Coleoptera: Chrysomelidae)

Population phenology and natural enemies of

Paropsis atomaria Olivier (Coleoptera: Chrysomelidae)

in South-East Queensland

Michael Patrick Duffy, B.Sc. (Hons)

Submitted in fulfilment of the requirements for the Master of Applied Science

School of Natural Resource Sciences

Queensland University of Technology

December 2006 Keywords

Paropsis atomaria, Neopolycystus sp., Eucalyptus cloeziana, parasitioids, natural enemies, habitat manipulation, conservation biological control

Abstract

Paropsis atomaria Olivier (Coleoptera: Chrysomelidae: Paropsini), is a major pest of commercially grown eucalypts in South-East Queensland. Current management of paropsine beetles involves regular inspection and the application of chemical sprays if defoliation is severe. However, non-chemical control of plantation pests is highly desirable given the requirement to certify forest practices for sustainability, and community concerns over the use of pesticides. One way of reducing pesticide use is through conservation biological control, which requires detailed knowledge of the life history of the pest and its natural enemies. This thesis documents aspects of P. atomaria phenology, including life tables, sex ratios and damage estimates; identifies the predators, parasites, and egg and larval parasitoids of P. atomaria; and examines the ecology of the most promising natural enemy, Neopolycystus Girault sp.

(Hymenoptera: Pteromalidae) in South-East Queensland.

P. atomaria adults are active from September until April and can complete up to four generations in a season. Field mortality between egg and fourth instar larvae is approximately 94%. A large proportion of this mortality can be attributed to natural enemies. The most abundant predators in eucalypt plantations were spiders, comprising 88% of all predators encountered.

Egg parasitoids exerted the greatest influence on P. atomaria populations, emerging from around 50% of all egg batches, and were responsible for mortality of almost one third of all eggs in the field. Only about one percent of larvae were parasitised in the field, in contrast to paropsine pests in temperate Australia, where egg parasitism rates are low and larval parasitism rates high.

Neopolycystus sp. was the only primary parasitoid reared from P. atomaria eggs, along with three hyperparasitoid species; Baeoanusia albifunicle Girault

(Encyrtidae), Neblatticida sp. (Encyrtidae) and Aphaneromella sp. (Platygasteridae).

This is the first record of B. albifunicle hyperparasitising Neopolycystus spp. B. albifunicle emerged from one-third of all parasitised egg batches and could pose a potential problem to the efficacy of Neopolycystus sp. as a biological control agent.

However, within egg batches, hyperparasitoids rarely killed all Neopolycystus sp. with only 9% of hyperparasitised egg batches failing to produce any primary parasitoids. Total field mortality of P. atomaria through direct and indirect effects of parasitism by Neopolycystus sp. was 28%. The proportion of egg batches parasitised increased with exposure time in the field, but within-batch parasitism rate did not. In general, there was no significant correlation between parasitism rates and distance from landscape features (viz. water sources and native forest).

iii

Table of Contents

Keywords ...... i Abstract ...... ii Table of Contents ...... iv List of Figures ...... vii List of Tables ...... xi Declarations ...... xiii Acknowledgements...... xiv

Chapter 1: Chrysomelids as pests of eucalypt plantations and their natural control ...... 1

1.1 Eucalypt forestry in Queensland...... 1 1.2 Paropsines as pests of eucalypt forestry...... 2 1.3 Paropsis atomaria...... 3 1.4 Conservation Biological Control ...... 4 1.5 Natural enemies...... 7 1.5.1 Predators...... 7 1.5.2 Parasitoids ...... 8 1.6 Conservation biological control in forestry ...... 9

Chapter 2: Phenology of Paropsis atomaria Olivier (Coleoptera: Chrysomelidae) in subtropical eucalypt plantations ...... 12

2.1 Introduction...... 12 2.2 Materials and Methods...... 14 2.2.1 Site details...... 14 2.2.2 Phenology...... 15 2.2.3 Voltinism estimates ...... 15 2.2.4 Adult sex ratios & size ...... 16 2.2.5 Mortality in the field ...... 16 2.2.6 Damage estimates & host plant foliage ...... 17 2.3 Results...... 18 2.3.1 General site comparisons...... 18

2.3.2 Phenology...... 18 2.3.3 Voltinism estimates ...... 22 2.3.4 Adult sex ratios & size ...... 22 2.3.5 Mortality in the field ...... 26 2.3.6 Damage estimates & host plant foliage ...... 28 2.4 Discussion ...... 30 2.4.1 Population Phenology...... 30 2.4.3 Population mortality ...... 31 2.4.4 Management implications...... 32

Chapter 3: Identifying the natural enemies of Paropsis atomaria Olivier (Coleoptera: Chrysomelidae) in South-East Queensland plantations ...... 34

3.1 Introduction...... 34 3.2 Materials and Methods...... 35 3.2.1 Predators...... 35 3.2.2 Parasitoids ...... 35 3.3 Results...... 38 3.3.1 Predators...... 38 3.3.2 Parasitoids ...... 40 3.3.3 Parasites...... 47 3.4 Discussion ...... 47 3.4.1 Predation...... 47 3.4.2 Larval parasitism ...... 49 3.4.3 Egg parasitism ...... 50 3.4.4 Parasites...... 50

Chapter 4: Ecology of Neopolycystus Girault sp. (Hymenoptera: Pteromalidae), an egg parasitoid of Paropsis atomaria Olivier (Coleoptera: Chrysomelidae) in South-East Queensland...... 52

4.1 Introduction...... 52 4.2 Materials and methods ...... 54 4.2.1 Sampling methods ...... 54 4.2.3 Indirect effects of egg parasitism ...... 55

4.2.4 Parasitism rates and host density ...... 56 4.2.5 Parasitism rate and exposure time in the field ...... 57 4.2.6 Parasitism rate in relation to plantation landscape features...... 58 4.2.7 Neopolycystus sp. longevity ...... 58 4.3 Results...... 59 4.3.1 Direct effects of egg parasitism: between- and within-batch parasitism rates across sites and seasons...... 59 4.3.2 Indirect effects of egg parasitism ...... 62 4.3.3 Parasitism rate and host density...... 64 4.3.4 Parasitism rates and exposure time in the field...... 66 4.3.5 Parasitism rate in relation to plantation landscape features...... 69 4.3.6 Neopolycystus sp. longevity ...... 70 4.4 Discussion ...... 74 4.4.1 Neopolycystus and P. atomaria mortality...... 74 4.4.2 Density dependent parasitism ...... 75 4.4.3 Site Effects...... 76 4.4.4 Hyperparasitoids...... 76

Chapter 5: General Discussion ...... 78

5.1 Introduction...... 78 5.2 Paropsis atomaria...... 79 5.3 Natural enemies...... 80 5.4 Neopolycystus sp...... 81 5.5 Implications to conservation biological control...... 82

References ...... 84

Appendix I ...... 98

A. Estimating egg batch size...... 98 B. Day-degree requirements and developmental thresholds for Paropsis atomaria...... 99

List of Figures

Figure 2.1: Mean ± s.e. number of Paropsis atomaria egg batches per branch at three sites in South-east Queensland [Site I (triangles), Site II (circles) and Site III (squares)] during 2004-2005 (Sites I & II) and 2005-2006 (Site III)……………… 19

Figure 2.2: Mean ± s.e. number of first and second (triangles), third (squares) and fourth (circles) instar Paropsis atomaria larvae per branch at three sites in South-east Queensland [Site I (A), Site II (B) and Site III (C)]. Note differences in scale of axes………………………………………………………………………………….20

Figure 2.3: Proportion of branches (n=144) occupied throughout the season by any developmental stage of Paropsis atomaria at three sites in South-east Queensland [Site I (A), Site II (B) and Site III (C)]. Note difference in scale of axes…………..21

Figure 2.4: Proportion of new (teneral) Paropsis atomaria adults collected throughout the 2004-2005 field season at two sites in South-east Queensland [Site I (A) and Site II (B)]. Numbers associated with each data point show the total number of adults collected on each sample date……………………………………………..23

Figure 2.5: Proportion of old (hard) female Paropsis atomaria adults collected from foliage at two sites in South-east Queensland [Site I (A) and Site II (B)] during the 2004-2005 season. Open symbols denote samples for which the sex ratio differed significantly from unity (Chi-square analysis, Bonferroni adjusted, P < 0.004). Numbers at each data point indicate the total number of old beetles collected for each (samples < 10 were not included in analyses). Dotted line represents 1:1 sex ratio. 24

Figure 2.6: Proportion of new (teneral) female Paropsis atomaria adults collected from foliage at two sites in South-east Queensland [Site I (A) and Site II (B)] during the 2004-2005 season. Open symbols denote samples for which the sex ratio differed significantly from unity (Chi-square analysis, Bonferroni adjusted, P < 0.004). Numbers at each data point indicate the total number of new beetles collected for each (samples < 10 were not included in analyses). Dotted line represents 1:1 sex ratio………………………………………………………………25

vii

Figure 2.7: Mean + s.e. damage scores (proportion of new growth foliage missing) (bars), and proportion of canopy in current season’s growth (lines) per tree at three sites in South-East Queensland [Site I (A), Site II (B) and Site III (C)]…………….29

Figure 3.1: Mean ± s.e. number of spiders (circles), P. atomaria eggs (triangles) and P. atomaria adults (squares) per branch in two plantation forestry sites in South-East Queensland in 2004 - 2005 (site data combined)……………………………………40

Figure 3.2: Monthly outcome frequencies for Paropsis atomaria eggs batches collected from two plantation forest sites in South-East Queensland [Site I (A) and Site II (B)] in 2004-2005. Larvae emerged (grey), parasitic wasps emerged (black) or eggs failed to hatch (white). Numbers above bars show the number of egg batches for each month’s data………………………………………………………………..41

Figure 3.3: Monthly outcome frequencies for Paropsis atomaria larvae (all instars) collected from two plantation forestry sites in South-East Queensland [Site I (A) and Site II (B)] in 2004-2005. Larvae died (white), pupated (grey), or were parasitised by tachinid flies (black). Numbers above bars denote the total number of larvae collected each month…………………………………………………………….…..44

Figure 3.4: Overall proportion of Paropsis atomaria larval instars parasitised by tachinid flies from two plantation forestry sites in South-East Queensland [Site I (white) and Site II (grey)]. Numbers above bars represent the number of larvae collected for each instar throughout the season…………………………………..…45

Figure 3.5: Average + s.e. proportion of Paropsis atomaria larval instars parasitised per field-collected batch from two plantation forestry sites in South-East Queensland [Site I (white) and Site II (grey)] in 2004 - 2005……………………………………46

Figure 4.1: Between (lines) and mean + se within (bars) batch parasitism rate of Paropsis atomaria egg batches by Neopolycystus sp. three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 [Site I 04/05 (A), Site II 04/05 (B), Site I 05/06 (C), and Site III 05/06 (D)]…………………………………61

viii

Figure 4.2: Mean + se proportion of Paropsis atomaria eggs per batch that produced neither wasps nor larvae from batches which were unparasitised (white) and parasitised (grey) from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006…………………………………………………………...62

Figure 4.3: Mean + se percentage Paropsis atomaria first instar larval mortality at different initial densities on new season’s expanded Eucalyptus cloeziana foliage after one week. ‘a’ designates the mean that differed significantly from all others (Fisher’s LSD post-hoc test)………………………………………………………...63

Figure 4.4: Relationship between the number of Paropsis atomaria egg batches collected within a plantation section,and the proportion of those egg batches that were parasitised for two eucalypt plantation sites in South-East Queensland in 2004/2005 [Site I (A) and Site II (B)]……………………………………………….65

Figure 4.5: Proportion of Paropsis atomaria egg batches parasitised after exposure to wasps in the field for different time periods at three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 [Site I 04/05(A), Site II 04/05 (B), Site III 05/06 (C), and Site I 05/06 (D)]………………………………………..67

Figure 4.6: Average + s.e. within-batch parasitism rate of Paropsis atomaria egg batches at different field exposure times from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 [Site I 04/05(A), Site II 04/05 (B), Site III 05/06 (C), and Site I 05/06 (D)]. Different letters above bars designate means that differed significantly (Fisher’s LSD post-hoc test)……………………..68

Figure 4.7: Kaplan-Meier survivorship curve for Neopolycycstus sp. in the laboratory……………………………………………………………………………70

Figure 4.8: Relative abundance of four parasitoid species associated with Paropsis atomaria egg batches from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006. Neopolycystus sp. (stippled), Baeoanusia albifunicle (grey), Neblatticida sp. (black), and Aphanomerella sp. (white). Different letters above bars designate means that differed significantly……………………………..71

Figure 4.9: Across-season species composition of wasps emerging from Paropsis atomaria egg batches from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 [Site I 04/05 (A), Site II 04/05 (B), Site I 05/06 (C), and Site III 05/06 (D)]. Neopolycystus sp. (stippled), Baeoanusia albifunicle (grey), Neblatticida sp. (black), and Aphanomerella sp. (white)…………………………………………………………………………….....72

List of Tables

Table 2.1: Mean ± s.e. length (range) (mm) of Paropsis atomaria adults collected from two field sites in South-east Queensland during the 2004 – 2005 season…….26

Table 2.2: Estimated mortality (proportion of eggs and larvae lost) at each immature life stage of Paropsis atomaria in the field at three sites (Li = first instar, Lii = second instar, Liii = third instar, Liv = fourth instar)……………………………….27

Table 2.3: Life table for Paropsis atomaria based on total number of each life stage collected in the field at three sites. Mortality rate between fourth instar and pupae is based on survival rate in laboratory (Li = first instar, Lii = second instar, Liii = third instar, Liv = fourth instar: x = age interval, nx = Number of individuals alive at start of age interval x. lx = Proportion of individuals surviving at the start of age interval x. dx = Number of individuals of a cohort dying during the age interval x to x+1. qx =

Finite rate of mortality during the age interval x to x+1. px = Finite rate of survival during the age interval x to x+1.)……………………………………………………27

Table 3.1: Abundance (mean ± s.e.) of predacious arthropods collected during the 2004/2005 season in two plantation forestry sites in South-East Queensland………39

Table 3.2: Whole-season sex ratios for hyperparasitoid species of Neopolycystus sp. emerged from Paropsis atomaria egg batches in South-East Queensland………….42

Table 4.1: Overall proportion of egg batches parasitised, mean ± se number of eggs parasitised within batches, and overall effective parasitism rate, by Neopolycystus sp. attacking Paropsis atomaria at three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006. Overlaid boxes indicate results of pairwise comparisons between those sites/seasons for between- and within-batch and total parasitism rates. The final row presents results of within-batch variation analyses throughout each sample period (see Figure 4.1)…………………………..60

Table 4.2: The proportion of egg batches parasitised and within batch parasitism rates of Paropsis atomaria, eggs from three eucalypt plantation sites in South-East

Queensland in 2004/2005 and 2005/2006 correlated against distance from the nearest native eucalypt forest and nearest source of water. (* designates significant correlation at 0.05 level)…………………………………………………………….69

Table 4.3: Emergence rates of primary (Neopolycystus sp.) and hyper-(Baoeanusia albifunicle, Neblatticida sp., and Aphanomerella sp.) parasitoids from Paropsis atomaria egg batches from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006………………………………………………………..73

xii

Declarations

The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

Signature

Date

xiii

Acknowledgements

I would like to thank the following people for their contributions to this thesis:

My supervisors Tony Clarke and Simon Lawson for their help and guidance throughout the project, and for their support when I decided to change from a PhD to Masters.

Helen Nahrung for invaluable assistance with fieldwork, labwork, data analysis and reviewing of chapters.

The staff at DPI&F Rebeccah Aigner, Jacinta Hodnett and Daniel Hancox for helping with fieldwork, as well as Janet McDonald who provided maps of plantations and assisted in the field.

From the Museum of Queensland Chris Burwell for wasp identifications, Brian Cantrell for tachinid identifications and Owen Seeman for spider identification and help with fieldwork.

Nikki Sims and Andrew Hulthen for assisting with endless counting of eggs and larvae in the lab as well as Mark Schutze and Alex Wilson for helping out at various stages over the years.

Amy Carmichael for bits of assistance here and there and for organising barbeques for as long as I can remember.

Tara Murray for providing helpful information on Neopolycystus identification and specimens of Neopolycystus insectifurax from New Zealand.

Last but not least, my family Annie, Josh, Bianca and Matthew, for putting up with the long hours that I have spent working on this project and supporting me in every possible way

This project was funded by an Australian Postgraduate Award Industry and Queensland Department of Primary Industries and Fisheries.

xiv

Chapter 1: Chrysomelids as pests of eucalypt plantations and their natural control

1.1 Eucalypt forestry in Queensland

The forestry and timber industry is a major component of the Australian economy, with Queensland alone having a direct annual turnover of $2.7 billion and employing more then 19000 people (Queensland Department of Primary Industries and

Fisheries 2004). Of major importance to Australian forestry is the plantation sector, which in 2004 consisted of a total area of 1.7 million hectares. An increasing proportion of the total estate comprises hardwood plantations; in 1994 the proportion of hardwood plantations was 15%, increasing to 42% in 2004 (Parsons &

Gavran 2005).

The majority of hardwood plantations are in Western Australia, Victoria and

Tasmania and as such most development research has concentrated in these more temperate regions. The introduction of large scale hardwood plantations to

Queensland is quite recent, as the plantation sector in this state has been traditionally dominated by softwoods. Since 2000 the total area planted with eucalypts has increased by approximately 80% to 35 000 hectares (Parsons & Gavran 2005). The predominant species planted are Eucalyptus cloeziana (F. Muell.) and Corymbia citriodora subsp. variegata (F. Muell.), as well as E. dunnii (Maiden) and E. argophloia (Blakely).

The recent and rapid increase in large scale planting of eucalypts has led to the emergence of new management problems, with one of the most serious being limiting the damage caused by insect pests (Ohmart 1990). Other pests that are likely

1 to emerge as significant threats to the establishment of a viable hardwood industry in

Queensland include fungal and bacterial infections (Bertus & Walker 1974) and vertebrate browsers (Bulinski & McArthur 2003). Among the insect pests, perhaps the greatest emerging threat to Queensland hardwood forestry are the paropsine chrysomelid beetles, whose adults and larvae feed on flush foliage of eucalypts.

Feeding by paropsines results in severe defoliation and can affect the growth rate, height, volume and possibly pulpwood quality of trees (Candy et al. 1992; Elek

1997; Elliott et al. 1998)

1.2 Paropsines as pests of eucalypt forestry

The tribe Paropsini belongs to the largest Australian subfamily of the family

Chrysomelidae, that of Chrysomelinae, which has more than 50 described genera and over 600 species (Lawrence & Britton 1996). Most species of paropsines feed on

Eucalyptus species and other Myrtaceae (de Little 1979; Selman 1994) and are major pests in all Australian states where eucalypts are grown commercially (de

Little 1989; Simmul & de Little 1999). A number of species have also been accidentally introduced overseas and where these introductions coincide with an established local eucalypt industry they have also become pests, eg in New Zealand

(Withers 2001) and South Africa (Tribe & Cillie 2000).

The eucalypt plantation industry in Queensland is relatively new and as such there has been, to date, only a limited amount of research carried out on its major pests. A number of paropsine species have been identified as potential pests of the Eucalyptus species currently favoured for subtropical plantations. Among these are Paropsis atomaria (Olivier), P. charybdis (Stål), Chrysophtharta1 cloelia (Stål) and C. “gold-

1 Chrysophtharta has recently been synonymised with Paropsisterna (Reid 2006) 2 shoulders” (an undescribed species) (Nahrung 2006). Paropsis atomaria is already recognised as a pest of Eucalyptus grandis (Hill ex Maiden), E. cloeziana and E. pilularis Smith in Queensland, as well as a number of other commercially grown

Eucalyptus species in NSW (Simmul & de Little 1999; Lawson & Ivory 2000),

Victoria (Collett 2001) and South Australia (Phillips 1993).

1.3 Paropsis atomaria

Until recently P. atomaria has not been considered a pest of commercial eucalypt plantations (Elliott et al. 1998), however, it is now recognised as a pest of a number of eucalypts (Nahrung 2006). In the Australian Capital Territory, adult flight activity and feeding commences shortly after the production of new eucalypt foliage in

October or November. Two generations are completed each year before the adults that emerge in autumn enter a reproductive diapause and hibernate in litter and soil beneath the host trees for the winter (Carne 1966a). Diapause commences in response to a reduction in day length and its termination is stimulated by the exhaustion of fat bodies (Carne 1966a).

Eggs are laid during spring and summer in batches of 20 - 100 (Elliott et al. 1998) and are deposited upright on young shoots, or sometimes leaf tips, forming a ringed cluster with each of the eggs projecting radially (Cumpston 1939). Larvae hatch from eggs after 10 - 14 days and, after consuming the remaining egg shell, are capable of moving large distances in search of suitable leaf material to consume.

The larval stage lasts for 3 - 4 weeks with a total of four instars. Like several other paropsine species (Simmul & de Little 1999), larvae of P. atomaria are highly gregarious, particularly in the earlier instars (Carne 1966a). Larvae defend themselves by elevating their posterior end and everting defensive glands from

3 between terminal segments. The defensive chemical secreted contains hydrogen cyanide, benzaldeyhyde and glucose (Moore 1967) and has been observed to kill attacking insects, e.g. ants, within a few minutes of contact (Carne 1966a). Towards the end of the fourth instar the larvae drop to the ground and form pupal cells several centimetres below the surface. Pupation occurs five days after cell formation and adults emerge approximately ten days later. Females are ready to oviposit three weeks after emergence (Carne 1966a)

In natural conditions, P. atomaria generally occurs in low numbers on a wide range of Eucalyptus species. However, in plantations, beetle populations occasionally outbreak, particularly on susceptible eucalypt species such as E. cloeziana. In the

Australian Capital Territory, where the majority of research on this species has been carried out, the preferred host is Eucalyptus blakelyi (Carne 1966a; Ohmart et al.

1985). In Queensland the preferred host appears to be E. cloeziana (Lawson & King

2002), but there are a number or other commercial eucalypt species on which P. atomaria is found including E. pilularis, E. grandis, E. dunnii and E. camaldulensis

(CABInternational 2005) and Corymbia citriodora subsp. variegata (Nahrung,

2006).

Currently, control of P. atomaria is effected by regular inspection of plantations, where if damage is considered significant insecticide sprays are applied. However, due to the requirement to certify forest practices for sustainability and increasing community concern over the use of pesticides, it is important that alternative methods of controlling forestry pests are developed.

1.4 Conservation Biological Control

Conservation biological control comprises “…manipulation of the environment to favour natural enemies, either by removing or mitigating adverse factors or by providing lacking requisites” (DeBach 1974). More recently, Eilenberg et al. (2001) defined it as “modification of the environment or existing practices to protect and enhance specific natural enemies or other organisms to reduce the effects of pests”.

Conservation biological control has received relatively little attention by pest managers (Ehler 1998), although its origins can be traced as far back as 900 AD to

Chinese citrus growers who promoted the activity of predaceous ants to reduce pest problems (Doutt 1964; Way & Khoo 1992). It is only in recent times, possibly due to government policies and growing public concern over the use of pesticides, that there has been an increased interest in conservation biological control (Barbosa 1998;

Pickett & Bugg 1998; Gurr et al. 2004b).

In conservation biological control, modification of the environment or existing practices can be achieved in two ways: 1. by careful use of pesticides to reduce the mortality of natural enemies; and 2. by habitat manipulation to improve natural enemy fitness and effectiveness (Gurr et al. 2004a). Although the judicious use of pesticides is considered an important component of conservation biological control, this review concentrates on habitat manipulation, as this is an area that demonstrates great potential but until recently has received very little attention.

Habitat manipulation to improve natural enemy fitness and effectiveness can be approached from two directions. The first is an applied approach that concentrates on economically significant plant crops and herbivores, with little effort applied to theoretical approaches. The second looks for a more general understanding of the response of arthropods to diversity (Andow 1991).

Most theory that addresses conservation biological control is related to the diversity- stability hypothesis (for review see McCann 2000), which predicts that the greater the biological diversity of a community, the greater the stability of that community

(ie there will fewer obvious perturbations, such as outbreaks of pest species). This idea has a long history of support in ecological literature (eg MacArthur 1955; Elton

1958) as it seems intuitively plausible and is “…the sort of thing that people like and want to believe” (Goodman 1975). One of the most important sources of evidence to support the hypothesis has been the study of arthropod response to vegetational diversity (Goodman 1975). Elton (1958), an advocate of the diversity-stability hypothesis, believed that complex communities comprising many predators and parasites prevented populations from undergoing explosive growth. This led to the idea that monocultures (such as hardwood plantations) are more vulnerable to outbreaks of arthropod pests (such as P. atomaria) than polycultures. However, the commonly observed phenomenon that agricultural monocultures are prone to pest outbreaks need not be explained by the diversity-stability hypothesis, but can be attributed to two, possibly linked, causes: (i) an abundance of resources for herbivorous pests (= the resource concentration hypothesis); or (ii) or reduced abundance of natural enemies ( = the enemies hypothesis) (Root 1973).

Both the enemies and the resource concentration hypotheses predict that vegetatively diverse systems are less prone to outbreaks of herbivorous pests, but the mechanisms and effects on generalist and specialist predators and herbivores vary between the two hypotheses. The enemies hypothesis suggests that predators and parasites are more effective in more complex environments. This greater effectiveness is predicted to result from a greater diversity of prey/host species and microhabitats within complex environments, which in turn promotes relatively stable and persistent

6 populations of generalised predators and parasites in such habitats (Root 1973). The outcome of such stable natural enemy populations is that any potential outbreak of pests is overcome by the community of natural enemies already present. In a review of 13 studies on the enemies hypothesis, Russell (1989) found nine supported it, two did not, and two showed no difference between monocultures and polycultures.

The resource concentration hypothesis predicts that herbivores, particularly those with a narrow host range, are more likely to find hosts that are concentrated.

Additionally, species that arrive in a clump of host plants and find conditions suitable will tend to remain in the area (Root 1973). This would result in a few specialised herbivores attaining greater relative densities in simple environments.

Several authors have proposed that it is more difficult for herbivores to locate annual plants than perennials, and that natural enemies have a greater effect on arthropod herbivores in perennial crop systems than in annual systems (see Andow (1991)).

This suggests that pest population density differences between monocultures and polycultures are the result of resource concentration in annual systems and natural enemies in perennial systems (Andow 1991). My target system is a perennial system

(hardwood forestry plantations), and therefore I focus on the promotion of natural enemies within plantations.

1.5 Natural enemies

1.5.1 Predators

Most classical biological control has concentrated on specialist natural enemies, particularly parasitoids, as it is often assumed that an effective biological control agent should be highly prey or host specific (Hoy 1994). This approach, while

7 appropriate for exotic introductions, may not be applicable to native natural enemies

(Symondson et al. 2002). Generalist predatory beetles, bugs and spiders are capable of exerting a significant influence on pest numbers, with spiders sometimes the dominant component of a natural enemy complex (Sunderland 1999). Ladybird and cantharid beetles, mirid bugs and spiders are all significant contributors to pest mortality in temperate hardwood plantations (Bashford 1999; Nahrung & Allen

2004). Interactions between generalist predators and specialist natural enemies are important because they are often complementary (Zhang 1992). Specialists can reduce the pest populations to levels where spiders and other generalists can prevent resurgence.

Spiders, predatory beetles and bugs have been used successfully as natural enemies in agriculture (Greenstone 1999; Naranjo 2001), but increases in abundance are often associated with increasing structural complexity, shelter, or overwintering sites

(Barbosa 1998; Sunderland & Samu 2000).

1.5.2 Parasitoids

Parasitoids are the most commonly used natural enemies in biological control programs (Gurr & Wratten 2000). They are also suitable for conservation biological control purposes as habitat manipulation has the potential to increase their effectiveness against pests (Barbosa & Benrey 1998; Coll 1998). A major justification of focusing on parasitoids as natural enemies is that there is a significant body of work demonstrating that their abundance, fecundity and longevity can be improved by the addition of supplementary food in the form of nectar and pollen from sources such as flowering plants (see review by Landis et al. 2000). There is still the risk that any supplementary nectar resources may provide benefits to

8 herbivores as well as parasitoids (Baggen et al. 1999), but careful selection of plants can reduce this possibility.

Determining the suitability of a parasitoid species as a candidate for conservation biological control is quite complex, especially if there is not much information available on the life histories of the species present in the habitat. However, if there are parasitoids known to effect significant control of the pest under non-manipulated conditions, and a limiting resource can be identified, that species should be given preference over those whose influence is weak or unknown (Jervis et al. 2004).

1.6 Conservation biological control in forestry

In agricultural monocultures, the suite of natural enemies available to control pests is drastically reduced (Letourneau 1998). This is also assumed to be the case in hardwood plantation forestry in Australia, as plantations support a less diverse community of insects than native remnant forests (Cunningham et al. 2005). Loss of natural enemies is caused by the simplification of the ecosystem as young eucalypt plantations have less floral diversity than native eucalypt forest (Wang et al. 2004), which may result in a failure to provide all the resources natural enemies require.

Although the enhancement of natural enemy activity through habitat manipulation is now an accepted, although still developing pest management strategy within the agricultural sector (Barbosa 1998; Pickett & Bugg 1998), extension of this technique to forestry is highly novel, with only one example of habitat manipulation in forestry in the international literature. Mensah and Madden (1994), with partial success, used artificial feeding stations to enhance activity of predators of the Tasmanian pest chrysomelid Chrysophtharta bimaculata. However, independent of manipulation, the potential of natural enemies to control forest pests in Australia has been

9 demonstrated in other situations. In Tasmania, natural enemies contribute to over

90% of the egg - larval mortality of pest paropsine populations in eucalypt plantations (de Little et al. 1990; Nahrung & Allen 2004). Further, ladybird beetles

(Coleoptera: Coccinellidae) have shown potential for augmentative release in high value eucalypt plantations (Bashford 1999; Baker et al. 2003), while Australian native parasitoids have been used successfully as classical biological control agents to control introduced Australian eucalypt pests overseas (Tribe & Cillie 2000; Jones

& Withers 2003; Tribe & Cillie 2004).

Although habitat manipulation literature in forestry is restricted to the one example cited above, a number of papers have demonstrated a positive relationship between habitat diversity and natural enemy abundance, or a negative correlation between habitat diversity and pest populations in plantation forests (e.g. Braganca et al. 1998;

Zanuncio et al. 1998; Jactel et al. 2002; Steinbauer et al. 2006). This, combined with the demonstrated success of other forms of biological control in forestry (eg.

Bashford 1999; Tribe & Cillie 2000; Baker et al. 2003) and the growing body of work involving habitat manipulation in agricultural systems (Andow 1991; Landis et al. 2000), demonstrates the potential for habitat management to promote natural enemies of plantation forestry pests.

1.7 Scope and structure of thesis

The focus of this research was originally on the enhancement of natural enemy mortality of P. atomaria in eucalypt plantations. This was to be achieved by manipulating the environment within and around plantations to make it more suitable for the natural enemies of P. atomaria. However, for this to be effective, highly detailed knowledge of the life history of the pest, its natural enemies and the

10 interactions between them is required. Preliminary research indicated that this knowledge was not available; little is known about the ecology of P. atomaria in

South-East Queensland and even less about its natural enemies. With this in mind the focus changed from enhancing natural enemy mortality of P. atomaria, to examining aspects of P. atomaria phenology, and identifying its most promising natural enemies and learning something of their life history. Chapter 2 presents a number of aspects of P. atomaria phenology, including voltinism estimates, life tables, sex ratios and damage estimates.

In Chapter 3 two Eucalyptus cloeziana plantations were sampled over one season to identify the natural enemies of P. atomaria in South-East Queensland. Abundance of natural enemies and their effect on P. atomaria populations is quantified with the aim of determining the natural enemies that have the most potential for conservation biological control.

After the initial assessment of natural enemies (Chapter 3), the natural enemies that appear to offer the most potential for biological control of P. atomaria in E. cloeziana plantations are the egg parasitoids, and in particular the primary egg parasitoid Neopolycystus sp.. Chapter 4 documents a number of aspects of the ecology of Neopolycystus sp., including the influence that it exerts on P. atomaria populations and its interactions with hyperparasitoids.

The final chapter (Chapter 5) discusses the general findings from each of the experimental chapters and the implications they have on the conservation biological control of P. atomaria in eucalypt plantations.

Chapter 2: Phenology of Paropsis atomaria Olivier (Coleoptera: Chrysomelidae) in subtropical eucalypt plantations

2.1 Introduction

The large scale planting of eucalypts in tropical and sub-tropical Australia is quite recent, with hardwood plantations in Queensland now covering almost 35000 ha

(Parsons & Gavran 2005), but with less than 1400 ha planted in early 1990s (Wylie

& Peters 1993). The predominant species planted are Eucalyptus cloeziana (F.

Muell.) and Corymbia citriodora subsp. variegata (F. Muell.), with significant areas also of E. dunnii (Maiden) and E. argophloia (Blakely). Because the eucalypt plantation industry is relatively new in Queensland, management problems are still emerging: one of the most serious problems to be faced by plantation managers is to reduce the damage caused by insect pests (Ohmart 1990).

One important group of eucalypt pests is the paropsine leaf beetles, a speciose group of native insects in which adults and larvae feed on the foliage of their Eucalyptus host plants. Because they target actively growing, current season’s foliage (Howlett et al. 2001), annual production losses of 50 - 100% may be experienced (Candy et al.

1992). Paropsis atomaria (Olivier) (Coleoptera: Chrysomelidae) is one species that poses a significant threat to eucalypt plantation productivity in northern New South

Wales and Queensland (Stone 1993; Lawson & King 2002; Nahrung 2006).

Paropsis atomaria feeds on a wide range of Eucalyptus species (CABInternational

2005). In the Australian Capital Territory (ACT), where most research on this pest has been carried out, its preferred host is E. blakelyi (Carne 1966a; Ohmart et al.

1985), however, in Queensland, the preferred host appears to be E. cloeziana

(Lawson & King 2002). As there is very little published work on this insect from

Queensland, I discuss the biology of P. atomaria below in relation to studies carried out in the ACT. Schutze et al. (2006) have demonstrated that P. atomaria populations from the ACT and Queensland demonstrate only minor genetic differences and so comparing life history characteristics between southern and northern populations is unlikely to be confounded by the presences of undetected cryptic species or high levels of local adaptation.

In the ACT, adult flight activity and feeding commences shortly after the production of new eucalypt foliage in October or November. Two generations are completed each year, before the adults that emerge in autumn enter reproductive diapause and hibernate in litter and soil beneath the host trees for the winter (Carne 1966a). Eggs are laid during spring and summer in batches of 20 - 100 (Elliott et al. 1998) on young shoots; females able to mature over 600 eggs in their lifetime. Larvae hatch from eggs after 10 - 14 days and, after consuming the remaining egg shell, can move large distances in search of suitable leaf material to consume. The gregarious larval stage lasts for 3 - 4 weeks and there is a total of four instars. Towards the end of the fourth instar the larvae drop to the ground and form pupal cells several centimetres below the soil surface. Pupation occurs five days after formation of the cell and adults emerge approximately ten days later. Females are ready to oviposit three weeks after emergence (Carne 1966a).

Despite individuals from northern and southern populations being genetically similar

(Schutze et al. 2006), the warmer and potentially longer breeding seasons in the subtropical South-East Queensland could lead to significantly different behavioural and phenological characteristics between individuals and populations from the two regions. Identifying phenological differences between the ACT and South-East

Queensland is important as knowledge of the phenology of insect pests may allow temporal prediction of attack, which in turn assists control (Nahrung & Allen 2004).

For example, the biological insecticide, Bacillus thuringiensis var tenebrionis. is most effective against early instar paropsine larvae (Elek et al. 1998; Beveridge &

Elek 2001), therefore for it to be applied effectively a knowledge of the temporal appearance and duration of larvae in the field is required (Clarke et al. 1997; Clarke

1998).

Presented here are the results of field monitoring at three sites over two seasons for

P. atomaria life stages and associated tree damage estimates. The aims of this study were to understand P. atomaria population dynamics in South-East Queensland, including the duration of life stages in the field, natural mortality, the number of generations each year, adult sex ratios and how these aspects of population dynamics relate to damage estimates of host plant foliage

2.2 Materials and Methods

2.2.1 Site details

Three South-East Queensland Eucalyptus cloeziana plantations were sampled for

Paropsis atomaria lifestages. All sites were planted on past grazing land with trees spaced at 5 x 2 metre intervals. Site details are as follows:

Site I (via Gympie) 26°04′30.72″S 152°44′8.88″E approx. 38 ha planted in May

2002. Altitude range 83.8 - 249 m. Total rainfall during the sample period (October

2004 – April 2005) was 686.5 mm. Mean average daily temperature was 22.5°C, mean maximum 28.8°C and mean minimum 13.8°C;

Site II (via Glastonbury) 26°11′20.4″S 152°29′40.2″E approx. 22 ha planted in

March 2002. Altitude range 67.7 - 162 m. Total rainfall during the sample period

(October 2004 – May 2005) was 620 mm. Mean average daily temperature was

22.7°C, mean maximum 29.5°C and mean minimum 13.8°C;

Site III (via Gympie) 26°05′97.2″S 152°43′7.54″E approx. 17 ha planted in March

2004. Altitude range 65.8 - 183 m. Rainfall during the 2005/2006 season (October to April) was 677.1 mm. Mean average daily temperature was 23.0°C, mean maximum 29.8°C and mean minimum 14.3°C.

2.2.2 Phenology

Sites I and II were monitored (every two weeks) during the 2004-2005 season

(September to April) and Site III (every two weeks) during the 2005-2006 season

(October to April). Eight different locations across each plantation were selected on each census date. The terminal 30cm of three branches from six trees within each location were visually searched for P. atomaria lifestages. The number of egg batches, larvae of each instar and adults on each of these 144 branches was counted and used to determine the average number of each lifestage present per branch throughout the field season. Counts of first and second instars were combined because of difficulty in differentiating them in the field.

2.2.3 Voltinism estimates

Sites I and II were quantitatively sampled for adult beetles every two weeks in the

2004-2005 season by collecting beetles for 15 minutes from each of eight areas throughout each plantation. Adult P. atomaria were collected from foliage using a beating tray and by visually searching branches. Beetles were taken to the laboratory

15 and classified into two developmental classes by compressing the beetle antero- posteriorly: elytral deformation indicated that the beetle was newly-emerged (teneral) and sexually immature (see Nahrung & Allen 2004). It is possible to use elytral deformation for approximately two weeks to classify beetles as teneral (H. Nahrung pers. comm.). The proportion of hard and soft beetles was used to estimate the number of generations by using peaks in the proportion of teneral beetles to signify emergence of new-generation adults, from which voltinism was inferred. This model assumes each new, mature adult cohort within a season is the predominant contributor to the next emergent cohort. As new adults constituted >80% of the total adult population sampled when such emergence peaks occurred (see results), this assumption is probably valid.

2.2.4 Adult sex ratios & size

The length of adults collected as described above was measured (tip of clypeus to posterior elytra) using a digital vernier calliper (±0.1 mm): body length is a very good predictor of overall body size in this species (M. Schutze pers com.). Sex was determined based on differences in the fore basitarsi, the male possesses uniform ventral discs of setae, while the female does not (CABInternational 2005). The sex ratios of sexually-mature and teneral beetles were determined and analysed for deviation from unity using a Chi-square contingency table (Bonferroni adjusted, P =

0.004). Only samples from which > 10 individuals were collected were included in analyses. Beetle size was compared between sites and sexes using t-tests. Sex ratio of beetles was only determined in the first field season (ie Sites I & II).

2.2.5 Mortality in the field

The population differential between egg and final instar larvae was calculated to estimate overall field mortality rates of immature stages. Mortality between each developmental stage was estimated using the total of each life-stage censused during each season at each site. Proportional mortality was calculated using the difference between each successional stage and the number of the previous stage. An average egg batch size of 76 eggs per batch was used to estimate subsequent mortality rates.

To construct a life table for P. atomaria, the mortality rate between fourth instar larvae and pupae was estimated using the survival rate of larvae reared in the laboratory (Chapter 3.).

2.2.6 Damage estimates & host plant foliage

Each tree sampled during the life-stage census was also scored for foliage damage using an estimate of damage incidence and severity based on the Crown Damage

Index (CDI- estimates damage using combination of proportion of damage to individual leaves and proportion of leaves damaged) (Stone et al. 2003). At the same time, the proportion of the crown of each tree in current season’s growth was visually estimated. Damage measurements and new host plant foliage estimates were used to quantify the damage caused by P. atomaria lifestages, and to relate damage peaks and foliage availability to the beetles’ population phenology.

2.3 Results

2.3.1 General site comparisons

Overall, the lowest Paropsis atomaria population was at Site II, while Site I and III had similar levels of P. atomaria infestation. Site III had a late season peak in number of egg batches and early instar larvae that was at least double that of the other plantations (Figures 2.2 and 2.3). Site II was not used in the second season of sampling due to the very low abundance of P. atomaria life stages. Paropsis atomaria was the predominate paropsine species at all sites but there were, however, a number of other paropsine species present including; Paropsis charybdis (Stål),

Paropsis variolosa (Marshall), Chrysophtharta cloelia (Stål) and Chrysophtharta

“goldshoulders” (an undescribed species).

2.3.2 Phenology

The number of egg batches collected varied throughout the field season and between sites (Figure 2.1), with Site II having less than a quarter of the number of eggs than at

Sites I and III. Site I demonstrated four peaks in egg numbers, while Sites II and III had two peaks. All sites showed egg peaks in mid-January and early March.

Generally, there were three peaks in larval populations at each site (Figure 2.2), with the lowest overall larval population at Site II. Peak branch occupancy by all P. atomaria life-stages coincided with peak early instar larvae (first and second instars) at all three sites (Figure 2.3). However, the proportion of branches occupied by these instars at Site III was almost double that at Site I and over three-fold that at Site II.

0.35

0.3

0.25

0.2

0.15

0.1 mean +- se egg batches +- mean 0.05

0 13-Oct 13-Apr 27-Apr 16-Mar 30-Mar 29-Sep 4-6 Jan 4-6 Mar 3-6 1-2 Feb 1-2 7-9 Dec7-9 9-11 Nov 9-11 25-28 Oct 25-28 17-19 Jan 17-19 15-17 Feb 15-17 21-23 Dec 21-23 23-25 Nov 23-25

4 A 3.5 3 2.5 2 1.5 1 mean +- se larvae +- mean 0.5 0 9/12/04 6/01/04 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 25/11/04 23/12/04 20/01/05 17/02/05 16/03/05 30/03/05 13/04/05 27/04/05

1.4 B 1.2 1 0.8 0.6 0.4

mean +- selarvae 0.2 0 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 25/11/04 23/12/04 20/01/04 17/02/05 16/03/05 30/03/05 13/04/05

14 C 12 10 8 6 4 mean +- se larvae 2 0 9/11/05 7/12/05 4/01/06 1/02/06 6/03/06 25/10/05 23/11/05 21/12/05 19/01/06 15/02/06 16/03/06 30/03/06 14/04/06 27/04/06

A 0.3

0.25 0.2

0.15

0.1 0.05

0 proportion of branches occupied 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 23/12/04 20/01/05 17/02/05 16/03/05 30/03/05 13/04/05 27/04/05

0.2 B

0.15

0.1

0.05

0 proportion of branchesoccupied 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 23/12/04 20/01/05 17/02/05 16/03/05 30/03/05 13/04/05

0.6 C 0.5

0.4 0.3 0.2

0.1 0 proportion of branchesoccupied 9/11/05 7/12/05 4/01/05 1/02/06 6/03/06 25/10/05 23/11/05 21/12/05 19/01/06 15/02/06 16/03/06 30/03/06 14/04/06 27/04/06

Figure 2.3: Proportion of branches (n=144 per sampling occassion) occupied throughout the season by any developmental stage of Paropsis atomaria at three sites in South-east Queensland [Site I (A), Site II (B) and Site III (C)]. Note difference in scale of axes.

2.3.3 Voltinism estimates

Consistent mid-season peaks in the proportion of new adults occurred at Sites I and

II on 23 December and 17 February (Figure 2.4), but Site I had two additional peaks

– one early season (11 November) and a smaller one late in the season (13 April).

Using these data to infer voltinism, Site I had 3-4 generations, while Site II had two generations following the emergence of overwintered adults in spring.

2.3.4 Adult sex ratios & size

Over the whole field season, sex ratios did not differ from unity at either site (Site I:

2 761 males, 722 females, χ 1=0.10, P = 0.31; Site II: 452 males, 464 females,

2 χ 1=0.16, P = 0.69), but there were some deviations in operational sex ratios (Figure

2.5). Operational sex ratios consider only reproductively competent individuals: teneral beetles are incapable of mating (Nahrung & Allen 2004). Notably, the operational sex ratio was significantly male-biased at the start of the field season at both sites, suggesting that males may emerge from overwintering earlier than females (protandry) as occurs in C. agricola (Nahrung & Reid 2002). There was one occasion when the operational sex ratio was female biased (Site I, 2/2/05). The sex ratio also tended towards male bias towards the end of the field season. Sex ratios of new (teneral) adults were normal (1:1) throughout the field season at both sites, except for one instance of male-bias late in the season at Site I (Figure 2.6).

A 1 38 118 0.8 101 0.6 124 158 148 0.4 57 206 0.2 90 90 112

proportion new adults new proportion 36 19 26 59 176 0

29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 09/12/04 23/12/04 06/01/05 20/01/05 02/02/05 17/02/05 03/03/05 16/03/05 30/03/05 13/04/05 27/04/05

B 28 1.0

102 0.8

0.6 132 128

0.4

5 71 53

proportion new adults new proportion 0.2 22 34 12 7 4 116 111 77 14 0.0 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 23/12/04 20/01/05 17/02/05 16/03/05 30/03/05 13/04/05 27/04/05

1 A

0.8 8 67 26 96 0.6 22 71 105

0.4 55 68 151 38 151

proportion female proportion 35 0.2 86 18 49 0 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 23/12/04 20/01/05 17/02/05 16/03/05 30/03/05 13/04/05 27/04/05

1.0 B

0.8 20 64 0.6 55 109 4 4 39 66 112 77 0.4 12 7

proportion female 0.2 14 34 22 1 0.0 29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 09/12/04 23/12/04 06/01/05 20/01/05 02/02/05 17/02/05 03/03/05 16/03/05 30/03/05 13/04/05 27/04/05

1 4 A 1

0.8 19 92 0.6 4 25 0.4 56 52 26 52 19 30 23 55

proportion female 0.2 53 1 0 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 23/12/04 20/01/05 17/02/05 16/03/05 30/03/05 13/04/05 27/04/05

1 1.0 2 B

0.8

0.6 68 62 82

0.4 14 27 proportion female 16 0.2 4 0 0 0 0 0 0 0 0.0 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 24/11/04 23/12/04 20/01/05 17/02/05 16/03/05 30/03/05 13/04/05 27/04/05

Female beetles were significantly larger than males at both sites (Table 2.1) (Site I: t- test, t1655= 49.6, P < 0.001; Site II: t967= 36.8, P < 0.001), while within sexes, Site II individuals were significantly larger than beetles collected from Site I (females: t1241=3.72, P < 0.001; males: t1381= 7.5, P < 0.001).

Table 2.1: Mean length ± s.e. (range) (mm) of Paropsis atomaria adults collected from two field sites in South-east Queensland during the 2004 – 2005 season.

Females Males

Site I 11.34 ± 0.02 (8.84 – 13.61) 9.82 ± 0.02 (7.4 – 11.85)

Site II 11.48 ± 0.03 (9.0 – 13.12) 10.07 ± 0.03 (8.25 – 11.98)

2.3.5 Mortality in the field

Less than 8% of eggs survived to become fourth instar larvae at all sites (Table 2.2).

The highest mortality occurred between egg and early instar larvae (first and second instars) at Site I, and between early instar larvae and third instar larvae at Sites II and

III. The mortality rate across all plantations from eggs to pupae was very high with only approximately 4% of eggs surviving to the pupal stage (Table 2.3). These data do not reflect loss from larval parasitoids or during the pre-pupal and pupal stage.

Life Stage Site I Site II Site III Average ± se Egg to Li+Lii 0.80 0.64 0.62 0.69 ± 0.07 Li+Lii to Liii 0.59 0.73 0.86 0.73 ± 0.10 Liii to Liv 0.37 0.00 0.33 0.23 ± 0.14 all larvae (Li 0.74 0.68 0.91 0.77 ± 0.08 to Liv) Egg - Liv 0.95 0.92 0.96 0.94 ± 0.02

Table 2.3: Life table for Paropsis atomaria based on total number of each life stage collected in the field at three sites. Mortality rate between fourth instar and pupae is based on survival rate in laboratory. Life table does not include mortality from larval and pupal parasitism. (Li = first instar, Lii = second instar, Liii = third instar, Liv = fourth instar: x = age interval, nx = Number of individuals alive at start of age interval x. lx = Proportion of individuals surviving at the start of age interval x. dx = Number of individuals of a cohort dying during the age interval x to x+1. qx = Finite rate of mortality during the age interval x to x+1. px = Finite rate of survival during the age interval x to x+1.)

Age interval (x) nx lx dx qx px eggs (1) 16796 1.000 11457 0.682 0.318 Li+Lii (2) 5339 0.318 4167 0.780 0.220 Liii (3) 1172 0.070 327 0.279 0.721 Liv (4) 845 0.050 218 0.258 0.742 Pupae (5) 627 0.037

2.3.6 Damage estimates & host plant foliage

At Site I and Site II, peak defoliation scores (Figure 2.7a,b) coincided with peaks in first and second instar larvae (Figure 2.2), although peak defoliation occurred about a month earlier at Site II than Site I. At Site III (Figure 2.7c), peak defoliation was recorded in the sample following the peak in first and second instar larvae. Across the 2004-2005 season, defoliation scores were around twice as high at Site I than at

Site II, but were similar between Site I and III.

30 A 100 currentproportion season's

25 80

20 growth 60 15 40 10 damage score damage 5 20 0 0 9/12/04 6/01/05 2/02/05 3/03/05 11/11/04 25/11/04 23/12/04 20/01/04 17/02/05 16/03/05 30/03/05 13/04/05 27/04/05

proportion current season's 10 B 100

8 80

6 60 growth

4 40

damage score damage 2 20

0 0 9/12/04 6/01/05 2/02/05 3/03/05 11/11/04 25/11/04 23/12/04 20/01/04 17/02/05 16/03/05 30/03/05 13/04/05

C proportion current season's 35 80 30 25 60 growth 20 40 15 10 damage score damage 20 5 0 0 9/11/05 7/12/05 4/01/05 1/02/06 6/03/06 25/10/05 23/11/05 21/12/05 19/01/06 15/02/06 16/03/06 30/03/06 14/04/06 27/04/06

2.4 Discussion

2.4.1 Population Phenology

In the ACT, P. atomaria adults are active from late October until late March (Carne

1966a), however, in South-East Queensland, the active season was longer with activity recorded from September until April. While P. atomaria is bivoltine in

South-East Australia (Carne 1966a), the longer season and higher mean temperatures in South-East Queensland means that more than two generations can occur in a season.

At the sites surveyed in this study there were two mid-season peaks in the proportion of new adults at both sites, indicating two generations. However, at Site I, where there was a higher proportion of flush foliage early in the season (Figure 2.7a), there were additional early and late season peaks (Figure 2.4a). The high abundance of flush foliage at Site I would have provided conditions that are preferred by paropsines for oviposition (Steinbauer et al. 1998) and neonate larval establishment

(Larsson & Ohmart 1988; Nahrung et al. 2001), far earlier in the season than at Site

II. Thus at Site I there was an early onset of oviposition corresponding to the earlier abundance of new growth, resulting in time for more generations of P. atomaria to be completed in the season. The onset of egg laying by other paropsine species, e.g.

Chrysophtharta agricola (Chapuis) and C. bimaculata (Olivier), also corresponds to the date when new growth becomes abundant (Paterson et al. 2004), and the availability of flush foliage may drive the population cycles of paropsine chrysomelids (Ohmart 1991).

2.4.2 Sex ratio and Adult size

Over the whole season sex ratios did not vary from unity, but when taking into account only reproductively competent individuals there was a higher proportion of males early in the season. This protandry, which has also been shown to occur in C. agricola (Nahrung & Reid 2002), means that there are receptive males present in the field ready to mate as soon as females emerge from overwintering.

One unexpected result of this study was the difference found in adult size between the two sites in 2004/05. Males and females were both significantly smaller at Site I and this is contrary to expected. Site I had higher proportions of flush foliage throughout the season and, as increased foliage toughness can reduce P. atomaria pupal weight (Larsson & Ohmart 1988), if there was to be a difference in size between sites it would be expected that the plantation with the highest abundance of soft flushing foliage should produce larger beetles. This was not the case in this study and warrants further investigation. Population density is another factor that may need to be considered as a possible influence on the size of beetles. The population density at Site I was greater then at Site II, but food resources did not appear to be limiting. However, due to the scope of this study it is difficult to draw any conclusion on the effect of population density on beetle size.

2.4.3 Population mortality

Field mortality of P. atomaria was greatest during the egg and first instar stages, as is also found in C. bimaculata (de Little et al. 1990) and C. agricola (Nahrung et al.

2001). A large percentage of this field mortality is attributable to egg parasitism (see

Chapter 4), but mortality is still high in the absence of parasitoids due to difficulties in feeding establishment of neonate larvae (Nahrung et al. 2001). The total field

31 mortality of 94% is comparable to that recorded for C. bimaculata (de Little et al.

1990) and C. agricola (Nahrung & Allen 2004).

2.4.4 Management implications

The highest number of branches occupied by P. atomaria life stages always coincided with the peak abundance of early instar larvae, reflecting the high mortality rate of these life stages (Figures 2.3 & 2.7). Peak abundance of early instar larvae in turn coincided with peak defoliation scores at Sites I and II (Figure 2.7), indicating that early instar larvae are responsible for a high percentage of defoliation.

This is at odds with accepted wisdom for other paropsine pests, where most damage is thought to be caused by the third and fourth instars (Carne 1966b; Elliott et al.

1992). The timing of damage has important implications for the timing of control measures as they should be applied before the appearance of the most damaging stages. To prevent peak defoliation, control at the egg or pre-reproductive adult stage would be ideal. Peak defoliation at Site III occurred in the sample following the peak in early instar larvae, which was at least four-fold of that at the other sites. The very high abundance of first and second instars may have resulted in a larger than usual proportion surviving to later instars. This would result in higher levels of foliage consumption as later instars consume greater amounts of foliage per day than early instars (Carne 1966b). At Site II peak defoliation occurred approximately one month before Site I, which would be due to there being less new foliage to begin with.

Even though P. atomaria experiences field mortality of up to 94%, the population levels in plantations and high adult fecundity means that even the small proportion of individuals that survive are capable of causing significant damage in plantations. The

32 longer active season combined with the ability of P. atomaria to have up to four generations in a season, when conditions are favourable, demonstrates that P. atomaria and possibly other paropsine beetles have the potential to become more significant pests in sub-tropical Queensland than in temperate areas.

Chapter 3: Identifying the natural enemies of Paropsis atomaria Olivier (Coleoptera: Chrysomelidae) in South-East Queensland plantations

3.1 Introduction

Current management of paropsine beetles in South-East Queensland involves regular inspection of plantations, particularly in late December and late March, when major defoliation events occur (Lawson & King 2002). If defoliation is severe enough

(>50% defoliation) chemical sprays (dimethoate) are applied. Reliance on chemicals to control pests has a number of negative consequences: many pesticides are toxic to non-target beneficial and endangered species, wildlife, and humans. Additionally, improper use of pesticides can result in the development of resistance in pests and lead to resurgence of target pests and outbreaks of secondary pests (DeBach 1974).

Further, the requirement to certify forest practices for sustainability (see Govender

2002), and community concerns caused by spraying, mean that achieving non- chemical control of plantation pests is highly desirable.

The potential of natural enemies to control forest pests in Australia has been demonstrated. In Tasmania, natural enemies contribute to over 90% egg - larval mortality of pest paropsine populations in eucalypt plantations (de Little et al. 1990;

Nahrung & Allen 2004). Further, coccinelid beetles have shown potential for augmentative release in high value Tasmanian eucalypt plantations (Bashford 1999;

Baker et al. 2003). Overseas, Australian native parasitoids have been used successfully as classical biological control agents to control introduced Australian plantation pests (Tribe 2000; Tribe & Cillie 2000; Withers 2001).

Here I sampled two Eucalyptus cloeziana plantations throughout one season to identify the natural enemies of P. atomaria in South-East Queensland. Abundance of natural enemies and their effect on P. atomaria populations is quantified with the aim of determining the natural enemies that have the most potential for conservation biological control.

3.2 Materials and Methods

3.2.1 Predators

During the same surveys of Site I and Site II as detailed in Chapter 2 (sections 2.2.1

& 2.2.2), three branches from each of six trees within each of the eight sections per site per sampling occasion were visually searched for potential P. atomaria predators. Predaceous insects and spiders were collected into plastic vials and taken to the laboratory for identification to family level. Unknown spiders were identified to family by Dr Owen Seeman (Queensland Museum).

3.2.2 Parasitoids

Egg parasitoids: On each sampling occasion above, all egg batches observed in each section were collected into separate plastic vials and returned to the laboratory in a cooled container. The number of eggs per batch was counted and then the egg batches were kept in individual cotton-wool-capped vials in a controlled-temperature cabinet at 24°C until larval or parasitoid emergence.

Emergent individuals (P. atomaria larvae, primary egg parasitoids and hyperparasitoids) were counted and representative specimens of each were identified to genus or species by Dr Chris Burwell (Queensland Museum). Average egg batch outcome frequencies (larvae emerged, wasps emerged, eggs failed to hatch) were

35 determined for each month. Wasp sex ratios were determined for species where sexual dimorphisms made sex determination possible without dissection.

Larval parasitoids: All larval batches observed on sampled foliage in the surveyed sections described above were collected and reared separately on field- collected E. cloeziana foliage in plastic containers (175mm x 120mm x 60mm

LxWxH). Later in the season, when larvae were more abundant (see Chapter 2), an additional 100 individuals of each instar from several larval batches throughout the plantation were collected into separate plastic bags for each instar and taken to the laboratory in a cooled container. Larvae were reared in plastic containers

(50/container for early instars; 20/container for late instars) on fresh E. cloeziana foliage that was changed every 3-4 days. Resultant P. atomaria larval parasitoids were counted, and parasitoids were identified to genus by Dr Brian Cantrell

(Queensland Museum). Average larval outcome frequencies (proportion that survived to pupation, were parasitised, or died) were determined for each sample date, and mortality was calculated for each instar, and compared to stage-specific field mortalities recorded from the field (Chapter 2). Within- and between-larval batch parasitism rates were calculated, and between-batch rates were arcsine-square root transformed and analysed using a 2-way ANOVA to compare sites and larval instars.

Larval hyperparasitoids: Hyperparasitoids that emerged from tachinid pupae above were identified to genus by Dr Chris Burwell. Only pupae from which either primary or hyperparasitoids emerged were used to calculate hyperparasitism rates

(number of wasps emerged/(number of flies emerged + number of wasps emerged)).

Perilampid planidia (Hymenoptera: Perilampidae) were visible through adult beetle

36 cuticle, after infesting P. atomaria larvae and being unable to develop further without a primary tachinid larval parasite host (Tanton & Epila 1984). An estimate of perilampid infestation rate was obtained by examining adult P. atomaria, collected for the parasitic mite survey (see below), under a x40 dissecting microscope for planidia visible externally.

3.2.3 Parasites

Parasitic mites: Adult P. atomaria were collected throughout the 2004-2005 field season (September – March) from Sites I and II using a beating tray and by visually searching host plant foliage. Beetles were collected individually into single plastic vials and taken to the laboratory where their sex was determined based on differences in the third tarsal segment of the foreleg (Baly 1862). They were further classified as “new” (teneral) or “old” (hard) beetles based on the rigidity of their elytra when compressed antero-posteriorly (see Chapter 2). All beetles (n = 2114) were assessed for the presence of sexually-transmitted, parasitic podapolipid mites

(Acari: Podapolipidae) under a dissecting microscope (x40) by raising the elytra and examining the elytral ventral surface, the dorsum, mesothorax, and spiracles for mite life stages. The proportion of beetles infected throughout the whole field season was determined and compared between sites using a Chi-square test. Mean seasonal infection rate (number of mites per host) was calculated and compared between sites with a t-test. Mites collected represented three previously undescribed species and taxonomic descriptions were subsequently prepared by Seeman & Nahrung (2005).

Nematodes: Nematodes (Mermithidae) were reported by Selman (1989;

1994) as “…the most important parasites of all in suppressing population explosions of paropsine beetles”; thus, a subsample of collected beetles was dissected (10 per

37 site per sample date, where available) to look for nematodes. Dissections (n = 160) were conducted in distilled water under a dissecting microscope (x40).

Fungi: Although not assessed in the field, larval and adult P. atomaria became infected with fungi in the laboratory and glasshouse (pers. obs.).

3.3 Results

3.3.1 Predators

Spiders were the most abundant natural enemy on surveyed branches at both sites and comprised 88% of the total number of natural enemies encountered (Table 3.1).

The spiders collected were members of the families Theridiidae, Oxyopidae,

Araneidae and Clubionidae. Abundance of other natural enemies observed preying on any life stage of P. atomaria was relatively low. There were significant positive correlations between the number of spiders and P. atomaria adults (r = 0.437 P =

0.014 n = 16) and egg batches (r = 0.521 P = 0.003 n = 16) per branch over the whole season at both plantations (Figure 3.1) (see Chapter 2 for P. atomaria data)

Table 3.1: Total number and mean (± SE) number of predacious arthropods collected during the 2004/2005 season in two plantation forestry sites in South-East Queensland.

Natural enemy Site I Site II Both

n Number / branch n Number / branch n Number / branch

Spiders 145 0.063± 0.005 113 0.052± 0.005 258 0.058± 0.004

Coccinellidae 10 0.004± 0.001 3 0.001± 0.001 13 0.003± 0.001

Pentatomidae 7 0.003± 0.001 2 0.001± 0.001 9 0.002± 0.001

Reduviidae 2 0.001± 0.001 4 0.002± 0.001 6 0.001± 0.001

Mantidae 3 0.001± 0.001 3 0.001± 0.001 6 0.001± 0.001

Chrysopidae 1 <0.001 1 <0.001 2 0.001± 0.001

All predators 168 0.081± 0.006 126 0.064± 0.005 294 0.073± 0.004

0.30

0.25

0.20

0.15

0.10

0.05 Mean +- se number per shoot

0.00

Sep Oct Nov Dec Jan Feb Mar Apr May

3.3.2 Parasitoids

Primary egg parasitoids: Neopolycystus sp. (Hymenoptera: Pteromalidae) was the only primary egg parasitoid reared from P. atomaria egg batches in large numbers, emerging from 49.3% of egg batches at Site I, and 51.7% of egg batches at

2 Site II (these rates were not statistically different (Chi-square test, χ 1= 0.45, P =

0.5)). Two egg batches yielded Pediobius sp. (Hymenoptera: Eulophidae), a primary parasitoid normally associated with Lepidoptera (C. Burwell pers com.). Monthly average egg batch outcomes are shown in Figure 3.2, and patterns of egg batch parasitism are discussed further in Chapter 4.

A 8 43 51 22 92 94 87 100%

80%

60%

40%

20% outcome frequency 0% Sep Oct Nov Dec Jan Feb Mar

11 20 8 70 88 47 B 100%

80%

60%

40%

20% outcome frequency

0% Oct Nov Dec Jan Feb Mar

Egg hyperparasitoids: Three hyperparasitic wasp species emerged from P. atomaria egg batches: Baeoanusia albifinicle Girault (Encyrtidae), Neblatticida sp.

(Encyrtidae) and Aphaneromella sp. (Platygasteridae). All are obligate hyperparasitoids (Tribe 2000). Collectively, these species represented 33% of emergent wasps from Site I (n = 3931), and 23% from Site II (n = 3665): the hyperparasitism rate was significantly higher at Site I than at Site II (Chi-square test,

2 χ 1= 92.9, P < 0.001). Two or more wasp species emerged from 38% of all

2 parasitised egg batches and this did not vary between sites (χ 1= 0.01, P = 0.92), while two or more hyperparasitoid species emerged per egg batch from 15% of egg

2 batches at Site I and 9% of batches at Site II (χ 1= 3.0, P = 0.08). Whole-season hyperparasitoid sex ratios were rarely even: Baeoanusia albifunicle showed a female bias in 75% of season’s samples; Neblatticida sp. was always male-biased, while

Aphanomerella exhibited even and female-biased sex ratios depending on the site/season (Table 3.2).

Table 3.2: Whole-season sex ratios for hyperparasitoid species of Neopolycystus sp. emerged from Paropsis atomaria egg batches in South-East Queensland.

Female Male Goodness of fit wasps wasps Site I 04/05 2 Baeoanusia 529 364 χ 1= 30.9, P < 0.001 albifunicle 2 Neblatticida sp. 2 17 χ 1= 9.9, P = 0.002 2 Aphanomerella sp. 234 140 χ 1= 23.6, P < 0.001 Site II 04/05 2 Baeoanusia 283 459 χ 1= 41.7, P < 0.001 albifunicle 2 Neblatticida sp. 4 28 χ 1= 18.0, P < 0.001 2 Aphanomerella sp. 42 9 χ 1= 21.4, P < 0.001

Primary larval parasitoids: Three tachinid fly species (Anagonia sp.,

Paropsivora sp., and Palexorixta sp.) were reared from P. atomaria larvae. The latter species was represented by only one specimen, and has not been previously associated with paropsine beetles (B. Cantrell, pers. comm.), while Anagonia

(=Froggattimyia) and Paropsivora are associated with P. atomaria in other regions in Australia (Tanton & Epila 1984). Parasitoid emergence rates were low: only 53.5

% of all tachinid pupae formed (n = 114) bore flies or wasps in the laboratory; the remainder failed to eclose. Parasitoid pupal numbers, not the number of flies emerged, were used to calculate primary parasitism levels. Only one parasitoid emerged per primary host. No primary parasitic wasps were reared from P. atomaria larvae, although they are commonly associated with P. atomaria in other regions

(Tanton & Epila 1984; CABInternational 2005).

Average larval parasitism rates were very low throughout the season (Figure 3.3), peaking at 11% at Site I and 3% at Site II. Overall parasitism rates did not differ significantly between Site I and Site II, with only 1.4% (n = 5464) and 1% (n =

2 3644) of all larvae collected parasitised, respectively (Chi-square test, χ 1= 3.4, P =

0.06). No larvae collected as first instars were parasitised (Figure 3.4), and parasitism rates generally increased with exposure time (larval age). Despite low rates of apparent parasitism, pupation of larvae returned to laboratory was still low, at only 36%. This may reflect undetected parasitoid induced mortality or inappropriate rearing techniques.

A 119 113 1539 716 1032 775 1167 100%

80%

60%

40%

20% outcome frequencyoutcome

0% Sep Oct Nov Dec Jan Feb Mar

B 813 584 823 572 445 100%

80%

60%

40%

20% outcome frequency 0% Nov Dec Jan Feb Mar

0.035

0.03

0.025

0.02

0.015 1552 936

0.01 1485 901 proportion parasitised 0.005 1824 1115

0 621 692 Li Lii Liii Liv

For larvae collected and subsequently reared as discrete batches from the field, within-batch parasitism rates were significantly negatively correlated with batch size at Site I (r = -0.66, P < 0.001), but not significantly so at Site II (r = -0.63, P = 0.06).

Within-batch parasitism rates ranged from 2 – 36% and did not vary according to instar or site (2-way ANOVA, instar: F2,28 = 1.11, P = 0.34; site: F,1,28 = 0.01, P =

0.91) (Figure 3.5), the interaction between instar and site was not examined.

0.3

0.25

0.2

0.15

0.1

parasitised/batch 0.05 mean + se proportion proportion se + mean 0 Lii Liii Liv

Figure 3.5: Mean + s.e. proportion of Paropsis atomaria larval instars parasitised per field-collected larval batch from two plantation forestry sites in South-East Queensland [Site I (white) and Site II (grey)] in 2004 - 2005.

Larval hyperparasitoids: Perilampid wasps (Perilampus sp.) represented 28

% of emergents from tachinid pupae at Site I and 7.4 % at Site II (Chi-square test,

2 χ 1= 4.5, P = 0.03). Although superparasitism by perilampid planidia was common, only one wasp emerged per primary host pupa. Perilampid planidia had a higher

2 prevalence in adult P. atomaria at Site I (53%) than at Site II (30%) (χ 1= 116.9, P <

0.001). The average number of planidia per infested beetle was also higher at Site I

(2.8 ± 0.1 (range 1 – 34)) than at Site II (1.9 ± 0.1 (range 1 – 11)) (t-test, t848 = 7.32,

P < 0.001). These figures probably represent an underestimate of actual perilampid planidial infestation rate since only those visible through the cuticle were counted; those lodged internally were not located.

3.3.3 Parasites

Parasitic mites: Three sexually-transmitted mite species were present on beetles from both sites: Parobia alipilus, Pb. gimlii, and Pb. lawsoni (Seeman &

Nahrung 2005), but were not considered separately in analyses. Only old (hard) beetles were infected, providing further evidence that teneral beetles are not capable of mating as these mites are sexually transmitted (Nahrung & Allen 2003).

Throughout the field season, 45.8% of old beetles were infected with podapolipid mites at Site I (n = 837), and 28.5% at Site II (n = 548): this difference was

2 statistically significant (χ 1= 41.7, P < 0.001). Further, average infection rate across the season was also significantly higher at Site I than Site II (t-test, t433= 5.56, P <

0.001), with mean ± se number of mites per beetle being 33.7 ± 1.6 (1 – 160) and

21.9 ± 1.6 (1 – 111), respectively.

Nematodes: There were no nematodes found in any beetles from any sample date at either site (Site I n = 101, Site II n = 59).

Fungi: an unidentified species of fungus was found infecting adult beetles under humid conditions in glasshouse and laboratory colonies.

3.4 Discussion

3.4.1 Predation

In many agricultural systems the strategy of enhancing background mortality of pests through assemblages of generalists is integrated into crop management regimes, and control of outbreak species is complemented by specific, tactical solutions such as inundative releases of biological control agents (Helenius 1998). If natural enemies can be identified that are amenable to manipulation then preference should be given

47 to species already known to effect significant control in unmanipulated conditions

(Jervis et al. 2004).

In this study it was found that the abundance of all predators, except spiders, was low and as such they are probably not capable of controlling outbreaks of P. atomaria, but still provide a desirable level of background mortality. Although predators are thought responsible for a high proportion of mortality of paropsines in the field (de

Little et al. 1990; Allen & Patel 2000), the possibility of enhancing the abundance of the generalist predators is limited in hardwood plantations (see for example Mensah

& Madden 1994; Baker et al. 2003). Much of the manipulation reported in the literature involves the provision of shelter, overwintering sites (usually in annual crops) and a general increase in vegetational structural complexity (Barbosa 1998;

Pickett & Bugg 1998; Gurr et al. 2004b). As hardwood plantations already provide most of these requirements, it would be difficult to manipulate the habitat to make it more suitable for generalists such as spiders.

This study showed a correlation between spider abundance and P. atomaria eggs and adults, suggesting that spiders respond to P. atomaria abundance in some way. But as spiders are univoltine (Helenius 1998) they are not likely to be able to increase in numbers in response to an outbreak of P. atomaria sufficient to effect control.

However, spiders and other generalist predators are still likely to have a part to play in the control of P. atomaria. Spiders are recognised as important predators of paropsine beetles in temperate eucalypt forestry (Allen & Patel 2000; Nahrung &

Allen 2004) and may indirectly interact with more specialist parasitoids. It has been demonstrated in some perennial systems that specialist natural enemies may reduce

48 the level of pests in one year, to a level at which generalists are able to prevent resurgence in later years (Mason & Torgersen 1987; Roland & Embree 1995).

3.4.2 Larval parasitism

In contrast to paropsine species in other regions of Australia (Tanton & Epila 1984;

Tribe 2000; Nahrung 2002), larval parasitism was not common. In the Australian

Capital Territory, Tanton & Epila (1984) reported overall larval parasitism rates of P. atomaria by tachinids at 19.2% and by Eadya paropsidis (Huddleston and Short)

(Hymenoptera: Braconidae) at 49.4%. In this study, E. paropsidis was not encountered at all and the 1% parasitism rate from the three species that were present was much lower than that reported by Tanton & Epila (1984). The observed parasitism rate could be an underestimation due to the high mortality of laboratory- reared larvae. Undetected parasitoid induced mortality and inappropriate rearing techniques could have reduced the observed parasitism rate. However the reason for the lower parasitism rates could also be related to the biology of the parasitoids.

Changes in an insect’s parasitoid fauna with geographic locality is common in many systems, particularly where the parasitoids are not one-on-one specialists (Hawkins

1993; Kruess 2003).

At both of the study sites (although only significant at one) there was a negative correlation between larval batch size and parasitism rate, which could be related to collective defence mechanisms of larval batches (Sillen-Tullberg & Hunter 1996).

Further investigation is warranted into this pattern, as an indirect effect of increasing egg parasitism rates could be increased mortality of larvae through a reduction in larval batch size (see Chapter 4).

3.4.3 Egg parasitism

In comparison to larval parasitism rates, the egg parasitism rates were substantially higher with parasitoids emerging from 13% of all eggs. This percentage is significantly increased when additional indirect effects of egg parasitism are taken into account (Chapter 4). Parasitism rates across the season were relatively constant, with the undescribed primary parasitoid Neopolycystus sp. and its obligate hyperparasitoid Baeoanusia albifunicle being the most common.

It has been previously demonstrated that that various aspects of parasitoid effectiveness can be increased by the addition of supplementary food resources (eg:

Baggen & Gurr 1998; Berndt et al. 2002; Berndt & Wratten 2005). This being the case, Neopolycystus sp. has great potential for use in conservation biological control as it already has a significant effect on P. atomaria populations. Of concern is the presence of the three species of hyperparasitoids (28% of all emergent wasps) as they have the potential to disrupt the effectiveness of any biological control, as has been experienced with the classical biological control of P. charybdis (Stål) (Coleoptera:

Chrysomelidae) by Enoggera nassaui (Girault) (Hymenoptera: Pteromalidae) in New

Zealand (Jones & Withers 2003). However, unlike the study in New Zealand, hyperparasitism rates did not increase over the season and only rarely were all

Neopolycystus sp. in a batch killed by hyperparasitoids (Chapter 4).

3.4.4 Parasites

Selman (1989; 1994) described nematodes as being ‘the most important of all in suppressing population explosions of paropsine beetles’, despite nematodes being rarely recorded from paropsines in Tasmania (Davies 1966; de Little 1979).

Although humid conditions may promote nematode infection (Selman 1989, 1994), none were found in the sub-tropical sites sampled here.

The infection rate of beetles by parasitic mites increased over the season and it has been demonstrated that podapolipid mite infection significantly decreased overwintering survival of Chrysophtharta cloelia (Stål), another subtropical paropsine beetle (Nahrung & Clarke 2006): a similar pattern of mite induced overwintering mortality was found in P. atomaria (H. Nahrung pers comm.).

Podapolipid mites have been considered as biological control candidates in other pest beetles (Drummond et al. 1985), but to determine if they are feasible options for biological control in hardwood forestry would require significant further study.

Similarly, the unidentified fungus species was found infecting adult beetles in humid conditions in laboratory colonies; whether it is present in the field or if it has any effect on P. atomaria populations is not known.

The natural enemies that appear to offer the most potential for biological control of

P. atomaria in E. cloeziana plantations are the egg parasitoids, and in particular the primary egg parasitoid Neopolycystus sp.. This study demonstrates that

Neopolycystus sp. already has a significant influence on P. atomaria populations in unmanipulated conditions. Whether it is amenable to habitat manipulation is yet to be tested: study of its ecology and biology is required to determine if and how its effectiveness as an egg parasitoid of P. atomaria can be increased. It is also important to remember that though Neopolycystus sp. is the most promising natural enemy for biological control, other natural enemies whose effect is weak or unknown could still be important in the overall control of P. atomaria.

Chapter 4: Ecology of Neopolycystus Girault sp. (Hymenoptera: Pteromalidae), an egg parasitoid of Paropsis atomaria Olivier (Coleoptera: Chrysomelidae) in South-East Queensland

4.1 Introduction

In South-East Queensland, the eucalyptus plantation pest species, Paropsis atomaria, becomes active in late September, when egg laying can begin depending on host quality (Chapter 2). Adults are active until mid April, completing up to four generations during this time (Chapter 2).

While current management of P. atomaria in South-East Queensland involves regular inspection of plantations, and if necessary insecticide sprays, a greater reliance on, and knowledge of, natural enemy induced mortality is required. An alternative to chemical control is the use of conservation biological control: manipulative actions that preserve or protect existing natural enemies in the environment. For conservation biological control to be successful the most suitable natural enemy must be selected and detailed knowledge of the life history of the target natural enemies is required (Gurr et al. 2004b). In the case of conservation biological control the key selection criteria are that the natural enemy must be amenable to habitat manipulation and secondarily that natural enemies should be chosen that exert significant control of pests in unmanipulated conditions (Jervis et al. 2004). The natural enemy that best meets these criteria for control of P. atomaria populations in plantation forestry is Neopolycystus Girault sp. (Hymenoptera:

Pteromalidae) (Chapter 3).

At least seven species of Neopolycystus are recorded from Australia and another from Papua New Guinea (Boucek 1988), however, only two species have been described: Neopolycystus insectifurax Girault and Neopolycystus abdominalis Girault

& Dodd. Very little is known of the genus except for N. insectifurax, which is a primary egg parasitoid that parasitises a wide range of paropsine species (Tribe

2000). This species has been recorded in Western Australia, Queensland, New South

Wales, the Australian Capital Territory and Victoria (Noyes 1998). It has also been introduced to New Zealand, possibly accidentally (Berry 2003), where it is now established as a promising biological control agent of P. charybdis Stål (Jones &

Withers 2003). Neopolycystus insectifurax has been recorded from P. atomaria eggs in the ACT (Tanton & Epila 1984), but was not found in South-East Queensland in this study (Chapter 3). Shepherd’s (2001) report of N. insectifurax from P. atomaria egg batches in South-East Queensland was almost certainly a case of misidentification (C. Burwell pers. comm.). An undescribed species of

Neopolycystus sp. was the only primary egg parasitoid recorded from P. atomaria eggs in this study.

Taking into account the little that is known about Neopolycystus, this chapter examines aspects of the ecology and biology of Neopolycystus sp. and the influence that it exerts on populations of P. atomaria in South-East Queensland. Presented here are the results of field monitoring of parasitism rates of P. atomaria eggs at three sites over two seasons, including the direct and indirect effects of egg parasitism on host populations, as well as the effect of host density, egg exposure time and plantation landscape features on parasitism rates. Neopolycystus spp are considered immune to hyperparasitism by the obligate hyperparasitoid Baeoanusia albifunicle

Girault (Hymenoptera: Encyrtidae) (Tribe 2000; Jones & Withers 2003), and here I

53 also examine the interaction between Neopolycystus sp. and its associated hyperparasitoid species.

4.2 Materials and methods

4.2.1 Sampling methods

In conjunction with Paropsis atomaria phenology monitoring (Chapter 2), egg batches were collected from the field for the natural enemy survey (Chapter 3) and to examine patterns of egg parasitism by Neopolycystus sp. (this chapter). Sampling details for egg batches follow those given in sections 2.2.1 and 2.2.2, with these additional details. In the 2004/2005 season, all egg batches were collected, while in the 2005/2006 season, a maximum of five egg batches per section (i.e. up to 40 per plantation) was collected on each sample date. Egg batches remained intact on a 2 –

3 cm length of host stem, and each was collected into a separate plastic vial and transported to the laboratory in a cooled box. In the laboratory, the exact number of eggs per batch was counted or estimated. Estimations used the following equations

(justified in Appendix I): 2004/2005 season, y = 0.556x + 45.071 where y = number of eggs in batch and x = sum of parasitoids and larvae emerging from batch;

2005/2006 season, y = 1.107x + 0.1643 where y = number of eggs in batch and x = length of egg batch multiplied by the number of eggs around base of egg collar).

Estimation of egg batch number, rather than absolute counts, was needed because of the simple time constraints imposed by an intensive field season with follow up laboratory work. Each egg batch was maintained separately in a cotton-wool capped vial in a controlled temperature cabinet at 24 °C. Egg batches were checked every 1

- 2 days, until larval or wasp emergence, the date of which was recorded, and the number of wasps and larvae that emerged from each batch was counted.

4.2.2 Direct effects of egg parasitism: between- and within-batch parasitism rates across sites and seasons

Between-batch parasitism rates (number of egg batches with parasitoids emerging/total number of egg batches collected) was determined for each sample date at each site. Within-batch parasitism rates (number of parasitoids emerged/number of eggs per parasitised batch) were also calculated for each sample date at each site. Total effective parasitism rates (number of parasitoids emerged/total number of all eggs) were calculated to provide an overall measure of the impact of Neopolycystus sp. on its host population. It was assumed that hyperparasitoids were obligate and therefore required primary parasitisation by

Neopolycystus sp., so included them in calculations.

4.2.3 Indirect effects of egg parasitism

The proportion of unhatched eggs per batch (those eggs that failed to produce either a wasp or a host larva) was compared between parasitised and unparasitised batches to determine whether the presence of egg parasitoids influenced egg mortality.

Proportions were arcsine-square root transformed and compared using t-tests. Total mortality rates (parasitised eggs + unhatched eggs) were compared with the stage- specific mortality rates reported in Chapter 2, to estimate the overall impact of egg parasitoids on egg mortality.

Egg parasitism may also indirectly impact on host mortality if, in obligate group feeders such as P. atomaria, smaller larval cohort sizes are less able to initiate feeding than larger groups (as found by Nahrung et. al. (2001) for C. agricola). This hypothesis was tested in the laboratory with Paropsis atomaria. Different egg parasitism rates (95, 90, 80, 60, 45, 30 and 15 % of mean number of eggs per batch)

55 were simulated by placing eleven replicates of five smaller larval groups (4, 8, 15, 30 and 42 larvae) and 6 replicates of larger groups (53, 65 and 76 larvae) onto freshly expanded, new season Eucalyptus cloeziana foliage in the laboratory. Initial densities were established using unfed (except for egg chorion consumption - see Ramsden &

Elek 1998) neonate larvae which were transferred to the E. cloeziana foliage using a fine paint brush. Replicates were held in a constant temperature cabinet (24°C;

16L:8D photoperiod) for one week, mortality for the preceding twenty-four hours was recorded each day when new foliage was added. Proportional mortality was arcsine-square root transformed and analysed using an ANOVA.

4.2.4 Parasitism rates and host density

Between-batch parasitism rates were examined for relationships with host density using Pearson correlation for the proportion of egg batches parasitised per sample date against the number of egg batches collected per sample. Within-batch parasitism rates were likewise examined against host density to test whether when fewer batches are available more eggs per batch are parasitised.

Further, the proportion of parasitised egg batches per site for each sample date was determined for Site I and Site II 2004/2005 and used to relate host density to parasitism rates. Only sections from which five or more egg batches were collected were considered in analyses. Linear regression (number of egg batches collected vs. proportion of egg batches parasitised) was used to determine whether there was a statistical relationship between these parameters. The hypothesis was that a significant relationship between host density and parasitism rates may indicate that wasps are better able to locate or utilise large patches of egg batches compared with small patches.

4.2.5 Parasitism rate and exposure time in the field

To ascertain the relationship between exposure time (egg batch age) and parasitism rate, and to ensure samples weren’t biased by collection frequency of different-aged egg batches, parasitism rate was examined as a function of egg batch age. The length of time that eggs were exposed to parasitoids in the field was estimated using lower temperature thresholds (Ltt) and day-degree (DD) development rates calculated from

Carne (1966b) (see Appendix I.B) and weather data for each site from the silo data drill (http://www.nrme.qld.gov.au/silo). The number of DD between egg collection and eclosion was determined using the equation

LabDD = (24 °C – 5.6 °C)*days to hatch in lab

where 24 °C is the constant temperature at which egg batches were housed in the laboratory, and 5.6 °C is the Ltt for egg development (Appendix I).

The number of DD required in the field was estimated by (178.6 – LabDD), where

178.6 is the number of DD above the Ltt that eggs require to develop. Field DD per day were estimated using mean daily field temperature minus Ltt (5.6 °C), and these were summed backward from each collection date until the required number DD was reached. The number of calendar days to obtain the required DD in the field was then counted. Field temperature data was supplied by the Queensland Department of

Primary Industries and Fisheries. Five egg batch field exposure times were used in analyses: 0-2 days, 3-4 days, 5-6 days, 7-8 days and 9-10 days. Between- and within-batch parasitism rates were compared between these age-classes for the whole season. Proportions were arcsine-square root transformed prior to ANOVA, and post-hoc comparisons were made using Fisher’s LSD test.

4.2.6 Parasitism rate in relation to plantation landscape features

To test the idea that key requirements of natural enemies such as the availability of alternative foods (Gurr & Nicol 2000; Berndt et al. 2002; Rebek et al. 2005), shelter or refugia (Wratten & van Emden 1995; Langellotto & Denno 2004) and alternative prey or hosts (van Emden 1990) can affect parasitoid populations, egg batch parasitism rates were related to distance from native eucalypt forest and nearest water source. These landscape features have the potential to provide a number of the key requirements listed above. The latitude and longitude of each sample site

(described in sampling methods) and landscape feature was recorded using Global

Positioning System (GPS) and plotted using Garmin Mapsource version 6.3.

Distances from each sample site to the nearest native eucalypt forest and water source were then measured using Mapsource. Correlations between distance from landscape features and proportion of egg batches parasitised were then carried out for

Sites I and II in 2004/2005 and Sites I and III in 2005/2006.

4.2.7 Neopolycystus sp. longevity

Field-collected egg batches were used to obtain wasps for laboratory longevity trials.

Wasps that emerged each day were transferred to trial colonies and held in a constant temperature cabinet at 24°C; 16L:8D photoperiod. Wasps were fed a five percent honey-water solution administered on a rolled tissue paper wick; fresh honey-water was added each day using an eye dropper. Five hundred and fifty-five wasps in 25 colonies (6 – 36 per group) were established and individual wasp deaths in each colony were recorded daily until all wasps died. Average wasp longevity per colony was determined, and all wasps’ total life spans were used to create a Kaplan-Meier survival curve.

4.2.8 Interactions with hyperparasitoids

All wasps that emerged from P. atomaria egg batches were identified to genus and guild (primary parasitoid or hyperparasitoid). Seasonal frequency of species composition was calculated, and patterns of hyperparasitism were examined. A 2- way ANOVA was used to test differences in species composition within and between sites and seasons, following arcsine-square root transformation. Post-hoc differences were identified using Fishers LSD test. Hyperparasitism rates were compared using contingency tables, and relationships between hyperparasitism rate and primary parasitoid density at the plantation scale were investigated using Pearson correlation.

4.3 Results

4.3.1 Direct effects of egg parasitism: between- and within-batch parasitism rates across sites and seasons

On average, approximately 45% of egg batches were parasitised at study sites between 2004 and 2006. The overall proportion of egg batches parasitised was lowest at Site III (around 30%) (Table 4.1), while about half of all egg batches were parasitised at Site I (both seasons) and Site II. Up to 100% of egg batches were parasitised on some sampling dates at Site I. Overall mean within-batch parasitism rates varied by a maximum of 9% (from 22-31%) and this difference was significant between sites and seasons (Table 4.1). Within-batch parasitism rates varied significantly throughout the season at Site I (04/05), Site II and Site III, but not at

Site I (05/06) (Table 4.1). Between-batch parasitism rates in the 04/05 season followed a similar trans-seasonal pattern at both sites (r = 0.70, P = 0.01), but there was no significant seasonal trend during the 05/06 season between sites (r = 0.51, P =

0.13), nor at Site I between seasons (r = -0.54, P = 0.11) (Figure 4.1). The effective

59 parasitism rate (the impact of egg parasitoids on the population as a whole i.e. total number of parasitoids over total number of eggs) averaged 13% for all sites and seasons, and differed significantly between sites and seasons (Table 4.1).

Site II 04/05 Site I 04/05 Site I 05/06 Site III 05/06 Overall 0.52 (0 – 0.87) 0.49 (0.1 – 0.67) 0.47 (0.2 – 1) 0.31 (0.05 – 0.73) between-batch 2 2 2 parasitism rate χ 1=0.45, P=0.50 χ 1=0.3, P=0.58 χ 1=16.7, P<0.001 (range across season) Mean ± se 0.31 ± 0.02 0.24 ± 0.02 0.30 ± 0.03 0.22 ± 0.02 within-batch (0.01 – 0.97) (0.01 – 0.87) (0.01 – 1) (0.01 – 0.74) parasitism rate (range) t415=3.4, P=0.001 t309=1.9, P=0.05 t151=2.0, P=0.04

Overall 0.18 (0 – 0.32) 0.13 (0.02 – 0.27) 0.13 (0.02 – 0.07 (0.001 – effective 0.27) 0.31) parasitism rate 2 2 2 (range across χ 1=91.7, P<0.001 χ 1=5.5, P=0.02 χ 1=478, P<0.001 season) Across-season Kruskall- ANOVA, ANOVA, ANOVA, within-batch Wallis, F11,236=3.23, F10,69=1.12, F9,63=2.0 parasitism H7=22.1, P < 0.001 P = 0.36 P = 0.047 rates P = 0.002

A 8 15 29 22 29 13 9 60 63 91 43 76 47 0.8 0.6 0.4 0.2 0 parasitism rate 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 25/11/04 23/12/04 20/01/05 17/02/05 16/03/05

B 8 3 12 8 6 2 55 62 67 36 60 8 1 0.8 0.6 0.4 0.2 0 parasitism rate 9/12/04 6/01/05 2/02/05 3/03/05 13/10/04 28/10/04 11/11/04 25/11/04 23/12/04 20/01/05 17/02/05 16/03/05

C 8 3 9 4 28 27 19 15 10 23 1 0.8 0.6 0.4 0.2 0 parasitism rate 9/11/05 4/01/06 1/02/06 1/03/06 26/10/05 23/11/05 21/12/05 19/01/06 15/02/06 16/03/06 D

10 19 29 11 10 21 14 32 27 33 39 0.8 0.6 0.4 0.2 0 parasitism rate 9/11/05 7/12/05 4/01/06 1/02/06 6/03/06 26/10/05 23/11/05 21/12/05 19/01/06 15/02/06 16/03/06

Figure 4.1: Between (lines) and mean + se within (bars) batch parasitism rate of Paropsis atomaria egg batches by Neopolycystus sp. at three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 [Site I 04/05 (A), Site II 04/05 (B), Site I 05/06 (C), and Site III 05/06 (D)].

4.3.2 Indirect effects of egg parasitism

The proportion of unhatched (into wasps or larvae) eggs was significantly higher from batches that were parasitised than for unparasitised batches at all sites in both seasons (Figure 4.2) (t-tests: Site I 04/05 t395=9.3 P< 0.001; Site II 04/05 t241=5.5 P<

0.001; Site I 05/06 t170=4.3 P< 0.001; Site III 05/06 t243=2.9 P= 0.004). On average, parasitism doubled egg failure, resulting in an additional 14.8 ± 3 % egg mortality to unparasitised eggs (normally 15 ± 3%).

0.50 0.40 0.30 0.20 0.10

eggs/batch 0.00 Site I Site II Site I Site 04/05 04/05 05/06 III 05/06 mean + proportionmean se unhatched

Overall P. atomaria egg-Li+Lii field mortality was about 70% (Chapter 2). About

15% of egg mortality can be attributed to average failure of eggs to hatch (see above). The estimated overall direct loss from egg parasitoids was 13% (Table 4.1), with an additional 15% lost as an indirect effect of parasitism (Figure 4.2). Thus, egg parasitoids contribute directly and indirectly to almost half of the estimated egg

62 mortality in the field and are responsible for mortality of about one-third of all eggs in the field. The remaining 22% egg-Li+Lii mortality probably arises from predation and failure of early instars to initiate feeding on host foliage (Fig 4.3).

Egg parasitism is unlikely to contribute to host mortality through indirect affects on larval feeding establishment, as overall there was no significant effect of initial batch size on neonate larval survival of P. atomaria (F7,70 = 1.962, P = 0.074). The only exception was the smallest batch size, where the mortality rate was approximately double that of any other batch (Figure 4.3) and was shown to be significantly different in post hoc tests.

a 80

Meanpercent+ se larval mortality 10

0 4 8 15 30 42 53 65 76 Initial larval density

4.3.3 Parasitism rate and host density

There was a significant, positive correlation between the proportion of eggs parasitised and the number of egg batches present on branches at Site I in the

2004/2005 season (r = 0.62, P = 0.02). At Site II in the same season, there was no significant relationship between parasitism rate and the number of egg batches censused (r = 0.21, P = 0.51). Site III, however, showed a significant, negative correlation between these parameters (r = -0.74, P = 0.01).

Host density (number of egg batches collected) was significantly correlated with the proportion of egg batches parasitised at Site I (Figure 4.4a: linear regression, R2=

0.48, P < 0.001), but not at Site II (Figure 4.4b: linear regression, R2= 0.008, P =

0.68). Because a maximum of five egg batches was collected per plantation section in the 2005/2006 season, these data were not tested in this way.

A 1 0.8 0.6 0.4 y = 0.0266x + 0.1701

parasitised 0.2 0 proportion of batches 0102030 number of batches collected

B y = 0.0033x + 0.5153 1 0.8 0.6 0.4

parasitised 0.2 0 proportion of batches 0 5 10 15 20 number of batches collected

Figure 4.4: Relationship between the number of Paropsis atomaria egg batches collected within a plantation section, and the proportion of those egg batches that were parasitised for two eucalypt plantation sites in South-East Queensland in 2004/2005 [Site I (A) and Site II (B)].

The proportion of eggs parasitised within batches (i.e. within-batch parasitism rates) increased with increases in the proportion of batches parasitised (i.e. between-batch parasitism rates) at Site I (04/05) and at Site III (05/06) (r = 0.59, P = 0.03; r = 0.74,

P = 0.01, respectively), but there was no such relationship at Site II (04/05) or Site I

(05/06) (r = 0.28, P = 0.47; r = -.22, P = 0.52). Within-batch egg parasitism rates were not related to host density at any site.

4.3.4 Parasitism rates and exposure time in the field

The number of egg batches parasitised generally increased up to 5-6 days after being laid, but few or no additional egg batches were parasitised after this time in the field at all sites (Figure 4.5). In contrast, within-batch parasitism rates did not significantly increase depending on exposure time, except at Site II during the

2004/2005 season (Figure 4.6) (ANOVA, Site I 04/05: F4,237= 1.6, P = 0.17; Site II

04/05: F4,160= 8.5, P < 0.001; Site I 05/06: F4,75= 0.52, P = 0.72; Site III 05/06: F4,76=

2.14, P = 0.09). Collection frequency of egg batches collected before and after the critical 5-6 day period was similar: 55% of batches were collected prior to maximum parasitisation, and 45% after; therefore collection data (as presented in this chapter) was not corrected for field exposure time.

A 0.8 81 161 173 43 28

0.6

0.4

0.2

proportion of egg 0 batches parasitised batches 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

B 1 61 136 75 26 18 0.8 0.6 0.4 0.2 proportion of egg egg of proportion

batches parasitised 0 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

C 0.5 57 93 65 24 8 0.4 0.3 0.2 0.1 proportion of egg egg of proportion

batches parasitised 0 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

D 28 66 47 33 8 0.6

0.4

0.2 proportion of egg egg of proportion

batches parasitised 0 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

A 0.5 0.4 0.3 0.2 0.1 parasitism rate 0.0

mean + se within-batch within-batch se + mean 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

b 0.6 B b

a a 0.4

0.2 parasitism rate 0

mean + se within-batch mean+ se within-batch 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

C 0.40

0.30

0.20

0.10 parasitism rate 0.00

mean + se within-batch within-batch se + mean 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

D 0.6 0.5 0.4 0.3 0.2 0.1 parasitism rate parasitism 0

mean + se within-batch mean+ se within-batch 0to2 3to4 5to6 7to8 9to10 egg batch age (days)

Figure 4.6: Mean + s.e. within-batch parasitism rate of Paropsis atomaria egg batches at different field exposure times from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 [Site I 04/05(A), Site II 04/05 (B), Site III 05/06 (C), and Site I 05/06 (D)]. Different letters above bars designate means that differed significantly (Fisher’s LSD post-hoc test).

4.3.5 Parasitism rate in relation to plantation landscape features

In general there was no correlation between distance from nearest native eucalypt forest or water and the proportion of P. atomaria egg batches parasitised or within batch parasitism rate at any of the three plantations (Table 4.2). The only exceptions were a weak negative correlation between distance from forest and within batch parasitism rate at Site III 2005/2006, and a weak positive correlation between distance from water and proportion of egg batches parasitised at Site I 2004/2005.

Table 4.2: The proportion of egg batches parasitised and within batch parasitism rates of Paropsis atomaria eggs from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 correlated against distance from the nearest native eucalypt forest and nearest source of water. (* designates significant correlation at 0.05 level).

Site I 04/05 Site II 04/05 Site I 05/06 Site III 05/06

Distance from forest and proportion r = -0.001 r = -0.157 r = -0.039 r = -0.034 of egg batches parasitised P=0.993 P= 0.316 P= 0.776 P= 0.773

Distance from forest and within r = 0.069 r = 0.114 r = 0.261 r = -0.320 batch Parasitism rate P= 0.649 P= 0.542 P= 0.083 P= 0.034 *

Distance from water and proportion r = 0.277 r = -0.113 r = 0.139 r = 0.144 of eggs parasitised P= 0.045 * P= 0.471 P= 0.307 P= 0.216

Distance from water and within r = 0.033 r = 0.088 r = -0.001 r = 0.136 batch parasitism rate P= 0.826 P= 0.638 P= 0.995 P= 0.380

4.3.6 Neopolycystus sp. longevity

Mean wasp longevity was 16.1 ± 0.3 days (range 1 – 66) in the laboratory. Highest mortality occurred within the first three days of emergence (Figure 4.7).

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

proportion of wasps surviving wasps of proportion 0 0 10203040506070 time (days)

Figure 4.7: Kaplan-Meier survivorship curve for Neopolycycstus sp. in the laboratory.

4.3.7 Interactions with hyperparasitoids

The species composition of wasps emerging from Paropsis atomaria egg batches was similar between sites and seasons (2-way ANOVA, site/season: F3,2112 = 0.018, P

= 0.99) (Figure 4.8), but differed within them (2-way ANOVA, wasp spp: F3,2112 =

770.6, P < 0.001). Neopolycystus sp. was the most abundant wasp, followed by

Baoeanusia albifunicle, while Aphanomerella sp. and Neblatticida sp. shared a similar level of abundance (Fisher’s LSD post-hoc test). However, there was also a significant interaction between site/season and species (2-way ANOVA,

70 site/season*wasp spp: F9,2112 = 4.11, P < 0.001). Neblatticida sp. was very rare at

Site I in 2004/2005, with only 19 individuals collected. Seasonal patterns of relative species abundance for each site are presented in Figure 4.9.

a a a 0.80 a

0.60

0.40 b b b b 0.20 c c c c c c c c 0.00 mean + se proportion of wasp spp. Site II 04/05 Site I 04/05 Site I 05/06 Site III 05/06

A 64 209 185 61 88 192 11 324 541 1129 69 853 198 100% 80% 60% 40% 20% 0% % ofwasps emerged 9/12/04 6/01/05 2/02/05 3/03/05 29/09/04 13/10/04 28/10/04 11/11/04 25/11/04 23/12/04 20/01/05 17/02/05 16/03/05 B 77 9 112 319 658 1595 205 616 74 100% 80% 60% 40% 20% 0% % of wasps emerged 9/12/04 6/01/05 2/02/05 3/03/05 13/10/04 11/11/04 20/01/05 17/02/05 16/03/05

C 106 34 122 65 272 295 422 366 17 191 112 100% 80% 60% 40% 20%

% waspsemerged 0% 9/11/05 4/01/06 1/02/06 1/03/06 26/10/05 23/11/05 21/12/05 19/01/06 15/02/06 16/03/06 30/03/06 D 88 142 177 279 9 143 132 207 261 24 5 1 100% 80% 60% 40% 20% 0% % of wasps emerged 9/11/05 7/12/05 4/01/06 1/02/06 6/03/06 26/10/05 23/11/05 21/12/05 19/01/06 15/02/06 16/03/06 30/03/06

Figure 4.9: Across-season species composition of wasps emerging from Paropsis atomaria egg batches from three eucalypt plantation sites in South-East Queensland in 2004/2005 and 2005/2006 [Site I 04/05 (A), Site II 04/05 (B), Site I 05/06 (C), and Site III 05/06 (D)]. Neopolycystus sp. (stippled), Baeoanusia albifunicle (grey), Neblatticida sp. (black), and Aphanomerella sp. (white). Numbers above columns are the number of wasps emerged.

Hyperparasitism rates were higher at Site I than Site II during the 2004/2005 season

2 (see Chapter 2), and also higher in that season than during 2005/2006 (Site I: χ 1=

2 34.5, P < 0.001), but similar between Site I and Site III in the 2005/2006 season (χ 1=

1.8, P = 0.18). On average, 27% of eggs parasitised by Neopolycystus were hyperparasitised (Table 4.3). However, within egg batches, hyperparasitoids rarely killed all Neopolycystus sp., with only an average 9 ± 2 % of parasitised egg batches failing to produce any primary parasitoids. Super-hyperparasitism (two or more hyperparasitoid species emerging from one egg batch) occurred in 11 % of egg batches. There was no relationship between hyperparasitism rate and primary parasitoid availability across the season at any site (Table 4.3).

Site II Site I 04/05 Site I Site III Mean ± s.e. 04/05 05/06 05/06 Overall hyperparasitism rate 0.23 0.33 0.28 0.25 0.27 ± 0.02 Proportion of parasitised egg batches with…. Neopolycystus sp. 0.94 0.88 0.92 0.89 0.91 ± 0.02 Baoeanusia albifunicle 0.27 0.40 0.32 0.28 0.32 ± 0.03 Neblatticida sp. 0.19 0.02 0.10 0.03 0.08 ± 0.05 Aphanomerella sp. 0.05 0.19 0.06 0.11 0.09 ± 0.02 Neopolycystus sp. only 0.58 0.55 0.63 0.66 0.60 ± 0.03 Baoeanusia albifunicle only 0.03 0.05 0.01 0.05 0.04 ± 0.01 Neblatticida sp. only 0 0 0 0 0 Aphanomerella sp. only 0.01 0.01 0 0 0 2 or more hyperparasitoid spp. 0.09 0.15 0.09 0.10 0.11 ± 0.02

No. primary parasitoids vs -0.09 0.23 0.52 0.24 n/a hyperparasitism rate (Pearson P=0.81 P = 0.45 P = 0.10 P= 0.46 correlation)

4.4 Discussion

4.4.1 Neopolycystus and P. atomaria mortality

Neopolycystus sp. exerts a significant influence on P. atomaria populations in E. cloeziana plantations in South-East Queensland. The total field mortality of P. atomaria from egg to first and second instar is approximately 70% (Chapter 2), of which almost half can be attributed to the direct and indirect effects of egg parasitism. Direct effects of parasitism are accounted for by the emergence of any species of parasitoid from the host egg, as Neopolycystus sp. was the only primary parasitoid encountered. The emergence of hyperparasitoids was used as an indication of initial parasitism by Neopolycystus. The additional indirect effect of increased proportion of unhatched eggs in parasitised egg batches is thought to be the result of eggs being parasitised but the parasitoid failing to develop, possibly the result of super parasitism by conspecifics or other species. Nahrung and Murphy (2002) likewise recorded high levels of collapsed eggs following exposure to egg parasitoids. Tribe (2000) found in dissections of collapsed eggs (after being exposed to multiple species of parasitoids) that eggs contained several parasitoid larvae, each with a quantity of ingested yolk, but insufficient for any larvae to complete their development

Egg parasitism can potentially affect host mortality through other indirect means: smaller larval groups of paropsines are less able to initiate feeding than larger groups

(Nahrung et al. 2001), thus, higher egg parasitism rates may increase mortality by reducing the group size of remaining larvae. This was indeed the case, but only at the highest parasitism rate tested (95%) was there any difference in larval mortality. The reason proposed for larger groups being more able to establish feeding is that the

74 probability of one larva successfully initiating feeding, thereby creating access to a suitable feeding site increases as group size increases (Nahrung et al. 2001). P. atomaria egg batches are much larger than those of the species studied by Nahrung et al. (2001), so any more then 5% larval emergence from an egg batch probably provides enough individuals to ensure feeding establishment in P. atomaria.

4.4.2 Density dependent parasitism

Many parasitoid species aggregate at places with higher host density (Meiners &

Obermaier 2004). However, this does not necessarily lead to positive density dependent parasitism (Lessels 1985), as many studies have failed to demonstrate direct density dependence due to limits in the number of mature eggs or time available for searching (Walde & Murdoch 1988). This can cause under exploitation of high density patches and can result in inverse density dependence, or domed relationships, in which parasitism increases at low host densities and then decreases at higher host densities (Lessels 1985). Although not tested here, Neopolycystus sp. is most likely synovigenic: over 98% of parasitoid wasp species are (Jervis et al. 2001), while the ovigeny index of females is negatively correlated to size across species

(Jervis et al. 2003) Therefore, it would be expected that Neopolycystus sp. might show density dependence at lower host densities and inverse density dependence at higher densities. This is demonstrated to some extent in this study, as at Site I there was evidence of density dependence and at Site III, which had a far greater host density (Chapter 2), there was evidence of inverse density dependence. There are also numerous examples where parasitism is independent of host density (Lessels

1985) and Site II in this study, which had the lowest density of hosts, showed no correlation between parasitism and host density.

4.4.3 Site Effects

Another factor that can influence parasitism rate is the characteristics of the surrounding habitat. A large body of literature exists demonstrating that the provision of supplementary resources can increase the abundance, fecundity and longevity of parasitoids, which in turn should lead to higher parasitism rates (see reviews by: Barbosa 1998; Pickett & Bugg 1998; Landis et al. 2000). Habitats in which sources of sugar are abundant are particularly likely to be good for parasitic wasps (Shaw 2006) and it has been demonstrated that parasitism decreases as distance from such habitats increases (Baggen & Gurr 1998). Native eucalypt forest and water features within, and adjoining, the study sites were chosen to test this hypothesis as they were thought to provide the best source of supplementary resources in the form of native plants and weedy vegetation. The results obtained were inconclusive, with only a few weak correlations between distance from landscape features and parasitism rate.

4.4.4 Hyperparasitoids

An important aspect of this study is the influence of hyperparasitoids on the effectiveness of Neopolycystus sp.. B. albifunicle has not been previously recorded hyperparasiting Neopolycystus (Tribe 2000; Jones & Withers 2003), and this could have possible implications towards the future selection of Neopolycystus spp. for biological control. Two of the other hyperparasitoid species encountered,

Neblatticida sp. and Aphanomerella sp., were quite rare, but B. albifunicle was present in sufficient numbers throughout the season that it could pose a potential problem to the efficacy of Neopolycystus sp.. However, as discussed by Jones &

Withers (2003) in their study of the parasitoid complex of P. charybdis,

76 hyperparasitism of parasitised eggs does still prevent the eggs from hatching. The only disruptive effect that B. albifunicle had on populations of the primary parasitoid in their system, Enoggera nassaui Girault (Hymenoptera: Pteromalidae), was that of increased mortality, which reduced the rate of E. nassaui population increase (Jones

& Withers 2003). This does not seem to apply in the system studied here, since in

New Zealand hyperparasitism rates of E. nassaui increased throughout the season until 100% of parasitised eggs were attacked by B. albifunicle, while in South-East

Queensland the proportion of Neopolycystus sp. and B. albifunicle emerging from egg batches remained relatively constant at all plantations. Within egg batches, hyperparasitoids rarely killed all Neopolycystus sp., with only 9 % of parasitised egg batches failing to produce any primary parasitoids.

This study examined a number of aspects of the ecology of Neopolycystus sp., expanding upon the almost nonexistent knowledge of the most important natural enemy of P. atomaria in South-East Queensland E. cloeziana plantations. Although the potential of the wasp to exert a significant influence on P. atomaria populations has been demonstrated, its suitability for conservation biological control has not.

Further studies on ecological and biological characteristics including fecundity, oviposition rate, why within batch parasitism rate was not affected by exposure time, host range and the effects of supplementary resources on parasitism rate, are required.

Chapter 5: General Discussion

5.1 Introduction

The original aim of this research was to enhance natural enemy mortality of Paropsis atomaria in eucalypt plantations, as an alternative to the economically and environmentally unsustainable practice of controlling pests with chemical pesticides.

This was to be achieved by manipulating the environment within and around plantations to make it more suitable for the natural enemies of P. atomaria. This in turn should have increased the abundance and efficacy of the natural enemies and result in a reduction in the abundance of P. atomaria. This aim was set, however, before I became aware of the void in knowledge of both the pest, P. atomaria, and its natural enemies.

For conservation biological control to be effective, highly detailed knowledge of the life history of the pest, its natural enemies (both generalist and specialist), and their interactions, are required (Verkerk et al. 1998; Jervis et al. 2004). When I determined that even the identities of the natural enemies of P. atomaria in South-East

Queensland were uncertain, it became apparent that the chance of developing any form of effective conservation biological control program was greatly reduced. With this in mind, the emphasis of this project changed from developing a method of controlling P. atomaria, to identifying the most promising natural enemies and learning something of their life history. Herein I summarise and discuss my results on the ecological characteristics of P. atomaria and its natural enemies, subsequently focussing on the most promising of the natural enemies identified, Neopolycystus sp..

Also discussed are the implications of the findings presented in this thesis to future

78 research into the conservation biological control of paropsines in eucalypt plantations.

5.2 Paropsis atomaria

Paropsine leaf beetles are major pests of eucalypts in all states where they are grown commercially (de Little 1989; Simmul & de Little 1999) and one species,

Chrysophtharta bimaculata, is regarded as Tasmania’s most important hardwood forestry pest (Elliott et al. 1992). In Queensland, paropsine beetles have the potential to pose an even greater threat to eucalypt forestry as the warmer weather enables a longer active season, within which multiple generations can be completed. P. atomaria is one of major paropsine pests of the currently favoured hardwood plantation species in South-East Queensland (Nahrung 2006) and, in comparison to the Australian Capital Territory where they are bivoltine (Carne 1966a), are capable of completing up to four generations in a single breeding season (Chapter 2).

Peak abundance of early instar larvae usually coincided with peak defoliation indicating that early instar larvae are responsible for a high percentage of defoliation.

This is in contrast to many other paropsine pests where most damage is thought to be caused by third and forth instars (Carne 1966b; Greaves 1966). The timing of peak damage has important implications for the timing of control measures as, to be most effective, they should be applied before the most damaging stage. In the case of P. atomaria control measures in the pre-reproductive adult or egg stage would be ideal.

By targeting the eggs the damage caused by early instar larvae can be reduced.

Egg parasitoids contributed to almost 30% of total egg mortality; directly, through successful parasitism (around 13 %), and indirectly, through failure of eggs to hatch

(around 15%). However, there were no flow-on indirect effects of increased

79 mortality due to larval feeding establishment, as demonstrated for C. agricola

(Nahrung et al. 2001). Due to the larger batch sizes of P. atomaria, any more than five percent emergence provides enough larvae to ensure feeding establishment

(Chapter 4).

5.3 Natural enemies

Natural enemies are responsible for a high proportion of paropsine mortality in the field in Tasmania (de Little et al. 1990; Allen & Patel 2000) and this may in fact be the case in South-East Queensland hardwood plantations. Several predators were identified in the field; although their effect on P. atomaria populations was not measured in this study. Abundance of all predators was low and it is doubtful whether their numbers could be increased sufficiently through habitat manipulation to deal with outbreaks of P. atomaria. Mortality provided by the predators, and in particular the spiders, is nevertheless important as it has been demonstrated in some perennial systems that generalist natural enemies can prevent resurgence of pests after outbreaks have been controlled by specialist natural enemies (Mason &

Torgersen 1987; Roland & Embree 1995).

Similarly, larval parasitoids appear not to be suitable for conservation biological control in this case as the impact that they had on P. atomaria populations in South-

East Queensland was negligible (Chapter 3). In the Australian Capital Territory, larval parasitism rates of up to 70% have been reported (Tanton & Epila 1984), but for reasons that remain unclear overall larval parasitism rates in South-East

Queensland were only around 1%.

5.4 Neopolycystus sp.

Of the natural enemies of P. atomaria that were identified, the species with the most potential for conservation biological control is Neopolycystus sp.. It was demonstrated that Neopolycystus sp. exerts a significant direct and indirect influence on P. atomaria populations (Chapter 4). However, in this study it was not demonstrated that there were any limiting resources that could be manipulated to increase the effectiveness of Neopolycystus sp.. The abundance, longevity and fecundity of parasitoids in agricultural systems can be increased by supplying supplementary food resources in the form of nectar and pollen, which in turn can increase the effectiveness of the parasitoids (Olson & Nechols 1995; Idris & Grafius

1996). If this is the case in eucalypt plantations it would be expected that there would be a correlation between distance from landscape features that provide supplementary resources and parasitism rate. There were no significant correlations observed (Chapter 4) which may be due to the relatively uniform distribution of weedy plants throughout plantations that may be capable of providing the supplementary resources required by parasitoids.

Unlike agricultural systems, where the majority of work in conservation biological control has been carried out, unmanipulated eucalypt plantations already provide quite a complex environment with a high level of structural complexity and diversity in the form of weedy undergrowth plants. Many of the requirements of natural enemies that are provided by increasing diversity in agricultural systems (eg pollen sources, nectar sources, habitat refuges and alternative prey) are probably already present in eucalypt plantations. If these resources are not limiting for the parasitoids

81 within plantations, then it is doubtful if manipulating the habitat will be able to provide an increase in parasitism rate.

Another area of this research that points to difficulties in increasing parasitism rates of P. atomaria by Neopolycystus sp. is indicated by the presence of a domed density dependence relationship, where parasitism rates increase at low P. atomaria densities and decrease at high densities. The under exploitation of high density patches tends to indicate that the parasitoids may be limited by their complement of mature eggs at any given time. In this case the provision of supplementary resources would provide no advantage as the parasitoids may already be operating near maximum efficiency.

5.5 Implications to conservation biological control

The data presented in this study indicate that the potential for effective conservation biological control through habitat manipulation is limited in South-East Queensland hardwood plantations. The guild of predators provides a useful level of background mortality but is not likely to be able to increase in numbers sufficiently to deal with pest outbreaks.

The primary egg parasitoid Neopolycystus sp. has a significant effect on P. atomaria populations, but whether parasitism rates can be increased by manipulating the habitat is yet to be demonstrated, but I consider it unlikely. To have any chance of success with a conservation biological control program involving Neopolycystus sp., it is vital to determine whether it is possible to increase parasitism rates. Within- batches parasitism rates generally did not increase over time with concomitant increases in between-batch parasitism rates, suggesting that Neopolycystus sp. may have behavioural or physiological characteristics limiting the number of eggs that are

82 parasitised in any single P. atomaria egg batch. Further study is required to determine if this is the case.

Although Neopolycystus sp. may not be suitable for conservation biological control in hardwood plantations, it may have potential for augmentative biological control within Australian plantations or classical biological control overseas. Further study into rearing techniques would determine if this was feasible.

References

Allen, G. R., & Patel, V. S. (2000). The spatial and temporal impact of endemic natural enemies on the mortality of chrysomelid beetles in eucalypt plantations. Paper presented at the XXI International Congress of Entomology, Brazil.

Andow, D. A. (1991). Vegetational diversity and arthropod population response. Annual Review of Entomology, 36, 561-586.

Baggen, L. R., & Gurr, G. M. (1998). The influence of food on Copidosoma koehleri (Hymenoptera: Encyrtidae), and the use of flowering plants as a habitat management tool to enhance biological control of Potato Moth, Phthorimaea operculella (Lepidoptera: Gelechiidae). Biological Control, 11, 9-17.

Baggen, L. R., Gurr, G. M., & Meats, A. (1999). Flowers in tri-trophic systems: mechanisms allowing selective exploitation by insect natural enemies for conservation control. Entomologia Experimentalis et Applicata, 91, 155-161.

Baker, S. C., Elek, J. A., Bashford, R., Paterson, S. C., Madden, J., & Battaglia, M. (2003). Inundative release of coccinellid beetles into eucalypt plantations for biological control of chrysomelid leaf beetles. Agricultural and Forest Entomology, 5, 97-106.

Baly, J. S. (1862). Descriptions of the species belonging to the genus Paropsis. Journal of Entomology, 2, 291-310.

Barbosa, P. (Ed.). (1998). Conservation Biological Control. San Diego: Academic Press.

Barbosa, P., & Benrey, B. (1998). The influence of plants on insect parasitoids:Implications for conservation biological control. In P. Barbosa (Ed.), Conservation Biological Control (pp. 55-82). San Diego: Academic Press.

Bashford, R. (1999). Predation by ladybird beetles (coccinellids) on immature stages of the Eucalyptus leaf beetle Chrysophtharta bimaculata (Olivier). Tasforests, 11, 77-86.

Berndt, L. A., & Wratten, S. D. (2005). Effects of alyssum flowers on the longevity, fecundity, and sex ratio of the leafroller parasitoid Dolichogenidea tasmanica. Biological Control, 32, 65-69.

Berndt, L. A., Wratten, S. D., & Hassan, P. G. (2002). Effects of buckwheat flowers on leafroller (Lepidoptera: Tortricidae) parasitoids in a New Zealand vineyard. Agricultural and Forest Entomology, 4, 39-45.

Berry, J. A. (2003). Neopolycystus insectifurax Girault (Hymenoptera: Pteromalidae) is established in New Zealand, but how did it get here? New Zealand Entomologist, 26, 113-114.

Bertus, A. W., & Walker, J. (1974). Ramularia on Eucalyptus and Angaphora. Australian Plant Pathology Newsletter, 3, 3.

Beveridge, N., & Elek, J. A. (2001). Insect and host-tree species influence the effectiveness of a Bacillus thuringiensis ssp. tenebrionis-based insecticide for controlling chrysomelid leaf beetles. Australian Journal of Entomology, 40, 386-390.

Boucek, Z. (1988). Australasian Chalcidoidea (Hymenoptera). A biosystematic revision of genera of fourteen families, with a reclassification of species. U.K: CAB International.

Braganca, M. A. L., Zanuncio, J., Picanco, M., & Laranjeiro, A. J. (1998). Effects of environmental heterogeneity on Lepidoptera and Hymenoptera populations in Eucalyptus plantations in Brazil. Forest Ecology and Management, 103, 287- 292.

Bulinski, J., & McArthur, C. (2003). Identifying factors related to the severity of mammalian browsing damage in eucalypt plantations. Forest Ecology and Management, 183, 239-247.

CABInternational. (2005). Forestry Compendium. Wallingford UK: CAB International.

Candy, S. G., Elliott, H. J., Bashford, R., & Greener, A. (1992). Modelling the impact of defoliation by the leaf beetle, Chrysophtharta bimaculata (Coleoptera: Chrysomelidae), on height growth of Eucalyptus regnans. Forest Ecology and Management, 54, 69-87.

Carne, P. B. (1966a). Ecological characteristics of the eucalypt-defoliating chrysomelid Paropsis atomaria Ol. Australian Journal of Zoology, 14, 647- 672.

Carne, P. B. (1966b). Growth and food consumption during the larval stages of Paropsis atomaria (Coleoptera: Chrysomelidae). Entomologia Experimentalis et Applicata, 9, 105-112.

Clarke, A. R. (1998). Day-degree estimates for the immature stages of the Tasmanian eucalyptus leaf beetle Chrysophtharta bimaculata (Oliver) (Coleoptera: Chrysomelidae). Australian Entomologist, 25, 35-42.

Clarke, A. R., Zalucki, M. P., Madden, J. L., Patel, V. S., & Paterson, S. C. (1997). Local dispersion of the Eucalyptus leaf-beetle, Chrysophtharta bimaculata (Coleoptera: Chrysomelidae), and implications for forest protection. Journal of Applied Ecology, 34, 807-817.

Coll, M. (1998). Parasitoid activity and plant species composition in intercropped systems. In C. H. Pickett & R. L. Bugg (Eds.), Enhancing Biological Control (pp. 85-119). Berkley and Los Angeles: University of California Press.

Collett, N. (2001). Insect pests of young eucalypt plantations (Agriculture Notes). Heidelberg: Department of Natural Resources and Environment.

Cumpston, M. (1939). Observations of the bionomics and morphology of seven species of the tribe Paropsini (Chrysomelidae). Proceedings of the Linnean Society NSW, 64, 353-366.

Cunningham, S. A., Floyd, R. B., & Weir, T. A. (2005). Do Eucalyptus plantations host an insect community similar to remnant Eucalyptus forest? Austral Ecol, 30, 103-117.

Davies, M. (1966). Neurosecretion in Chrysophtharta bimaculata. Unpublished Honours, University of Tasmania, Hobart. de Little, D. W. (1979). Taxonomic and ecological studies of the Tasmanian Eucalyptus-defoliating paropsids. The University of Tasmania, Hobart. de Little, D. W. (1989). Paropsine chrysomelid attack on plantations of Eucalyptus nitens in Tasmania. New Zealand Journal of Forestry Science, 19, 223-227. de Little, D. W., Elliott, H. J., Madden, J. L., & Bashford, R. (1990). Stage-specific mortality in two field populations of immature Chrysophtharta bimaculata (Oliver) (Coleoptera: Chrysomelidae). Journal of the Australian Entomological Society, 29, 51-55.

DeBach, P. (1974). Biological Control by Natural Enemies. London: Cambridge University Press.

Doutt, R. L. (1964). The historical development of biological control. In P. DeBach (Ed.), Biological Control of Insect Pests and Weeds (pp. 21-42). London: Chapman and Hall.

Drummond, F., Logan, P., Casagrande, R., & Gregson, F. (1985). Host specificity tests of Chrysomelobia labidomerae, a mite parasitic on the Colorado potato beetle. International Journal of Acarology, 11, 169-172.

Ehler, L. E. (1998). Conservation biological control: past, present and future. In P. Barbosa (Ed.), Conservation Biological Control. San Diego: Academic Press.

Eilenberg, J., Hajek, A., & Lomer, C. (2001). Suggestions for unifying the terminology in biological control. Biocontrol, 46, 387-400.

Elek, J., Beveridge, N., & Ramsden, N. (1998). To Bt or not to Bt: is a Bacillus thuringiensis-based insecticide suitable for IPM to control leaf beetles in

eucalypt plantations? Paper presented at the the Sixth Australasian Applied Entomological Research Conference, Brisbane.

Elek, J. A. (1997). Assessing the impact of leaf beetles in eucalypt plantations and exploring options for their management. Tasforests, 9, 139-154.

Elliott, H. J., Bashford, R., Greener, A., & Candy, S. G. (1992). Integrated pest management of the Tasmanian Eucalyptus beetle, Chrysophtharta bimaculata (Oliver) (Coleoptera: Chrysomelidae). Forest Ecology and Management, 53, 29-38.

Elliott, H. J., Ohmart, C. P., & Wylie, F. R. (1998). Insect Pests of Australian Forests, Ecology and Management. Melbourne: Reed International Books/Inkata Press.

Elton, C. S. (1958). Ecology of Invasions by Animals and Plants. London: Chapman and Hall.

Goodman, D. (1975). The theory of diversity-stability relationships in ecology. Quarterly Review of Biology, 50, 237-266.

Govender, P. (2002). Management of insect pests: have the goalposts changed with certification? Southern African Forestry Journal, 39-45.

Greaves, R. (1966). Insect defoliation of eucalypt regrowth in the Florentine Valley, Tasmania. Appita, 19, 119-126.

Greenstone, M. H. (1999). Spider predation: how and why we study it. Journal of Arachnology, 27, 333-342.

Gurr, G. M., & Nicol, H. I. (2000). Effect of food on longevity of adults of Trichogramma carverae Oatman and Pinto and Trichogramma nr brassicae Bezdenko (Hymenoptera: Trichogrammatidae. Australian Journal of Entomology, 39, 185-187.

Gurr, G. M., Scarratt, S. L., Wratten, S., Berndt, L. A., & Irvin, N. (2004a). Ecological engineering, habitat manipulation and pest management. In G. M.

Gurr, S. Wratten & M. A. Altieri (Eds.), Ecological Engineering for Pest Management (pp. 1-12). Collingwood: CSIRO Publishing.

Gurr, G. M., & Wratten, S. (2000). Biological Control: Measures of Success. Netherlands: Kluwer Academic Publishers.

Gurr, G. M., Wratten, S., & Altieri, M. A. (Eds.). (2004b). Ecological Engineering for Pest Management. Collingwood: CSIRO Publishing.

Hawkins, B. A. (1993). Parasitoid species richness, host mortality, and biological control. American Naturalist, 141, 634-641.

Helenius, J. (1998). Enhancement of Predation Through Within-Field Diversification. In C. H. Picket & R. L. Bugg (Eds.), Enhancing Biological Control (pp. 122-160). Berkely and Los Angeles: University of California Press.

Howlett, B. G., Clarke, A. R., & Madden, J. L. (2001). The influence of leaf age on the oviposition preference of Chrysophtharta bimaculata (Olivier) and the establishment of neonates. Agricultural and Forest Entomology, 3, 121-127.

Hoy, M. A. (1994). Parasitoids and predators in management of arthropod pests. In R. L. Metcalf & W. H. Luckmann (Eds.), Introduction to Insect Pest Management (pp. 129-198). New York: Wiley.

Idris, A., & Grafius, E. (1996). Effects of wild and cultivated host plants on oviposition, survival, and development of diamondback moth(Lepidoptera: Plutellidae) and its parasitoid Diadegma insulare(Hymenoptera: Ichneumonidae). Environmental Entomology, 25, 825-833.

Jactel, H., Goulard, M., Menassieu, P., & Goujon, G. (2002). Habitat diversity in forest plantations reduces infestations of the pine stem borer Dioryctria sylvestrella. Journal of Applied Ecology, 39, 618-628.

Jervis, M. A., Ferns, P. N., & Heimpel, G. E. (2003). Body size and the timing of egg production in parasitoid wasps: a comparative analysis. Functional Ecology, 17, 375-383.

Jervis, M. A., Heimpel, G. E., Ferns, P. N., Harvey, J. A., & Kidd, N. A. C. (2001). Life-history strategies in parasitoid wasps: a comparative analysis of 'ovigeny'. Journal of Animal Ecology, 70, 442-458.

Jervis, M. A., Lee, J. C., & Heimpel, G. E. (2004). Use of behavioural and life- history studies to understand the effects of habitat manipulation. In G. M. Gurr, S. Wratten & M. A. Altieri (Eds.), Ecological Engineering for Pest Management (pp. 65-100). Collingwood: CSIRO Publishing.

Jones, D. C., & Withers, T. M. (2003). The seasonal abundance of the newly established parasitoid complex of the Eucalyptus tortoise beetle (Paropsis charybdis). New Zealand Plant Protection, 56, 51-55.

Kruess, A. (2003). Effects of landscape structure and habitat type on a plant- herbivore-parasitoid community. Ecography, 26, 283-290.

Landis, D. A., Wratten, S. D., & Gurr, G. M. (2000). Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology, 45, 175-201.

Langellotto, G. A., & Denno, R. F. (2004). Responses of invertebrate natural enemies to complex-structured habitats: a meta-analytical synthesis. Oecologia, 139, 1-10.

Larsson, J., & Ohmart, C. P. (1988). Leaf age and performance of the leaf beetle Paropsis atomaria. Ecological Entomology, 13, 19-24.

Lawrence, J. F., & Britton, E. B. (1996). Coleoptera (Beetles). In The Insects of Australia: A textbook for students and research workers. Melbourne: Melbourne University Press.

Lawson, S., & Ivory, M. H. (2000). State and company overviews of eucalypt plantation estates and pest problems: Queensland. Paper presented at the Workshop on Managing Pests of Eucalypt Plantations, CRC for Sustainable Production Forestry, ANU, Canberra.

Lawson, S. A., & King, J. (2002). Leaf beetles (Paropsines). Queensland Forest Research Institute, Agency for Food and Fibre Sciences, DPI. Brisbane: DPI Note FOR0027.

Lessels, C. M. (1985). Parasitoid foraging: should parasitism be density dependent? Journal of Animal Ecology, 54, 27-41.

Letourneau, D. K. (1998). Conservation Biology: lessons for conservation of natural enemies. In P. Barbosa (Ed.), Conservation Biological Control (pp. 9-38). San Diego: Academic Press.

MacArthur, R. (1955). Fluctuations of animal populations and a measure of community stability. Ecology, 36, 533-536.

Mason, R., & Torgersen, T. (1987). Dynamics of a nonoutbreak population of the Douglas-fir tussock moth (Lepidoptera: Lymantriidae) in southern Oregon. Environmental Entomology, 16, 1217-1227.

McCann, K. S. (2000). The diversity-stability debate. Nature, 405, 228-233.

Meiners, T., & Obermaier, E. (2004). Hide and seek on two spatial scales - vegetation structure effects herbivore oviposition and egg parasitism. Basic and Applied Ecology, 5, 87-94.

Mensah, R. K., & Madden, J. L. (1994). Conservation of two predator species for biological control of Chrysophtharta bimaculata (Coleoptera: Chrysomelidae) in Tasmanian forests. Entomophaga, 39, 71-83.

Moore, B. P. (1967). Hydrogen cyanide in the defensive secretions of larval Paropsini (Coleoptera: Chrysomelidae). Journal of the Australian Entomological Society, 6, 36-38.

Nahrung, H. (2002). Biological differences between mainland and Tasmanian Chrysophtharta agricola, a Eucalyptus leaf beetle. Tasmanian Naturalist, 124, 56–64.

Nahrung, H., & Reid, C. (2002). Reproductive development of the Tasmanian eucalypt-defoliating beetles Chrysophtharta agricola (Chapuis) and C.

bimaculata (Olivier)(Coleoptera: Chrysomelidae: Paropsini). The Coleopterists Bulletin, 56, 84–95.

Nahrung, H. F. (2006). Paropsine beetles (Coleoptera: Chrysomelidae) in South-East Queensland hardwood plantations: identifying potential pest species. Australian Forestry, 69, 270-274.

Nahrung, H. F., & Allen, G. R. (2003). Intra-plant host selection, oviposition preference and larval survival of Chrysophtharta agricola (Chapuis) (Coleoptera: Chrysomelidae: Paropsini) between foliage types of a heterophyllous host. Agricultural and Forest Entomology, 5, 155-162.

Nahrung, H. F., & Allen, G. R. (2004). Population dynamics of the chrysomelid leaf beetle Chrysophtharta agricola (Chapius), a pest of Eucalyptus nitens plantations in Tasmania. Tasforests, 15, 67-84.

Nahrung, H. F., & Clarke, A. R. (2007). Sexually-transmitted disease in a subtropical eucalypt beetle: infection of the fittest? Evolutionary Ecology, 21, 143-156.

Nahrung, H. F., Dunstan, P. K., & Allen, G. R. (2001). Larval gregariousness and neonate establishment of the eucalypt-feeding beetle Chrysophtharta agricola (Coleoptera : Chrysomelidae : Paropsini). Oikos, 94, 358-364.

Nahrung, H. F., & Murphy, B. D. (2002). Differences in egg parasitism of Chrysophtharta agricola (Chapuis) (Coleoptera: Chrysomelidae) by Enoggera nassaui Girault (Hymenoptera: Pteromalidae) in relation to host and parasitoid origin. Australian Journal of Entomology, 41, 267-271.

Naranjo, S. E. (2001). Conservation and evaluation of natural enemies in IPM systems for Bemisia tabaci. Crop Protection, 20, 835-852.

Noyes, J. S. (1998). Catalogue of the Chalcidoidea of the World. Biodiversity catalogue database and image library CD-ROM series: ETI, Amsterdam and The Natural History Museum, London.

Ohmart, C. P. (1990). Insect pests in intensively-managed eucalypt plantations in Australia: some thoughts on this challenge to a new era in forest management. Australian Forestry, 53, 7-12.

Ohmart, C. P. (1991). Role of food quality in the popluation dymanics of chrysomelid beetles feeding on Eucalyptus. Forest Ecology and Management, 39, 35-46.

Ohmart, C. P., Stewart, L. G., & Thomas, J. R. (1985). Effects of food quality, particularly nitrogen concentrations, of Eucalyptus blakelyi foliage on the growth of Paropsis atomaria larvae (Coleoptera: Chrysomelidae). Oecologia, 65, 543-549.

Olson, D., & Nechols, J. (1995). Effects of squash leaf trichome exudates and honey on adult feeding, survival, and fecundity of the squash bug (Heteroptera: Coreidae) egg parasitoid Gryon pennsylvanicum(Hymenoptera: Scelionidae). Environmental Entomology, 24, 454-458.

Parsons, M., & Gavran, M. (2005). National Forest Inventory 2005. National Plantation Inventory 2005 Update. Canberra: Bureau of Rural Sciences.

Paterson, S. C., Fitzgerald, H., & Matsuki, M. (2004). Technical Report 144: Seasonal pattern of leaf production and its effects on population dynamics of Chrysophtharta bimaculata, C. agricola and their natural enemies: Cooperative Research Centre for Sustainable Production Forestry.

Phillips, C. L. (1993). Insect pest problems of eucalypt plantations in Australia 5. South Australia. Australian Forestry, 56, 378-380.

Pickett, C. H., & Bugg, R. L. (Eds.). (1998). Enhancing Biological Control. Berkley and Los Angeles: University of California Press.

Queensland Department of Primary Industries and Fisheries. (2004). The Queensland Forest Industry: An overview of the commercial growing, management and processing of forest products in Queensland.

Ramsden, N., & Elek, J. (1998). Life cycle and development rates of the leaf beetle Chrysophtharta agricola (Chapuis) (Coleoptera: Chrysomelidae) on Eucalyptus nitens at two temperature regimes. Australian Journal of Entomology, 37, 238-242.

Rebek, E. J., Sadof, C. S., & Hanks, L. M. (2005). Manipulating the abundance of natural enemies in ornamental landscapes with floral resource plants. Biological Control, 33, 203-216.

Reid, C. A. M. (2006). A taxonomic revision of the Australian Chrysomelinae, with a key to the genera (Coleoptera: Chrysomelidae). Zootaxa, 1292, 1-119.

Roland, J., & Embree, D. (1995). Biological control of the winter moth. Annual Review of Entomology, 40, 475-492.

Root, R. B. (1973). Organization of a plant-arthropod association in simple and diverse habitats: the fauna of collards (Brassica Oleracea). Ecological Monographs, 43, 95-124.

Russell, E. P. (1989). Enemies hypothesis: a review of the effect of vegetational diversity on predatory insects and parasitoids. Environmental Entomology, 18, 590-599.

Schutze, M. K., Mather, P. B., & Clarke, A. R. (2006). Species status and population structure of the Australian Eucalyptus pest Paropsis atomaria Olivier (Coleoptera: Chrysomelidae). Agricultural and Forest Entomology, 8, 323- 332.

Seeman, O. D., & Nahrung, H. F. (2005). New Parobia (Acari: Podapolipidae) parasitic on pestiferous paropsine beetles (Coleoptera: Chrysomelidae) in eucalypt plantations. Systematic & Applied Acarology, 10, 111-135.

Selman, B. J. (1989). Some factors including nematodes influencing the population levels of paropsine chrysomelid beetles. Entomography, 6, 403-406.

Selman, B. J. (1994). The biology of paropsine eucalyptus beetles in Australia. In P. H. Jolivet, M. L. Cox & E.Petitpierre (Eds.), Novel aspects of the biology of Chrysomelidae: Kluwer Academic Publishers.

Shaw, M. (2006). Habitat considerations for parasitic wasps (Hymenoptera). Journal of Insect Conservation, 10, 117-127.

Shepherd, B. W. (2001). Chrysomelid leaf-eating beetle Paropsis atomaria (Olivier): its damage to Eucalyptus cloeziana (F.Muell), with identification of potential parasitic control agents in recently established southeast Queensland commercial hardwood plantations forests. Unpublished B.Sc Honors thesis, Sunshine Coast University.

Sillen-Tullberg, B., & Hunter, A. (1996). Evolution of larval gregariousness in relation to repellent defences and warning coloration in tree-feeding Macrolepidoptera: a phylogenetic analysis based on independent contrasts. Biological Journal of the Linnean Society, 57, 253-276.

Simmul, T. L., & de Little, D. W. (1999). Biology of the Paropsini (Chrysomelidae: Chrysomelinae). In M. L. Cox (Ed.), Advances in Chrysomelidae Biology (pp. 463-477). Leiden: Backhuys Publishers.

Steinbauer, M., Clarke, A., & Madden, J. (1998). Oviposition preference of a Eucalyptus herbivore and the importance of leaf age on interspecific host choice. Ecological Entomology, 23, 201-206.

Steinbauer, M., Short, M. W., & Schimdt, S. (2006). Preliminary insights into the influence of age and vegetational composition of eucalypt arboreta on the diversity and abundance of Ichneumonoidea (Hymenoptera) native to southeast mainland Australia. Austral Ecology.

Stone, C. (1993). Insect problems of eucalypt plantations in Australia. Australian Forestry, 56, 363-369.

Stone, C., Matsuki, M., & Carnegie, A. (2003). Pest and disease assessment in young eucalypt plantations: field manual for using the Crown Damage Index, editor Parsons, M. Canberra: National Forest Inventory, Bureau of Rural Sciences.

Sunderland, K. (1999). Mechanisms underlying the effects of spiders on pest populations. Journal of Arachnology, 27, 308-316.

Sunderland, K., & Samu, F. (2000). Effects of agricultural diversification on the abundance, distribution, and pest control potential of spiders: a review. Entomologia Experimentalis et Applicata, 95, 1-13.

Symondson, W. O. C., Sunderland, K. D., & Greenstone, M. H. (2002). Can generalist predators be effective biocontrol agents? Annual Review of Entomology, 47, 561-594.

Tanton, M. T., & Epila, J. S. O. (1984). Parasitization of the larvae of Paropis atomaria Ol. (Coleoptera: Chrysomelidae) in the Australian Capital Territory. Australian Journal of Zoology, 32, 251-259.

Tribe, G. D. (2000). Ecology, distribution and natural enemies of the Eucalyptus- defoliating beetle Trachymela tincticollis (Blackburn) (Chrysomelidae: Chrysomelini: Paropsina) in southwestern Australia, with reference to its biological control in South Africa. African Entomology, 8, 23-45.

Tribe, G. D., & Cillie, J. J. (2000). Biological control of the Eucalyptus-defoliating Australian tortoise beetle Trachymela tincticollis (Blackburn) (Chrysomelidae : Chrysomelini : Paropsina) in South Africa by the egg parasitoid Enoggera reticulata Naumann (Hymenoptera : Pteromalidae : asaphinae). African Entomology, 8, 15-22.

Tribe, G. D., & Cillie, J. J. (2004). The spread of Sirex noctilio Fabricius (Hymenoptera : Siricidae) in South African pine plantations and the introduction and establishment of its biological control agents. African Entomology, 12, 9-17. van Emden, H. F. (1990). Plant diversity and natural enemy efficiency in agrosystems. In M. Mackauer, L. E. Ehler & J. Roland (Eds.), Critical Issues in Biological Control (pp. 63-80). Andover: Intercept.

Verkerk, R. H. J., Leather, S. R., & Wright, D. J. (1998). The potential for manipulating crop-pest-natural enemy interactions for improved insect pest management. Bulletin of Entomological Research, 88, 493-501.

Walde, S., & Murdoch, W. (1988). Spatial density dependence In parasitoids. Annual Review of Entomology, 33, 441-466.

Wang, J., Borsboom, A. C., & Smith, G. C. (2004). Flora diversity of farm forestry plantations in southeast Queensland. Ecological Management and Restoration, 5, 43-51.

Way, M. J., & Khoo, K. C. (1992). Role of ants in pest management. Annual Review of Entomology, 37, 479-503.

Withers, T. (2001). Colonization of eucalypts in New Zealand by Australian insects. Austral Ecology, 26, 467-476.

Wratten, S. D., & van Emden, H. F. (1995). Habitat Management for Enhanced Activity of Natural Enemies of Insect Pests. In D. M. Glen, M. P. Greaves & H. M. Anderson (Eds.), Ecology and Integtated Farming Systems (pp. 117- 145). Bristol: John Wiley & Sons Ltd.

Wylie, F. R., & Peters, B. C. (1993). Insect pest problems of eucalypt plantations in Australia 1. Queensland. Australian Forestry, 56, 358-362.

Zanuncio, J. C., Mezzomo, J. A., Guedes, R. N. C., & Oliveira, A. C. (1998). Influence of strips of native vegetation on Lepidoptera associated with Eucalyptus cloeziana in Brazil. Forest Ecology and Management, 108, 85-90.

Zhang, Z. Q. (1992). The natural enemies of Aphis-gossypii Glover (Hom, Aphididae) in China. Zeitschrift Fur Angewandte Entomologie, 114, 251-262.

Appendix I

A. Estimating egg batch size Due to time constraints involved with counting the number of eggs in a batch it was decided for the 2004/2005 season that only eggs collected off shoots that were part of the natural enemy survey and a random sample of other eggs collected would be counted. Uncounted batches were placed in vials similarly to the counted eggs and were checked daily for emergence. Batches that only had P. atomaria larvae emerge were discarded, while batches with parasitoids present had the number of parasitoids and larvae emerging counted. It was initially planned that this data would be used to calculate the percentage of egg batches parasitised. But due to the correlation between the number of larvae plus parasitoids emerging from egg batches, and egg batch size in counted batches (r= 0.687 n= 378 p< 0.000) the equation of the regression line (y = 0.556x + 45.071 was used to estimate egg batch size). This data was then used in calculations of within batch parasitism rates.

160

120

y = 0.5555x + 45.071

Number of eggs Number 40 R2 = 0.4721

0 0 50 100 150 Number of larvae + parasitoids emerging from batch

Figure 1: Regression of the sum of parasitoids and larvae emerging from each Paropsis atomaria egg batch and the number of eggs counted, used to estimate the number of eggs in batches that were not counted before larval or parasitoid emergence during 2004/2005 season (Chapter 4.).

For the 2005/2006 season the number of eggs in a batch was estimated by using the equation of the regression line when the number of eggs in a batch was plotted against the length of the egg batch multiplied by the number of eggs around the base of the egg collar (y = 1.107x + 0.1643 R2 = 0.7296 p<0.001, where y = number of eggs in batch and x = length of batch * number of eggs around base of collar)

150 y = 1.1017x + 0.1643 R2 = 0.7296 100

Number of eggs in batch of Number 0 20 40 60 80 100 120 140 Length of batch (mm) * eggs around base

Figure 2: Regression of length of Paropsis atomaria egg batches multiplied by number of eggs around base of egg collar against the number of eggs in batch. Used to estimate the number of eggs in a batch during 2005/2006 season

B. Day-degree requirements and developmental thresholds for Paropsis atomaria Carne (1966b) provided data on immature development rates of Paropsis atomaria under different constant temperatures. From the egg stage data presented (stage duration in days at six constant temperatures (°F)) I determined development rate as the reciprocal of development time (1/days). Temperatures were converted to °C using the equation ((temperature°F – 32)/1.8). Development rate was then plotted against temperature (°C), and the linear portion of the curve was used to obtain a regression line y = mx + c, where y = development rate, and x = temperature (Figure 1). The lower temperature threshold (Ltt) for egg development was estimated by solving the regression equation for development rate = 0 (x-intercept). The number of day-degrees (DD) required for eggs to hatch was estimated as 1/m.

0.12 y = 0.0056x - 0.0314 0.1 R2 = 0.9865 0.08 0.06 0.04 0.02

development rate 0 0102030 temperature (C)

Figure 3: Development rate (1/days) of Paropsis atomaria eggs at five constant temperatures (after Carne (1966)) used to estimate lower temperature threshold and day-degree requirements (Chapter 4).

100