 25-37.

References 259

Olivieri I, Couvet D, Gouyon PH (1990) The genetics of transient populations: research at

the metapopulation level. Trends in Ecology & Evolution 5, 207-210.

Ormerod SJ (2003) Current issues with fish and fisheries: editor's overview and

introduction. Journal of Applied Ecology 40, 204-213.

Østbye K, Bernatchez L, Naesje TF, Himberg KJM, Hindar K (2005) Evolutionary history

of the European whitefish Coregonus lavaretus (L.) species complex as inferred from

mtDNA phylogeography and gill-raker numbers. Molecular Ecology 14, 4371-4387.

Ovenden JR, Peel D, Street R, Courtney AJ, Hoyle SD, Peel SL, Podlich H (2007) The

genetic effective and adult census size of an Australian population of tiger prawns

(Penaeus esculentus). Molecular Ecology 16, 127-138.

Paetkau D, Calvert W, Stirling I, Strobeck C (1995) Microsatellite analysis of population

structure in Canadian polar bears. Molecular Ecology 4, 347-354.

Paetkau D, Slade R, Burden M, Estoup A (2004) Genetic assignment methods for the

direct, real time estimation of migration rate: a simulation based exploration of

accuracy and power. Molecular Ecology 13, 55-65.

Palm D, Lindberg M, Brännäs E, Lundqvist H, Östergren J, Carlsson U (2009) Influence of

European sculpin, Cottus gobio, on Atlantic salmon Salmo salar, recruitment and the

effect of gravel size on egg predation–implications for spawning habitat restoration.

Fisheries Management and Ecology 16, 501-507. References 260

Palsbøll PJ, Bérubé M, Allendorf FW (2007) Identification of management units using

population genetic data. Trends in Ecology & Evolution 22, 11-16.

Palstra FP & Ruzzante DE (2008) Genetic estimates of contemporary effective population

size: what can they tell us about the importance of genetic stochasticity for wild

population persistence? Molecular Ecology 17, 3428-3447.

Palstra AP, Crespo D, Van den Thillart GE, Planas JV (2010) Saving energy to fuel

exercise: swimming suppresses oocyte development and downregulates ovarian

transcriptomic response of rainbow trout Oncorhynchus mykiss. American Journal of

Physiology-Regulatory, Integrative and Comparative Physiology 299, 486-499.

Palstra FP & Ruzzante DE (2010) A temporal perspective on population structure and gene

flow in Atlantic salmon (Salmo salar) in Newfoundland, Canada. Canadian Journal

of Fisheries and Aquatic Sciences 67, 225-242.

Palumbi SR, Martin A, Romano S, McMillan WO, Stice L, Grabowski G (1991) The

simple fool’s guide to PCR, version 2.0. University of Hawaii, Honolulu.

Parker PG, Snow AA, Schug MD, Booton GC, Fuerst PA (1998) What molecules can tell

us about populations: Choosing and using a molecular marker. Ecology 79, 361-382.

Peck DR & Congdon BC (2004) Reconciling historical processes and population structure

in the sooty tern Sterna fuscata. Journal of Avian Biology 35, 327-335.

References 261

Pemberton JM (2008) Wild pedigrees: the way forward. Proceedings of the Royal Society

B: Biological Sciences 275, 613.

Perna C & Pearson RG (2008) Temporal dynamics of fish assemblages in small seasonal

streams in the Queensland tropics. Australian Journal of Zoology 56, 65-73.

Pertoldi C, Bijlsma R, Loeschcke V (2007) Conservation genetics in a globally changing

environment: present problems, paradoxes and future challenges. Biodiversity and

Conservation 16, 4147-4163.

Peterson BJ & Fry B (1987) Stable Isotopes in Ecosystem Studies. Annual Reviews in

Ecology and Systematics 18, 293-320.

Petheram C, McMahon TA, Peel MC (2008) Flow characteristics of rivers in northern

Australia: Implications for development. Journal of Hydrology 357, 93-111.

Petit RJ & Excoffier L (2009) Gene flow and species delimitation. Trends in Ecology &

Evolution 24, 386-393.

Phillips DL & Gregg JW (2003) Source partitioning using stable isotopes: coping with too

many sources. Oecologia 136, 261-269.

Phillips RD, Storey AW, Johnson MS (2009) Genetic structure of Melanotaenia australis

at local and regional scales in the east Kimberley, Western Australia. Journal of Fish

Biology 74, 437-451. References 262

Piry S, Alapetite A, Cornuet JM, Paetkau D, Baudouin L, Estoup A (2004) GENECLASS2:

a software for genetic assignment and first-generation migrant detection. Journal of

Heredity 95, 536-539.

Posada D & Crandall KA (2001) Intraspecific gene genealogies: trees grafting into

networks. Trends in Ecology & Evolution 16, 37-45.

Preston BL & Jones RN (2008) Screening Climatic and Non Climatic Risks to Australian

Catchments. Geographical Research 46, 258-274.

Primmer CR, Moller AP, Ellegren H (1995) Resolving genetic relationships with

microsatellite markers: a parentage testing system for the swallow Hirundo rustica.

Molecular Ecology 4, 493-498.

Primmer CR, Veselov AJ, Zubchenko A, Poututkin A, Bakhmet I, Koskinen MT (2006)

Isolation by distance within a river system: genetic population structuring of Atlantic

salmon, Salmo salar, in tributaries of the Varzuga River in northwest Russia.

Molecular Ecology 15, 653-666.

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using

multilocus genotype data. Genetics 155, 945-959.

Prugnolle F & De Meeus T (2002) Inferring sex-biased dispersal from population genetic

tools: a review. Heredity 88, 161-165.

References 263

Pusey A & Wolf M (1996) Inbreeding avoidance in animals. Trends in Ecology &

Evolution 11, 201-206.

Pusey BJ & Kennard MJ (1996) Species richness and geographical variation in assemblage

structure of the freshwater fish fauna of the Wet Tropics region of northern

Queensland. Marine and Freshwater Research 47, 563-573.

Pusey BJ & Arthington AH (2003) Importance of the riparian zone to the conservation and

management of freshwater fish: a review. Marine and Freshwater Research 54, 1-16.

Pusey BJ, Kennard MJ, Arthington AH (2004) 'Freshwater Fishes of North-eastern

Australia.' (CSIRO Publishing: Melbourne).

Queller DC, Strassmann JE, Hughes CR (1993) Microsatellites and kinship. Trends in

Ecology & Evolution 8, 285-288.

R Development Core Team (2009). R: A language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-

project.org.

Raeymakers JAM, Maes GE, Geldof S, Hontis I, Nackaerts K & Volchaert FAM (2008)

Modeling genetic connectivity in sticklebacks as a guideline for river restoration.

Evolutionary Applications 1, 475-488.

References 264

Raeymaekers JAM, Raeymaekers D, Koizumi I, Geldof S, Volckaert FAM (2009)

Guidelines for restoring connectivity around water mills: a population genetic

approach to the management of riverine fish. Journal of Applied Ecology 46, 562-571.

Ramos-Onsins SE & Rozas J (2002) Statistical Properties of New Neutrality Tests Against

Population Growth. Molecular Biology and Evolution 19, 2092-2100.

Rannala B & Mountain JL (1997) Detecting immigration by using multilocus genotypes.

Proceedings of the National Academy of Sciences 94, 9197-9201.

Real KM, Schmidt DJ, Hughes JM (2009) Mogurnda adspersa microsatellite markers:

multiplexing and multi-tailed primer tagging. Conservation Genetics Resources 1,

411-414.

Reed DH & Frankham R (2003) Correlation between Fitness and Genetic Diversity.

Conservation Biology 17, 230-237.

Revaldaves E, Renesto E & Machado MFPS (1997) Genetic variability of Prochilodus

lineatus (Characiformes, Prochilodontidae) in the upper Paraná river. Brazilian

Journal of Genetics 20, DOI: S0100-84551997000300005.

Rhymer JM & Simberloff D (1996) Extinction by hybridization and introgression. Annual

Review of Ecology and Systematics 27, 83-109.

Rice WR (1989) Analyzing tables of statistical tests. Evolution 43, 223-225. References 265

Riginos C & Nachman MW (2001) Population subdivision in marine environments: the

contributions of biogeography, geographical distance and discontinuous habitat to

genetic differentiation in a blennioid fish, Axoclinus nigricaudus. Molecular Ecology

10, 1439-1453.

Robledo Arnuncio JJ & Rousset F (2010) Isolation by distance in a continuous population

under stochastic demographic fluctuations. Journal of Evolutionary Biology 23, 53-

71.

Rogers AR & Harpending H (1992) Population growth makes waves in the distribution of

pairwise genetic differences. Molecular Biology and Evolution 9, 552-69.

Ronce O (2007) How does it feel to be like a rolling stone? Ten questions about dispersal

evolution. Annual Review of Ecology, Evolution and Systematics 38, 231-253.

Rosenfeld JS & Hatfield T (2006) Information needs for assessing critical habitat of

freshwater fish. Canadian Journal of Fisheries and Aquatic Sciences 63, 683-698.

Rousset F (1997) Genetic Differentiation and Estimation of Gene Flow from F-Statistics

Under Isolation by Distance. Genetics 145, 1219-1228.

Rozas J & Rozas R (1999) DnaSP version 3: an integrated program for molecular

population genetics and molecular evolution analysis. Bioinformatics 15, 174-5.

References 266

Rubenstein DR & Hobson KA (2004) From birds to butterflies: animal movement patterns

and stable isotopes. Trends in Ecology & Evolution 19, 256-263.

Rubidge E, Corbett P, Taylor EB (2001) A molecular analysis of hybridization between

native westslope cutthroat trout and introduced rainbow trout in southeastern British

Columbia, Canada. Journal of Fish Biology 59, 42-54.

Russell DJ, McDougall AJ, Kistle SE, Ryan TJ, Aland G (2003) Species diversity and

spatial variation in fish assemblage structure of streams in connected tropical

catchments in northern Australia with reference to the occurrence of translocated and

exotic species. Marine and Freshwater Research 54, 813-824.

Ryman N & Palm S (2006) POWSIM: a computer program for assessing statistical power

when testing for genetic differentiation. Molecular Ecology Notes 6, 600-602.

Santos MCF, Ruffino ML, Farias IP (2007) High levels of genetic variability and panmixia

of the tambaqui Colossoma macropomum (Cuvier, 1816) in the main channel of the

Amazon River. Journal of Fish Biology 71, 33-44.

Saunders DL, Meeuwig JJ, Vincent ACJ (2002) Freshwater protected areas: strategies for

conservation. Conservation Biology 16, 30-41.

Schaefer J (2001) Riffles as barriers to interpool movement by three cyprinids (Notropis

boops, Campostoma anomalum and Cyprinella venusta). Freshwater Biology 46, 379-

388. References 267

Schluter D (2009) Evidence for ecological speciation and its alternative. Science 323, 737.

Schultheis AS, Marchant R, Hughes JM (2008) Contrasting patterns of genetic structure

and disequilibrium in populations of a stone-cased caddisfly (Tasimiidae) from

northern and southern Australia. Marine and Freshwater Research 59, 235-245.

Schwartz MK, Luikart G, Waples RS (2007) Genetic monitoring as a promising tool for

conservation and management. Trends in Ecology & Evolution 22, 25-33.

Schwenk K, Brede N, Streit B (2008) Introduction. Extent, processes and evolutionary

impact of interspecific hybridization in animals. Philosophical Transactions of the

Royal Society B: Biological Sciences 363, 2805-2811.

Segelbacher G & Storch I (2002) Capercaillie in the Alps: genetic evidence of

metapopulation structure and population decline. Molecular Ecology 11, 1669-1677.

Segelbacher G, Cushman SA, Epperson BK, Fortin MJ, Francois O, Hardy OJ,

Holderegger R, Taberlet P, Waits LP, Manel S (2010) Applications of landscape

genetics in conservation biology: concepts and challenges. Conservation Genetics 11,

375-385.

Selkoe KA & Toonen RJ (2006) Microsatellites for ecologists: a practical guide to using

and evaluating microsatellite markers. Ecology Letters 9, 615-629.

References 268

Sexton JP, McIntyre PJ, Angert AL, Rice KJ (2009) Evolution and ecology of species

range limits. Annual Review of Ecology, Evolution, and Systematics 40, 415-436.

Shaklee JB & Salini JP (1985) Genetic variation and population subdivision in Australian

barramundi, Lates calcarifer. Marine and Freshwater Research 36, 203-218.

Shaw PW, Arkhipkin AI, Al Khairulla H (2004) Genetic structuring of Patagonian

toothfish populations in the Southwest Atlantic Ocean: the effect of the Antarctic

Polar Front and deep water troughs as barriers to genetic exchange. Molecular

Ecology 13, 3293-3303.

Sheldon F, Bunn SE, Hughes JM, Arthington AH, Balcombe SR, Fellows CS (2010)

Ecological roles and threats to aquatic refugia in arid landscapes: dryland river

waterholes. Marine and Freshwater Research 61, 885-895.

Sih A & Wooster DE (1994) Prey behavior, prey dispersal, and predator impacts on stream

prey. Ecology, 1200-1207.

Sivasundar A, Bermingham E & Orti G (2001) Population structure and biogeography of

migratory freshwater fishes (Prochilodus: Characiformes) in major South American

rivers. Molecular Ecology 10, 407-417.

Slatkin M (1977) Gene flow and genetic drift in a species subject to frequent local

extinctions. Theoretical Population Biology 12, 253-262.

References 269

Slatkin M (1985) Gene Flow in Natural Populations. Annual Reviews in Ecology and

Systematics 16, 393-430.

Slatkin M (1987) Gene flow and the geographic structure of natural populations. Science

236, 787-792.

Slatkin M & Maddison WP (1990) Detecting Isolation by Distance Using Phylogenies of

Genes. Genetics 126, 249-260.

Slatkin M & Hudson RR (1991) Pairwise Comparisons of Mitochondrial DNA Sequences

in Stable and Exponentially Growing Populations. Genetics 129, 555-562.

Slatkin M (1993) Isolation by distance in equilibrium and non-equilibrium populations.

Evolution 47, 264-279.

Slatkin M (1995) A Measure of Population Subdivision Based on Microsatellite Allele

Frequencies. Genetics 139, 457-462.

Smith GR (1992) Introgression in fishes: significance for paleontology, cladistics, and

evolutionary rates. Systematic Biology 41, 41-57.

Snaddon CD, Wishart MJ, Davies BR (1998) Some implications of inter-basin water

transfers for river ecosystem functioning and water resources management in southern

Africa. Aquatic Ecosystem Health and Management 1, 159-182.

References 270

So N, Van Houdt JKJ, Volckaert FAM (2006(a)) Genetic diversity and population history

of the migratory catfishes Pangasianodon hypophthalmus and Pangasius bocourti in

the Cambodian Mekong River. Fisheries Science 72, 469-476.

So N, Maes GE, Volckaert FAM (2006(b)) High genetic diversity in cryptic populations of

the migratory sutchi catfish Pangasianodon hypophthalmus in the Mekong River.

Heredity 96, 166-174.

Song M, Zhang Z, Neumann K, Gattermann R (2005) Sex-biased dispersal of greater long-

tailed hamster (Tscherskia triton) revealed by microsatellites. Canadian Journal of

Zoology 83, 773-779.

Soule ME (1985) What is conservation biology? BioScience 35, 727-734.

Sparks RE (1995) Need for ecosystem management of large rivers and their floodplains.

BioScience 45, 168-182.

Stewart-Koster B, Olden JD, Kennard MJ, Pusey BJ, Boone EL, Douglas M, Jackson S

(2011) Fish response to the temporal hierarchy of the natural flow regime in the Daly

River, northern Australia. Journal of Fish Biology, doi:10.1111/j.1095-

8649.2011.03072.x

Stiver KA, Fitzpatrick J, Desjardins JK, Balshine S (2006) Sex differences in rates of

territory joining and inheritance in a cooperatively breeding cichlid fish. Animal

Behaviour 71, 449-456. References 271

Storfer A, Murphy MA, Evans JS, Goldberg CS, Robinson S, Spear SF, Dezzani R,

Delmelle E, Vierling L, Waits LP (2007) Putting the ‘landscape’ in landscape

genetics. Heredity 98, 128-142.

Stoll S & Fischer P (2011) Three different patterns of how low-intensity waves can affect

the energy budget of littoral fish: a mesocosm study. Oecologia 165, 567-576.

Strobeck C, Smith JM, Charlesworth B (1976) The effects of hitchhiking on a gene for

recombination. Genetics 82, 547-58.

Sugg DW, Chesser RK, Stephen Dobson F, Hoogland JL (1996) Population genetics meets

behavioral ecology. Trends in Ecology & Evolution 11, 338-342.

Sunnucks P (2000) Efficient genetic markers for population biology. Trends in Ecology &

Evolution 15, 199-203.

Tajima F (1989) The effect of change in population size on DNA polymorphism. Genetics

123, 597-601.

Templeton AR (1980) The theory of speciation via the founder principle. Genetics 94,

1011-1038.

Templeton AR, Shaw K, Routman E, Davis SK (1990) The genetic consequences of habitat

fragmentation. Annals of the Missouri Botanical Garden 77, 13-27.

References 272

Templeton AR (1995) A Cladistic Analysis of Phenotypic Associations with Haplotypes

Inferred from Restriction Endonuclease Mapping or DNA Sequencing. V. Analysis of

Case/Control Sampling Designs: Alzheimer's Disease and the Apoprotein E Locus.

Genetics 140, 403-409.

Thuesen PA, Pusey BJ, Peck DR, Pearson R, Congdon BC (2008) Genetic differentiation

over small spatial scales in the absence of physical barriers in an Australian rainforest

stream fish. Journal of Fish Biology 72, 1174-1187.

Trakhtenbrot A, Nathan R, Perry G, Richardson DM (2005) The importance of long-

distance dispersal in biodiversity conservation. Diversity & Distributions 11, 173-181.

Turner TF, Osborne MJ, Moyer GR, Benavides MA, Alò D (2006) Life history and

environmental variation interact to determine effective population to census size ratio.

Proceedings of the Royal Society B: Biological Sciences 273, 3065-3073.

Turner TF, Collyer ML, Krabbenhoft TJ (2010) A general hypothesis-testing framework

for stable isotope ratios in ecological studies. Ecology 91, 2227-2233.

Väli U, Einarsson A, Waits L, Ellegren H (2008) To what extent do microsatellite markers

reflect genome-wide genetic diversity in natural populations? Molecular Ecology 17,

3808-3817.

References 273

Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P (2004) MICRO-CHECKER:

software for identifying and correcting genotyping errors in microsatellite data.

Molecular Ecology Notes 4, 535-538.

Vitousek PM, Aber JD, Howarth RW, Likens GE, Matson PA, Schindler DW, Schlesinger

WH, Tilman DG (1997) Human alteration of the global nitrogen cycle: sources and

consequences. Ecological Applications 7, 737-750.

Voigt CC, Matt F, Michener R, Kunz TH (2003) Low turnover rates of carbon isotopes in

tissues of two nectar-feeding bat species. Journal of Experimental Biology 206, 1419-

1427.

Vrijenhoek RC (1998) Conservation genetics of freshwater fish. Journal of Fish Biology

53, 394-412.

Wahlund S (1928) Zusammensetzung von Populationen und Korrelationserscheinungen

vom Standpunkt der Vererbungslehre aus betrachtet. Hereditas 11, 65-106.

Waples RS (1998) Separating the wheat from the chaff: patterns of genetic differentiation

in high gene flow species. Journal of Heredity 89, 438-450.

Waples RS & Gaggiotti O (2006) What is a population? An empirical evaluation of some

genetic methods for identifying the number of gene pools and their degree of

connectivity. Molecular Ecology 15, 1419-1439.

References 274

Waser PM & Strobeck C (1998) Genetic signatures of interpopulation dispersal. Trends in

Ecology & Evolution 13, 43-44.

Wasson RJ, Furlonger L, Parry D, Pietsch T, Valentine E, Williams D (2010) Sediment

sources and channel dynamics, Daly River, Northern Australia. Geomorphology 114,

161-174.

Waters JM, Craw D, Youngson JH, Wallis GP (2001) Genes meet geology: fish

phylogeographic pattern reflects ancient, rather than modern, drainage connections.

Evolution 55, 1844-1851.

Waters JM, Allibone RM, Wallis GP (2006) Geological subsidence, river capture, and

cladogenesis of galaxiid fish lineages in central New Zealand. Biological Journal of

the Linnean Society 88, 367-376.

Webster MS, Marra PP, Haig SM, Bensch S, Holmes RT (2002) Links between worlds:

unravelling migratory connectivity. Trends in Ecology & Evolution 17, 76-83.

Werner RA & Brand WA (2001) Referencing strategies and techniques in stable isotope

ratio analysis. Rapid Communications in Mass Spectrometry 15, 501-519.

West JB, Bowen GJ, Cerling TE, Ehleringer JR (2006) Stable isotopes as one of nature's

ecological recorders. Trends in Ecology & Evolution 21, 408-414.

References 275

Willi Y, Van Buskirk J, Schmid B, Fischer M (2007) Genetic isolation of fragmented

populations is exacerbated by drift and selection. Journal of Evolutionary Biology 20,

534-542.

Wilkinson MJ, Marshall LG, Lundberg JG (2006) River behavior on megafans and

potential influences on diversification and distribution of aquatic organisms. Journal

of South American Earth Sciences 21, 151-172.

Wishart MJ & Davies BR (2003) Beyond catchment considerations in the conservation of

lotic biodiversity. Aquatic Conservation: Marine and Freshwater Ecosystems 13,

429-437.

Wong BB, Keogh JS, McGlashan DJ (2004) Current and historical patterns of drainage

connectivity in eastern Australia inferred from population genetic structuring in a

widespread freshwater fish Pseudomugil signifier (Pseudomugilidae). Molecular

Ecology 13, 391-401.

Woodford DJ & McIntosh AR (2010) Evidence of source-sink metapopulations in a

vulnerable native galaxiid fish driven by introduced trout. Ecological Applications 20,

967-977.

Wright S (1943) Isolation by Distance. Genetics 28, 114-38.

Wright S (1949) The genetical structure of populations. Annals of Human Genetics 15, 323-

354. References 276

Wright S (1951) Fisher and Ford on “The Sewall Wright Effect”. American Scientist 39,

452-479.

Wunder MB & Norris DR (2008) Improved estimates of certainty in stable-isotope-based

methods for tracking migratory animals. Ecological Applications 18, 549-559.

Yamamoto S, Morita K, Koizumi I, Maekawa K (2004) Genetic Differentiation of White-

Spotted Charr (Salvelinus leucomaenis) Populations After Habitat Fragmentation:

Spatial–Temporal Changes in Gene Frequencies. Conservation Genetics 5, 529-538.

Young MJ, Simmons LW, Evans JP (2009) Isolation and characterization of 12 novel DNA

microsatellites in the western rainbowfish, Melanotaenia australis. Molecular

Ecology Resources 9, 1252.

Young MJ, Simmons LW, Evans JP (2010) Pre-and post-mating sexual selection both favor

large males in a rainbowfish. Behavioral Ecology and Sociobiology 64, 915-925.

Young MJ, Evans JP, Simmons LW (2011) Population genetic structure and a possible role

for selection in driving phenotypic divergence in a rainbowfish (Melanotaeniidae).

Biological Journal of the Linnean Society 102, 144-160.

Zeug SC & Winemiller KO (2007) Ecological correlates of fish reproductive activity in

floodplain rivers: a life-history-based approach. Canadian Journal of Fisheries and

Aquatic Sciences 64, 1291-1301.

References 277

Zhang DX & Hewitt GM (1996) Nuclear integrations: challenges for mitochondrial DNA

markers. Trends in Ecology & Evolution 11, 247-251.

Zhang DX & Hewitt GM (2003) Nuclear DNA analyses in genetic studies of populations:

practice, problems and prospects. Molecular Ecology 12, 563-584.

Zhu D, Jamieson BG, Hugall A, Moritz C (1994) Sequence evolution and phylogenetic

signal in control-region and cytochrome b sequences of rainbow fishes

(Melanotaeniidae). Molecular Biology and Evolution 11, 672-683.

Zollner PA & Lima SL (2005) Behavioral tradeoffs when dispersing across a patchy

landscape. Oikos 108, 219-230.

References 278

7.0 Appendix

Table 7.1 Latitude and longitude for sites at which M. australis was sampled.

Site Site Code Latitude Longitude Beeboom DA -13.8616 131.0745 Claravale/Sawmill DM -14.3638 131.5570 Middle Creek DB -13.8093 131.3411 Fergusson River DFF -14.0710 131.9744 U/S Centipede Dreaming and Snowdrop Creek DE -13.4201 133.2025 UF Daly River Road DEE -13.7705 130.6277 Flora River DGG -14.7820 131.4772 Fish River DJ -14.2337 130.9090 Big Bird Billabong DJJ -13.6292 130.4931 Umbrawarrah Gorge DKK -13.9650 131.6940 Galloping Jack's DL -14.5557 132.1208 Sandy/Hermit DLL -14.0208 130.4437 Mathison Creek DP -14.8353 131.5274 Stray Creek DR -14.1176 131.4414 Douglas River DS -13.7670 131.4470 Oolloo Crossing (upstream) DT -14.1084 131.2840 Oolloo Crossing DX -14.0725 131.2540

Chapter 7.0 ∙ Appendix 279

Table 7.2 Latitude and longitude for sites at which M. s. inornata was sampled.

Site Site Code Latitude Longitude Ten Mile Lagoon MB -16.3400 143.0417 Mitchell River @ Gamboola MC -16.5350 143.6767 Cairo Lagoon ME -16.4617 143.2183 Fishhole Creek MF -15.4667 141.7850 Twelve Mile Lagoon MG -16.2933 143.0317 Koolatah Gauging Station MH -15.9500 142.3833 Mitchell River @ Koolatah MI -15.9648 142.3785 Magnificent Creek MJ -15.4883 141.7617 Saltwater Creek MK -17.8167 144.4167 Walsh River @ Nullinga ML -17.1733 145.3033 Palmer River @ Drumduff MM -16.0415 143.0381 Kingfish Lagoon MN -16.1705 142.8242 Dickson's Hole MO -17.4957 143.9599 Tate River @ Ootan MP -17.4722 144.6207 Emu Creek @ Petford MQ -17.3330 144.9495 Mitchell River @ St. George MR -16.4741 144.2863

Chapter 7.0 ∙ Appendix 280

Table 7.3 Latitude and longitude for sites at which H. fuliginosus was sampled in the Daly River.

Site Site Code Latitude Longitude Middle Creek DB -13.8093 131.3411 Claravale/Sawmill DM -14.3638 131.5570 Flora River (upstream) DQ -14.7819 131.4774 Mt Nancar DN -13.8358 130.7351 Mission Hole DC -13.7619 130.4474 Bottom of Centipede Dreaming Gorge DD -13.7584 133.1313 U/S Centipede Dreaming and Snowdrop Creek DE -13.4201 133.2025 Beeboom DA -13.8616 131.0745 Oolloo Crossing (upstream) DT -14.1084 131.2840 Stray Creek DR -14.1176 131.4414 Mathison Creek DP -14.8353 131.5274 Upstream of Manyallaluk DU -14.0798 132.7218 Fish River DJ -14.2337 130.9090 Fergusson River DFF -14.0710 131.9744 Douglas River DS -13.7670 131.4470 Oolloo Crossing DX -14.0725 131.2540 Galloping Jack's DL -14.5411 132.1389 Downstream of Galloping Jacks DH -14.5557 132.1208 ID1 Green Ant Creek ID1 -13.7748 131.0991 ID2 Green Ant Creek ID2 -13.7452 131.0978 Station Creek ID3 -13.6994 131.0582 Edith River ID4 -14.1714 132.1180 ID5 Stray Creek ID5 -14.0696 131.4801

Chapter 7.0 ∙ Appendix 281

Table 7.4 Latitude and longitude for sites at which H. fuliginosus was sampled in the Mitchell River.

Site Site Code Latitude Longitude Lynd River @ Torwood 58 -17.3673 143.7358 Lynd River @ Rocky Creek 59 -17.5516 144.0752 Lynd River/Pinnacle Creek 60 -17.4514 143.8422 Mitchell River (10 Mile Creek) 61 -16.3400 143.0409 Lynd River (1 Mile Creek) 62 -16.8408 143.4542 Mitchell River (Kingfish) 63 -16.1701 142.8216 Walsh River (Leadingham Creek Road Crossing) 118 -17.1268 145.1515 Bruce Weir MA -17.1109 145.1159 Mitchell River @ Gamboola MC -16.5350 143.6767 Walsh River @ Rookwood MD -16.9817 144.2867 Magnificent Creek MJ -15.4883 141.7617 Saltwater Creek MK -17.8167 144.4167 Walsh River @ Nullinga ML -17.1733 145.3033 Dickson's Hole MO -17.4957 143.9599 Tate River @ Ootan MP -17.4722 144.6207 Emu Creek @ Petford MQ -17.3330 144.9495

Table 7.5 Latitude and longitude for sites at which M. s. inornata and primary sources of organic material was sampled.

Site Site Code Latitude Longitude Cooktown Crossing SIS1/6 -16.5637 144.8902 McLeod River SIS2/8 -16.4984 145.0019 Rifle Creek SIS3/7 -16.6742 145.3239 Bushy Creek SIS4/5 -16.5901 145.3367

Chapter 7.0 ∙ Appendix 282

Table 7.6 Haplotype distributions for M. australis in the Daly River. Haplotype frequency and sample sizes shown for all sites.

Haplotype DA DM DB DFF DE DEE DGG DJ DJJ DKK DLL DP DR DS DT DX DL Total/Haplotype A1 3 22 4 3 3 7 4 4 1 17 68 A2 4 1 4 1 1 4 3 2 20 A3 2 1 3 A4 1 1 A5 2 3 5 A6 1 1 A7 1 3 3 26 8 2 10 4 57 A8 2 2 4 1 9 A9 2 2 A10 2 2 A11 2 1 3 A12 1 1 A13 1 1 A14 1 1 A15 1 1 A16 1 1 A17 1 1 2 A18 1 1 A19 1 1 A20 1 1 A21 1 1 Total/Site 9 35 4 27 5 4 3 4 2 8 5 7 10 12 17 4 26 182

Chapter 7.0 ∙ Appendix 283

Table 7.7 Haplotype distributions for M. s. inornata in the Mitchell River. Haplotype frequency and sample sizes shown for all sites.

Haplotype MB MC ME MF MG MH MI MJ MK ML MM MN MO MP MQ MR Total/Haplotype I1 1 1 I2 3 3 19 25 I3 2 1 3 4 1 7 18 I4 3 3 5 2 2 4 7 26 I5 1 1 I6 1 1 I7 1 1 I8 1 5 6 I9 2 2 1 25 1 2 13 7 53 I10 1 1 2 I11 1 1 I12 1 1 2 I13 3 3 I14 1 1 I15 1 1 2 I16 1 1 2 8 12 I17 1 2 1 1 5 I18 3 1 4 I19 1 1 I20 1 4 1 6 I21 1 1 2 I22 2 2 I23 1 1 I24 1 1 I25 3 3 I26 1 3 4 I27 1 1 Total/Site 8 2 19 1 13 3 1 7 25 4 19 18 14 11 20 20 185

Chapter 7.0 ∙ Appendix 284

Table 7.8 Mantel test distance matrix for populations of M. australis in the Daly River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims.

DA DM DB DFF DE DEE DGG DJ DJJ DKK DL DLL DP DR DS DT DX DA 0.00 DM 121.50 0.00 DB 35.87 144.43 0.00 DFF 202.85 81.35 225.78 0.00 DE 464.27 342.77 487.20 391.11 0.00 DEE 109.60 231.10 145.47 312.45 573.87 0.00 DGG 187.93 66.43 210.86 114.71 312.87 297.53 0.00 DJ 81.84 203.34 117.71 284.69 546.11 95.35 269.77 0.00 DJJ 120.72 242.22 156.59 323.57 584.99 34.05 308.65 119.34 0.00 DKK 149.27 89.02 172.20 170.37 431.79 258.87 155.45 231.11 269.99 0.00 DL 229.52 108.02 252.45 156.36 234.75 339.12 78.12 311.36 350.24 197.04 0.00 DLL 179.69 301.19 215.56 382.54 643.96 93.02 367.62 60.37 58.97 328.96 409.21 0.00 DP 200.50 79.00 223.43 127.28 325.44 310.10 12.57 282.34 321.22 168.02 90.69 380.19 0.00 DR 102.98 42.73 125.91 124.08 385.50 212.58 109.16 184.82 223.70 46.29 150.75 282.67 121.73 0.00 DS 49.11 160.61 16.18 241.96 503.38 158.71 227.04 130.95 169.83 188.38 268.63 228.80 239.61 142.09 0.00 DT 64.60 56.90 87.53 138.25 414.27 174.20 137.93 146.44 185.32 84.67 179.52 244.29 150.50 38.38 103.71 0.00 DX 57.30 64.20 80.23 145.55 406.97 166.90 130.63 139.14 178.02 91.97 172.22 236.99 143.20 45.68 96.41 7.30 0.00

Chapter 7.0 ∙ Appendix 285

Table 7.9 Mantel test distance matrix for populations of M. s. inornata in the Mitchell River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims.

MB MC ME MF MG MH MI MJ MK ML MM MN MO MP MQ MR MB 0.00 MC 88.62 0.00 ME 33.35 63.30 0.00 MF 313.44 368.71 313.44 0.00 MG 31.21 121.75 62.76 277.22 0.00 MH 98.05 186.67 131.40 182.04 95.18 0.00 MI 96.45 185.07 129.80 183.64 93.58 1.60 0.00 MJ 278.84 367.46 312.19 16.83 275.97 180.79 182.39 0.00 MK 261.12 272.00 227.77 541.21 263.99 359.17 357.57 539.96 0.00 ML 357.66 269.04 332.34 637.75 390.79 455.71 454.11 636.50 541.04 0.00 MM 88.56 177.18 113.88 270.73 85.69 68.11 87.09 269.48 345.00 446.22 0.00 MN 32.22 118.56 63.29 250.15 32.34 68.11 66.51 248.90 291.65 387.60 53.35 0.00 MO 189.51 200.39 156.16 469.60 207.65 287.56 285.96 468.35 71.61 469.43 273.39 220.04 0.00 MP 270.04 280.08 236.69 550.13 348.92 368.09 366.49 548.88 212.88 549.12 353.24 361.31 141.27 0.00 MQ 323.86 235.24 290.51 603.95 356.99 421.91 420.31 602.70 507.24 157.59 412.42 353.80 435.63 515.50 0.00 MR 180.69 92.07 155.37 460.78 213.82 278.74 277.14 459.53 364.07 357.11 269.25 210.63 292.46 372.15 323.31 0.00

Chapter 7.0 ∙ Appendix 286

Table 7.10 Average allelic diversity estimates for microsatellite loci in populations of M. australis in the Daly River, and M. s. inornata in the Mitchell River.

Site Sample Average allelic Site Code Size diversity Beeboom DA 9 0.797 Claravale/Sawmill DM 35 0.758 Middle Creek DB 4 0.685 Fergusson River DFF 27 0.000 U/S Centipede Dreaming and Snowdrop Creek DE 5 0.369 UF Daly River Road DEE 4 0.000 Flora River DGG 3 0.617 Fish River DJ 4 0.644 Big Bird Billabong DJJ 2 0.808 Umbrawarrah Gorge DKK 8 0.752 Galloping Jack's DL 26 0.812 Sandy/Hermit DLL 5 0.817 Mathison Creek DP 7 0.712 Stray Creek DR 10 0.722 Douglas River DS 12 0.618 Oolloo Crossing (upstream) DT 17 0.768 Oolloo Crossing DX 4 0.720 Ten Mile Lagoon MB 8 0.772 Mitchell River @ Gamboola MC 2 0.792 Cairo Lagoon ME 19 0.339 Fishhole Creek MF 2 0.556 Twelve Mile Lagoon MG 13 0.329 Koolatah Gauging Station MH 3 0.783 Mitchell River @ Koolatah MI 2 0.778 Magnificent Creek MJ 7 0.567 Saltwater Creek MK 25 0.482 Walsh River @ Nullinga ML 4 0.692 Palmer River @ Drumduff MM 19 0.701 Kingfish Lagoon MN 18 0.781 Dickson's Hole MO 14 0.661 Tate River @ Ootan MP 11 0.411 Emu Creek @ Petford MQ 20 0.550 Mitchell River @ St. George MR 20 0.921

Chapter 7.0 ∙ Appendix 287

Table 7.11 Haplotype distributions for H. fuliginosus in the Daly River. Haplotype frequency and sample sizes shown for all sites.

Haplotype DB DM DC DFF DD DA DE DJ DP DQ DR DS DU DN DT DX DH DL ID1 ID2 ID3 ID4 ID5 Total/Haplotype FI 1 6 4 7 1 7 2 2 2 1 4 2 4 2 3 1 1 2 1 53 F2 1 1 2 F3 1 1 2 F4 1 1 1 4 7 F5 1 1 F6 3 2 1 1 7 F7 2 3 1 3 1 3 2 3 1 19 F8 1 1 F9 3 3 1 1 1 1 1 11 F10 1 5 1 2 1 2 4 2 1 1 1 21 F11 1 1 2 F12 1 1 F13 1 1 F14 1 1 2 F28 1 1 F30 4 2 1 2 2 1 3 2 17 F31 2 1 4 1 1 1 1 2 2 1 1 1 18 F32 1 1 F33 1 1 F34 2 1 2 5 F35 1 1 1 3 F36 1 1 F37 1 1 F38 1 1 F39 1 1 F40 1 2 1 1 2 2 2 1 5 1 18 F41 1 1 F42 1 1 F43 1 1 2 F44 1 1 1 3 F45 1 1 F46 1 1 Total/Site 7 21 6 27 2 26 4 5 13 9 6 5 6 2 12 12 21 8 3 3 4 2 3 207

Chapter 7.0 ∙ Appendix 288

Table 7.12 Haplotype distributions for H. fuliginosus in the Mitchell River. Haplotype frequency and sample sizes shown for all sites.

Haplotype 58 59 60 61 62 63 118 MA MC MD MJ MK ML MO MP MQ Total/Haplotype F15 1 1 F16 1 1 F17 1 2 1 2 4 1 1 2 1 15 F18 1 2 2 5 F19 1 1 2 F20 1 1 2 1 1 6 F21 1 2 3 1 1 2 2 12 F22 1 1 2 F24 1 1 F25 1 1 F27 1 1 F29 1 1 Total/Site 2 1 4 1 1 1 10 11 2 2 1 4 1 1 2 4 48

Chapter 7.0 ∙ Appendix 289

Table 7.13 Mantel test distance matrix for populations of H. fuliginosus in the Daly River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims.

DB DM DC DFF DD DA DE DJ DP DQ DR DS DU DN DT DX DH DB 0 DM 144.4 0 DC 179.8 265.5 0 DFF 225.8 81.35 346.8 0 DD 438.6 294.1 559.6 342.5 0 DA 35.87 121.5 144 202.9 415.6 0 DE 487.2 342.8 608.2 391.1 48.63 464.3 0 DJ 117.7 203.3 151.8 284.7 497.5 81.84 546.1 0 DP 223.4 79 344.5 127.3 276.8 200.5 325.4 282.3 0 DQ 210.9 66.43 331.9 114.7 264.2 187.9 312.9 269.8 12.57 0 DR 125.9 42.73 247 124.1 336.9 103 385.5 184.8 121.7 109.2 0 DS 16.18 160.6 193.1 242 454.8 49.11 503.4 131 239.6 227 142.1 0 DU 357.6 213.2 478.7 261.6 80.93 334.7 129.6 416.6 195.9 183.3 255.9 373.8 0 DN 98.54 184.2 81.3 265.5 478.3 62.67 526.9 70.5 263.2 250.6 165.7 111.8 397.4 0 DT 87.53 56.9 208.6 138.3 365.6 64.6 414.3 146.4 150.5 137.9 38.38 103.7 284.7 127.3 0 DX 80.23 64.2 201.3 145.6 358.3 57.3 407 139.1 143.2 130.6 45.68 96.41 277.4 120 7.3 0 DH 249.7 105.3 370.8 153.6 188.9 226.8 237.5 308.6 87.96 75.39 148 265.9 107.9 289.5 176.8 169.5 0 DL 252.5 108 373.5 156.4 186.1 229.5 234.8 311.4 90.69 78.12 150.8 268.6 105.2 292.2 179.5 172.2 2.73 ID1 38.34 124 158 205.3 418.1 14.06 466.7 95.9 203 190.4 105.4 51.67 337.2 76.73 67.06 59.76 229.3 ID2 42.12 127.7 161.8 209.1 421.9 17.84 470.5 99.68 206.7 194.2 109.2 55.45 341 80.51 70.84 63.54 233 ID3 50.06 135.7 169.8 217 429.8 25.78 478.5 107.6 214.7 202.1 117.2 63.39 348.9 88.45 78.78 71.48 241 ID4 227.8 83.34 348.8 62.66 312.2 204.8 360.8 286.7 129.2 116.6 126.1 244 231.3 267.5 140.2 147.5 123.3 ID5 135.9 52.7 256.9 134.1 346.8 113 395.5 194.8 131.7 119.1 9.97 152.1 265.9 175.6 48.35 55.65 158

Chapter 7.0 ∙ Appendix 290

DL ID1 ID2 ID3 ID4 ID5

0 232 0 235.8 3.78 0 243.7 11.72 11.41 0 126.1 207.3 211.1 219 0 160.7 115.4 119.2 127.1 136 0

Chapter 7.0 ∙ Appendix 291

Table 7.14 Mantel test distance matrix for populations of H. fuliginosus in the Mitchell River. Distance is in kilometres, and is measured via waterway distance i.e. as the fish swims.

58 59 60 61 62 63 118 MA MC MD MJ MK ML MO MP MQ 58 0 59 45.9 0 60 19.96 34.07 0 61 166.91 212.81 186.87 0 62 73.96 119.86 93.92 92.95 0 63 191.8 237.7 211.76 32.89 117.84 0 118 415.14 461.04 435.1 398.1 341.18 422.99 0 MA 410.95 456.85 430.91 393.91 336.99 418.8 4.19 0 MC 170.12 216.02 190.08 96.66 96.16 121.55 301.44 297.25 0 MD 288.46 334.36 308.42 215 214.5 239.89 126.69 122.49 118.34 0 MJ 438.08 483.98 458.04 275.88 364.12 245.91 612.48 608.29 367.46 485.8 0 MK 101.88 60.47 90.05 268.79 175.84 293.68 517.02 512.83 272 390.34 539.96 0 ML 439.16 485.06 459.12 365.7 365.2 390.59 32.4 28.21 269.04 150.7 636.5 541.04 0 MO 30.27 15.63 18.44 197.18 104.23 222.07 445.41 441.22 200.39 318.73 468.35 71.61 469.43 0 MP 111 156.9 130.96 277.91 184.96 302.8 525.1 520.91 280.08 398.42 548.88 212.88 549.12 141.27 0 MQ 405.36 451.26 425.32 331.9 331.4 356.79 133.54 129.35 235.24 116.9 602.7 507.24 157.59 435.63 515.5 0

Chapter 7.0 ∙ Appendix 292

Table 7.15 Standard lengths (mm) and life stage for M. s. inornata individuals analysed for both genetic and isotopic variation in the upper Mitchell River catchment.

Individual Standard Length Life Site Code (mm) Stage Cooktown Crossing (pre-wet) MSISI1-1 39.0 Adult Cooktown Crossing (pre-wet) MSISI1-2 43.0 Adult Cooktown Crossing (pre-wet) MSISI1-3 51.0 Adult Cooktown Crossing (pre-wet) MSISI1-4 46.0 Adult Cooktown Crossing (pre-wet) MSISI1-5 25.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-6 48.0 Adult Cooktown Crossing (pre-wet) MSISI1-7 40.0 Adult Cooktown Crossing (pre-wet) MSISI1-8 59.0 Adult Cooktown Crossing (pre-wet) MSISI1-9 48.0 Adult Cooktown Crossing (pre-wet) MSISI1-10 25.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-11 25.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-12 23.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-13 42.0 Adult Cooktown Crossing (pre-wet) MSISI1-14 43.0 Adult Cooktown Crossing (pre-wet) MSISI1-15 19.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-16 18.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-17 23.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-18 19.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-19 46.0 Adult Cooktown Crossing (pre-wet) MSISI1-20 35.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-21 37.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-22 27.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-23 18.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-24 23.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-25 19.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-26 22.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-27 32.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-28 35.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-29 22.0 Juvenile Cooktown Crossing (pre-wet) MSISI1-30 23.0 Juvenile McLeod River (pre-wet) MSISI2-1 48.0 Adult McLeod River (pre-wet) MSISI2-2 55.0 Adult McLeod River (pre-wet) MSISI2-3 48.0 Adult McLeod River (pre-wet) MSISI2-4 32.0 Juvenile McLeod River (pre-wet) MSISI2-5 42.0 Adult McLeod River (pre-wet) MSISI2-6 40.0 Adult McLeod River (pre-wet) MSISI2-7 38.0 Adult McLeod River (pre-wet) MSISI2-8 37.0 Juvenile McLeod River (pre-wet) MSISI2-9 51.0 Adult McLeod River (pre-wet) MSISI2-10 33.0 Juvenile McLeod River (pre-wet) MSISI2-11 42.0 Adult McLeod River (pre-wet) MSISI2-12 35.0 Juvenile McLeod River (pre-wet) MSISI2-13 40.0 Adult McLeod River (pre-wet) MSISI2-14 47.0 Adult McLeod River (pre-wet) MSISI2-15 43.0 Adult McLeod River (pre-wet) MSISI2-16 56.0 Adult McLeod River (pre-wet) MSISI2-17 45.0 Adult McLeod River (pre-wet) MSISI2-18 30.0 Juvenile McLeod River (pre-wet) MSISI2-19 50.0 Adult McLeod River (pre-wet) MSISI2-20 53.0 Adult

Chapter 7.0 ∙ Appendix 293

McLeod River (pre-wet) MSISI2-21 51.0 Adult McLeod River (pre-wet) MSISI2-22 35.0 Juvenile McLeod River (pre-wet) MSISI2-23 37.0 Juvenile McLeod River (pre-wet) MSISI2-24 60.0 Adult McLeod River (pre-wet) MSISI2-25 38.0 Adult McLeod River (pre-wet) MSISI2-26 33.0 Juvenile McLeod River (pre-wet) MSISI2-27 54.0 Adult McLeod River (pre-wet) MSISI2-28 36.0 Juvenile McLeod River (pre-wet) MSISI2-29 47.0 Adult McLeod River (pre-wet) MSISI2-30 59.0 Adult Rifle Creek (pre-wet) MSISI3-1 44.0 Adult Rifle Creek (pre-wet) MSISI3-2 26.0 Juvenile Rifle Creek (pre-wet) MSISI3-3 45.0 Adult Rifle Creek (pre-wet) MSISI3-4 55.0 Adult Rifle Creek (pre-wet) MSISI3-5 53.0 Adult Rifle Creek (pre-wet) MSISI3-6 61.0 Adult Rifle Creek (pre-wet) MSISI3-7 32.0 Juvenile Rifle Creek (pre-wet) MSISI3-8 30.0 Juvenile Rifle Creek (pre-wet) MSISI3-9 60.0 Adult Rifle Creek (pre-wet) MSISI3-10 43.0 Adult Rifle Creek (pre-wet) MSISI3-11 53.0 Adult Rifle Creek (pre-wet) MSISI3-12 64.0 Adult Rifle Creek (pre-wet) MSISI3-13 52.0 Adult Rifle Creek (pre-wet) MSISI3-14 55.0 Adult Rifle Creek (pre-wet) MSISI3-15 72.0 Adult Rifle Creek (pre-wet) MSISI3-16 82.0 Adult Bushy Creek (pre-wet) MSISI4-1 77.0 Adult Bushy Creek (pre-wet) MSISI4-2 68.0 Adult Bushy Creek (pre-wet) MSISI4-3 63.0 Adult Bushy Creek (pre-wet) MSISI4-4 50.0 Adult Bushy Creek (pre-wet) MSISI4-5 44.0 Adult Bushy Creek (pre-wet) MSISI4-6 45.0 Adult Bushy Creek (pre-wet) MSISI4-7 50.0 Adult Bushy Creek (pre-wet) MSISI4-8 50.0 Adult Bushy Creek (pre-wet) MSISI4-9 77.0 Adult Bushy Creek (pre-wet) MSISI4-10 45.0 Adult Bushy Creek (pre-wet) MSISI4-11 54.0 Adult Bushy Creek (pre-wet) MSISI4-12 60.0 Adult Bushy Creek (pre-wet) MSISI4-13 49.0 Adult Bushy Creek (pre-wet) MSISI4-14 44.0 Adult Bushy Creek (pre-wet) MSISI4-15 55.0 Adult Bushy Creek (pre-wet) MSISI4-16 54.0 Adult Bushy Creek (pre-wet) MSISI4-17 48.0 Adult Bushy Creek (pre-wet) MSISI4-18 51.0 Adult Bushy Creek (pre-wet) MSISI4-19 42.0 Adult Bushy Creek (pre-wet) MSISI4-20 46.0 Adult Bushy Creek (pre-wet) MSISI4-21 58.0 Adult Bushy Creek (pre-wet) MSISI4-22 44.0 Adult Bushy Creek (pre-wet) MSISI4-23 50.0 Adult Bushy Creek (pre-wet) MSISI4-24 51.0 Adult Bushy Creek (pre-wet) MSISI4-25 47.0 Adult Bushy Creek (pre-wet) MSISI4-26 66.0 Adult Bushy Creek (pre-wet) MSISI4-27 67.0 Adult Bushy Creek (pre-wet) MSISI4-28 48.0 Adult Bushy Creek (pre-wet) MSISI4-29 50.0 Adult Bushy Creek (pre-wet) MSISI4-30 59.0 Adult Bushy Creek (during wet) MSISI5-1 49.0 Adult Chapter 7.0 ∙ Appendix 294

Bushy Creek (during wet) MSISI5-2 54.0 Adult Bushy Creek (during wet) MSISI5-3 17.0 Juvenile Bushy Creek (during wet) MSISI5-4 23.0 Juvenile Bushy Creek (during wet) MSISI5-5 39.0 Adult Bushy Creek (during wet) MSISI5-6 29.0 Juvenile Bushy Creek (during wet) MSISI5-7 23.0 Juvenile Bushy Creek (during wet) MSISI5-8 39.0 Adult Bushy Creek (during wet) MSISI5-9 31.0 Juvenile Bushy Creek (during wet) MSISI5-10 26.0 Juvenile Bushy Creek (during wet) MSISI5-11 23.0 Juvenile Bushy Creek (during wet) MSISI5-12 47.0 Adult Bushy Creek (during wet) MSISI5-13 23.0 Juvenile Bushy Creek (during wet) MSISI5-14 26.0 Juvenile Bushy Creek (during wet) MSISI5-15 27.0 Juvenile Bushy Creek (during wet) MSISI5-16 17.0 Juvenile Bushy Creek (during wet) MSISI5-17 29.0 Juvenile Bushy Creek (during wet) MSISI5-18 26.0 Juvenile Bushy Creek (during wet) MSISI5-19 41.0 Adult Bushy Creek (during wet) MSISI5-20 48.0 Adult Bushy Creek (during wet) MSISI5-21 30.0 Juvenile Bushy Creek (during wet) MSISI5-22 57.0 Adult Bushy Creek (during wet) MSISI5-23 57.0 Adult Bushy Creek (during wet) MSISI5-24 71.0 Adult Bushy Creek (during wet) MSISI5-25 47.0 Adult Bushy Creek (during wet) MSISI5-26 43.0 Adult Bushy Creek (during wet) MSISI5-27 35.0 Juvenile Bushy Creek (during wet) MSISI5-28 22.0 Juvenile Bushy Creek (during wet) MSISI5-29 39.0 Adult Bushy Creek (during wet) MSISI5-30 72.0 Adult Cooktown Crossing (during wet) MSISI6-1 22.0 Juvenile Cooktown Crossing (during wet) MSISI6-2 20.0 Juvenile Cooktown Crossing (during wet) MSISI6-3 21.0 Juvenile Cooktown Crossing (during wet) MSISI6-4 19.0 Juvenile Cooktown Crossing (during wet) MSISI6-5 47.0 Adult Cooktown Crossing (during wet) MSISI6-6 18.0 Juvenile Cooktown Crossing (during wet) MSISI6-7 27.0 Juvenile Cooktown Crossing (during wet) MSISI6-8 31.0 Juvenile Cooktown Crossing (during wet) MSISI6-9 25.0 Juvenile Cooktown Crossing (during wet) MSISI6-10 21.0 Juvenile Cooktown Crossing (during wet) MSISI6-11 20.0 Juvenile Cooktown Crossing (during wet) MSISI6-12 25.0 Juvenile Cooktown Crossing (during wet) MSISI6-13 29.0 Juvenile Cooktown Crossing (during wet) MSISI6-14 28.0 Juvenile Cooktown Crossing (during wet) MSISI6-15 37.0 Juvenile Cooktown Crossing (during wet) MSISI6-16 26.0 Juvenile Cooktown Crossing (during wet) MSISI6-17 25.0 Juvenile Cooktown Crossing (during wet) MSISI6-18 27.0 Juvenile Cooktown Crossing (during wet) MSISI6-19 27.0 Juvenile Cooktown Crossing (during wet) MSISI6-20 29.0 Juvenile Cooktown Crossing (during wet) MSISI6-21 24.0 Juvenile Cooktown Crossing (during wet) MSISI6-22 15.0 Juvenile Cooktown Crossing (during wet) MSISI6-23 18.0 Juvenile Cooktown Crossing (during wet) MSISI6-24 30.0 Juvenile Cooktown Crossing (during wet) MSISI6-25 58.0 Adult Cooktown Crossing (during wet) MSISI6-26 29.0 Juvenile Cooktown Crossing (during wet) MSISI6-27 34.0 Juvenile Cooktown Crossing (during wet) MSISI6-28 19.0 Juvenile Chapter 7.0 ∙ Appendix 295

Cooktown Crossing (during wet) MSISI6-29 44.0 Adult Cooktown Crossing (during wet) MSISI6-30 67.0 Adult Rifle Creek (during wet) MSISI7-1 38.0 Adult Rifle Creek (during wet) MSISI7-2 28.0 Juvenile Rifle Creek (during wet) MSISI7-3 33.0 Juvenile Rifle Creek (during wet) MSISI7-4 24.0 Juvenile Rifle Creek (during wet) MSISI7-5 35.0 Juvenile Rifle Creek (during wet) MSISI7-6 36.0 Juvenile Rifle Creek (during wet) MSISI7-7 18.0 Juvenile Rifle Creek (during wet) MSISI7-8 40.0 Adult Rifle Creek (during wet) MSISI7-9 40.0 Adult Rifle Creek (during wet) MSISI7-10 27.0 Juvenile Rifle Creek (during wet) MSISI7-11 33.0 Juvenile Rifle Creek (during wet) MSISI7-12 36.0 Juvenile Rifle Creek (during wet) MSISI7-13 33.0 Juvenile Rifle Creek (during wet) MSISI7-14 29.0 Juvenile Rifle Creek (during wet) MSISI7-15 46.0 Adult Rifle Creek (during wet) MSISI7-16 30.0 Juvenile Rifle Creek (during wet) MSISI7-17 39.0 Adult Rifle Creek (during wet) MSISI7-18 52.0 Adult Rifle Creek (during wet) MSISI7-19 16.0 Juvenile Rifle Creek (during wet) MSISI7-20 39.0 Adult Rifle Creek (during wet) MSISI7-21 42.0 Adult Rifle Creek (during wet) MSISI7-22 23.0 Juvenile Rifle Creek (during wet) MSISI7-23 25.0 Juvenile Rifle Creek (during wet) MSISI7-24 60.0 Adult Rifle Creek (during wet) MSISI7-25 63.0 Adult McLeod River (during wet) MSISI8-1 21.0 Juvenile McLeod River (during wet) MSISI8-2 25.0 Juvenile McLeod River (during wet) MSISI8-3 20.0 Juvenile McLeod River (during wet) MSISI8-4 22.0 Juvenile McLeod River (during wet) MSISI8-5 20.0 Juvenile McLeod River (during wet) MSISI8-6 20.0 Juvenile McLeod River (during wet) MSISI8-7 25.0 Juvenile McLeod River (during wet) MSISI8-8 25.0 Juvenile McLeod River (during wet) MSISI8-9 17.0 Juvenile McLeod River (during wet) MSISI8-10 22.0 Juvenile McLeod River (during wet) MSISI8-11 33.0 Juvenile McLeod River (during wet) MSISI8-12 19.0 Juvenile McLeod River (during wet) MSISI8-13 20.0 Juvenile McLeod River (during wet) MSISI8-14 20.0 Juvenile McLeod River (during wet) MSISI8-15 15.0 Juvenile McLeod River (during wet) MSISI8-16 12.0 larvae McLeod River (during wet) MSISI8-17 32.0 Juvenile McLeod River (during wet) MSISI8-18 30.0 Juvenile McLeod River (during wet) MSISI8-19 35.0 Juvenile McLeod River (during wet) MSISI8-20 49.0 Adult McLeod River (during wet) MSISI8-21 21.0 Juvenile McLeod River (during wet) MSISI8-22 19.0 Juvenile McLeod River (during wet) MSISI8-23 23.0 Juvenile McLeod River (during wet) MSISI8-24 34.0 Juvenile McLeod River (during wet) MSISI8-25 29.0 Juvenile McLeod River (during wet) MSISI8-26 38.0 Adult McLeod River (during wet) MSISI8-27 42.0 Adult McLeod River (during wet) MSISI8-28 37.0 Juvenile McLeod River (during wet) MSISI8-29 32.0 Juvenile McLeod River (during wet) MSISI8-30 25.0 Juvenile Chapter 7.0 ∙ Appendix 296

Table 7.16 Two-tailed t-test statistics between M. s. inornata adult and juvenile isotope signatures (C13 and N15) at each sampling time. Significant values are denoted by *, where α = 0.05.

Site Isotope df t Statistic P Cooktown Crossing (Dry season) C13 23 8.212 0.000* Cooktown Crossing (Dry season) N15 20 5.877 0.000* Cooktown Crossing (Wet season) C13 3 0.086 0.937 Cooktown Crossing (Wet season) N15 2 -0.610 0.604 Rifle Creek (Dry season) C13 11 4.134 0.002* Rifle Creek (Dry season) N15 1 -1.048 0.485 Rifle Creek (Wet season) C13 8 0.366 0.724 Rifle Creek (Wet season) N15 14 0.081 0.937 Bushy Creek (Wet season) C13 19 2.794 0.012* Bushy Creek (Wet season) N15 23 -0.452 0.655 McLeod River (Dry season) C13 25 4.076 0.000* McLeod River (Dry season) N15 17 -0.920 0.370 McLeod River (Wet season) C13 2 -4.182 0.053 McLeod River (Wet season) N15 17 8.244 0.000*

Chapter 7.0 ∙ Appendix 297

Figure 7.1 δ13C and δ15N for all samples at Bushy Creek in the upper Mitchell catchment, with all samples represented for each sampled season.

Bushy Creek

14.0

12.0 Wet Season Fish Dry Season Fish 10.0 CPOM Dry Season CPOM Wet Season 13 8.0 δ C Riparian Veg Dry Season 6.0 Riparian Veg Wet Season Filamentous Algae Dry Season 4.0 Seston (75) Wet Season 2.0 Seston (250) Wet Season FPOM Dry Season 0.0 FPOM Wet Season -35.0 -33.0 -31.0 -29.0 -27.0 -25.0 -23.0 -21.0 -19.0 -2.0 15 δ N

Chapter 7.0 ∙ Appendix 298

Figure 7.2 δ13C and δ15N for all samples at Cooktown Crossing in the upper Mitchell catchment, with all samples represented for each sampled season.

Cooktown Crossing

14.0

12.0 Fish Dry Season Fish Wet Season 10.0 FPOM Dry Season FPOM Wet Season 8.0 13 CPOM Dry Season δ C CPOM Wet Season 6.0 Riparian Veg Dry Season Riparian Veg Wet Season 4.0 Filamentous Algae Dry Season Seston (75) Wet Season 2.0 Seston (250) Wet Season

0.0 -35.0 -33.0 -31.0 -29.0 -27.0 -25.0 -23.0 -21.0 -19.0 -2.0 δ15N

Chapter 7.0 ∙ Appendix 299

Figure 7.3 δ13C and δ15N for all samples at the McLeod River in the upper Mitchell catchment, with all samples represented for each sampled season. McLeod River

14.0

Fish Wet Season 12.0 Fish Dry Season FPOM Dry Season 10.0 FPOM Wet Season CPOM Dry Season 8.0 δ13C CPOM Wet Season Riparian Veg Dry Season 6.0 Riparian Veg Wet Season Seston (75) Wet Season 4.0 Seston (250) Wet Season Biofilm Dry Season 2.0 Biofilm Wet Season

0.0 -35.0 -33.0 -31.0 -29.0 -27.0 -25.0 -23.0 -21.0 -19.0 -2.0 δ15N

Chapter 7.0 ∙ Appendix 300

Figure 7.4 δ13C and δ15N for all samples at Rifle Creek in the upper Mitchell catchment, with all samples represented for each sampled season.

Rifle Creek

16.0

14.0 Fish Wet Season 12.0 Fish Dry Season Seston (75) Wet Season 10.0 Seston (250) Wet Season δ13C 8.0 Riparian Veg Dry Season Riparian Veg Wet Season 6.0 CPOM Dry Season 4.0 CPOM Wet Season FPOM Dry Season 2.0 FPOM Wet Season 0.0 -35.0 -30.0 -25.0 -20.0 -2.0 δ15N

Chapter 7.0 ∙ Appendix 301

Figure 7.5 Initial investigation into the variability of the marker ATPase 6 and 8 in

H. fuliginosus. Black dots denote haplotypes that have not been sampled, or have become extinct from the population. Circle size corresponds to sample size.

HF1 HF2 HF3 HF4

Mitchell River Daly River

n = 21 - 39 n = 2 - 20 n = 1

Chapter 7.0 ∙ Appendix 302