Paranephrops Zealandicus and Odonata Species in Semi-Natural Aquaculture Environments

Predator - Prey Interactions Between Paranephrops zealandicus and Odonata Species in Semi-natural Aquaculture Environments.

John Ward.

A thesis submitted in partial fulfilment for the degree of Master of Science, Zoology Degree Programme University of Otago, Dunedin, New Zealand.

25th August 2020. Front piece:

Abstract: Interactions between predator and prey contribute towards shaping food webs and can be responsible for the composition of species in an ecosystem. In semi-natural aquaculture ponds, interactions between predators and competitors can be unfavorable for farmed species as they might reduce biomass in ponds and therefore economic growth. Juvenile farmed species are significantly more vulnerable in pond aquaculture as size is usually the factor that dictates trophic levels in aquatic habitats. Odonata are a family of pond invertebrates that are considered opportunistic predators and are known to feed on the prey of many juvenile vertebrate and invertebrate ponds species. Their ability to successfully capture prey makes them top invertebrate predators and how they interact with farmed species, specifically juveniles in aquaculture pond systems has been speculated. In New Zealand (NZ), semi-natural aquaculture ponds are used to cultivate the freshwater water crayfish species Paranephrops zealandicus yet how this species interacts with other pond invertebrates, particularly odonate larvae has not been studied. It was therefore the aim of this thesis to further investigate the predator-prey interactions between NZ odonate larvae and the farmed species P. zealandicus in semi-natural aquaculture ponds. This thesis examines predator-prey interactions based on species spatial patterns and habitat use in ponds, as well as stomach content and stable isotope analysis on study species.

Results indicate odonate larvae have no significant effect on the abundance of P. zealandicus in ponds. Vegetation and the densities of P. zealandicus in ponds are more likely to have a greater impact on odonate larvae and juvenile P. zealandicus abundance. Desirable habitat was occupied by larger P. zealandicus which forced juvenile P. zealandicus to seek refuge in areas of the pond that were not attractive for larger territorial individuals. Based on stomach content analysis, odonate larvae were found to primarily feed on small sedentary pond invertebrates such as Chironomidae P. zealandicus tissue was exceptionally low in odonate larvae stomachs. Stable isotope analysis further showed all odonate larvae had lower trophic levels to juvenile P. zealandicus indicating odonate larvae are unlikely to be major predators of juveniles. P. zealandicus diets reflected their omnivorous nature feeding on both plant detritus and invertebrate tissue, however P. zealandicus were highly cannibalistic with adults found to have high volumes of P. zealandicus tissue in their stomachs. Stable isotope analysis showed between P. zealandicus and juvenile crayfish there was no significant difference

3 in trophic levels. Year 1 crayfish represented much of the P. zealandicus sample used for stable isotope analysis and adults were poorly represented in the sample which could be the reason for this result. Therefore Year 1 P. zealandicus are unlikely to be feeding on juveniles, however more research needs to be carried out looking at the trophic interactions between adult P. zealandicus and year of young.

Ultimately, odonate larvae are not likely to be major predators of juvenile P. zealandicus, and cannibalism is more likely to be the main factor for low juvenile success in ponds. Future research investigating other pond invertebrates and ways to mitigate cannibalism in ponds is speculated.

Acknowledgements: Firstly, I would like to thank my supervisor Gerry Closs for all his guidance, and patience over the duration of this thesis. His time and energy gone into this thesis was much appreciated.

John Hollows for his guidance and allowing me to carry out my study in his kōura ponds. Also, his assistance with sampling and data collection was appreciated. I hope these results may assist him in future management of the ponds.

Ernslaw one LTD for allowing me access into the forestry and providing me with assistance from Callum. His assistance in the field made sampling go by a lot quicker, and I could not have done it without him.

Russell Frew and the chemistry department for all your help surrounding stable isotope analysis. As I am not from a chemistry background my knowledge of stable isotope analysis was limited, however the chemistry department took me under their wing and provided me with the correct equipment and talked me through in detail the stable isotope analysis process.

The technical staff at zoology specifically Nat for providing me with equipment for out in the field and making my sampling process a whole lot easier.

Statistical advice that was given from the stats cadet support programme. Specifically, a big thank you to Stefanie Neupert, and Sam Macaulay for your guidance with univariate analysis and ordination plots.

Ella Sussex for your assistance with analysing samples. Without your help I would still be processing them.

Michael Fox for your support and advice through the editing process.

Support from my peers through challenges with my thesis and all their help carrying me through university has not gone unnoticed. Particularly, the Vogel St fam for your support, and wisdom in the final stages of this thesis.

Lastly mum and dad for your constant support. No matter what the situation is, you guys have constantly supported me through the process and for that I am grateful.

Table of Contents: Front piece: ...... 2 Abstract: ...... 3 Acknowledgements: ...... 5 List of figures:...... 8 List of tables: ...... 10 Chapter 1: General Introduction ...... 13 Food web interactions: ...... 13 Aquaculture: ...... 14 Kōura: ...... 15 Odonata: ...... 16 Odonate study species:...... 17 Thesis overview: ...... 18 Study objectives: ...... 18 Chapter 2: The Spatial Patterns of Paranephrops zealandicus and Odonate Larvae Species in Semi-natural Aquaculture Ponds ...... 19 Introduction: ...... 20 Aims/hypothesis: ...... 21 Methods: ...... 22 Study site:...... 22 Study design: ...... 26 Statistical Analysis: ...... 27 Results: ...... 29 Environmental factors influencing pond heterogeneity: ...... 29 Odonate composition: ...... 34 Effects of vegetation on odonate abundance: ...... 35 Habitat use and odonate assemblages: ...... 37 Paranephrops zealandicus composition: ...... 39 Ponds ...... 40 Effects of environmental factors and original pond stocking on Paranephrops zealandicus abundance: ...... 40 Effects of odonates on Paranephrops zealandicus abundance: ...... 42 Habitat use and Paranephrops zealandicus assemblage: ...... 44 Discussion: ...... 46 Odonates: ...... 46 Paranephrops zealandicus: ...... 47 Conclusion: ...... 48 6

Chapter 3: Diets of Odonate Species and Paranephrops zealandicus in Semi-natural Aquaculture Ponds: An Analysis of Stomach Contents and Stable Isotopes ...... 49 Introduction: ...... 50 Aims/hypothesis: ...... 52 Methods: ...... 53 Sample collection: ...... 53 Dissections: ...... 53 Analytic approach to stomach content analysis:...... 53 Stable isotope analysis sample collection: ...... 54 Sample preparation: ...... 55 Statistical analysis: ...... 56 Results: ...... 58 Odonate stomach content analysis: ...... 58 Volume percentages of organic matter in stomachs of Paranephrops zealandicus: ...... 64 Stomach content analysis of Paranephrops zealandicus size groups: ...... 65 Volume percentages of organic matter in stomachs of Paranephrops zealandicus size groups: ...... 66 Stable isotope analysis of odonate larvae and Paranephrops zealandicus: ...... 70 Discussion: ...... 77 Odonates: ...... 77 Parenphrops zealandicus: ...... 78 Food web interactions: ...... 80 Conclusion: ...... 81 Chapter 4: General Discussion...... 82 Introduction: ...... 82 Review of findings: ...... 82 Management implications: ...... 84 Future research: ...... 85 Conclusion: ...... 85 References: ...... 87 Appendix 1: Invertebrate composition in ponds...... 98 Appendix 2: Frequency of organic matter present in study species stomachs...... 99 Appendix 3: Frequency of organic matter present in stomachs of Paranephrops zealandicus in different habitats sampled...... 100 Appendix 4: Frequency of organic matter present in stomachs of Paranephrops zealandicus size groups...... 101

List of figures: Figure 2.1 Generalised location of sample ponds, Dusky Forest, Tapanui, New Zealand.

Figure 2.2 Images of ponds. (a) pond 1, (b) pond 2, (c) pond 3, (d) pond 4, (e) pond 5, (f) pond 6, (g), pond 7, (h) pond 8, (i) pond 9, (j) pond 10. Figure 2.3 Mean dissolved oxygen concentrations (mean ± SE) in ponds (mg/L). n = 4. Figure 2.4 Principle component analysis (PCA) of log transformed environmental factors in ponds. Co- variance represents the different environmental factors. (A) pH, (B) water clarity, (C) tree canopy overhang (%), (D) bank vegetation cover (%), (E) average phosphate concentration(mg/L), (F) average nitrogen concentration (mg/L), (G) average dissolved oxygen concentration (mg/L), (H) pond area (m2).

Figure 2.5 Canonical correspondence analysis (CCA) average abundance of Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, Procordulia grayi, and Paranephrops zealandicus in (a) benthic bracken bags and (b) littoral kick net samples. Environmental variables; (A) pH, (B) water clarity, (C) tree canopy overhang (%), (D) bank vegetation cover (%), (E) average phosphate concentration(mg/L), (F) average nitrogen concentration (mg/L), (G) average dissolved oxygen concentration (mg/L), (H) pond area (m2). Figure 2.6 Mean total abundance of odonate species (mean ± SE) Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia. smithii, and Procordulia grayi in ponds. n=3. Figure 2.7 Bivariate analysis of odonate abundance and bank vegetation cover (%) in two sampling methods. (a) Benthic bracken bag samples. Slope = -0.03, p= 0.219. (b) Littoral kick net samples. Slope = 0.19, p= 0.051.

Figure 2.8 Average abundance of odonate species (mean ± SE) Austolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, and Procordulia grayi in bracken bag and kick net samples with all ponds combined. n=30. Figure 2.9 Mean abundance of Paranephrops zealandicus size groups (mean ± SE), small (≥20mm), medium (21-40mm), and large (≤41mm) in ponds. n=3.

Figure 2.10 Bivariate analysis of Paranephrops zealandicus abundance and original stocking in ponds. Slope = 0.27, P>0.05. Figure 2.11 Bivariate analysis of Paranephrops zealandicus original stocking and bank vegetation cover (%) in ponds. Slope = 0.05 P> 0.05. Figure 2.12 Bivariate analysis of Paranephrops zealandicus abundance and bank vegetation cover (%) in ponds. Slope = 0.41, P<0.05. Figure 2.13 Bivariate analysis of odonate abundance and Paranephrops zealandicus abundance in two sampling methods. (a) Benthic bracken bag samples. Slope = -0.0953, p= 0.0601. (b) Littoral kick net samples. Slope = 0.169, p=0.427. Figure 2.14 Average abundance of Paranephrops zealandicus size groups (mean ± SE) Small, Medium, Large in bracken bag and kick net samples with all ponds combined. n=30.

Figure 3.1 Images of invertebrates identified in the stomachs of odonate larvae. (a) Cyclopoida sp., (b) terrestrial invertebrate (unknown), (c) Parenephrops zealandicus chelipeds, (d) Ostracoda sp., (e) Chironomidae sp., (f) Diptera sp.

Figure 3.2 Mean volume percentages of different stomach contents categories in odonate species stomachs. (a) Austrolestes colensonis stomach percentage, (b) Xanthocnemis zealandica stomach percentages, (c) Hemicordulia australiae stomach percentages, (d) Procordulia smithii stomach percentages, (e) Procordulia grayi stomach percentages.

Figure 3.3 NMDS showing the mean percentage volume of 10 categories present in stomachs of Austrolestes colensonis (circle), Xanthocnemis zealandica (cross), Hemicordulia australiae (fill square), Procordulia smithii (triangle), Procordulia grayi (plus), and Paranephrops zealandicus (square). Bray- Curtis similarity index, 3D dimensionality, stress = 0. (a) Represents plot axes 1 & 2. (b) Represents plot axes 2 & 3.

Figure 3.4 Images of invertebrates identified in the stomachs of Parenephrops zealandicus. (a) Austrolestes colensonis caudal lamella, (b) Xanthocnemis zealandica caudal lamella, (c) Cyclopoida sp., (d) Ostracoda sp. shell, (e) Parenephrops zealandicus end of cheliped, (f) terrestrial invertebrate (unknown), (g) Diaprepocoris sp., (h) Chironomidae sp.

Figure 3.5 Mean volume percentage of plant detritus, invertebrates and Paranephrops zealandicus tissue found in the stomachs of Paranephrops zealandicus. n=99.

Figure 3.6 Mean volume percentage of plant detritus, invertebrates, and Paranephrops zealandicus tissue found in the stomachs of Paranephrops zealandicus size groups (mean ± SE), Small (n=37), Medium (n=33), and Large (n=30).

Figure 3.7 Mean volume percentages of different stomach contents categories in stomachs of Paranephrops zealandicus size groups. (a) Small, (b) Medium, (c) Large.

Figure 3.8 Mean isotopic values of stable carbon (δ13C) and nitrogen (δ15N) for Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, Paranephrops zealandicus, and YOY. (a) pond 1, (b) pond 2, (c) pond 6. D. rhantus, A. wakefieldi, and Chironomidae sp. were used to help portray food web structure.

List of tables: Table. 2.1 Physical characteristics of ponds; Pond (pond reference number) Pond code (for forestry reference), Area (m2), Ph, Water clarity (m), Sediment (SB = soft bottom), Canopy overhang cover (%), Bank vegetation cover (%), Total phosphate concentration start of sampling (µg/L), Total phosphate concentration end of sampling (µg/L), Total nitrogen concentration start of sampling (µg/L), Total nitrogen concentration end of sampling (µg/L). Table. 2.2 Bivariate analysis displaying correlation coefficient and P-values of odonate species abundance and bank vegetation cover (%). Significant values are highlighted. Table. 2.3 Bivariate analysis displaying correlation coefficient and P-values of odonate species abundance and Paranephrops zealandicus abundance. Significant values are highlighted. Table 2.4 Results of Welch F test for total abundance of odonate species; Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, and Procordulia grayi between bracken bag and kick net sampling methods in all ponds combined. F represents the F-value, Omega2 represents the effect size, and P represents the P-value. Significant values are highlighted. Table 2.5 Paranephrops zealandicus originally stocked in ponds and the population densities in ponds (m2). Table 2.6 Results of Welch F test for total abundance of Paranephrops zealandicus size classes; small, medium, and large between Kick net and bracken bag sampling methods (n=30). F represents the F ratio, Omega2 represents the effect size, and P represents the P value. Significant values are highlighted. Table 3.1 Mean invertebrate taxon richness in stomachs of study species (mean ± SE), Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, Procordulia grayi, Paranephrops zealandicus.

Table 3.2 Bonferroni post hoc test of stomach volume percentages in stomach of Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, Procordulia grayi, Paranephrops zealandicus. Significant values are highlighted.

Table 3.3 ANOVA and Dunn’s post hoc test of mean volume percentage of plant detritus, invertebrates, and Paranephrops zealandicus tissue found in the stomachs of Paranephrops zealandicus.

Table 3.4 Mean taxon richness in stomachs of Paranephrops zealandicus size groups (mean ± SE), Small (n=37), Medium (n=33), and Large (n=30).

Table 3.5 ANOVA and Tukey post-hoc test of log transformed taxon richness in stomachs of Paranephrops zealandicus size groups small, medium, and large. Significant values are highlighted.

Table 3.6 One-way ANOVA and Dunn’s post hoc analysis of mean volume percentage of plant detritus, invertebrates, and Paranephrops zealandicus tissue in stomachs of Paranephrops zealandicus size groups small, medium, and large. Significant values are highlighted.

Table 3.7 One-way ANOVA and Dunn’s post hoc analysis of mean volume percentage of plant detritus, invertebrates, and Paranephrops zealandicus tissue in stomachs of Paranephrops zealandicus size groups small, medium, and large. Significant values are highlighted.

Table 3.8 ANOVA and Tukey post hoc analysis giving fractionation values of stable carbon (δ13C) and nitrogen (δ15N) for Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, and Paranephrops zealandicus in pond 1. Significant values are highlighted. Table

3.9 ANOVA and Tukey post hoc analysis giving fractionation values of stable carbon (δ13C) and nitrogen (δ15N) for Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, and Paranephrops zealandicus in pond 2. Significant values are highlighted.

Table 3.10 ANOVA and Tukey post hoc analysis giving fractionation values of stable carbon (δ13C) and nitrogen (δ15N) for Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, and Paranephrops zealandicus in pond 6. Significant results are highlighted.

Chapter 1: General Introduction.

Interactions involving top predators in semi-natural aquaculture ponds are important for commercial production of farmed species. How farmed species interact with other organisms in these partly artificially assembled communities can provide insights into the assembly of natural communities, and the life history responses of species, advancing both pond aquaculture management and theory related to community assembly.

Naturally, ponds play a functional role in their environment benefiting metacommunity dynamics. De Meester et al., (2005), refers to ponds as being great “stepping stones” for species between terrestrial and aquatic environments. The specific abiotic characteristics and physiochemistry in ponds provides ideal conditions for many plants and animals (Janssen et al., 2018). Without macrophytes invertebrate species would not be as diverse in ponds as vegetation provides habitat and food for many species (Hassall, Hollinshead, & Hull, 2011). Even though ponds are smaller water bodies, healthy pond systems usually support higher amounts of invertebrate diversity compared to lakes and rivers (Céréghino et al., 2007; Williams et al., 2004). As a result, ponds are ideal for studying invertebrate community dynamics and predator-prey interactions. Despite this, little research has been conducted on ponds in NZ and their ecological and economic value they hold.

Food web interactions: Food webs are a useful tool in helping understand the ecological roles of organisms in pond communities (Allesina, & Pascual, 2009). Whether it is to better understand a species role in its environment, determine energy flows, or to analyse the diversity of species within a community, food webs can clearly show patterns of community interactions in ponds. Food webs can be complex structures consisting of many different species with interactions between top, intermediate, and basal species creating connections that shape pond communities. Predation is a key driver of food web dynamics, influencing prey survival (Stamp, & Bowers, 1991), reproductive recruitment (Miller et al; Tonn et al. 1992), prey behaviour (Stein, 1979; Lima, & Dill, 1990), distribution,

and foraging patterns (Dill, & Fraser, 1984). In turn, prey can also influence the demographics of predators affecting feeding rates (Sih, 1982; Benoit-Bird, & Au, 2003), growth, and reproductive success, survivorship (Mayntz, & Toft, 2001), and distribution (Bernstein, Auger, & Poggiale, 1999).

The size of a predator usually dictates the range of prey sizes that are available to them, regulating predator-prey interactions into size structured cohorts (Olson, 1996). Warren and Lawton (1987), suggested that many predator-prey interactions in food webs systems are controlled by body size, where predators that are significantly larger than their prey can influence community structure and food web interactions (Jennings & Mackinson, 2003; Shurin et. al, 2005; Sheldon, 1969). Large top predators can increase an individual’s connectivity with other species in its environment, influencing how other species may interact. (Cohen et al.1993). Whilst large predators tend to dominate higher trophic positions in food webs, for each individual, surviving to adulthood can be challenging, given that even large polyphagous predators start off as small juveniles that are predated on by other predators. Studies have shown that whilst body size in food web systems is strongly correlated to trophic level, mean body size of specific species is weakly correlated to their trophic level (Fry, & Quiñones, 1994; Jennings, Warr, & Mackinson, 2002). This suggests size is more important when looking at tropic position in food webs compared to allocating species to a specific trophic level.

Aquaculture: Over the last 30 years aquaculture has increased in global recognition, and the range of species reared for aquaculture has diversified (Jeffs, & Hooker, 2000). World protein shortages leading to nutritional deficiencies has catalysed this awareness for aquaculture, with lakes, rivers, and ponds becoming recognised for their potential (Bardach et al., 1972). Freshwater species contribute towards a significant proportion of global aquaculture systems with many macroinvertebrate species recognised for their potential (Bardach, Ryther, & McLarney, 1972). Freshwater crayfish aquaculture is now a major industry usually carried out in intensive or semi-intensive pond systems. Farming crayfish species in aquaculture can have rapid success as they have polytrophic larvae, have a relatively high fecundity, rapid growth rates, mature relatively quickly (Huner, & Lindqvist, 1995), and can be farmed in relatively high densities (Jones, Medley, & Avault Jr, 1994). The main countries farming freshwater crayfish are USA, Australia, and Europe which produce between 40,000 – 60,000 tonnes annually (Holdich, 2002). Intensive practises can slow the growth of crayfish in ponds, and increases cannibalism

(Manor et al., 2002; Romano, & Zeng, 2017). At higher densities, the period between moults increases which ultimately decreases the rate of growth in crayfish ponds (Ramalho, Correia, & Anastácio, 2008). This is a behavioural response among crayfish as at higher densities crayfish become more territorial and are known to have cannibalistic tendencies (Taugbøl, & Skurdal, 1992). During moulting periods crayfish are not able to defend themselves against other crayfish, so by limiting the number of moults in the presence of high densities of crayfish the risk of predation is reduced. For juvenile crayfish, longer inter-moult periods make them more vulnerable to predation from larger crayfish and pond invertebrates, as their size restricts them to a lower trophic level. Juvenile crayfish could therefore have increased mortality rates as a result of more predators. Also associated with intense crayfish aquaculture are potential problems, such as habitat degradation and the spread of disease to wild populations (Folke, & Kautsky, 2014). Correct management developing semi-natural conditions mimicking natural ecosystems can lead to a lower amount of environmental effects mitigating the susceptibility of disease and environmental damage (Folke, & Kautsky, 1992). Semi-natural aquaculture ponds that develop self-sustaining ecosystems for farmed species is an effective way to carry out aquaculture with low environmental impacts (Hutchinson, 2005). Freshwater crayfish aquaculture has recently become a topic of interest in NZ with endemic species successfully reared in semi-natural aquaculture pond systems.

Kōura: There are two species of freshwater crayfish also known as kōura in NZ, Paranephrops planifrons located in the North Island and west coast of the South Island, and Paranephrops zealandicus located in eastern and southern regions in the South Island. P. zealandicus can be easily distinguished from P. planifrons as the chelipeds are hairier and larger, as well as P. zealandicus on average being the slightly larger species (Hopkins, 1970; Whitmore et al., 2000). Both species are endemic to NZ and are culturally significant among Māori. For centuries kōura were caught in vegetated traps called whakaweku traditionally made from bundles of bracken fern, which are still an effective way to trap kōura (Kusabs 2015). The kōura species that will be investigated in this study is P. zealandicus.

Size is important for survival in aquaculture environments and dictates the type of interaction between farmed crayfish and other pond invertebrates. In order to grow, P. zealandicus must shed their exoskeleton making them extremely vulnerable to predation. Moults are usually more frequent when younger and decrease with age. Hopkins (1967) recorded kōura moulting up to 9 times in their first year, 3 times in their second year, and twice in their third year. Many

factors influence the moulting and hardening process in crayfish with temperature being a significant contributing factor (Hammond et al., 2006). In cooler temperatures (1.8C-11.9C) of headwater streams P. zealandicus were found to be much slower growing (Whitmore et al., 2000), whereas in studies with warmer temperatures (14C – 22C), inter- moult periods decreased (Hammond et al., 2006). Nutrient concentrations especially calcium is another limiting factor on growth rates in aquatic environments. Kōura like most crayfish species have high calcium demands due to their calcified exoskeletons and constant need to moult. In environments with low calcium concentrations crayfish can undergo metabolic stress decreasing their survivability (Cairns, & Yan, 2009). A study by Hammond et al (2006), found survivability to increase in high calcium concentrations over 10 mg 1-1. Having conditions that prompts shorter moult periods and faster growth rates is particularly crucial for the survivability of juvenile crayfish. In aquaculture environments conditions such as stock densities, predation, competition, and nutrient input can all be controlled enhancing faster recruitment of juvenile stock (Westin, & Gydemo, 1986). Understanding the predator prey interactions of P. zealandicus in semi-natural aquaculture ponds can provide insights into community dynamics and help boost productivity in aquaculture ponds.

Odonata: One group of invertebrates in aquatic food webs that are considered top predators is odonate larvae. Odonates can be found in a variety of different habitats including lakes, streams, and ponds (Wahizatul-Afzan, Julia, & Amirrudin, 2006). Being polyphagous predators, they feed on a wide variety of pond invertebrates (Johansson, 1993). Odonates use an active or sit and wait approach when seeking out their prey. This ability to alternate foraging behaviours increases the interactions with other pond invertebrates making them heavily intertwined in food webs (Johansson, 1993). Jara (2008), suggested prey size selectivity in odonates may be strongly related to their size and ability to successfully capture prey. The late instar of many odonate larvae have been found to predate on aquatic vertebrates such as juvenile fish species (Kasimov, 1956), and tadpoles (Jara, 2008). The ability to successfully capture much larger prey is concerning for farmed species in freshwater hatcheries and aquaculture farms (Gydemo, Westin, & Nissling, 1990). Whilst they do provide supplementary feed for farmed species (Santos, Costa, & Pujol-Luz, 1988), intense predation from top opportunistic predators like odonates on farmed species could be detrimental to juveniles resulting in economic loss (De Marco Jr, Latini, & Reis, 1999). Few studies have investigated the interactions between late instar odonate larvae and aquaculture species.

Odonate study species:

Austrolestes colensonis:

Known as the blue damselfly and is the largest endemic damselfly species found in NZ. Fully grown larvae are between 17 – 21mm. Young are active swimmers and are quite cannibalistic whereas later instars prefer moving between aquatic vegetation where they actively hunt prey (Rowe, 1987).

Xanthocnemis zealandica:

This endemic species is known as the red damselfly and is found throughout NZ. Fully grown larvae are between 15 – 18mm and are usually found in aquatic vegetation in ponds. X. zealandica are less active compared to other damselfly larvae and are known to be territorial occupying some form of aquatic vegetation waiting for prey to drift pass. They will often defend a desirable piece of vegetation from others if required (Rowe, 1992).

Hemicordulia australiae:

This is an introduced species to NZ and is dominant in freshwater communities of the North Island. Southern populations are very unstable as temperature is thought to effect activity rates, however with warming temperatures, populations further south have increased in status (Armstrong, 1978). Fully grown larvae are usually between 16 – 20mm in length. The larvae are found in benthic detritus but unlike P. smithii and P. grayi, they are highly active. H. australiae hunt prey through vegetation and appear to be more active at night (Rowe, 1987).

Procordulia grayi:

Endemic to NZ, Procordulia grayi is commonly called the yellow spotted dragonfly. Fully grown larvae are between 22 -27mm in length. Larvae are very inactive when exposed to light. They hunt at night chasing prey and bury themselves in the sediment during the day usually in benthic detritus beds to avoid predators (Rowe, 1987).

Procordulia smithii:

Another endemic dragonfly known as the ranger dragonfly, Procordulia smithii is found throughout NZ. Fully grown larvae are between 19 – 22mm in length. They are slow moving, and burrow into benthic detritus beds much like P. grayi hunting for prey at night before

burrowing back into the sediment. There is no information on their invertebrate predators (Rowe, 1987).

Thesis overview: The purpose of this research was to investigate the spatial distribution and diets of Paranephrops zealandicus and five odonate species in semi natural aquaculture ponds. This research was in association with the company Keewai managed by John Hollows. Keewai manage ponds in forestry areas throughout the Southland region of NZ, surrounded by pine forestry. The secluded forestry areas provide isolation and security from humans making it the perfect area for ponds to naturally thrive without the risk of kōura being taken.

There is a problem with low juvenile survival in ponds as the juvenile demographic of kōura is significantly underrepresented. This could mean that after harvesting, recruitment of the juvenile demographic is delayed leading to gaps in harvests. Because the ponds are semi- natural ecosystems, they support a diverse community of aquatic pond macroinvertebrates, including damselfly and dragonfly larvae. Based on John Hollows knowledge of the ponds and what previous literature suggests, odonate larvae could be predating on the juvenile biomass of kōura and be the reason why juveniles are not significantly represented in the crayfish population in ponds.

Study objectives: My objectives for this study are (1) investigate the spatial patterns of P. zealandicus and odonate species among different microhabitats to examine how environmental factors and interactions between P. zealandicus and odonates influence habitat use. (2) Investigate the diets of P. zealandicus and odonates in semi natural aquaculture ponds using stomach content and stable isotope analysis to determine likely trophic relationships and determine whether odonate species are predators of juvenile P. zealandicus contributing towards their low biomass in ponds. From these results I hope to propose a better understanding of community dynamics between top predators (odonates) and farmed species (crayfish) in semi-natural aquaculture pond systems that hopefully can be applicable to crayfish aquaculture.

Chapter 2: The Spatial Patterns of Paranephrops zealandicus and Odonate Larvae Species in Semi-natural Aquaculture Ponds.

Introduction:

The spatial patterns of farmed species in aquaculture pond systems can be significantly affected by both environmental factors and predator-prey interaction (Jones, Lawton, & Shachak, 1994; Scheffer et al. 2006). In semi-natural pond systems, juveniles are more vulnerable to predation from invertebrates, and may be restricted to habitats that provide refugia from predators (Heck, & Crowder, 1991). Because of this, the densities of farmed juvenile species may spatially vary between different pond habitats.

Maintaining vegetative habitat in and surrounding ponds is important as vegetation can increase biodiversity in pond communities (Hassall, Hollinshead, & Hull, 2011). In the littoral zone, bank side vegetation can provide shelter for juvenile species from predators (Sychra, Adámek, & Petřivalská, 2010), and can also function as a buffer zone limiting the amount of nutrient runoff from entering into ponds (Sollie, & Verhoeven, 2008). Vegetation in the benthic zone increases habitat complexity, providing opportunities for invertebrate species to colonise new areas within ponds (Yuan, Lu, Liu, 2002). Pond physiochemistry, and the density of detritivores in ponds can both influence the amount of vegetation (Scholz, 1990; Wissinger, Perchik, & Klemmer, 2018). Ponds with low vegetation could increase niche overlap between species as there is less desirable habitat and predator-prey interactions may be more frequent, influencing spatial patterns among pond invertebrates (Brown, 1982). As a result, habitat complexity within ponds can significantly influence the spatial distribution of macroinvertebrate communities (Ulehlova, & Přibil, 1978).

Crayfish are regarded as keystone species and can significantly impact food web dynamics in aquatic ecosystems (Collier, Parkyn, & Rabeni, 1997; Crandall, & Buhay, 2007; Reynolds, & Souty-Grosset, 2011). They carry out many functional roles in pond communities which includes the processing of allochthonous and autochthonous organic carbon (Fisher, 1973), and facilitating nutrient cycling (Grimm, 1988). Being omnivores, crayfish can directly affect other species through predation and indirectly by feeding on macrophytes which reduces desirable habitat for a variety of invertebrate species (Parkyn et al., 2001). In a study by Nyström (2017), invertebrate species richness in ponds decreased significantly as crayfish abundance increased. It is the combination of all these roles that crayfish carry out that makes them such an important piece to functioning ecosystems compared to other invertebrate species.

Larger crayfish can also influence spatial distribution patterns among juvenile crayfish. Stewart and Tabak (2011), found a size-based dominance hierarchy exists among P. zealandicus, with smaller individuals avoiding larger individuals based on the territorial nature of the larger individuals seeking more desirable habitat. Size is usually the variable dictating the trophic structure of food webs (Warren, & Lawton, 1987), and whilst crayfish normally occupy a higher trophic level in pond food webs, juvenile crayfish are smaller and have a lower trophic position than adults. This makes juvenile crayfish more vulnerable to predation influencing spatial distribution patterns in ponds. Larger crayfish have been found to prefer substrates such as cobbles, whereas juveniles prefer finer substrates which are easier to burrow into (Collier et al., 1997). P. zealandicus are spatially territorial in their natural habitat, and the presence of other crayfish or predatory invertebrates could influence the distribution of juvenile P. zealandicus. In semi-natural aquaculture ponds, P. zealandicus have not been well studied. Therefore, our understanding of inter-cohort interactions and potential interactions with other invertebrate species, like predatory odonates, is unknown.

Odonate larvae have adapted to living in different pond habitats which has facilitated their success as predators in pond ecosystems (Wahizatul-Afzan, Julia, & Amirrudin, 2006; Sheldon, & Walker, 1998). Some species are better adapted to living amongst vegetation (Wahizatul- Afzan et al., 2006), whereas others are more suited to benthic environments (Layton, & Voshell Jr, 1991). Because of this, several odonate species can coexist in the same pond occupying different habitats (Rowe, 1987). Even in the presence of higher trophic predators, odonates have been found to thrive in ponds (Nyström et al., 1996). In NZ, the distribution of odonate species has been well documented in their natural aquatic habitats (Rowe, 1987). However, their spatial distribution in aquaculture systems is unknown, as is how they interact with different life history stages of crayfish in aquaculture systems.

Aims/hypothesis: It was the objective of this chapter to investigate the spatial patterns of P. zealandicus and late instar odonate species in multiple crayfish aquaculture ponds from different microhabitats to examine how environmental factors and interactions between P. zealandicus and odonates can influence habitat use. Based on previous studies, vegetation in ponds can significantly influence the diversity and abundance of invertebrate species (Ulehlova, & Přibil, 1978). I hypothesize the spatial patterns of P. zealandicus and odonate species to be significantly influenced by vegetation, as it will affect their abundance in ponds. Large P. zealandicus have a territorial 21 nature when it comes to defending desirable habitat and can significantly influence invertebrate communities directly and indirectly (Stewart, & Tabak, 2011 Nyström, Brönmark, & Graneli, 1996). Because of this I hypothesize larger P. zealandicus will significantly affect the abundance and spatial distribution of odonate species, and juvenile P. zealandicus.

Methods:

Study site: This study was carried out in Dusky Forest, Tapanui, south West Otago (Fig 2.1). Tapanui has a cool and temperate climate with consistent precipitation throughout the year, receiving on average 82.3mm of rainfall over the summer period, and 71.3 mm on average during the winter period. Temperatures fluctuate greatly in summer and winter periods, with temperatures reaching above 20 degrees in summer and below zero in winter months (Climate-Data, 2019). Dusky Forest lies approximately 500m above sea level and is a functioning forestry owned by Ernslaw One LTD, with pine trees (Pinus radiata) the main species cultivated. There are no public access ways into the forest, with only forestry workers having access to the area. Forestry ponds are scattered in clusters throughout the forestry block in close proximity, usually within three meters of each other. The 10 ponds selected were previously stocked with different densities of P. zealandicus, however all ponds had relatively low densities of crayfish originally stocked in them compared to densities normally stocked in crayfish aquaculture pond systems (Brown et al., 1995).

All ponds were measured using tape measures and were of a similar size (approx. 50m2). Substrate primarily comprises of anoxic mud with fringing shingle rock around the littoral zone. Ponds had been stocked with vegetation including, bracken fern (Pteridium aquilinum), pine branches, and scotch broom (Cytisus scoparius) to increase habitat complexity, but the exact amounts of vegetation to ponds was unknown. Surrounding the ponds, the percentage of bank vegetation cover and tree canopy cover across ponds varied from a low amount (>10%) to a high amount (<90%) (Fig 2.2). This was estimated at each pond using vegetation survey methods (Freshwater Habitats Trust, 2015). Common predators of P. zealandicus such as eels and brown trout, were not present in the ponds due to previous monitoring and removal. Shags (Cormorants) are also common predators and can access the ponds, however none were observed over the duration of the study.

Figure 2.1 Generalised location of sample ponds, Dusky Forest, Tapanui, New Zealand.

23 a b

c d

e f

Figure 2.2 Images of ponds. (a) Pond 1, (b) Pond 2, (c) Pond 3, (d) Pond 4, (e) Pond 5, (f) Pond 6, (g), Pond 7, (h) Pond 8, (i) Pond 9, (j) Pond 10.

Study design: Four bracken bags (1m x 0.5m onion bags stuffed with bracken fern and weighted with two large rocks) were placed in each of the 10 ponds on 9th November 2018 to sample P. zealandicus and other invertebrates from the benthic habitat. Five holes were made on each side of the bag to allow invertebrates to better access the bags. The bracken bags were chosen as a sampling method to replicate a 0.5m2 area of benthic habitat that occurs in ponds. Bags were placed in ponds one month prior to sampling to allow time for macroinvertebrates to colonise them. Four bags were placed in each pond and were randomly allocated to either the left, centre, or right-hand side of the pond at distances 1m, 2m, 3m, and 4m from a pond bank edge. To ensure bags would remain submerged over the summer period if water levels decreased, bags were placed at a distance of 30cm from the waterline. Ponds were sampled 3 times over the summer period. Due to the extensive sampling protocol, only 5 ponds could be accurately sampled in one day, so there were six sampling dates as follows: 21st December 2018, 27th December 2018, 15th January 2019, 23rd January 2019, 28th February 2019, and 8th March 2019. Animal ethics approval was also obtained before sampling commenced to ensure the correct handling of crayfish.

Pond physiochemical factors and pond vegetation was assessed to determine whether they influence differences in crayfish density between ponds. Total nitrogen and phosphate concentrations were measured at the beginning (21st December 2018) and the end of the sampling dates (8th March 2019). Water samples were collected in 50 mL falcon tubes to be analysed in the laboratory using the San++ Automated Continuous Flow Analyzer. Water clarity was measured using a Secchi water clarity tube, and pH was taken using a HI98127 pHep®4 pH/Temperature Tester (HANNA instruments). Bag depth was measured where the four bags in each pond were situated, and the dissolved oxygen levels (mg/L) at each bag depth was taken with a YSI Professional Plus instrument (Professional Series - Instrument 6050000, YSI Incorporated, Yellow Springs, Ohio, USA). Bank vegetation cover (% area covering pond) and tree canopy overhang (% area covering pond) were assessed. Environmental assessments were completed before invertebrate sampling commenced.

Each bag was collected using a scoop net and placed into a fish bin to ensure all invertebrates were collected. Bracken was washed thoroughly, and samples inspected for all odonate species and P. zealandicus. odonates were collected and put into 500 mL containers in ethanol to be identified in the laboratory. P. zealandicus were counted, and orbit carapace length (OCL) and measured with Vernier Callipers - this method of assessing size was chosen as it is a

standard crayfish measurement, measuring from behind the eye to the back of the carapace (Kusabs et al., 2005). Following measurement, P. zealandicus were released back into the pond they were sampled from. Bracken bags were repacked and placed back in the ponds in the same position.Kick net samples were also taken from around the edge of the pond. A one metre long stretch of a pond’s bank was kick netted for five seconds scooping contents into a fish bin. This was repeated until the whole perimeter of the pond was sampled. Samples were checked for the presence of P. zealandicus – if present, the OCL of each individual was measured then specimens were released. The remainder of the sample was preserved in ethanol for later taxonomic processing. The sampling protocol was repeated for all 10 ponds.

In the laboratory, all samples were processed and identified using the Guide to the Aquatic Insects of New Zealand (Winterbourn et al., 1989) and the online Manaaki Whenua Freshwater invertebrates guide (Moore, 2001). All odonates were measured and counted, whereas only presence/absence was recorded for other invertebrate taxa (see Appendix 1). Crayfish were allocated into small, medium, and large size groups based on OCL measurements. Data was recorded in Microsoft Excel (2018).

Statistical Analysis: Prior to analysis, average phosphate and nitrogen concentrations were converted to mg/L to standardise data. Environmental variables were then log transformed to better approximate for multivariate normality. Pond sediment type was left out of this analysis as all ponds had the same sediments. A principal component analysis (PCA) with a correlation matrix was used to summarise the similarities between ponds based on the environmental data. A canonical correspondence analysis (CCA) was used to examine the effects environmental factors have on study species mean abundance. Before carrying out a CCA, species average abundances were log transformed to normalise the data. CCA is a multivariate analysis used to show clear relationships between species and environmental data usually along an environmental gradient by combining ordination and regression techniques. Three bivariate analyses were carried out to show the relationship between odonate larvae abundance and bank vegetation cover, and P. zealandicus abundance and bank vegetation. The third bivariate analysis showed the relationship between odonate larvae abundance and P. zealandicus abundance. Habitat preferences between littoral and benthic samples were investigated for odonate species and P. zealandicus size groups. Mean abundance of odonate species were compared between

27 littoral kick net and benthic bracken bag samples. Data was highly skewed even with log transformations, so a Welch’s F test was carried out to compare means. It should be noted that the assumptions of data normality and homogeneity of variance were violated, so the results should be interpreted with caution. Crayfish mean abundance from littoral kick net and benthic bracken bag samples was analysed to see if different size groups of crayfish have a preference between habitats. Data was highly skewed even with transformations, so a Welch’s F test was carried out to compare means. It should be noted that the assumptions of data normality and homogeneity of variance were violated so these results should also be interpreted with caution. All analyses were carried out using the statistical program Past (version 3.26, 2019),

Results: Environmental factors influencing pond heterogeneity: Phosphate concentrations in ponds were relatively high with ponds 4, 7, and 8 having a significant increase in phosphate by the end of sampling. Nitrogen concentrations were consistently high, with ponds 7, and 8 also having a significant increase by the end of sampling (Table 2.1). Ph varied across all ponds ranging from 5.47 – 6.93 making all ponds slightly acidic to relatively neutral (Table 2.1). Dissolved oxygen levels did not vary significantly between depths in the same pond, but between ponds mean dissolved oxygen levels did. Ponds 4, and 9 were nearly hypoxic with mean dissolved oxygen levels of 1.25, and 1.89 mg/L. Pond 10 was well oxygenated with a mean dissolved oxygen level of 11.24 mg/L. Other ponds were not as oxygenated but were not considered to be within the hypoxic range (Fig 2.3). It should be mentioned that dissolved oxygen levels between ponds were taken at different times of the day which could contribute to the variation in dissolved oxygen levels existing between ponds.

14 12 10 8 6 4 2 0 1 2 3 4 5 6 7 8 9 10 Mean DO concentration (mg/L) concentration DO Mean Ponds

Figure 2.3 Mean dissolved oxygen concentrations (mean ± SE) in ponds (mg/L). n = 4.

Table 2.1 Physical characteristics of ponds; Pond (pond reference number) Pond code (for forestry reference), Area (m2), pH, Water clarity (m), Sediment (SB = soft bottom), Canopy overhang cover (%), Bank vegetation cover (%), Total phosphate concentration start of sampling (µg/L), Total phosphate concentration end of sampling (µg/L), Total nitrogen concentration start of sampling (µg/L), Total nitrogen concentration end of sampling (µg/L).

Pond Pond code Area pH Water Sediment Canopy Bank Total Phosphate Total Phosphate Total Nitrogen Total Nitrogen Clarity overhang vegetation concentration concentration concentration concentration cover cover start end start end

#1 D126 34 6.16 0.21 SB 20 80 83.52 69.64 1766.60 1264.55

#2 D125 48 5.94 0.23 SB 10 70 75.14 81.13 1478.07 1342.54

#3 D124 48 6.14 0.22 SB 10 60 81.67 43.46 1559.11 1006.12

#4 D123 40 5.95 0.19 SB 20 95 77.26 138.49 1634.32 1579.21

#5 D120 30 5.47 0.2 SB 20 60 56.32 75.99 1502.56 1357.50

#6 D119 42 6.93 0.15 SB 5 70 60.03 75.47 1572.99 1476.02

#7 D121 32 5.93 0.13 SB 10 20 51.73 155.57 1184.93 1942.35

#8 D122 36 6.07 0.15 SB 5 5 70.14 146.44 1182.16 1775.73

#9 D118 42 6.08 0.21 SB 15 90 96.80 115.50 1546.30 1678.91

#10 D134 40 6.25 0.19 SB 10 30 99.90 97.63 1329.34 1507.18

The PCA plot showed ponds relatively close to each other had similar environmental conditions, and Ponds at greater distances apart had much greater environmental heterogeneity (Fig 2.4). The first two PCA axis accounted for 56.4% of the total variance in environmental data. Because variance explained was relatively low, we cannot say with confidence that environmental heterogeneity between ponds was influenced by any of these environmental variables.

In the CCA plot, the mean abundance of study species from benthic bracken bag samples showed the strongest relationship with tree canopy overhang (C), and bank vegetation cover (D) along the first axis (p-value, <0.05) and with average phosphate concentration (E), and average dissolved oxygen concentration (G) along the second axis (p-value, <0.05). The first two axis account for 71.9% of total variation in species data in bracken bag samples (Fig 2.5a). The mean abundance of study species from littoral kick net samples also showed the strongest relationship with tree canopy over hang (C), and bank vegetation cover (D) along the first axis (p-value, <0.05), and with average phosphate concentration (E) along the second axis (p-value, <0.05). The first two axis account for 93.0% of total variation in species data in kick net samples (Fig 2.5b). For both sampling methods there were no significant values between environmental variables and odonate species, so we cannot conclusively determine what environmental variables significantly influenced odonate densities.

Figure 2.4 Principal component analysis (PCA) of log transformed environmental factors in ponds. Co-variance represents the different environmental factors. (A) pH, (B) water clarity, (C) tree canopy overhang (%), (D) bank vegetation cover (%), (E) average phosphate concentration(mg/L), (F) average nitrogen concentration (mg/L), (G) average dissolved oxygen concentration (mg/L), (H) pond area (m2).

Odonate composition: The abundance of odonate species varied between ponds. X. zealandica, H. australiae, P. smithii and P. grayi all had a relatively low mean abundance whereas A. colensonis had a relatively high mean abundance in all ponds. A. colensonis mean abundance did vary considerably between ponds - Ponds 4, 7, 8, and 10 all had an low mean abundance, however compared to other odonate species, A. colensonis was still the most abundant species in these ponds (Fig 2.6). The high abundance of A. colensonis could be the outcome of sampling error as different sampling methods were used between the two habitats. A larger area of the vegetative littoral zone was sampled compared to the benthic vegetative zone, so comparisons between the two need to be tentative. Pond 8 had no crayfish and compared to other ponds; it also had a low mean abundance of odonates. This showed factors other than crayfish abundance could be influencing odonate abundance. Nevertheless, all ponds (except pond 8) had the same odonate species present indicating some form of interaction between all odonate species and P. zealandicus may occur in all ponds.

A. colensonis X. zealandica H. australiae P.smithii P. grayi

180 160

140

120

100 80 60

Mean abundance Mean 20 0 -20 1 2 3 4 5 6 7 8 9 10 Pond

Figure 2.6 Mean total abundance of odonate species (mean ± SE) Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia. smithii, and Procordulia grayi in ponds. n=3.

Effects of vegetation on odonate abundance: Odonate species were correlated against bank vegetation cover to see if increased habitat could influence their abundance. In bracken bag samples, the overall spread of data showed no relationship (slope = -0.03) and the p-value was not significant (p >0.05) (Fig 2.7a). The abundance of individual odonate species from bracken bag samples also had a very weak correlation against bank vegetation cover and was not statistically significant (Table 2.2).

In littoral kick net samples, there was an overall weak relationship between odonate abundance and bank vegetation cover (slope = 0.20) which was not statistically significant (p >0.05) (Fig 2.7b). The abundance of individual odonate species changed significantly between species indicating that certain odonate species could be more affected by vegetation in the littoral zone compared to other species. The abundance of all Anisoptera species had no relationship to bank vegetation (Table 2.2). A. colensonis abundance had a strong correlation to bank vegetation (slope = 0.84), and the p-value was significant (p <0.05), which indicates the significance of vegetation contributing positively to A. colensonis abundance.

Figure 2.7 Bivariate analysis of average odonate abundance and bank vegetation cover (%) from two different habitats. (a) Benthic bracken bag samples. Slope = -0.03, p >0.05. (b) Littoral kick net samples. Slope = 0.20, p >0.05

Table. 2.2 Bivariate analysis displaying correlation coefficient and P-values of odonate species average abundance and bank vegetation cover (%). Significant values are highlighted.

Species/sample Correlation co-efficient P-value A. colensonis benthic -0.11 <0.05 X. zealandica benthic -0.10 >0.05 H. australiae benthic 0.12 >0.05 P. smithii benthic -0.07 >0.05 P. grayi benthic 0.0004 >0.05 A. colensonis littoral 0.84 <0.05

X. zealandica littoral 0.15 <0.05

H. australiae littoral 0.003 >0.05

P. smithii littoral -0.001 >0.05 P. grayi littoral 0.004 >0.05

Habitat use and odonate assemblages: Habitat preferences were evident between odonate species from benthic and littoral samples. A. colensonis had a significantly higher mean abundance in littoral kick net samples compared to benthic bracken bag samples (Welch F test, p <0.001). There was no significance difference in mean abundance of X. zealandica between benthic bracken bags and littoral kick net samples (Welch F test, p >0.05). Amongst Anisoptera species, H. australiae had a significantly higher mean abundance in benthic bracken bags compared to littoral kick net samples (Welch F test, p <0.001). P. smithii also had a significantly higher mean abundance in benthic bracken bags compared to littoral kick net samples with all ponds combined (Welch F test, p <0.05) (Fig 2.8). P. grayi had no significant difference in mean abundance between benthic bracken bag samples and littoral kick net samples (Welch F test, p >0.05) (Table 2.4).

80 Bag Kick net

0 Meanabundance of Odonate A. colensonis X. zealandica H. australiae P. smithii P. grayi Species

Table 2.4 Results of Welch F test for average abundance of odonate species; Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, Procordulia grayi between bracken bag and kick net sampling methods in all ponds combined. F-value represents the F-ratio, Omega2 represents the effect size, and P-value represents the level of significance. Significant values are highlighted.

F-value Omega2 P-value A.colensonis 21.49 0.255 <0.05 X. zealandica 0.426 0 >0.05 H. australiae 22.03 0.260 <0.05 P. smithii 10.83 0.141 <0.05 P. grayi 2.458 0.024 >0.05

Paranephrops zealandicus composition: All ponds were stocked with relatively low densities of crayfish in relation to the size of ponds. Low stockage in ponds may have potentially concealed other potential factors that are driving patterns in P. zealandicus densities. Between ponds stocking sizes varied with ponds 3 and 8 having significantly lower amounts stocked compared to other ponds. Pond 7 had the highest amount with 90 crayfish (Table 2.5).

Table 2.5 Paranephrops zealandicus originally stocked in ponds and the population densities in ponds (m2).

Pond Original stocking Original population density (m2)

1 61 1.8 2 51 1.1 3 6 0.13 4 51 1.3 5 79 2.6 6 58 1.4 7 90 2.8 8 11 0.3 9 54 1.3 10 72 1.8

Mean abundance of P. zealandicus represented a sample size proportional to the amount that was originally stocked in ponds. On average pond 5, 6, and 7 had a low mean abundance and no crayfish were collected from pond 8 over all three sampling periods indicating there was exceptionally low numbers or all the crayfish have since died. Size demographics also varied between ponds. Ponds 1, 2, 3, 4, 9, and 10 had a high mean abundance of small crayfish compared to medium and large crayfish in ponds. On average small crayfish was the size demographic most abundant in ponds. Ponds 5, and 6 had a similar mean abundance of all size groups. No small crayfish were found in Pond 7 (Fig 2.9).

Small Medium Large

60 50 40 30 20

10 Meanabundance 0 -10 1 2 3 4 5 6 7 8 9 10 Ponds

Figure 2.9 Mean abundance of Paranephrops zealandicus size groups (mean ± SE), Small (≥20mm), Medium (21-40mm), and Large (≤41mm) in ponds. n=3.

Effects of environmental factors and original pond stocking on Paranephrops zealandicus abundance:

There was a weak correlation between the average abundance of crayfish and original stocking of crayfish in ponds, however this correlation was not significant (p >0.05) (Fig 2.10). Whilst the amount of P. zealandicus originally stocked would influence population densities, ponds have had a minimum of 2-3 years to establish a larger crayfish population. This suggests that other factors could have had a greater influence on their abundance in ponds. The original stocking of crayfish was also correlated against bank vegetation cover; however, there was no correlation, suggesting when crayfish were first added to ponds, bank vegetation cover had no significant impact on crayfish abundance (Fig 2.11). Mean abundance of crayfish was moderately correlated against bank vegetation cover (p >0.001) (Fig 2.12). Based on these results, bank vegetation and the original amount of crayfish stocked in ponds are factors that could influence the current crayfish abundance in ponds.

Figure 2.10 Bivariate analysis of Paranephrops zealandicus abundance and original stocking in ponds. Slope = 0.27, P >0.05. n=3.

Figure 2.11 Bivariate analysis of Paranephrops zealandicus original stocking and bank vegetation cover (%) in ponds. Slope = 0.05 P >0.05. n=3.

Figure 2.12 Bivariate analysis of Paranephrops zealandicus abundance and bank vegetation cover (%) in ponds. Slope = 0.41, P <0.05. n=3.

Effects of odonates on Paranephrops zealandicus abundance: Odonate abundance was correlated against P. zealandicus abundance to see if odonates may significantly affect their abundance. The average abundance of odonate species in benthic bracken bag samples was shown to have no relationship when correlated against average P. zealandicus abundance (slope = -0.095) and the p-value was not significant (p >0.05) (Fig 2.13a). In littoral kick net samples, there was also no relationship between the average abundance of odonate species and average P. zealandicus abundance (slope = 0.169) (Fig 2.13b), and the p-value was not statistically significant (Table 2.3).

Figure 2.13 Bivariate analysis of average odonate abundance and average Paranephrops zealandicus abundance from two different habitats. (a) Benthic bracken bag samples. Slope = -0.14, p= 0.0601. (b) Littoral kick net samples. Slope = 0.14, p=0.427.

Table 2.3 Bivariate analysis displaying correlation coefficient and P-values of odonate species average abundance and Paranephrops zealandicus average abundance. Significant values are highlighted.

Species/sample Correlation co-efficient P-value A. colensonis benthic -0.14 >0.05 X. zealandica benthic -0.20 >0.05 H. australiae benthic -0.27 >0.05 P. smithii benthic -0.002 >0.05 P. grayi benthic -0.11 >0.05 A. colensonis littoral 0.52 >0.05 X. zealandica littoral 0.18 >0.05 H. australiae littoral -0.02 >0.05 P. smithii littoral -0.0003 >0.05 P. grayi littoral -0.001 >0.05

Habitat use and Paranephrops zealandicus assemblage: Contrasting habitat preferences occurred between size cohorts of crayfish from benthic and littoral samples. Medium and large crayfish both had a significantly higher mean abundance in littoral kick net samples compared to benthic bracken bag samples (Table 2.6). Small crayfish however had no significant difference in mean abundance between benthic bracken bags and littoral kick net samples (Welch F test, p >0.05) (Fig 2.14). It should be mentioned different sampling methods were used in each habitat which would influence relative proportions of size classes to some degree.

18 Bag Kick net

12 10 8

6 zealandicus 4

Mean abundance ofP. abundance Mean 2 0 Small Medium Large Size groups

Figure 2.14 Average abundance of Paranephrops zealandicus size groups (mean ± SE) Small, Medium, Large in bracken bag and kick net samples with all ponds combined. n=30.

Table 2.6 Results of Welch F test for average abundance of Paranephrops zealandicus size classes; small, medium, and large between Kick net and bracken bag sampling methods. F- value represents the F ratio, Omega2 represents the effect size, and P-value represents the level of significance. Significant values are highlighted.

F-value Omega2 P-value Small 3.494 0.04122 >0.05 Medium 7.069 0.09186 <0.05 Large 19.37 0.2344 <0.05

Discussion: It was clear that the amount of crayfish originally stocked has undoubtedly had an effect on the current population of stock in ponds, potentially masking other factors that are significantly driving patterns. In light of this, spatial patterns of crayfish between ponds could also be influenced by environmental factors, and food web interactions. pH levels were relatively neutral across all ponds. Low pH is problematic for organisms with exoskeletons as the hardening process during molting takes longer making them more vulnerable to predation (Beaune et al. 2018). Many crayfish species can cope with lower pH ranges (Reynolds, Souty-Grosset, & Richardson, 2013), higher nutrient concentrations, and fluctuating dissolved oxygen levels (Broughton et al., 2018; Moore, & Burn, 1968). Water temperature and calcium concentrations were two factors not measured in this study which are known to affect growth and survivability (Hammond et al., 2006). Based on the physiochemical factors investigated it is unlikely they have had a significant impact on odonate and crayfish densities across ponds as physio chemical factors were within standard ranges. Other factors could have a stronger influence.

Odonates: In ponds it was apparent that a combination of factors could influence the abundance of odonates. The factor that had the strongest influence on odonate abundance was bank vegetation cover. Bank vegetation cover affected the abundance of odonate species differently with Zygoptera (damselfly) species being more affected than Anisoptera (dragonfly) species. Zygoptera diversity and richness has been linked to increased macrophyte complexity (Butler, & Demaynadier, 2008). A. colensonis and X. zealandica spatial distributions are strongly associated with littoral vegetation, whereas Anisoptera species are more commonly found in benthic vegetation and the benthos (Rowe, 1987). These findings were also found in the current study as Anisoptera species were more commonly found in benthic bracken bag samples compared to littoral kick net samples. A. colensonis was the opposite and were more commonly found in littoral kick net samples. Kick net sampling and habitat trapping in studies has been shown to be effective sampling methods when collecting macroinvertebrate samples (Hawking, & New, 1999; Raebel et al., 2010), however when comparing between habitats, different sampling methods could generate sampling bias, affecting interpretation of odonate larval distribution patterns. Regardless, habitat use varies between odonate species as certain species could be better adapted to a specific habitat, and

therefore, bank vegetation cover in the littoral zone may not directly affect the abundance of all odonate species. Because of this we can be confident that differences in spatial distribution patterns among odonate species occur between the different pond habitats.

The density of crayfish was another factor that could be significantly influencing odonate abundance, however based on the current study this seems unlikely. In benthic bracken bag samples and littoral kick net samples, there was no statistically significant relationship between odonate species and crayfish abundance. As opportunistic predators, crayfish consume invertebrates that are readily available to them which are generally small sedentary invertebrates (Usio, & Townsend, 2004). Because of this, odonates may not be considered essential prey for crayfish as odonates may have efficient avoidance/camouflage strategies, or even P. zealandicus prefer using valuable energy to successfully capture slower invertebrates. P. zealandicus could therefore have a stronger influence on odonate larvae as a competitor rather than a predator, indirectly influencing their abundance (Nystrom et al., 1996). However, based on these results we cannot conclusively say crayfish densities had any influence on the abundance of odonate larvae.

Paranephrops zealandicus: Factors that are likely to contribute to the abundance of crayfish, apart from the original stocking sizes of crayfish in ponds, are presence of bank vegetation and the density of crayfish in ponds. Bank vegetation had a moderately positive relationship with crayfish abundance. The presence of vegetation in and around ponds increases habitat complexity for pond invertebrates (Van de Meutter, Cottenie, & De Meester, 2008). In the current study it was known that ponds were stocked with additional vegetation (bracken, pine, etc.), however, the densities of vegetation placed in ponds was not recorded. This could have further influenced macroinvertebrate abundance between ponds, as higher habitat complexity can decrease niche overlap (Nyström, & Pérez, 1998) and increase invertebrate richness and diversity (Harrison, & Harris, 2002; McAbendroth et al. 2005). Whilst bankside vegetation did influence crayfish abundance, it only had a moderate impact and other factors in accordance may also contribute to crayfish abundance (Stiers et al., 2011).

Crayfish densities may have influenced habitat use among crayfish size groups and may also be a contributing factor to crayfish abundance in ponds. It was evident that medium and large crayfish size groups preferred the littoral vegetative zone compared to the benthic vegetative

47 zone suggesting the littoral zone was the favored habitat. P. zealandicus are known to be aggressive and territorial when it comes to defending desirable habitat (Stewart, & Tabak, 2011). In semi-natural aquaculture ponds with high densities of crayfish, aggressive behavior may have increased. As we also know pond communities are size-structured, so in the presence of larger crayfish smaller individuals would more likely lose desirable habitat to larger individuals forcing them to seek refuge wherever they may find it. In the current study small crayfish had no preference between the littoral vegetative zone and benthic vegetative zone as mean abundance was similar in both habitats. Kick netting allows more chances for crayfish to successfully escape from the littoral zone to deeper waters during sampling which may result in an under-estimate of crayfish in the littoral zone. Bracken bags were gathered efficiently giving little opportunity for crayfish to escape however bracken bags are biased towards younger demographics (Parkyn, 2015). Because of this, the error between sampling methods could influence the ability to detect spatial patterns in habitat use between crayfish size groups. Studies involving a combination of sampling methods have proven useful at collecting different crayfish sizes (Byrne, Lynch, & Bracken, 1999; Usio et al., 2006). Whilst these two sampling methods chosen were efficient at gathering samples, the error between sampling methods cannot be ignored when interpreting the results. Small crayfish did however have a higher mean abundance in bracken bags compared to large crayfish. Juvenile crayfish have been found to occupy different habitats to adults to reduce interspecific competition (Reynolds, Souty- Grosset, & Richardson, 2013). Habitat traps are very efficient at capturing young crayfish (Kusabs et al., 2018; Parkyn, 2015). In the current study, adult crayfish were not well represented in bracken bags indicating larger crayfish may not see it as desirable habitat.

Conclusion: When investigating complex food webs, the combination of factors can all affect the abundance of species in ponds and some factors can have more of an impact on spatial patterns than others. In this study, vegetation and the densities of crayfish were hypothesized to influence odonate species abundance. The main factors believed to affect crayfish abundance were the original stocking of crayfish, vegetation, and the densities of crayfish in ponds. Habitat use between crayfish size cohorts may be influenced by larger territorial crayfish occupying more desirable habitat. In higher densities of crayfish smaller individuals are forced to seek refuge from larger individuals in benthic bracken bags. It was unlikely that odonate larvae significantly influenced crayfish abundance and spatial patterns in ponds.

Chapter 3: Diets of Odonate Species and Paranephrops zealandicus in Semi-natural Aquaculture Ponds: An Analysis of Stomach Contents and Stable Isotopes.

Introduction: Stomach content and stable isotope analysis have been widely used to show trophic and food web linkages between pond invertebrate species. Predator-prey interactions in semi-natural aquaculture systems are important to understand as the interactions between invertebrates and farmed crayfish species could be detrimental to crayfish stock in aquaculture ponds. Juvenile crayfish stock is far more likely to be vulnerable to predation because of the size structured nature of food webs (Blake, & Hart, 1995; Dethier et al., 2019). Odonate species in crayfish aquaculture ponds are a potential food source for larger crayfish, however, they could also be detrimental to juvenile crayfish if they are predators.

Odonates are amongst the top predators in pond environments with a specialised prey- grasping labium that allows them to capture a wide variety of prey (Corbet, 1999). Optimal foraging theory suggests that odonate larvae would ideally capture larger prey if possible, as it provides more calories and requires less energy (Corbet, 1999). Odonate larvae usually feed on smaller invertebrates like gastropods, cladocerans, chironomids, planktonic crustaceans, oligochaetes, and even scavenging dead invertebrates (Rowe, 1985; Thompson, 1978). They have also been observed capturing terrestrial invertebrates from the water’s surface (Spence, 1986). As odonate larvae become larger, the size range of their prey can also increase, with some species attacking prey much larger than themselves. Some odonate larvae have been observed catching juvenile fish (Pritchard, 1964), tadpoles (Crump, 1984), and even young of year crayfish during their first molt (Jonsson, 1992). In a study by Witzig, Huner, & Avault Jr, (1986), odonate larvae were observed predating on juvenile crayfish up to 30mm in total length. Odonates in semi- natural aquaculture ponds could therefore predate on young P. zealandicus, reducing the juvenile biomass in ponds. Whilst studies including stomach content and stable isotope analysis have shed light on odonate species’ diets (Hamada, & Oliveira, 2003; Kraus, 2010), there are limited studies showing predator prey relationships between NZ odonate species and pond macroinvertebrates in semi-natural aquaculture environments.

Crayfish have an omnivorous diet obtaining energy from a variety of different trophic levels. Being opportunistic feeders, they forage for pond invertebrates including snails, nymphs, mayflies, cladocerans, and chironomid larvae, as well as scavenging for dead organisms

(Gutiérrez-Yurrita et al., 1998; Hollows, Townsend, & Collier, 2002). Prey selection and diet preferences can vary depending on the size/age of many crayfish species (Reynolds, Souty- Grosset, & Richardson, 2013). There is evidence that adult crayfish being larger, are much slower than juveniles and might not be as efficient at capturing invertebrates (Momot, 1995). Rather than wasting energy chasing invertebrates, adults consume comparatively more vegetation instead. In an experiment conducted by Parkyn, Collier, & Hicks, (2001), crayfish greater than or equal to 20mm (OCL) were observed to feed comparatively more on detritus rather than macroinvertebrates. Crayfish may also have preferences over what type of detritus they consume. France (1996), reported that adult crayfish (>20mm) prefer to feed on terrestrial detritus rather than algae, which could be linked to the greater nutritional value of terrestrial detritus. Due to the active nature of juvenile crayfish, it is assumed that juveniles may feed on significantly greater amounts of invertebrates compared to detritus (Momot, 1995).

Many crayfish species are cannibalistic, particularly if conditions are crowded, and alternative food is scarce (Alcorlo, Geiger, & Otero, 2004; Holdich, 2002). Larger crayfish are usually more so towards juveniles or moulting individuals as they are more vulnerable to cannibalism (Abrahamsson, 1966). Being strongly territorial, crayfish will often show aggression towards other crayfish over desirable habitat, and on some occasions kill and eat smaller victims (Stewart, & Tabak, 2011). When habitat complexity is high and food is plentiful, cannibalistic behaviours have been shown to significantly reduce (Nyström, & Granéli, 1996). Studies have been done on intensive crayfish rearing operations showing both aggression and cannibalism increase at relatively higher densities (Baird, Patullo, & Macmillan, 2006), however no studies have examined the intensity of cannibalism in semi-natural aquaculture ponds. Studies also investigating the diet of P. zealandicus have shown both cannibalism (Parkyn, Collier, & Hicks, 2001; Hollows, Townsend, & Collier, 2002), and ontogenetic shifts (Whitmore, 1997), occur naturally. However, P. zealandicus has only been investigated in its natural habitat and not in a semi-natural aquaculture environment.

Stomach content analysis is effective at identifying recent composition of diet, however certain limitations are associated with this analytical approach. The rate of digestion and type of tissue ingested are factors that determine the likelihood of being able to correctly identify organic matter in species stomachs. Material such as intractable detritus or hard bodied organisms are likely to take longer to break down, which can misrepresent certain proportions of consumed material in species’ diets (Momot, 1995). Stable isotope analysis is useful alongside stomach

content analysis as it can show the long- term energy flows in food webs between species based on the isotope ratios of 13C/12C and 15N/14N in the tissue of organisms (Gu et al., 1996). The ratio of carbon isotopes is used to determine energy flows in food webs from primary producers to consumers. The ratio of nitrogen isotopes is an indicator for the trophic position of organisms as 15N accumulates in organisms from predators consuming their prey, so a higher 15N/14N ratio indicates a predator feeding at a relatively higher trophic level in the food web. Using these two sampling methods alongside each other can provide insights into the diets of invertebrates in pond communities. (Stenroth et al., 2006).

Aims/hypothesis: In this chapter I aim to improve our understanding of the diets of P. zealandicus and odonate species using stomach content and stable isotope analysis to determine likely trophic relationships and ascertain whether odonate species are predators of juvenile P. zealandicus contributing towards their low biomass in ponds. Several hypotheses were formulated regarding diets of P. zealandicus and odonate species. Many odonate species are found to feed on a similar variety of invertebrates and some species are even found to predate on crayfish up to 30mm in length (Corbet, 1999; Witzig, Huner, & Avault Jr, 1986). Because of this I hypothesize all odonate species will have similar diets and are predators of juvenile P. zealandicus. I also hypothesize cannibalism to be intense in ponds based on the aggressive behaviour of crayfish in overcrowded conditions (Stewart, & Tabak, 2011)

Methods:

Sample collection: Odonate larvae collected during invertebrate surveys were readily available for dissections in ethanol-preserved samples. As many odonate larvae were dissected as were available and in suitable condition in previously collected samples. The sample size of species are as follows: A. colensonis (n=100), X. zealandica (n=83), H. australiae (n=76), P. smithii (n=61), P. grayi (n=22). Larger specimens (>20mm) were selected for dissections as larger odonates are more likely to predate on juvenile crayfish compared to smaller ones. Taking only larger individuals also eliminates the bias of size as a variable for prey selectivity.

Animal ethics approval was completed prior to taking any crayfish from ponds. This ensure the correct handling and euthanizing was carried out. In this study, crayfish were humanely euthanised by being placed on ice and taken back to the laboratory to be dissected. A sample size of 100 crayfish were selected from ponds for stomach content analysis on 15th January 23rd January, and 8th March 2019. From each pond, 10 crayfish varying in size were selected for analysis. Because pond 8 had no crayfish, 20 crayfish were selected from pond 10, as it had a younger demographic, and the juvenile sample size collected for stomach content analysis was not well represented.

Dissections: Odonate larvae were pinned to a wax board and the abdominal segments from the upper thorax to the lower abdomen were removed to reveal the gut tract. The gut tract was then cut at the oesophagus and above the rectum. The gut tract was then removed and put on a glass slide with an eye drop of water to prevent desiccation.

Crayfish were placed on a dissecting board and the carapace was cut open exposing the internal cavity. Incisions through the top of the cephalothorax along the cervical groove and down the carapace were made. The cardiac stomach and pyloric stomach were removed cutting through the oesophagus and just below the pyloric stomach. The stomach was then cut open and rinsed with ethanol to remove organic matter. The stomach contents was then stored in ethanol in 10 mL vials ready for analysis.

Analytic approach to stomach content analysis: Quantifying diets using stomach contents can be difficult due to fragmentation and partial

digestion of stomach contents (Buckland et al. 2017; Hyslop, 1980). Mohan & Sankaran (1988), recommends using multiple methods to provide a robust assessment of diet. In this study the frequency of occurrence and volume percentage of organic matter were two quantitative methods used. The frequency of occurrence was carried out recording the presence/absence of organic matter identified in stomachs. The volume percentage of organic matter was recorded in the following 10 categories. Plant detritus, Odonata, Chironomidae, P. zealandicus, Ostracoda, Copepoda, terrestrial invertebrates, other invertebrates (aquatic), inorganic, and unidentified. A stomach fulness scale from 0 – 100% was used to measure the amount of contents in stomachs. Individuals with 0% stomach fullness were left out of the analysis as there was a large number of individuals with empty guts which would have skewed results. The sample size of volume percentage for each species is as follows: A. colensonis (n=56), X. zealandica (n=45), H. australiae (n=44), P. smithii (n=46), P. grayi (n=16), and P. zealandicus (n=99). Percentage volumes of each of the 10 categories was measured in stomachs and was adjusted depending on the overall gut fullness value. E.g. 50% stomach fullness with 30% plant detritus: 50% of 30% = 15% of stomach contents is plant detritus. Adjusting for stomach fullness between individuals is also used in similar studies (Hollows, Townsend, & Collier, 2002; Musgrove, 1988). The 10mL vials of P. zealandicus stomach contents and glass slides with odonate stomach contents was examined under a dissecting microscope at 40x magnification, and if necessary, under a binocular microscope at 400x magnification. A 5mm grid sheet was placed underneath samples on the dissecting microscope to estimate volume percentages accurately. Body parts of invertebrates were identified using the Guide to Aquatic Insects of New Zealand (Winterbourn et al., 1989), and the online Manaaki Whenua Freshwater Invertebrates Guide (Moore 2001). Data was recorded in Microsoft Excel (2018).

Stable isotope analysis sample collection: Odonate larvae collected during invertebrate surveys were also used for stable isotope analysis. It was suitable to use odonate larvae preserved in samples as carbon and nitrogen isotope values are not affected by ethanol (Syväranta et al., 2008). Of the five odonate species, four were used for the analysis (Austrolestes colensonis, Xanthocnemis zealandica, Procordulia smithii, and Hemicordulia australidae). Procordulia grayi was left out of the analysis because of limited numbers. Odonate larvae from ponds 1, 2, & 6 were chosen for this analysis as these ponds all had a similar abundance of odonate larvae and P. zealandicus. In each pond, a sample size of 10 individuals from each species were selected. Larger odonates were selected for analysis

because larger specimens are more likely to predate on P. zealandicus juveniles.

P. zealandicus used in stomach content analysis were also used in stable isotope analysis to utilize as much of the organism as possible and to reduce waste. A sample size of 10 individuals were selected from each of the three chosen ponds. A range of different sizes (10 27mm) were chosen for the analysis. Larger individuals would have been preferred but 27 mm (OCL) was the maximum size from these three ponds. Because of this YOY and 1-year old P. zealandicus are the age groups represented in the analysis. Abdominal tissue of P. zealandicus was used for analysis to ensure consistency between specimens as different parts of the anatomy store carbon for different periods of time, which can lead to bias in the analysis (Hobson, Gloutney, & Gibbs, 1997; Stenroth et al., 2006). The abdominal tissue was removed from the tergum dorsal plates prior to final sample preparation.

Additional young of year (YOY) crayfish samples were collected on the 7/11/2019 as there were limited numbers in the original samples. These were collected from pond 4 due to the limited numbers found in other ponds at the time. YOY were humanely euthanized on ice and brought back for sample preparation. The size range of YOY all varied between 10 – 13 mm (OCL). Abdominal tissue of YOY crayfish was also used for the analysis.

To gain a better understanding of the relative trophic positions of crayfish and odonates in the pond foodwebs, other pond invertebrates were also used in the analysis. These samples were used as reference points helping portray the overall trophic structure of pond food webs. The invertebrate taxa used for the analysis were Chironomidae larvae, Rhantus sp., and Anisops wakefieldi. A sample size of three for each invertebrate taxon was used.

Sample preparation: 169 samples in total were prepared for the analysis. All tissue samples were put into an oven for 36 hours at 50 C˚ and then crushed into powder. Powdered samples were then weighed out into tin (Sn) capsules at 1mg ± 0.2mg on the Excellence XPR Ultra-Microbalance scales. All samples were then analysed in the Mellor chemistry laboratory at the University of Otago using a Circon 20:20 isotope ratio mass spectrometer, Crewe UK, and CarloErba, Mulan Italy.

Standardization of the carbon, and nitrogen ratios are as follows:

15 15 14 3 Δ N = [( Nsample / Nstandard) – 1] × 10

13 13 12 3 Δ C = [( Csample / Cstandard) – 1] × 10

Isotope ratio mass spectrometry instruments (IRMS) are designed to accurately measure the ratio of ions that correspond to the different isotopic forms of a gas in a sample. For this experiment, 13C/12C and 15N/14N differences were measured. Samples were first converted to

CO2 and N2 by combustion usually between 900-1050 °C, but the heat of combustion of the tin capsules raises the sample temperature to around 1800 °C. Ions are then focused through an electrostatic field with a high voltage. The strength of the magnetic field and the accelerating voltage determines the trajectories of the ions and, therefore, which ions will enter which of the collectors. The use of multiple collectors allows the simultaneous measurement of ion intensity ratios, negating the smallest fluctuations in the overall intensity of the ion beam. Each collector is connected to a dedicated amplifier which is then sent and recorded in the IRMS software. The IRMS uses a three-point calibration method with glutamic acid (40, 41), and ethylenediaminetetracetic acid (EDTA) to standardise for 13C/12C and 15N/14N mass. The instrumental error of the machine has an uncertainty better than 0.02 ‰ for isotopic values. Computer software used to calibrate the data was the computer software programme Calisto.

Statistical analysis: All analyses were carried out in the statistical program Past (version 3.26, 2019). Diets of odonate species and P. zealandicus were investigated to compare similarities in stomach volume percentages. Diets of P. zealandicus size groups were also analysed to compare similarities in stomach volume percentages. Nine different stomach content categories were analysed using an analysis of similarities (ANOSIM) with a Bonferroni post hoc test. Unidentified organic tissue was left out of the analysis due to uncertainty in the data.

Total number of invertebrate taxa found in stomachs was also compared between the different size groups of P. zealandicus. A one-way random-effect ANOVA with ponds as a block factor was used comparing the total number of taxa found in the stomachs of small, medium, and large P. zealandicus. Data was log transformed to normalise its distribution, and populations shared a common variance. A Tukey post hoc test was run to see if there was a significant difference.

The average relative volume of stomach content categories of P. zealandicus size groups was also investigated using a one-way ANOVA. Three factors were compared. Plant detritus, invertebrates, and crayfish tissue. The 6 invertebrate categories (Odonata, Chironomidae, Ostracoda, Copepoda, terrestrial invertebrates, and other invertebrates) were combined to represent the invertebrate group. Log transforming the data was tried but it did not normalise the distribution. Homogeneity of variance was also not observed between size groups. Because of 56 this a Dunn’s post hoc test was carried out on significant differences. It should be noted to treat the results of this part of the analysis with caution as assumptions of normality and homogeneity of variance were not met.

The 13C/12C and 15N/14N ratios were compared between invertebrates in ponds using a one- way ANOVA with a Tukey post hoc test. It should be noted that the sample size of species in each pond was low with 10 individuals per species except in pond #1 where P. zealandicus has only nine samples as one sample failed to run in the isotope analyzer. The low sample size and unequal sample sizes across species can increase the likelihood of a type ll error (occurring) giving a false positive. The homogeneity of variances was uneven in pond 6 for 13C/12C values among species, so the results should be interpreted with appropriate caution when referring to this section of data in the analysis. Chironomidae spp., Rhantus sp., and A. wakefieldi only had one sample each and were used to show trophic positions in ponds and were not referred to in the analysis.

Results: Odonate stomach content analysis: The diet of odonate species was characterised by a limited number of similar invertebrate taxa indicating odonate species feed on a common pool of invertebrates. Mean invertebrate taxon richness in stomachs was also particularly low between odonate species, and P. zealandicus (Table. 3.1). Chironomidae, Copepoda, and Ostracoda were the three most commonly recorded invertebrate taxa in X. zealandica, H. australiae, P. smithii, and P. grayi stomachs (see appendix 2). In A. colensonis, the three most common invertebrate taxa in stomachs were Chironomidae (20%), Copepoda (16%), and terrestrial invertebrates (14%), however Ostracoda was still abundant and was found in 11% of A. colensonis species stomachs (see appendix 2). Crayfish tissue was exceptionally low in stomachs of odonates with P. grayi, H. australiae and A. colensonis being the only odonates found with any crayfish tissue in their stomachs. H. austaliae had the highest amount with 4% of individuals dissected found to have crayfish tissue in stomachs (see appendix 2).

Table 3.1 Mean invertebrate taxon richness in stomachs of study species (mean ± SE), Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, Procordulia grayi, Paranephrops zealandicus.

Species Mean SE ± Austrolestes colensonis 0.86 0.118 Xnathonemis zealandica 0.54 0.086 Hemicordulia australiae 0.41 0.071 Procordulia smithii 0.51 0.095 Procordulia grayi 0.5 0.143 Paranephrops zealandicus 1.57 0.161

a b

c d

e f

Figure 3.1 Images of invertebrates identified in the stomachs of odonate larvae. (a) Cyclopoida sp., (b) terrestrial invertebrate (unknown species), (c) Parenephrops zealandicus chelipeds, (d) Ostracoda sp., (e) Chironomidae sp., (f) Diptera sp.

Volume percentages of organic matter in stomachs of odonate larvae: Chironomidae, Ostracoda, and Copepoda were present at relatively high-volume percentages in all odonate species stomachs (Fig. 3.2), indicating small and slow sedentary pond invertebrates seem to be the primary targets of odonate species. A. colensonis and X. zealandica had higher percentages of other invertebrates represented in stomachs compared to the three Anisoptera species (Fig. 3.2). A. colensonis also had higher percentages of terrestrial invertebrates compared to other species (Fig. 3.2a). H. australiae had the highest volume percentage of crayfish tissue in stomach out of all odonate species, however the mean percentage only accounted for 2% of stomach volume (Fig. 3.2c). Overall, odonate species were found to have similar volume percentages of different invertebrate taxa in their stomachs with no significant difference between odonate species diets (ANISOM, r = 0.003, p = 0.394). Diet comparisons between odonate species and P. zealandicus were significantly different (ANISOM, r = 0.04, p = 0.036) (Table 3.2).

Species A.colensonis X. zealandica H. australiae P. smithii P. grayi P. zealandicus A.colensonis >0.05 >0.05 >0.05 >0.05 <0.001 X. zealandica >0.05 >0.05 >0.05 <0.001 H. australiae >0.05 >0.05 <0.001 P. smithii >0.05 <0.001 P. grayi <0.001

A. colensonis a X. zealandica b

H. australiae c P. smithii d

P. grayi e P. zealandicus f

Paranephrops zealandicus stomach content analysis: P. zealandicus was found to feed on a wide variety of organic matter. Plant detritus was the most frequent item in stomachs and was found in 87% of individuals. Of invertebrate taxa, P. zealandicus tissue was the highest found in 45% of individuals stomachs indicating a high rate of cannibalism. The three most frequently occurring invertebrates other than P. zealandicus tissue were Chironomidae (21%), Copepoda (18%), and terrestrial invertebrates (14%). The fibres used to construct the bracken bags were also a frequent occurrence in gut contents, with 29% individuals having fibres present in their stomachs (see appendix 2).

Habitat differences were evident across ponds, with variation in P. zealandicus diets from the littoral zone and benthic vegetative zone. 56% of individuals from littoral kick net samples had P. zealandicus tissue in stomachs, whereas only 34% of individuals from benthic bracken bags had it in their stomachs. Terrestrial invertebrates were also higher in littoral kick net samples (20%), compared to benthic bracken bag samples (8%). Both benthic bracken bags and littoral kick net samples had relatively equal amounts of individuals with plant detritus present in their stomachs with 90% found in individuals from benthic bracken bag samples and 84% found in individuals from littoral kick net samples (see appendix 3).

a b

c d

e f

g h

Volume percentages of organic matter in stomachs of Paranephrops zealandicus: Based on the mean percentage volume of organic matter in P. zealandicus stomachs, organic matter was not significantly greater compared to invertebrates, showing diet consists of similar quantities of vegetation and invertebrates (Fig. 3.5). There was a significant difference in the amount of P. zealandicus tissue found in stomachs. P. zealandicus tissue had a significantly lower mean percentage volume in stomachs compared to plant detritus (Dunn’s post hoc, p <0.001), and invertebrates (Dunn’s post hoc, p <0.001) (Table 3.2).

Plant detritus Invertebrates P. zealandicus tissue

Mean volume percdentage (%) volumepercdentage Mean 0 Paranephrops1 zealandicus

Figure 3.5 Mean volume percentage of plant detritus, invertebrates and Paranephrops zealandicus tissue found in the stomachs of Paranephrops zealandicus. n=99.

Table 3.3 ANOVA and Dunn’s post hoc test of mean volume percentage of plant detritus, invertebrates, and Paranephrops zealandicus tissue found in the stomachs of Paranephrops zealandicus.

Organic matter F-value Omega2 P-value 8.918 0.05062 <0.001

Dunn’s post hoc Plant detritus Invertebrates P. zealandicus tissue Plant detritus >0.05 <0.001 Invertebrates <0.001

Stomach content analysis of Paranephrops zealandicus size groups: Diets of P. zealandicus were relatively similar between the different size groups, however there were some noticeable differences worth mentioning. Plant detritus was high in all size groups occurring in 73% of small individuals, 94% in medium individuals and 97% of large individuals. The frequency of P. zealandicus tissue in stomachs was found to increase as size of P. zealandicus increased indicating larger crayfish were more cannibalistic. Crayfish tissue was found in 11% of small individuals, 46% of medium individuals, and in 87% of large individuals. Large P. zealandicus also had a higher frequency of terrestrial invertebrates compared to small and medium sized P. zealandicus with 43% of large individuals found to have terrestrial invertebrates in their stomachs (see Appendix 4).

Mean invertebrate taxon richness in stomachs changed between size groups with large P. zealandicus found on average to have a significantly higher number of different invertebrate taxa present in stomachs compared to small (Tukey’s pairwise, p <0.05), and medium (Tukey’s pairwise, p < 0.001) size groups (Table. 3.6). There was no significant difference between small and medium P. zealandicus groups (Table 3.4).

Table 3.4 Mean taxon richness in stomachs of Paranephrops zealandicus size groups (mean ± SE), Small (n=37), Medium (n=33), and Large (n=30).

Size Mean ± SE Small 1.22 0.242 Medium 0.94 0.204 Large 2.53 0.306

Table 3.5 ANOVA and Tukey post-hoc test of log transformed taxon richness in stomachs of Paranephrops zealandicus size groups small, medium, and large. Significant values are highlighted.

P. zealandicus size F-value Omega2 P-value Levene’s test 10.8 0.1481 <0.001 0.3355

Tukey post-hoc Small Medium Large Small >0.05 <0.05 Medium <0.001

Volume percentages of organic matter in stomachs of Paranephrops zealandicus size groups: Between all size groups, plant detritus and invertebrate categories in stomachs were similar, with no significant differences in mean volume percentages (Fig. 3.6). There was a significant difference in P. zealandicus tissue in stomachs of small P. zealandicus having a significantly lower mean volume percentage to medium size group (Dunn’s post hoc, p <0.05) and large size group (Dunn’s post hoc, p <0.001) (Table 3.5). Medium P. zealandicus also had a significantly lower mean volume percentage of crayfish tissue compared to large P. zealandicus (Dunn’s post hoc, p <0.05) (Table 3.5).

Within each size group there was only a significant difference in mean volume percentages of organic matter in small P. zealandicus stomachs (p <0.001) (Table 3.6). P. zealandicus tissue in the small size group was significantly lower compared to plant detritus (p <0.001) and invertebrate categories (p <0.001). There was no significant difference between plant detritus and invertebrate categories in small P. zealandicus stomachs (Table 3.6).

Plant detritus Invertebrates P. zealandicus tissue

18 16 14 12 10 8 6

volume percentage (%) volume

2 0 Mean Small Medium Large Size groups

Between size groups F-value Omega2 P-value

Plant detritus 2.518 0.02976 >0.05 Invertebrates 0.9511 0 >0.05 P. zealandicus tissue 6.314 0.09695 <0.05

P. zealandicus tissue (Dunn’s post hoc) Small Medium Large

Small <0.05 <0.001 Medium <0.05

Within size groups F-value Omega2 P-value Small 7.838 0.1124 <0.001 Medium 2.993 0.03871 >0.05 Large 1.407 0.008956 >0.05

Small size group (Dunn’s post hoc) Plant detritus Invertebrates P. zealandicus tissue

Plant detritus >0.05 <0.001 Invertebrates <0.001

P. zealandicus size groups had slight differences in volume percentages of invertebrates in stomachs. Volume percentages of terrestrial invertebrates were higher in large crayfish compared to medium and small crayfish. Volume percentages of Ostracoda were higher in small crayfish compared to medium and large crayfish (Fig. 3.7). Overall, there was a significant difference in diets between small and large size groups of P. zealandicus (ANISOM, r = 0.036, p = 0.018).

Small a Medium b

Large c

Figure 3.7 Mean volume percentages of different stomach contents categories in stomachs of Paranephrops zealandicus size groups. (a) Small, (b) Medium, (c) Large.

Stable isotope analysis of odonate larvae and Paranephrops zealandicus: In pond 1 there was a significant difference in 13C/12C values of P. zealandicus between all species apart from P. smithii (Tukey’s post hoc, p >0.05). P. zealandicus also had a significantly higher mean 15N/14N value of 5.11 ± 0.07 compared to all odonate species (Fig 3.8a). A. colensonis, X. zealandica, H. australiae and P. smithii all had similar 13C/12C values suggesting that they obtain carbon from a similar source. There was no significant difference in 15N/14N values between H. australiae and X. zealandica (Tukey post hoc, p >0.05). YOY also had a significantly higher mean 15N/14N value compared to all odonate species (Table 3.7).

In pond 2 there was no significant difference in mean 15N/14N values between P. zealandicus adults and YOY (Tukey’s post hoc, p >0.05). All odonate species had significantly lower mean 15N/14N values to P. zealandicus adults and YOY (Table 3.8). A. colensonis and X. zealandica had similar mean 15N/14N values, whereas P. smithii and H. australiae both had similar mean 15N/14N values. This showed between Zygoptera, and Anisoptera species there were clear trophic levels (Fig 3.8b). P. smithii and H. australiae had no significant difference between both 13C/12C and 15N/14N values indicating both species to feed at comparable trophic levels (Table 3.8). The 13C/12C values were significantly different between A. colensonis and X. zealandica (Tukey post hoc, p <0.001), but 15N/14N values were not significantly different (Tukey’s post hoc, p>0.05) implying they have similar trophic levels but sourced energetically from a different basal carbon source.

In pond 6 there was no significant difference in mean 15N/14N values between P. zealandicus and YOY (Tukey post hoc, p >0.05). All odonate species had significantly lower mean 15N/14N values compared to P. zealandicus and YOY (Table 3.9). No significant differences in 15N/14N values occurred between P. smithii and H. australiae (tukey post hoc p >0.05) and also there was no significant difference in 15N/14N values between A. colensonis and X. zealandica (Tukey post hoc, p >0.05). The mean 13C/12C value of P. zealandicus was significantly difference in 13C/12C values between P. smithii (Tukey post hoc, p <0.001) and H. australiae (Tukey post hoc, p <0.05), whereas there was no significant difference in 13C/12C values between P. zealandicus and A. colensonis and X. zealandica (Fig 3.8c).

Other invertebrates included in the analysis provided a broader understanding of trophic structure. Rhantus sp. had the highest trophic position across all ponds with a mean 15N/14N value of 6.69 ± 0.16 and had a mean 13C/12C value range of 32.3 ± 0.88. Chironomidae also had a large mean 13C/12C value range (-30.4 ± 1.25). 15N/14N values were relatively consistent with a low mean 70

15N/14N value (2.0 ± 0.12) indicating low trophic status. A. wakefieldi had a relatively small mean 13C/12C value range (-30.05 ± 0.39) and 15N/14N values were also similar to odonate species (3.05 ± 0.22) (Fig 3.8).

71 a

Pond 1 F-value Omega2 P-value Levene’s test δ15N (‰) 117.3 0.908 <0.001 0.3707 δ13C (‰) 12.12 0.4851 <0.001 0.4169

δ15N (‰) A.colensonis X.zealandica H. australiae P. smithii P. zealandicus YOY

A. colensonis >0.05 <0.001 <0.001 <0.001 <0.001 X. zealandica >0.05 <0.001 <0.001 <0.001 H. australiae <0.05 <0.001 <0.001 P. smithii <0.001 <0.001 P. zealandicus <0.001

δ13C (‰) A. colensonis X. zealandica H. australiae P. smithii P. zealandicus YOY

A. colensonis >0.05 <0.05 >0.05 <0.05 >0.05 X. zealandica >0.05 <0.05 <0.001 >0.05 H. australiae <0.001 <0.001 <0.001 P. smithii >0.05 >0.05 P. zealandicus <0.05

Table 3.9 ANOVA and Tukey post hoc analysis giving fractionation values of stable carbon (δ13C) and nitrogen (δ15N) for Austrolestes colensonis, Xanthocnemis zealandica, Hemicordulia australiae, Procordulia smithii, and Paranephrops zealandicus in pond 2. Significant values are highlighted.

Pond 2 F-value Omega2 P-value Levene’s test δ15N (‰) 260.7 0.956 <0.001 0.1404 δ13C (‰) 34.36 0.735 <0.001 0.166

δ15N (‰) A.colensonis X. zealandica H. australiae P. smithii P. zealandicus YOY

A.colensonis >0.05 <0.001 <0.001 <0.001 <0.001 X. zealandica <0.001 <0.001 <0.001 <0.001 H. australiae >0.05 <0.001 <0.001 P. smithii <0.001 <0.001 P.zealandicus >0.05

δ13C (‰) A.colensonis X. zealandica H. australiae P. smithii P. zealandicus YOY

A.colensonis <0.001 >0.05 >0.05 <0.001 <0.001 X. zealandica <0.001 <0.001 >0.05 >0.05 H. australiae >0.05 <0.001 <0.001 P. smithii <0.001 <0.001 P. zealandicus 0.541

Pond 6 F-value Omega2 P-value Levene’s test δ15N (‰) 105.4 0.8969 <0.001 0.711 δ13C (‰) 34.97 0.739 <0.001 0.0003

15 X. zealandica H. australiae P. smithii P. zealandicus YOY δ N (‰) A. colensonis

A.colensonis >0.05 <0.001 <0.001 <0.001 <0.001 X. zealandica <0.001 <0.001 <0.001 <0.001 H. australiae >0.05 <0.001 <0.001 P. smithii <0.001 <0.001 P. zealandicus >0.05

δ13C (‰) A.colensonis X. zealandica H. australiae P. smithii P. zealandicus YOY

A.colensonis >0.05 <0.001 >0.05 >0.05 <0.001 X. zealandica <0.001 <0.001 >0.05 <0.001 H. australiae >0.05 <0.001 <0.001 P. smithii <0.001 <0.001 P. zealandicus <0.001

Discussion: Integrating stomach content and stable isotope analysis: The key findings of this chapter found odonate species to have similar diets with slight differences in percentage volumes of invertebrate taxa in stomachs. It also found crayfish size groups feed on a similar variety of invertebrates, percentage volumes of organic matter vary slightly between size groups, and crayfish are become increasingly cannibalistic as they grow.

Odonates: Similarities in dietary overlaps between odonate species showed the three most common invertebrate taxa consumed were Chironomidae, Copepoda, and Ostracoda. Mean Percentages of these invertebrates were also high in all odonate stomachs indicating that the five odonate species all feed on similar invertebrates. Similar findings in a study by Lawton (1970), investigating the damselfly genus Pyrrhosoma found the most common invertebrate taxa in stomachs were Chironomidae, Copepoda, and Ostracoda. As far as we know the current study represents the first quantitative patterns in diets making comparisons between NZ odonate species in a semi-natural environment. Given these findings on odonates, these three invertebrate groups may be more vulnerable to predation compared to other pond invertebrates with better avoidance mechanisms, as odonates may expend less energy to search and successfully capture these invertebrates. Chironomidae and zooplankton are important prey for many pond invertebrate species accounting towards large proportions of their diets (Armitage, 1995; Herwig, & Schindler, 1996), so it is not surprising all odonate species in this study consumed similar invertebrate taxa.

Diet percentages between odonate species were not significantly different implying odonates feed on similar quantities of invertebrates, however odonate suborders also shared common similarities and differences in diet. Zygoptera species did however have higher percentages of the ‘other invertebrates’ category in stomachs compared to Anisoptera species. Stable isotope analysis also showed Zygoptera species to have similar trophic positions as 15N/14N and 13C/12C values were similar. Anisoptera species also had similar trophic positions but 15N/14N values were much lower compared to Zygoptera species. 13C/12C values again were similar between Anisoptera species. This suggests that odonates of the same suborder are assimilating almost the same food items and are at the same trophic level. Prey selectivity in odonates may occur between species based on habitat complexity, and where they are found in their environments (Folsom & Collins 1984). Findings in Chapter 2 indicated that Anisoptera species are more abundant in the benthic vegetative zone, whereas A. colensonis species were more abundant in 77

the littoral vegetative zone. Prey species are also likely to vary spatially with higher densities occurring where in preferred micro-habitats (Rennie, & Jackson, 2005). Odonates may therefore be feeding on invertebrate taxa that occur in higher densities in their preferred habitat. To gain better understanding of the predator prey interactions between invertebrate taxa and late instar odonate species, prey densities would need to be examined more closely.

The low amount of crayfish tissue in stomachs suggests that odonates are not major predators of P. zealandicus. All odonate species had either no crayfish tissue in stomachs or exceptionally low frequencies and percentages of crayfish tissue in stomachs. Stable isotope analysis revealed odonates had a lower trophic level to both P. zealandicus and YOY crayfish as 15N/14N values were significantly lower. For the crayfish that were found in the stomachs of odonate species, it is unknown if they were captured alive, or if odonates found and consumed crayfish tissue by scavenging already dead individuals. Odonates are opportunistic feeders and have been found to scavenge on dead organisms in their environments (Anderson et al., 2019; Rowe, 1987). Crayfish mortality rates are generally higher among juveniles in aquaculture ponds (Momot, 1984; Taugbøl, & Skurdal, 1992), so it is possible odonates may scavenge on them. Regardless, in terms of crayfish productivity in ponds, it is unlikely that odonate species contribute significantly towards low juvenile crayfish survival in the ponds.

Parenphrops zealandicus: The diet of P. zealandicus was strongly omnivorous consisting of similar percentages of plant detritus and invertebrates (excluding crayfish tissue). Between size groups, percentages of plant detritus and invertebrates did not significantly vary indicating across all size groups, similar amounts of plant detritus and invertebrate tissue is consumed relative to each size categories stomach volumes. Large crayfish also had a high percentage of invertebrates in their stomachs. On average, the 13C/12C values were also significantly different between larger crayfish and YOY crayfish in the three ponds. This indicates different size groups assimilate different carbon sources and that diets may also vary between size groups. This contradicts findings in similar studies that found the percentage of invertebrates in stomachs were higher in smaller individuals compared to larger individuals (Hollows, Townsend, & Collier, 2002; Momot, 1995; Whitmore, 1997). Momot (1995), suggested crayfish are not indiscriminate omnivores preferring animal protein to detritus. Invertebrates are higher in both calories and protein providing crayfish with more energy (Bowen, Lutz, & Ahlgren, 1995). Momot (1995), further goes on to say that adult crayfish are larger and slower which restricts their abilities to successfully capture prey resulting in a higher consumption of detritus. All these published

78 studies were carried out in river systems and not in aquaculture ponds like the current study. Crayfish densities in aquaculture ponds are much higher compared to natural environments. This would increase competition between individuals and would influence the consumption rates of different types of organic matter. Given competition is higher, juveniles may have to consume larger proportions of plant detritus to meet their carbon demands. Ponds are also rich in biodiversity and often have a greater species richness compared to other aquatic environments (Williams et al., 2004). Because of this, the density of prey may mean larger crayfish are able to successfully capture prey and therefore is why invertebrates are represented so highly in large crayfish diets.

Similarities and differences in invertebrates found in stomachs occurred between size groups. High frequencies of Chironomidae, Copepoda, Ostracoda, and terrestrial invertebrates were consumed by all size groups. In similar studies, Whitmore (1997), and Hollows Townsend, & Collier (2002) also found high amounts of chironomids in stomachs of P. zealandicus. Small crayfish were found to have higher percentages of Ostracoda and Copepoda in their stomachs compared to medium and large crayfish. For smaller crayfish it might be easier to capture zooplankton with compared to other larger pond invertebrates. Large crayfish had a higher frequency and percentage of terrestrial invertebrates in stomachs compared to small P. zealandicus. Larger crayfish are thought to consume more (most likely dead or dying) terrestrial invertebrates as they are easier to capture compared to aquatic invertebrates (Momot, 1995).

Crayfish cannibalism was high in the ponds based on stomach content analysis and was found to increase as size increased. Medium and large size groups had a significantly higher percentage of crayfish tissue in their stomachs compared to small crayfish. The crayfish tissue in stomachs were parts of the exoskeleton, which crayfish are known to consume when they are in more acidic environments. They consume the newly shed exoskeleton to utilise calcium reabsorption which helps develop their new exoskeleton faster (Hammond et al., 2006; Reynolds, 2002). However, some exoskeletons in stomachs were found to have tissue attached which indicated cannibalism was occurring. Cannibalism is common among many crayfish species (Kusabs, 2015), with predation on smaller individuals (Holdich, 2002). Previous studies involving P. zealandicus recorded lower rates of cannibalism compared to findings in the current study (Hollows, Townsend, & Collier, 2002; Momot, 1995; Whitmore, 1997). As this study was carried out in a semi-natural aquaculture environment, the densities of crayfish

are increased substantially compared to naturally occurring populations. Higher densities influence aggressive behaviours in P. zealandicus (Stewart, & Tabak, 2011), and could therefore create more unfavourable interactions between the different size demographics in ponds (Romano, & Zeng, 2017). With a size-structured food web, smaller P. zealandicus could be more vulnerable to predation from larger more territorial crayfish. Semi-natural aquaculture with increased densities of crayfish may therefore lead to higher rates of cannibalism.

Food web interactions: Many pond invertebrates and organic detritus in ponds was not analysed, making it harder to speculate on species interactions in pond food webs. Based on the interactions between odonates and P. zealandicus, it is clear that odonates are not likely to be major predators of juvenile P. zealandicus. The variation in isotope ratios between species showed all odonates maintain a lower trophic position to juvenile P. zealandicus which indicates that the nature of odonate interactions on juvenile P. zealandicus is unlikely to be predacious. In contrast P. zealandicus may not be major predators of odonates but are trophically higher than them, suggesting they could feed on odonates if opportunities arise. Nystrom (1996), mentions the role of crayfish being able to directly (through predation) and indirectly (by manipulating habitat) influence invertebrate populations considerably, signifying the importance of their interaction within food webs. They are also opportunistic feeders, consuming a wide variety of different invertebrates with juveniles shown to consume higher quantities of invertebrates to adults (Momot, 1995). Interactions between larger P. zealandicus and YOY were not clearly shown from isotope ratios making trophic structure between size groups hard to infer. These results were shown in similar studies with minimal separation in trophic positions between size groups of P. zealandicus (Hollows, Townsend, & Collier, 2002). The considerably high proportion of P. zealandicus tissue in stomachs did indicate that crayfish diets are highly cannibalistic, which suggests that interactions between larger P. zealandicus and juveniles is likely to have a negative impact on the biomass of juvenile crayfish in ponds.

Isotope ratios of Rhantus sp. revealed that it had the highest trophic position out of all invertebrates and held a significantly higher trophic position to crayfish. This indicated that other pond invertebrates other than odonates may be more of a threat to juvenile P. zealandicus. The 13C/12C values did not overlap with crayfish suggesting they are not a major food source for Rhantus sp., however the isotope ratios of other species within the dytiscidae family could show stronger signs of interaction with juvenile P. zealandicus. Adult diving beetles are

80 predators of a variety of different pond invertebrates (Culler, Ohba, & Crumrine, 2014; Yee, 2010), and could potentially be predators of juvenile crayfish in ponds. Future studies carrying out stomach content analysis on adult dytiscidae could show insight as to why they maintain such high trophic positions.

Conclusion: The results from this study suggest that all odonate species feed on similar invertebrates, and slight differences in diet preferences and trophic positions may occur based on either morphological difference between species or the spatial variation of odonate species in their preferred habitat. The intensity of predation on juvenile crayfish from odonate species was low and odonate trophic levels were significantly lower than crayfish indicating predation from odonates is not likely to be a contributing factor to low juvenile crayfish survival in the semi- natural aquaculture ponds. P. zealandicus in semi-natural aquaculture ponds had a high trophic position in pond food webs but looked to obtain their carbon from different sources like chironomids and did not rely on odonates as a primary source. Larger crayfish did have cannibalistic tendencies based on stomach content analysis. Apart for larger individuals being more cannibalistic, dietary overlap was similar between crayfish size groups with similar percentages of plant detritus, and invertebrates in stomachs. Regardless, the intensity of cannibalism in ponds is a likely factor contributing towards high mortality rates for juvenile P. zealandicus in semi-natural aquaculture ponds.

Chapter 4: General Discussion. Introduction: Interactions between pond invertebrates in an aquaculture environment is important to understand as it provides insight into how certain species could interact with farmed species. Predator-prey relationships continuously shape pond communities and in semi-managed aquaculture ponds, predators can reduce juvenile stock of farmed species. The predatory relationship between odonates and farmed aquaculture species has been mentioned in many sectors of aquaculture (De Marco Jr, Latini, & Reis, 1999; Pritchard, 1964), yet the relationship in ponds with NZ species has been little studied. In the current study, my aim was to examine the relationships between the freshwater crayfish species Pararenphrops zealandicus and five odonate species in semi-natural aquaculture ponds. The species spatial patterns and diets were used to investigate this relationship. Juvenile crayfish were thought to be vulnerable to predation from odonate species and their reduced biomass in ponds was possibly influenced by late instar odonate larvae in ponds. This study was the first to analyze diets of five NZ odonate species using quantitative methods, and the findings will hopefully benefit understanding of odonate larval ecology and contribute to future studies on feeding patterns in NZ odonate species. In this chapter the main findings from this research are summarized expanding on the applications of this research for management approaches in crayfish aquaculture, and recommendations for future studies.

Review of findings: The spatial distribution patterns of odonate species and P. zealandicus in ponds is important to understand as it can provides insights into how species may interact with each other based on where they are distributed in ponds. In Chapter Two I investigated the spatial distribution patterns of odonate species and P. zealandicus in ponds and examined whether environmental factors or inter-specific interactions might influence P. zealandicus abundance in ponds. Two different micro-habitats within ponds were sampled, the littoral zone, and the vegetative benthic zone. Based on the observed patterns of distribution and abundance, it seemed unlikely that any odonate species could have a significant impact on P. zealandicus

82 abundance. The factors most likely to influence P. zealandicus abundance were vegetation, and the original stocking densities of P. zealandicus in ponds.

Habitat use between different P. zealandicus size groups was also investigated. My findings indicate that large P. zealandicus dominate the littoral zone more than the benthic vegetative zone. Small P. zealandicus were found in both habitats, but their spatial distribution pattern is likely influenced by avoidance of larger individuals. Stewart, & Tabak (2011), describe the nature of crayfish to be aggressive and territorial, defending desirable habitat. Larger P. zealandicus are likely to occupy more desirable habitat and may defend it by killing other individuals. In this case smaller P. zealandicus are more vulnerable as they are smaller and cannot defend themselves, so in overcrowded areas of ponds, smaller individuals seek refuge in benthic vegetation away from larger individuals.

Stomach content and stable isotope analysis can help provide insight into predator-prey interactions in pond food webs. Interactions between late instar odonates and other pond species have been investigated showing odonate species have the ability to capture prey larger than themselves (Pritchard, 1964), making them a potential threat towards juvenile species in aquaculture systems. Whilst they are top predators in pond environments, dietary analysis of odonate species in NZ is limited with no information surrounding their impact on crayfish aquaculture. Crayfish interact broadly with other species in food and community interaction webs and can significantly influence various species, both directly and indirectly (Nyström, & Graneli, 1996). In an aquaculture environment the densities of crayfish are a lot higher which may increase competition between individuals. The diets of P. zealandicus have been well studied in their natural environment, showing interactions between species (Parkyn, Rabeni, & Collier, 1997), however there is limited research on P. zealandicus diets in an aquaculture environment.

In Chapter Three the interactions between odonate species and P. zealandicus was further investigated using stomach content and stable isotope analysis. Diet analysis of the odonate species suggested they were not major predators of crayfish as crayfish tissue was rarely recorded in their stomachs, and stable isotope analysis indicated their trophic levels were lower than both crayfish and juvenile crayfish. In crayfish, detritus, and invertebrates both made up a significant part of their diet. Cannibalism was also high in ponds suggesting this could be a cause of low juvenile crayfish survival in the ponds, which is the opposite pattern to what had been observed in similar studies (Hollows, Townsend, & Collier, 2002; Momot, 1995).

Management implications: My results indicate that late instar odonate species are not likely to be major predators of juvenile crayfish and are probably not a major influence on crayfish abundance and distribution in the studied aquaculture ponds. Instead intra-specific competition between crayfish and cannibalism more likely to influence spatial distribution patterns and abundance of crayfish in ponds. In crayfish aquaculture, aggressive behavior and cannibalism has been documented in ponds if conditions get too crowded (Baird, Patullo, & Macmillan, 2006; Savolainen, Ruohonen, & Railo, 2004). This would influence the behavior of smaller crayfish delaying their molt phases (Lutz, & Wolters, 1986). As a result, it is appropriate that the correct management practices are carried out to ensure the growth of juvenile crayfish stock in ponds.

The main threat towards young crayfish in ponds is the density of crayfish. If densities were controlled in ponds, aggressive behaviors may be reduced. A study by Brown et al., (1995), found densities of one crayfish per 2.5m2 resulted in both high rates of growth and survival. The ideal density of crayfish does vary between species, as every species’ behavior are different, however lower densities in general do reduce aggressive behavior (Yu et al., 2020). Studies involving stocking densities of P. zealandicus in ponds would need to be investigated further to determine suitable densities.

Separating ponds into size cohorts is another effective way to manage stock in ponds. In crustacean aquaculture, size grading in ponds can reduce cannibalism and lead to higher survival rates among juveniles, which in turn allows for increased stock density (Daly, Swingle & Eckert, 2012; Ahvenharju et al., 2005). Without pressures from larger individuals, juveniles grow faster reaching a larger size faster, placing them higher in the pond’s food web (Barki, & Karplus, 2004). If juveniles reach larger sizes more quickly, recruitment will also increase as crayfish recently harvested for consumption will be replaced with on-growing individuals of a size close to that required for commercial harvest.

Providing additional vegetation and planting along the backs of ponds may reduce pressures on juvenile stock. Increasing vegetation and habitat complexity in ponds can result in fewer interactions between crayfish and other pond invertebrates. Crayfish in areas with more complex habitat also reduces aggressive behavior towards other crayfish (Baird, Patullo, & Macmillan, 2006). Increased habitat complexity also increases pond invertebrate diversity which in turn provides more prey for crayfish also reducing cannibalism (Corkum & Cronin, 2004). Vegetation in ponds also provides additional food for crayfish as they are detritivores

and rely heavily on detritus (Corkum & Cronin, 2004). Invertebrate species richness would also increase in ponds as invertebrate diversity and richness is strongly correlated to increased vegetation in ponds (Hassall, Hollinshead, & Hull, 2011; Van de Meutter, Cottenie, & De Meester, 2008). Additional invertebrates and vegetation would therefore provide more feed for P. zealandicus relieving cannibalistic pressures on juvenile crayfish (Andriahajaina & Hockley, 2007).

Future research: My research suggests food web interactions in aquaculture ponds should be investigated further to gain a better understanding of the interactions between species. The interaction between odonate species and crayfish has been well researched in this study, however other species may also affect juvenile crayfish stock. One family of invertebrates that could be investigated more closely in future studies is dytiscidae. Adult diving beetles are very efficient predators in pond ecosystems, and feed on a wide range of pond invertebrates (Culler, Ohba, & Crumrine, 2014; Yee, 2010) – stable isotope analysis of the pond food web suggested they are the top predator in the ponds.

Future research should also focus on the rate of cannibalism in aquaculture ponds given that my dietary analysis clearly indicated it was occurring at a relatively high rate. Many studies involving crayfish have investigated cannibalism in ponds, with high densities, low habitat complexity, and low supplementary feed being the most likely cause of increased cannibalism in ponds (Mills, & McCloud, 1983; Naranjo-Páramo, Hernandez-Llamas, & Villarreal, 2004; Taugbøl, & Skurdal, 1992). Investigating ways to mitigate cannibalism in P. zealandicus in the semi-natural aquaculture ponds would be beneficial for better productivity of the aquaculture operation.

Conclusion: Crayfish aquaculture involving Pararenphrops zealandicus is a relatively new venture that has become a topic of interest in NZ. Because of this, the life history traits of crayfish in ponds and how they interact with other pond invertebrates is understudied. In this study I assessed the interactions between P. zealandicus and five odonate species based on spatial distribution patterns in ponds and dietary analysis. It was evident that odonates are not likely to have a significant impact on juvenile crayfish, whereas interactions between small and relatively larger crayfish are more likely to be a more important driver of crayfish survival. Whilst cannibalism was identified as the likely primary factor driving low juvenile abundance in 85 ponds, future studies involving the life history traits of crayfish in ponds could provide insights into approaches that will result in the better management of crayfish stock.

References:

Abrahamsson, S. A. A. (1966). Dynamics of an isolated population of the crayfish Astacus astacus Linné. Oikos, 96-107.

Adámek, Z., & Maršálek, B. (2013). Bioturbation of sediments by benthic macroinvertebrates and fish and its implication for pond ecosystems: a review. Aquaculture International, 21(1), 1-17.

Ahvenharju, T., Savolainen, R., Tulonen, J., & Ruohonen, K. (2005). Effects of size grading on growth, survival and cheliped injuries of signal crayfish (Pacifastacus leniusculus Dana) summerlings (age 0+). Aquaculture Research, 36(9), 857-867.

Alcorlo, P., Geiger, W., & Otero, M. (2004). Feeding preferences and food selection of the red swamp crayfish, Procambarus clarkii, in habitats differing in food item diversity. Crustaceana, 435-453.

Anderson, G. S., Barton, P. S., Archer, M., & Wallace, J. R. (2019). Invertebrate Scavenging Communities. In Carrion Ecology and Management (pp. 45-69). Springer, Cham.

Andriahajaina, & Hockley. (2007). The potential of native species aquaculture to achieve conservation objectives: freshwater crayfish in Madagascar. The International Journal of Biodiversity Science and Management, 3(4), 217-222.

Armitage, P. D. (1995). Chironomidae as food. In the Chironomidae (pp. 423-435). Springer, Dordrecht.

Armstrong, J. S. (1978). Colonisation of New Zealand by Hemicordulia australiae, with notes on its displacement of the indigenous Procordulia grayi (Odonata: Corduliidae). New Zealand Entomologist, 6(4), 381-384.

Baird, H. P., Patullo, B. W., & Macmillan, D. L. (2006). Reducing aggression between freshwater crayfish (Cherax destructor Clark: Decapoda, Parastacidae) by increasing habitat complexity. Aquaculture Research, 37(14), 1419-1428.

Bardach, J. E., Ryther, J. H., & McLarney, W. O. (1972). Aquaculture. The farming and husbandry of freshwater and marine organisms. John Wiley & Sons, Inc.

Barki, A., & Karplus, I. (2004). Size rank and growth potential in redclaw crayfish (Cherax quadricarinatus): are stunted juveniles suitable for grow‐out? Aquaculture research, 35(6), 559-567.

Beaune, D., Sellier, Y., Luquet, G., & Grandjean, F. (2018). Freshwater acidification: an example of an endangered crayfish species sensitive to pH. Hydrobiologia, 813(1), 41-50.

Blake, M. A., & Hart, P. J. B. (1995). The vulnerability of juvenile signal crayfish to perch and eel predation. Freshwater Biology, 33(2), 233-244.

Bowen, S. H., Lutz, E. V., & Ahlgren, M. O. (1995). Dietary protein and energy as determinants of food quality: trophic strategies compared. Ecology, 76(3), 899-907.

Broughton, R. J., Marsden, I. D., Hill, J. V., & Glover, C. N. (2018). Interactive effects of hypoxia and dissolved nutrients on the physiology and biochemistry of the freshwater crayfish Paranephrops zealandicus. Marine and Freshwater Research, 69(6), 933-941.

Brown, K. M. (1982). Resource overlap and competition in pond snails: an experimental analysis. Ecology, 63(2), 412-422.

Brown, P. B., Wilson, K. A., Wetzel, J. E., & Hoene, B. (1995). Increased Densities Result in Reduced Weight Gain of Crayfish Orconectes virilis 1. Journal of the World Aquaculture Society, 26(2), 165-171.

Buckland, A., Baker, R., Loneragan, N., & Sheaves, M. (2017). Standardising fish stomach content analysis: The importance of prey condition. Fisheries Research, 196, 126-140.

Butler, R. G., & Demaynadier, P. G. (2008). The significance of littoral and shoreline habitat integrity to the conservation of lacustrine damselflies (Odonata). Journal of Insect Conservation, 12(1), 23-36.

Byrne, C. F., Lynch, J. M., & Bracken, J. J. (1999, October). A sampling strategy for stream populations of white-clawed crayfish, Austropotamobius pallipes (Lereboullet)(Crustacea, Astacidae). In Biology and environment: proceedings of the Royal Irish Academy (pp. 89-94). Royal Irish Academy.

Collier, K. J., Parkyn, S. M., & Rabeni, C. F. (1997). Kōura: a keystone species. Water and Atmosphere, 5(1), 18-20.

Corbet, P. S. (1999). Dragonflies: behaviour and ecology of Odonata. Harley books.

Corkum, L. D., & Cronin, D. J. (2004). Habitat complexity reduces aggression and enhances consumption in crayfish. Journal of Ethology, 22(1), 23-27.

Crandall, K. A., & Buhay, J. E. (2007). Global diversity of crayfish (Astacidae, Cambaridae, and Parastacidae—Decapoda) in freshwater. In Freshwater animal diversity assessment (pp. 295- 301). Springer, Dordrecht.

Crump, M. L. (1984). Ontogenetic changes in vulnerability to predation in tadpoles of Hyla pseudopuma. Herpetologica, 265-271.

Culler, L. E., Ohba, S. Y., & Crumrine, P. (2014). Predator-prey interactions of dytiscids. In Ecology, systematics, and the natural history of predaceous diving beetles (Coleoptera: Dytiscidae) (pp. 363-386). Springer, Dordrecht.

Daly, B., Swingle, J. S., & Eckert, G. L. (2012). Increasing hatchery production of juvenile red king crabs (Paralithodes camtschaticus) through size grading. Aquaculture, 364, 206-211.

De Marco Jr, P., Latini, A. O., & Reis, A. P. (1999). Environmental determination of dragonfly assemblage in aquaculture ponds. Aquaculture Research, 30(5), 357-364.

Dethier, M. N., Dobkowski, K., Noreen, A., Yun, M., & Moosmiller, A. (2019). Vulnerability of juvenile clams to predation by shore crabs. Aquaculture, 506, 350-354.

Fisher, S. G.; Likens, G. E. 1973: Energy flow in Bear Brook, New Hampshire: an integrative approach to stream ecosystem metabolism. Ecological monographs 43: 421-139.

Folsom, T. C., & Collins, N. C. (1984). The diet and foraging behavior of the larval dragonfly Anax junius (Aeshnidae), with an assessment of the role of refuges and prey activity. Oikos, 105-113.

France, R. (1996). Ontogenetic shift in crayfish δ13C as a measure of land-water ecotonal coupling. Oecologia, 107(2), 239-242.

Freshwater Habitats Trust. (2015). Pond Habitat Survey Method Booklet. Freshwater Habitats

Trust, Oxford.

Grimm, N. B. (1988). Role of macroinvertebrates in nitrogen dynamics of a desert stream. Ecology, 69(6), 1884-1893.

Gu, B., Schell, D. M., Huang, X., & Yie, F. (1996). Stable isotope evidence for dietary overlap between two planktivorous fishes in aquaculture ponds. Canadian Journal of Fisheries and

Aquatic Sciences, 53(12), 2814-2818.

Gutiérrez-Yurrita, P. J., Sancho, G., Bravo, M. A., Baltanas, A., & Montes, C. (1998). Diet of the red swamp crayfish Procambarus clarkii in natural ecosystems of the Donana National Park temporary fresh-water marsh (Spain). Journal of Crustacean Biology, 18(1), 120-127.

Hamada, N., & Oliveira, S. J. (2003). Food items of larvae of Rimanella arcana (Needham, 1933) (Odonata: Amphipterygidae) in central Amazonia, Brazil. Entomotropica, 18(2), 153- 155.

Hanson, J. M., Chambers, P. A., & Prepas, E. E. (1990). Selective foraging by the crayfish Orconectes virilis and its impact on macroinvertebrates. Freshwater Biology, 24(1), 69-80.

Hammond, K. S., Hollows, J. W., Townsend, C. R., & Lokman, P. M. (2006). Effects of temperature and water calcium concentration on growth, survival and moulting of freshwater crayfish, Paranephrops zealandicus. Aquaculture, 251(2-4), 271-279.

Hassall, C., Hollinshead, J., & Hull, A. (2011). Environmental correlates of plant and invertebrate species richness in ponds. Biodiversity and Conservation, 20(13), 3189-3222.

Hawking, J. H., & New, T. R. (1999). The distribution patterns of dragonflies (Insecta: Odonata) along the Kiewa River, Australia, and their relevance in conservation assessment. Hydrobiologia, 392(2), 249-260.

Heck, K. L., & Crowder, L. B. (1991). Habitat structure and predator—prey interactions in vegetated aquatic systems. In Habitat structure (pp. 281-299). Springer, Dordrecht.

Herwig, B. R., & Schindler, D. E. (1996). Effects of aquatic insect predators on zooplankton in fishless ponds. Hydrobiologia, 324(2), 141-147.

Hirvonen, H. (1992, January). Effects of backswimmer (Notonecta) predation on crayfish (Pacifastacus) young: autotomy and behavioural responses. In Annales Zoologici Fennici (pp. 261-271). Finnish Zoological Publishing Board, formed by the Finnish Academy of Sciences, Societas Biologica Fennica Vanamo, Societas pro Fauna et Flora Fennica, and Societas Scientiarum Fennica

Hobson, K. A., Gloutney, M. L., & Gibbs, H. L. (1997). Preservation of blood and tissue samples for stable-carbon and stable-nitrogen isotope analysis. Canadian Journal of Zoology, 75(10), 1720-1723.

Holdich, D. M. (Ed.). (2002). Biology of freshwater crayfish (p. 702). Oxford: Blackwell Science.

Hollows, J. W., Townsend, C. R., & Collier, K. J. (2002). Diet of the crayfish Paranephrops zealandicus in bush and pasture streams: insights from stable isotopes and stomach analysis. New Zealand Journal of Marine and Freshwater Research, 36(1), 129-142.

Hopkins, C. L. (1970). Systematics of the New Zealand freshwater crayfish Paranephrops (Crustacea: Decapoda: Parastacidae). New Zealand journal of marine and freshwater research, 4(3), 278-291.

Hutchinson, L. (2005). Ecological Aquaculture. Hampshire, Permanent Publications.

Hyslop, E. J. (1980). Stomach contents analysis—a review of methods and their application. Journal of fish biology, 17(4), 411-429.

Janssen, A., Hunger, H., Konold, W., Pufal, G., & Staab, M. (2018). Simple pond restoration measures increase dragonfly (Insecta: Odonata) diversity. Biodiversity and Conservation, 27(9), 2311-2328.

Jara, F. G. (2008). Tadpole–odonate larvae interactions: influence of body size and diel rhythm. Aquatic Ecology, 42(3), 503-509.

Jennings, S., & Mackinson, S. (2003). Abundance–body mass relationships in size‐structured food webs. Ecology Letters, 6(11), 971-974.

Jennings, S., Warr, K. J., & Mackinson, S. (2002). Use of size-based production and stable isotope analyses to predict trophic transfer efficiencies and predator-prey body mass ratios in food webs. Marine Ecology Progress Series, 240, 11-20.

Johansson, F. (1993). Intraguild predation and cannibalism in odonate larvae: effects of foraging behaviour and zooplankton availability. Oikos, 80-87.

Jones, C. G., Lawton, J. H., & Shachak, M. (1994). Organisms as ecosystem engineers. In Ecosystem management (pp. 130-147). Springer, New York, NY.

Jonsson, A. N. D. E. R. S. (1992). Shelter selection in YOY crayfish Astacus astacus under predation pressure by dragonfly larvae. Nordic journal of freshwater research. Drottningholm, 67, 82-87.

Kasimov, T. (1956). The feeding of the dragonfly larva Anax imperator Leach. Izvestia Akademiya Nauk SSSR, 11, 91-92.

Kraus, J. M. (2010). Diet shift of lentic dragonfly larvae in response to reduced terrestrial prey subsidies. Journal of the North American Benthological Society, 29(2), 602-613.

Kusabs, I. A. (2015). Kōura (Paranephrops planifrons) populations in the Te Arawa lakes: An ecological assessment using the traditional Māori tau kōura harvesting method and recommendations for sustainable management (Doctoral dissertation, University of Waikato).

Kusabs, I. A., Emery, W., Parkyn, S., Hickey, C., & Quinn, J. (2005). Tau kōura sample collection and processing protocol.

Kusabs, I. A., Hicks, B. J., Quinn, J. M., Perry, W. L., & Whaanga, H. (2018). Evaluation of a traditional Māori harvesting method for sampling kōura (freshwater crayfish, Paranephrops planifrons) and toi toi (bully, Gobiomorphus spp.) populations in two New Zealand streams. New Zealand Journal of Marine and Freshwater Research, 52(4), 603-625.

Lawton, J. H. (1970). Feeding and food energy assimilation in larvae of the damselfly Pyrrhosoma nymphula (Sulz.) (Odonata: Zygoptera). The Journal of Animal Ecology, 669-689.

Layton, R. J., & Voshell Jr, J. R. (1991). Colonization of new experimental ponds by benthic macroinvertebrates. Environmental Entomology, 20(1), 110-117.

Lutz, C. G., & Wolters, W. R. (1986). The effect of five stocking densities on growth and yield of red swamp crawfish Procambarus clarkii. Journal of the World Aquaculture Society, 17(1‐ 4), 33-36.

Manor, R., Segev, R., Leibovitz, M. P., Aflalo, E. D., & Sagi, A. (2002). Intensification of redclaw crayfish Cherax quadricarinatus culture: II. Growout in a separate cell system. Aquacultural Engineering, 26(4), 263-276.

McAbendroth, L., Foggo, A., Rundle, S. D., & Bilton, D. T. (2005). Unravelling nestedness and spatial pattern in pond assemblages. Journal of Animal Ecology, 74(1), 41-49.

Mills, B. J., & McCloud, P. I. (1983). Effects of stocking and feeding rates on experimental pond production of the crayfish Cherax destructor Clark (Decapoda: Parastacidae). Aquaculture, 34(1-2), 51-72.

Mohan, M. V., & Sankaran, T. M. (1988). Two new indices for stomach content analysis of fishes. Journal of Fish Biology, 33(2), 289-292.

Momot, W. T. (1995). Redefining the role of crayfish in aquatic ecosystems. Reviews in Fisheries Science, 3(1), 33-63.

Moore, S. (2001) Identification guide: What freshwater invertebrates is this? Retrieved August 19, 2019, from https://www.landcareresearch.co.nz/resources/identification/animals/freshwater- invertebrates/guide.

Moore, W. G., & Burn, A. (1968). Lethal oxygen thresholds for certain temporary pond invertebrates and their applicability to field situations. Ecology, 49(2), 349-351.

Musgrove, R. J. (1988). The diet, energetics and distribution of the freshwater crayfish Paranephrops zealandicus (White) in Lake Georgina, South Island, New Zealand.

Naranjo-Páramo, J., Hernandez-Llamas, A., & Villarreal, H. (2004). Effect of stocking density on growth, survival and yield of juvenile redclaw crayfish Cherax quadricarinatus (Decapoda: Parastacidae) in gravel-lined commercial nursery ponds. Aquaculture, 242(1-4), 197-206.

Nyström, P. (2017). Ecological impact of introduced and native crayfish on freshwater communities: European perspectives. In Crayfish in Europe as alien species (pp. 63-85). Routledge.

Nyström, P., & Granéli, W. (1996). The effect of food availability on survival, growth, activity and the number of mature females in crayfish populations. Freshwater Crayfish, 11(4), 170- 181.

Nyström, P., & Pérez, J. R. (1998). Crayfish predation on the common pond snail (Lymnaea stagnalis): the effect of habitat complexity and snail size on foraging efficiency. Hydrobiologia, 368(1-3), 201-208.

Nyström, P. E. R., Brönmark, C., & Graneli, W. (1996). Patterns in benthic food webs: a role for omnivorous crayfish?. Freshwater Biology, 36(3), 631-646.

Parkyn, S. M. (2015). A review of current techniques for sampling freshwater crayfish. Freshwater Crayfish. CRC Press, Boca Raton, 205-220.

Parkyn, S., Collier, K., & Hicks, B. (2001). New Zealand stream crayfish: functional omnivores but trophic predators? Freshwater Biology, 46(5), 641-652. 93

Parkyn, S. M., Rabeni, C. F., & Collier, K. J. (1997). Effects of crayfish (Paranephrops planifrons: Parastacidae) on in‐stream processes and benthic faunas: A density manipulation experiment. New Zealand journal of marine and freshwater research, 31(5), 685-692.

Pritchard, G. (1964). The prey of dragonfly larvae (Odonata; Anisoptera) in ponds in northern Alberta. Canadian Journal of Zoology, 42(5), 785-800.

Raebel, E. M., Merckx, T., Riordan, P., Macdonald, D. W., & Thompson, D. J. (2010). The dragonfly delusion: why it is essential to sample exuviae to avoid biased surveys. Journal of Insect Conservation, 14(5), 523-533.

Rennie, M. D., & Jackson, L. J. (2005). The influence of habitat complexity on littoral invertebrate distributions: patterns differ in shallow prairie lakes with and without fish. Canadian Journal of Fisheries and Aquatic Sciences, 62(9), 2088-2099.

Reynolds, J. D. (2002). Growth and reproduction. Biology of freshwater crayfish, 152-191.

Reynolds, J., & Souty-Grosset, C. (2011). Management of freshwater biodiversity: crayfish as bioindicators. Cambridge University Press.

Reynolds, J., Souty-Grosset, C., & Richardson, A. (2013). Ecological roles of crayfish in freshwater and terrestrial habitats. Freshwater Crayfish, 19(2), 197-218.

Romano, N., & Zeng, C. (2017). Cannibalism of decapod crustaceans and implications for their aquaculture: a review of its prevalence, influencing factors, and mitigating methods. Reviews in Fisheries Science & Aquaculture, 25(1), 42-69.

Rowe, R. J. (1985). Intraspecific interactions of New Zealand damselfly larvae I. Xanthocnemis zealandica, Ischnura aurora, and Austrolestes colensonis (Zygoptera: Coenagrionidae: Lestidae). New Zealand journal of zoology, 12(1), 1-15.

Rowe, R. J. (1987). The dragonflies of New Zealand. The dragonflies of New Zealand.

Rowe, R. J. (1992). Ontogeny of agonistic behaviour in the territorial damselfly larvae, Xanthocnemis zealandica (Zygoptera: Coenagrionidae). Journal of Zoology, 226(1), 81-93.

Saoud, I. P., Rodgers, L. J., Davis, D. A., & Rouse, D. B. (2008). Replacement of fish meal with poultry by‐product meal in practical diets for redclaw crayfish (Cherax quadricarinatus). Aquaculture Nutrition, 14(2), 139-142.

Savolainen, R., Ruohonen, K., & Railo, E. (2004). Effect of stocking density on growth, survival and cheliped injuries of stage 2 juvenile signal crayfish Pasifastacus leniusculus Dana. Aquaculture, 231(1-4), 237-248.

Scholz, O. (1990). Physicochemistry and vegetation of piccaninnie ponds, a coastal aquifer- fed pond in south-eastern South Australia. Marine and Freshwater Research, 41(2), 237-246.

Sollie, S., & Verhoeven, J. T. (2008). Nutrient cycling and retention along a littoral gradient in a Dutch shallow lake in relation to water level regime. Water, air, and soil pollution, 193(1-4), 107-121.

Spence, J. R. (1986). Relative impacts of mortality factors in field populations of the waterstrider Gerris buenoi Kirkaldy (Heteroptera: Gerridae). Oecologia, 70(1), 68-76.

Stenroth, P., Holmqvist, N., Nyström, P., Berglund, O., Larsson, P., & Granéli, W. (2006). Stable isotopes as an indicator of diet in omnivorous crayfish (Pacifastacus leniusculus): the influence of tissue, sample treatment, and season. Canadian Journal of Fisheries and Aquatic Sciences, 63(4), 821-831.

Stewart, C., & Tabak, J. (2011). Size-based dominance hierarchies in the New Zealand freshwater crayfish (kōura) Paranephrops zealandicus. New Zealand Journal of Marine and Freshwater Research, 45(2), 281-285.

Stiers, I., Crohain, N., Josens, G., & Triest, L. (2011). Impact of three aquatic invasive species on native plants and macroinvertebrates in temperate ponds. Biological Invasions, 13(12), 2715-2726.

Sychra, J., Adámek, Z., & Petřivalská, K. (2010). Distribution and diversity of littoral macroinvertebrates within extensive reed beds of a lowland pond. In Annales de Limnologie- International Journal of Limnology (Vol. 46, No. 4, pp. 281-289). EDP Sciences.

Syväranta, J., Vesala, S., Rask, M., Ruuhijärvi, J., & Jones, R. I. (2008). Evaluating the utility of stable isotope analyses of archived freshwater sample materials. Hydrobiologia, 600(1), 121- 130.

Taugbøl, T., & Skurdal, J. (1992). Growth, mortality and moulting rate of noble crayfish, Astacus astacus L., juveniles in aquaculture experiments. Aquaculture Research, 23(4), 411- 420.

Thompson, D. J. (1978). Prey size selection by larvae of the damselfly, Ischnura elegans (Odonata). The Journal of Animal Ecology, 769-785.

Ulehlova, B., & Přibil, S. (1978). Water chemistry in the fishpond littorals. In Pond Littoral Ecosystems (pp. 126-140). Springer, Berlin, Heidelberg.

Usio, N., Nakajima, H., Kamiyama, R., Wakana, I., Hiruta, S., & Takamura, N. (2006). Predicting the distribution of invasive crayfish (Pacifastacus leniusculus) in a Kusiro Moor marsh (Japan) using classification and regression trees. Ecological Research, 21(2), 271-277.

Usio, N., & Townsend, C. R. (2004). Roles of crayfish: consequences of predation and bioturbation for stream invertebrates. Ecology, 85(3), 807-822.

Van de Meutter, F., Cottenie, K., & De Meester, L. (2008). Exploring differences in macroinvertebrate communities from emergent, floating-leaved and submersed vegetation in shallow ponds. Fundamental and Applied Limnology/Archiv Für Hydrobiologie, 173(1), 47- 57.

Wahizatul-Afzan, A., Julia, J., & Amirrudin, A. (2006). Diversity and distribution of dragonflies (Insecta: Odonata) in Sekayu recreational forest, Terengganu. Journal of Sustainability Science and Management, 1(2), 97-106.

Warren, P. H., & Lawton, J. H. (1987). Invertebrate predator-prey body size relationships: an explanation for upper triangular food webs and patterns in food web structure? Oecologia, 74(2), 231-235.

Whitmore, N. (1997). Population ecology of the freshwater crayfish Paranephrops zealandicus and its effect on the community structure of a lowland bush stream (Doctoral dissertation, University of Otago).

Whitmore, N., Huryn, A. D., Arbuckle, C. J., & Jansma, F. (2000). Ecology and distribution of the freshwater crayfish Paranephrops zealandicus in Otago. Science for Conservation, 148, 5- 42.

Winterbourn, M. J., Gregson, K. L., & Dolphin, C. H. (1989). Guide to the aquatic insects of New Zealand.

Wissinger, S. A., Perchik, M. E., & Klemmer, A. J. (2018). Role of animal detritivores in the breakdown of emergent plant detritus in temporary ponds. Freshwater Science, 37(4), 826- 835.

Witzig, J. F., Huner, J. V., & Avault Jr, J. W. (1986). Predation by dragonfly naiads Anax junius on young crawfish Procambarus clarkii. Journal of the world aquaculture society, 17(1‐ 4), 58-63.

Yee, D. A. (2010). Behavior and aquatic plants as factors affecting predation by three species of larval predaceous diving beetles (Coleoptera: Dytiscidae). Hydrobiologia, 637(1), 33-43.

Yu, J., Xiong, M., Ye, S., Li, W., Xiong, F., Liu, J., & Zhang, T. (2020). Effects of stocking density and artificial macrophyte shelter on survival, growth and molting of juvenile red swamp crayfish (Procambarus clarkii) under experimental conditions. Aquaculture, 521, 735001.

Yuan, X., Lu, J., & Liu, H. (2002). Influence of characteristics of Scirpus mariqueter community on the benthic macro-invertebrate in a salt marsh of the Changjiang Estuary. Acta Ecologica Sinica, 22(3), 326-333.

Appendix 1: Invertebrate composition in ponds.

Pond 1 Pond 2 Pond 3 Pond 4 Pond 5 Pond 6 Pond 7 Pond 8 Pond 9 Pond 10 Total Trichoptera Economidae 1 1 0 1 0 0 0 0 1 1 5 Triplectides 1 1 1 1 1 1 0 1 1 0 8 Triplectidina 1 1 0 1 0 1 0 0 1 0 5 Hudsonema 0 1 0 0 0 1 0 0 0 0 0 Hemiptera A.wakefieldi 1 1 1 1 1 1 1 1 1 1 10 Microvelia 1 1 1 1 1 1 1 1 1 1 10 Misovelia 0 0 0 0 0 0 0 0 0 0 0 C. sigara 1 1 1 1 0 1 1 0 1 1 8 C. Diaprepocoris 1 0 0 1 0 0 0 0 1 0 3 coleoptera Hydrophilidae 1 1 1 0 0 1 0 0 1 0 5 Dytiscidae (beetle) 0 1 0 1 0 1 0 0 1 0 4 Onychohydrus 1 1 1 1 1 1 0 0 1 0 7 Lancetes 0 1 0 0 0 0 0 1 1 0 3 Antiporus 1 1 1 1 1 0 1 0 1 1 8 Dytiscidae (larvae) 0 0 0 0 0 0 0 0 0 0 0 Antiporus (larvae) 0 0 0 0 0 0 1 1 1 1 4 Hydrophilid (larvae) 1 0 0 1 1 1 0 0 0 0 4 Rhantus 1 1 0 1 0 1 0 0 1 1 6 Staphylinidae 0 0 0 1 1 0 0 0 0 1 3 Odonata P. grayii 1 1 0 0 1 1 1 1 0 1 7 P. smithi 1 1 1 1 1 1 1 1 1 1 10 H. austalidae 1 1 1 1 1 1 1 1 1 1 10 A. colensis 1 1 1 1 1 1 1 1 1 1 10 X. zealandica 1 1 1 1 1 1 1 1 1 1 10 Diptera Culicidae 1 1 0 0 1 1 0 1 1 1 7 Chironomidae 1 1 1 1 1 1 1 1 1 1 10 Oligochaete Eiseniella 1 1 1 1 1 1 1 1 1 1 10 Salifidae Barbronia 0 0 0 0 1 0 0 0 1 0 2 Mollusca Ampullariidae 0 0 0 0 1 0 0 0 1 0 2 Terrestrial insects 1 1 1 1 1 1 1 1 1 1 10 Crustacea P. zealandicus 1 1 1 1 1 1 1 0 1 1 9 Ostracoda 1 1 1 1 1 1 1 1 1 1 10 Cyclopoids 1 1 1 1 1 1 1 1 1 1 10 Bosmina 1 1 1 1 1 1 1 1 1 1 10 Entognatha Collembola 1 1 1 0 1 1 0 0 1 0 6 Araneae P. dolomedes 1 1 1 1 1 1 1 1 1 1 10 Acari Temnocephala 0 1 1 1 1 1 1 1 1 0 8 mites 0 1 0 1 0 1 0 0 0 0 3 Nematoda 1 1 1 1 1 1 1 1 1 1 10 Hydroids 0 1 0 1 0 1 1 0 1 0 5 Total 28 32 22 29 26 30 21 20 33 23

Appendix 2: Frequency of organic matter present in study species stomachs.

Stomach contents A. colensonis X. zealandica H. australiae P. smithii P. grayi P. zealandicus Wood detritus 0 0 0 0 0 34 Plant detritus 0 0 0 0 0 87 P. zealandicus 1 0 3 0 1 45 Copepoda 16 11 5 2 2 18 Odonata 3 0 2 0 1 7 Bosmina 8 12 0 1 0 5 Chironomidae 20 16 11 16 5 21 Diptera 0 2 0 1 0 3 Ostracoda 11 4 8 10 2 11 Tipulidae 1 0 0 0 0 0 Dytiscidae 1 0 0 0 0 0 Microvelia 11 0 0 1 0 4 Terrestrial invertebrates 14 0 2 0 0 14 Trichoptera 0 0 0 0 0 4 Oligochaete 0 0 0 0 0 11 Collembola 0 0 0 0 0 4 Coleoptera 0 0 0 0 0 4 C. sigara 0 0 0 0 0 6 Onion sack 0 0 0 0 0 29

Appendix 3: Frequency of organic matter present in stomachs of Paranephrops zealandicus in different habitats sampled.

Stomach contents Bracken bags Kick net Wood detritus 17 17 Plant detritus 45 42 P. zealandicus 17 28 Copepoda 10 8 Odonata 3 4 Bosmina 3 2 Chironomidae 9 12 Culicidae 0 3 Ostracoda 7 4 Tipulidae 0 0 Dytiscidae 0 0 Microvelia 0 4 Terrestrial Invertebrates 4 10 Trichoptera 0 4 Oligochaete 7 4 Collembola 0 4 Coleoptera 3 1 C. sigara 1 5 Onion sack 21 8

100

Appendix 4: Frequency of organic matter present in stomachs of Paranephrops zealandicus size groups.

Stomach contents Small Medium Large Wood detritus 6 12 16 Plant detritus 27 31 29 P. zealandicus 4 15 26 Copepoda 9 2 7 Odonata 2 2 3 Bosmina 5 0 0 Chironomidae 8 4 9 Culicidae 2 1 0 Ostracoda 8 2 1 Tipulidae 0 0 0 Dytiscidae 0 0 0 Microvelia 0 1 3 Terrestrial Invertebrates 4 0 10 Trichoptera 0 1 3 Oligochaete 5 0 6 Collembola 0 1 3 Coleoptera 1 1 2 C. sigara 2 1 3 Onion sack 10 12 7

101

102