APPLICATION FORM Release

To obtain approval to release new organisms (Through importing for release or releasing from containment)

Send to Environmental Protection Authority preferably by email ([email protected]) or alternatively by post (Private Bag 63002, Wellington 6140) Payment must accompany final application; see our fees and charges schedule for details.

Figure 1. Horehound growing in Mackenzie Basin

Application Number

APP203542

Date

2 May 2018

www.epa.govt.nz 2

Application Form Approval to release a new organism

Completing this application form

1. This form has been approved under section 34 of the Hazardous Substances and New Organisms (HSNO) Act 1996. It covers the release without controls of any new organism (including genetically modified organisms (GMOs)) that is to be imported for release or released from containment. It also covers the release with or without controls of low risk new organisms (qualifying organisms) in human and veterinary medicines. If you wish to make an application for another type of approval or for another use (such as an emergency, special emergency, conditional release or containment), a different form will have to be used. All forms are available on our website.

2. It is recommended that you contact an Advisor at the Environmental Protection Authority (EPA) as early in the application process as possible. An Advisor can assist you with any questions you have during the preparation of your application including providing advice on any consultation requirements.

3. Unless otherwise indicated, all sections of this form must be completed for the application to be formally received and assessed. If a section is not relevant to your application, please provide a comprehensive explanation why this does not apply. If you choose not to provide the specific information, you will need to apply for a waiver under section 59(3)(a)(ii) of the HSNO Act. This can be done by completing the section on the last page of this form.

4. Any extra material that does not fit in the application form must be clearly labelled, cross- referenced, and included with the application form when it is submitted.

5. Please add extra rows/tables where needed.

6. You must sign the final form (the EPA will accept electronically signed forms) and pay the application fee (including GST) unless you are already an approved EPA customer. To be recognised by the EPA as an “approved customer”, you must have submitted more than one application per month over the preceding six months, and have no history of delay in making payments, at the time of presenting an application.

7. Information about application fees is available on the EPA website.

8. All application communications from the EPA will be provided electronically, unless you specifically request otherwise. Commercially sensitive information

9. Commercially sensitive information must be included in an appendix to this form and be identified as confidential. If you consider any information to be commercially sensitive, please show this in

` December 2013 EPA0322 3

Application Form Approval to release a new organism

the relevant section of this form and cross reference to where that information is located in the confidential appendix.

10. Any information you supply to the EPA prior to formal lodgement of your application will not be publicly released. Following formal lodgement of your application any information in the body of this application form and any non-confidential appendices will become publicly available.

11. Once you have formally lodged your application with the EPA, any information you have supplied to the EPA about your application is subject to the Official Information Act 1982 (OIA). If a request is made for the release of information that you consider to be confidential, your view will be considered in a manner consistent with the OIA and with section 57 of the HSNO Act. You may be required to provide further justification for your claim of confidentiality. Definitions

Restricting an organism or substance to a secure location or facility to prevent Containment escape. In respect to genetically modified organisms, this includes field testing and large scale fermentation

Any obligation or restrictions imposed on any new organism, or any person in relation to any new organism, by the HSNO Act or any other Act or any Controls regulations, rules, codes, or other documents made in accordance with the provisions of the HSNO Act or any other Act for the purposes of controlling the adverse effects of that organism on people or the environment

Any organism in which any of the genes or other genetic material: • Have been modified by in vitro techniques, or Genetically Modified • Are inherited or otherwise derived, through any number of replications, from Organism (GMO) any genes or other genetic material which has been modified by in vitro techniques

As defined in section 3 of the Medicines Act 1981 Medicine http://www.legislation.govt.nz/act/public/1981/0118/latest/DLM53790.html?src= qs

A new organism is an organism that is any of the following: • An organism belonging to a species that was not present in New Zealand immediately before 29 July 1998; • An organism belonging to a species, subspecies, infrasubspecies, variety, strain, or cultivar prescribed as a risk species, where that organism was not present in New Zealand at the time of promulgation of the relevant regulation; New Organism • An organism for which a containment approval has been given under the HSNO Act; • An organism for which a conditional release approval has been given under the HSNO Act; • A qualifying organism approved for release with controls under the HSNO Act; • A genetically modified organism;

` December 2013 EPA0322 4

Application Form Approval to release a new organism

• An organism belonging to a species, subspecies, infrasubspecies, variety, strain, or cultivar that has been eradicated from New Zealand; • An organism present in New Zealand before 29 July 1998 in contravention of the Animals Act 1967 or the Plants Act 1970. This does not apply to the organism known as rabbit haemorrhagic disease virus, or rabbit calicivirus A new organism does not cease to be a new organism because: • It is subject to a conditional release approval; or • It is a qualifying organism approved for release with controls; or • It is an incidentally imported new organism

Qualifying Organism As defined in sections 2 and 38I of the HSNO Act

To allow the organism to move within New Zealand free of any restrictions Release other than those imposed in accordance with the Biosecurity Act 1993 or the Conservation Act 1987

As defined in section 2 of the Biosecurity Act 1993 Unwanted Organism http://www.legislation.govt.nz/act/public/1993/0095/latest/DLM314623.html?src =qs

As defined in section 2(1) of the Agricultural Compounds and Veterinary Medicines Act 1997 Veterinary Medicine http://www.legislation.govt.nz/act/public/1997/0087/latest/DLM414577.html?se arch=ts_act%40bill%40regulation%40deemedreg_Agricultural+Compounds+a nd+Veterinary+Medicines+Act+_resel_25_a&p=1

` December 2013 EPA0322 5

Application Form Approval to release a new organism

1. Applicant details

1.1. Applicant

Company Name: (if applicable) Horehound Biocontrol Group

Contact Name: Gavin Loxton

Job Title: Chair

Physical Address: Sawdon Station, Lake Tekapo

Postal Address: PO Box 9, Lake Tekapo, 7945

Phone (office and/or mobile): 03 680 6757 / 0274 412 214

Fax: N/A

Email: [email protected]

1.2. New Zealand agent or consultant (if applicable)

Company Name: Richard Hill & Associates

Contact Name: Richard Hill

Job Title: Principal

Physical Address: 237 Ashgrove Tce, Christchurch 8024

Postal Address: N/A

Phone (office and/or mobile): 0211 376 919

Fax: N/A

Email: [email protected]

` December 2013 EPA0322 6

Application Form Approval to release a new organism

2. Information about the application

2.1. Brief application description Approximately 30 words about what you are applying to do

An application to introduce two moths, Wheeleria spilodactylus and Chamaesphecia mysiniformis, as biological control agents for the weed horehound (Marrubium vulgare).

2.2. Summary of application Provide a plain English, non-technical description of what you are applying to do and why you want to do it

This is an application by the Horehound Biocontrol Group to introduce two biological control agents for the weed horehound.

Horehound (Marrubium vulgare) is a small, woody herb belonging to the mint family. It is related to sage, mint, thyme, and a range of other herb species. Extracts of horehound have commonly been used in cough remedies, and it is thought to have other medicinal benefits. It was probably introduced to New Zealand for medicinal purposes and was recorded in the wild as early as 1867. Horehound is associated with dry environments and is abundant in Canterbury and Otago.

Horehound establishes easily in disturbed sites, from which it encroaches into adjacent farmland, displacing desirable forage. It also produces burs that mat wool, devaluing the fleece of sheep and causing distress in lambs. It establishes strongly in hill- and high-country lucerne forage crops and is exceptionally hard to control. Not only does horehound reduce the productivity of lucerne, but herbicide treatments are highly detrimental to the survival and regeneration of lucerne. Farmers report that horehound is spreading at an alarming rate into lucerne crops and other pasture types in hill- and high-country farms. Overall, horehound is a serious threat to the viability of some farming enterprises. A recent survey conservatively estimated the cost of horehound to dryland farming in New Zealand at $6.85 million per annum.

A viable solution is biological control, which involves introducing a natural enemy in order to control a weed such as horehound. Introduced natural enemies become permanently established in the environment, and their effects on a target weed persist from year to year. This means these natural enemies must be safe if such a weed management tactic is to be environmentally acceptable. Horehound plume moth (Wheeleria spilodactylus) and horehound clearwing moth (Chamaesphecia mysiniformis) have been released in Australia. The larvae of the plume moth feed on horehound leaves, and the larvae of the clearwing moth feed in the roots. Approval to introduce the two moths to

` December 2013 EPA0322 7

Application Form Approval to release a new organism

Australia was obtained following detailed research to determine the range of plants these species could live on in the wild. Both moths have very narrow host ranges in their native Europe, and this was confirmed in the experiments. These conclusions can also be applied in New Zealand. A degree of horehound control has now been achieved in Australia, and no adverse effects have been reported.

Section 5 of the application explains the expected positive effects of the biological control of horehound in New Zealand, including:

• reduced risk of invasion of horehound into new sites • reduced costs and adverse environmental consequences of controlling horehound using herbicides • improved productivity of pastures and lucerne on dry hill- and high-country farms, especially in the South Island • reduced damage to wool from matting with horehound burs.

This application summarises the results of tests to confirm the range of plants susceptible to the two control agents, an assessment of the beneficial and adverse effects of the proposal, and the results of consultation with the community. The potential effects on Māori cultural values are also presented.

Experimental evidence indicates that neither control agent will persist on or damage any native plant or desirable ornamental. Horehound is valued as a medicinal herb, and it is harvested for this purpose. Successful biological control would adversely affect the value of this harvest. Otherwise, the application concludes that introduction of the control agents will have no adverse effect on ecological, environmental, social or community values in New Zealand.

Overall, the potential benefits of introducing Wheeleria spilodactylus and Chamaesphecia mysiniformis to New Zealand outweigh the potential risks and costs. Other information can be found on the Manaaki Whenua – Landcare Research website at http://www.landcareresearch.co.nz/science/plants- animals-fungi/plants/weeds/biocontrol/approvals/current-applications.

2.3. Background and aims of application This section is intended to put the new organism(s) in perspective of the wider activitie(s) that they will be used in. You may use more technical language but all technical words must be included in a glossary.

2.3.1 Purpose

This application seeks approval to introduce two new organisms, the horehound plume moth (HPM), Wheeleria spilodactylus, and the horehound clearwing moth (HCM), Chamaesphecia mysiniformis, for the biological control of horehound, Marrubium vulgare. The application is submitted by the Horehound Biocontrol Group (HBG), a collective of farmers adversely affected by the weed, supported by the Sustainable Farming Fund of the Ministry for Primary Industries. The HBG has determined that biological control is the most environmentally acceptable and sustainable means of permanently

` December 2013 EPA0322 8

Application Form Approval to release a new organism containing the adverse economic impacts of horehound in New Zealand. Manaaki Whenua – Landcare Research (MWLR) is the science provider for this development, and has contracted Richard Hill & Associates to prepare this application and to manage the application process on behalf of the HBG.

Biocontrol would complement other control methods used to mitigate the negative impacts of this weed because biocontrol agents will persist once established, offering the potential to:

• suppress horehound plants in areas where herbicide is not used, reducing competition with forage species • reduce the damage caused by herbicides used for horehound control in forage lucerne and other vegetation types • reduce seed production, leading to a reduced rate of spread and possibly to population decline.

However, if this outcome is achieved, opportunities for harvesting wild horehound for medicinal purposes may be compromised or reduced (see section 5.3.2). Manaaki Whenua – Landcare Research (2018b) describes the potential risks, costs and benefits of the proposal.

2.3.2 Biology and pest status of horehound

Biology of horehound Horehound is an erect perennial herb or sub-shrub in the mint family, Labiatae (the preferred name for the family formerly called Lamiaceae). The heart-shaped, wavy-edged leaves are silvery-green, arranged in opposite pairs. Leaves are wrinkled on the upper surface and densely covered with short, white, globular hairs on the lower surface (Figure1). The fruit form brown burrs with small hooked spines (Figure 2). Each burr contains up to four spear-shaped seeds, 1–2 mm long. Seeds are dispersed when burs attach to passing animals.

Figure 2. Seeds in horehound bur

Burs also lodge in the socks of walkers brushing past the weed, and there is a significant risk of invasion of the conservation estate mediated by humans. Horehound has become a significant weed

` December 2013 EPA0322 9

Application Form Approval to release a new organism of the walking tracks of Mt Maunganui, and burs mat the coats of family dogs (https://www.stuff.co.nz/business/farming/83781970/the-horehound-is-loose-in-paradise).

Horehound invasion of stream margins indicates that seeds are also carried by water. Each plant can produce over 1,000 seeds per annum, and in Australia there can be over 15,000 seeds per square metre in the soil under horehound (Weiss & Sagliocco, 2000). Horehound is tolerant of drought, and its seeds can germinate at low temperatures.

The family Labiatae also contains plant species valued for culinary use, such as thyme, sage, mint, oregano, basil, rosemary, and many others. The family is represented in New Zealand’s native flora by five species in five genera, all relatively distantly related to the genus Marrubium. Two species are endemic, and three species are indigenous but not endemic.

Groenteman et al. (2017) have reviewed the biology, distribution and worldwide pest status of horehound in detail. The information summarised in this application is based on that study.

Horehound is native to temperate Eurasia, Europe, the Middle East and the Mediterranean region, including North Africa. It has become a weed in southern parts of North America (California, Texas), in South America (Argentina, Chile, Peru, Uruguay), in Australia, and in New Zealand. Marrubium vulgare is presumed to have been introduced to New Zealand deliberately for its medicinal properties, and was first recorded as naturalised in 1867. Horehound can now be found throughout New Zealand, but it is relatively rare north of the volcanic plateau in the North Island. It is associated with moderately fertile sites in dry environments and is abundant in inland Canterbury and Otago.

Figure 3. Occurrence of horehound in National Vegetation Survey plots (nvs.landcareresearch.co.nz/data/search)

` December 2013 EPA0322 10

Application Form Approval to release a new organism

Horehound has been the subject of a long-standing biological control programme in Australia (Sagliocco, 2000; Weiss & Sagliocco, 2000, 2012). A first consignment of a French ecotype of HPM collected at Cap d’Agde, France, was approved for release in Australia in late 1993, and by early 1995 the insect had been released at a number of sites. In 1996 a population originating from dry regions of Spain with long, hot summers was shipped to Australia. In 1997 both ecotypes were released at over 100 sites in the four south-eastern Australian states. By 1998 HPM had established at 90% of sites and the insect had dispersed over 20 km, and 41% of the sites were used for redistribution programmes (Weiss & Sagliocco, 2012). HCM was released in Australia in 1997, and by 2003, 50.3% of plants at the original release site were attacked (Sagliocco & Weiss, 2004). Larvae were found in 29.2% of plants sampled over an extensive area (Weiss & Sagliocco, 2012).

Adverse effects of horehound

Horehound establishes easily in disturbed sites, such as roadsides and sheep camps, from which it encroaches into adjacent farmland. It can form virtual monocultures in sparsely vegetated environments such as high-country pastures (Webb et al. 1988). The aggressiveness of horehound varies from place to place. Many farms record an explosion of the weed in the past few years, while others can manage infestations simply by grubbing isolated plants. The long tap-root makes horehound a strong competitor in dry conditions. When horehound is treated in other vegetation, soil is often bared, resulting in soil erosion. Horehound also tends to be the first plant to regenerate following chemical control. It is a specialist invader of dry environments, which may become more prevalent with climate change (Gerard et al., 2013).

We find every time we break the ground, the horehound is the fastest thing to grow back, beating all other seeds. We have tried to put in lawns and a track, and now these are covered in horehound, we have only been at our current site 1 year. Problem increasing exponentially over the last 3 years.

I am shocked at how it is appearing in other areas of my farm so quickly. Currently found around river banks but spreading quickly into pasture. The last two years of dry-drought has seen the horehound on our property grow exponentially. After multiple attempts at control we have lost the war in the most productive areas of our farm. We have tried all methods commercially available, from ground to air application, sought professional advice (on and off label recommendations) and used contractors where required. The results have been very disappointing. We look forward to the findings of your study and survey. [It is] easily controlled by grubbing if you never let it get away Been Horehound here as long as I can remember, I am 65. Mainly confined to fringes of shelter belts, stock camps. Not a problem for us in Lucerne. Incidentally can be fermented and brewed. Horehound beer, not a bad drop!!!

Intensive cattle grazing controls most of the hore hound on this property. (Sample of survey responses in Groenteman, 2018)

I have been farming here for 20 years. Over this time, we have seen an exponential increase in the horehound plant population … When we began farming, 20 years ago, horehound was just considered a

` December 2013 EPA0322 11

Application Form Approval to release a new organism

“nuisance weed”, we could do very little about. Over the last 20 years we have been developing dry land pastures to more drought tolerant species such as lucerne, lupins and cocksfoot. With this management change there has been significant increase in the horehound plant population. Both in the developed and undeveloped areas. (G. Loxton, Sawdon Station)

Lucerne is a key element in the viability of dry hill- and high-country farming in New Zealand. Not only does horehound directly reduce the productivity of lucerne, but treatment with metsulfuron-methyl, the only herbicide capable of containing horehound, is highly detrimental to the survival and persistence of lucerne (Groenteman, 2018; section 5.3.1. Young lucerne stands are susceptible to the chemical in their first 3–4 years, so if horehound becomes established in newly sown lucerne, chemical control is not an option. Metsulfuron-methyl can remain active in the soil for several seasons following application in the cool, dry regions where horehound is invasive. This means legumes cannot be sown in the sprayed area, but horehound often regrows from the seedbank during this period. Horehound occupies space and so also reduces the productivity of other pasture types (Groenteman et al., 2017).

Not only does horehound displace productive forage; its bitter foliage is distasteful to grazing animals and it can be toxic to stock in its own right.

Horehound is a significant problem in that the seeds attach to the fleece and downgrade the value of the wool, and long-term exposure to horehound plants can cause ill thrift in sheep as the toxic alkaloids build up in the liver of the animal. Spraying horehound from developed areas of the farm has become integrated into our farm management program. The undeveloped areas which grow horehound are impossible to spray with any positive economic impact, so have been abandoned to the horehound weed. (G. Loxton, Sawdon Station).

Because horehound is so hard to kill and over the last 5 years is taking over our lucerne it is making farming lucerne too expensive for us to farm. Our main concern is in dry steep hill country where control is difficult. It is also encroaching on areas of chicory-clover pastures where control is as difficult as in Lucerne. Aerial or boom spraying has had initially ok results but reinvasion occurs rapidly. You are very wrong stating that the problem is mainly in lucerne. Large areas of otherwise productive hill country are being lost to horehound. And there is no real control mechanism available. Thus encroachment on the hill country is a much bigger problem because we can’t do anything about it. Our horehound problem is not a large area however we are finding it very difficult to control and stop spreading over our property. Also have mixed results from spraying. We have large areas on hill we spray by Heli, this really only contains at best, it is encroaching large areas. We have a challenge on new pasture that does not include lucerne (in Fescue clover paddocks). Any chemical control has a detrimental effect to existing sward and really only contains it for a couple of years. (Samples of survey responses in Groenteman, 2018)

Mechanical control has a place in horehound management, but its use is limited by challenging terrain and large areas of infestation. Although frequent cultivation can remove horehound, it is not economically viable in low-intensity farming and/or is technically impossible on much of the steep terrain that is infested.

` December 2013 EPA0322 12

Application Form Approval to release a new organism

Not only lucerne but grasses are affected. Biggest cost is the limiting of crops and things that can be grown. Having to prep paddocks with Metsulferon and stand down period have great cost and seed bank in soil is heavy. [Metsulfuron-methyl] provides 90% control initially but doesn't provide a solution that lasts. Chemical effect on lucerne is not only killing, but also slows production. The rate at which it is coming into our young paddocks is major concern. (Sammple of survey responses in Groenteman, 2018)

Would it be cheaper, more sustainable and also provide employment opportunities to have students from Otago manually grub out horehound instead of the current spray regime or the introduction of a bio control? This option would connect urban people with rural people and perhaps foster some understanding into each other’s beliefs. (S. Parker, medical herbalist, in Groenteman et al., 2017)

Marrubium vulgare fruits form burrs with small hooked spines (Figure 2) that attach easily to wool. Heavy contamination is common, more than doubling the cost of processing the fleece (Groenteman et al., 2017). In the absence of M. vulgare contamination, New Zealand wool is globally renowned for cleanliness and near-freedom from organic matter, but the presence of M. vulgare downgrades wool to the lowest grade, which has to be manually separated from non-contaminated wool and shipped to Australia for processing. Grazing animals also contribute to seed dispersal via movement of burrs and seeds attached to their fleeces.

In my situation it’s the combination of being highly rabbit prone with low rainfall and the primary sites for its establishment being non arable creating the seed bank to gradually contaminate the rest of the property. There are high costs in lost income due to wool contamination that leads to discounting. (Sample of survey responses in Groenteman, 2018)

Groenteman (2018) reports the results of a survey to estimate the value of some of the adverse effects of horehound on hill- and high-country farming in New Zealand. The combined annual costs of horehound to the 95 survey respondents can be calculated reliably at $316,250. If their experience is typical of all dryland hill- and high-country farmers, production losses and costs of control can be estimated at $6.85 million (Groenteman, 2018). This is a conservative estimate because it does not account for: • the increased costs of treating wool contaminated with horehound burs • production losses in forage types other than lucerne • opportunity costs of changes in farming practice to avoid the effects of horehound.

In Victoria and South Australia horehound is a significant weed, both in pastures and in disturbed natural habitats (Sagliocco, 2000). Horehound is more common on land with little or no competition from other plants (Sagliocco, 2000). Its dominance in native habitats in Australia is related to disturbance (Weiss & Wills, 2000), and the plant spreads easily in areas that experience frequent tourist visits, such as camping grounds and along creek lines (Clarke et al., 2000). It is likely that horehound will become an increasing problem in similar amenity areas in southern New Zealand, such as track margins, bike tracks, roadsides and carparks.

` December 2013 EPA0322 13

Application Form Approval to release a new organism

Beneficial attributes of horehound

Marrubium vulgare leaves contain marrubiin, an alkaloid that imparts medicinal properties to the plant. Groentemann et al. (2017 reviewed much of the literature concerning these properties, and found that horehound has been described as beneficial in the treatment of respiratory disorders (such as bronchitis, colds and asthma), diabetes, inflammatory disorders, gastro-enteric disorders, as well as treatment of appetite loss, indigestion, gut and liver conditions, leprosy, foot ache, and other conditions. In addition, the plant is reported to possess hypoglycaemic, hyperglycaemic, and antihypertensive qualities. It has been said to be an analgesic, an antioxidant, an anti-ulcer agent, an antispasmodic, to have antifungal and antimicrobial properties, and to reduce cholesterol, protect against cardiovascular disease, stimulate the immune system, and treat tumours. The plant has also been studied for dementia-preventing properties and was shown to have an abortifacient effect (causing abortion) in rats.

The evidence for these claims is varied (Groenteman et al., 2017), but there is no question that horehound is highly valued as a medicinal herb in its native range, its introduced range and in New Zealand.

Our business model is based on using New Zealand naturalized and native plants wherever possible. Our plant medicine products are sought after which is reflected in our double-digit growth year on year. (S. Clair, Artemis, in Groenteman et al., 2017) The issue at play here seems to be what, we as a culture, term to be a weed or pest species and what other introduced species we view as not being pests. …. As a herbalist, I view Horehound not as a pest species, or a weed, but as a valuable economic and also culturally significant medicinal plant. …. Harvesting wild horehound, to me, is also important to my livelihood. (S. Parker, in Groenteman, 2018) It really seems a terrible and tragic thing to eradicate such a wonderful, useful, available herb that hopefully in future, can be entrenched in the public mindset, as free, available, and valuable medicine. (C. Epps, in Groenteman et al., 2017)

Wild horehound is harvested for medicinal use in New Zealand, and the potential impact of biological control on this value is discussed in section 5.3.2.

The value of sales of horehound in the US through mainstream channels was US$106 million in 2014. This positions the herb as the top-selling herbal dietary supplement in that market, and double the sales value of the next-best-selling supplement.

A voluntary online survey of herbalists in Australia (where horehound control agents have been present for over 20 years) was initiated through websites and social media in January 2018 (R. Groenteman, pers. comm.). The purpose of the survey was to find out how horehound is sourced there, and to elicit comments about the impact of HPM and HCM on wild harvest in Australia. No responses have been received. As yet there is no evidence that wild horehound was harvested by Australian herbalists before control agents were introduced or that it is harvested now. No issues were raised in Australia when HPM and HCM were approved for release.

` December 2013 EPA0322 14

Application Form Approval to release a new organism

We had very little resistance to the declaration of horehound as a weed for biological control, no comments from herbalists – but some minor comment from apiarists who thought it was ‘another nail in the coffin’ for the bee industry due to its spring/early summer flowering which could be used by bees to quickly build up the colony/hive. (J. Weiss, horehound biocontroller, Australia, pers. comm.)

The plant has also been studied for its potential to rehabilitate soils from salinity and heavy metal contamination, as a pesticide, as an improver for cereal crops, as a food preservative, as an ingredient in beer brewing, and for inhibiting corrosion. Horehound does not appear to be highly valued as a nectar source for bees in New Zealand (Manaaki Whenua – Landcare Research, 2018e).

3. Information about the new organism(s)

3.1. Name of organism Identify the organism as fully as possible

Non-GMOs - Provide a taxonomic description of the new organism(s).

GMOs – Provide a taxonomic description of the host organism(s) and describe the genetic modification.

Both - • Describe the biology and main features of the organism including if it has inseparable organisms. • Describe if the organism has affinities (e.g. close taxonomic relationships) with other organisms in New Zealand. • Could the organism form an undesirable self-sustaining population? If not, why not? • How easily could the new organism be recovered or eradicated if it established an undesirable self- sustaining population?

3.1.1 Taxonomy

Agent 1

Order: Lepidoptera

Family: Pterophoridae

Genus: Wheeleria

Species: spilodactylus (Curtis, 1827) (https://fauna- eu.org/cdm_dataportal/taxon/52559b97-5349-4e3f-b93b-8a08bd4410a1)

Common name: Horehound plume moth (HPM)

` December 2013 EPA0322 15

Application Form Approval to release a new organism

Agent 2

Order: Lepidoptera

Family: Sesiidae

Genus: Chamaesphecia

Species: mysiniformis (Boisduval, 1840) (https://fauna- eu.org/cdm_dataportal/taxon/1cca2295-ef91-4426-b1c1-4293932a0b62#synonymy)

Common name: Horehound clearwing moth (HCM)

Founding populations of the two agents will be obtained from Australia. 3.1.2 Biology and native range

Horehound plume moth

Pterophoridae are small moths with long, fragile bodies, which are covered by small scales (Figure 4). They are mainly active at dusk, Each female moth lays about 100 eggs on the underside of the wrinkled horehound leaves over a two-week period, singly or in groups of up to four. Eggs hatch after about one week and young caterpillars feed in the developing shoot tip. Older larvae feed on the shoot tips and leaves, progressively defoliating the stem. They also feed on the stem’s other soft tissues and shoot apices, often causing abortion of buds, thus affecting plant vigour and reducing the number of flowers and seeds produced. The generation time in summer in Australia is approximately 48 days, and so the plume moth is likely to have at least two generations each year in New Zealand. The insect overwinters as small larvae in leaf buds, completing development in the spring.

HPM has been recorded feeding on M. vulgare and Ballota nigra in Europe (Groenteman et al., 2017).

Figure 4. Horehound plume moth, Wheeleria spilodactylus. (photo: inaturalist.org)

` December 2013 EPA0322 16

Application Form Approval to release a new organism

Horehound clearwing moth

The horehound clearwing moth, C. mysiniformis, has one generation per year (univoltine). Adult moths (Figure 5) emerge from the soil in late spring. In Europe, Sagliocco and Coupland (1995) found that females lay eggs at the base of plants over 7 to 12 days, with an average of 96 eggs per female. Larvae crawl to the base of the plant and chew a gallery, allowing entry to the root. Larvae then develop through summer, autumn and winter within the root. The last instar larva burrows an exit hole at the collar level or at the base of an erect stem in early spring, prior to pupating. Adults emerge from the root in late spring. Normally only one larva develops per root. The moth causes plants to die both directly, by disrupting vascular flow, and indirectly, through secondary infection by pathogens. In its native range HCM was observed to attack 50–90% of two- to three-year-old plants (Sagliocco & Coupland, 1995).

Experience in Australia suggests that HCM requires warm summer conditions to develop significant populations.

Clearwing moths are generally host specific, and HCM has not been recorded from any plant other than M. vulgare in its native Europe.

Figure 5. Horehound clearwing moth, Chamaesphecia mysiniformis (photo: Daniel Morel, http://www.leps.it/indexjs.htm?SpeciesPages/ChamaMysin.htm)

` December 2013 EPA0322 17

Application Form Approval to release a new organism

3.1.3 Affinities with the New Zealand fauna

Horehound plume moth

New Zealand has a rich fauna of native pterophorid moths (plume moths). For example, Dugdale (1988) listed four genera and 18 species of pterophorids present in New Zealand. Most of the specimens listed by Dugdale were collected from the North Island but some native pterophorid species occur in the high-country environments invaded by horehound. The genus Wheeleria does not appear to be present in New Zealand. Like many plume moths, the larvae of HPM feed on leaves and are physically exposed to parasitism and predation. The native parasitoids of the native pterophorid fauna have not been summarised, but two exotic species, Trigonospila brevifacies (Tachiidae) and Meteorus pulchricornis (Braconidae), have been recorded attacking pterophorid larvae.

The ragwort plume moth (Platyptilia isodactyla) was approved for release in 2005 (NOR05002) and is now contributing to substantial control of ragwort on the West Coast and in Southland. No parasitoids have been recorded from field-collected larvae in New Zealand.

There have been no reports of indigenous natural enemies regulating populations of HPM larvae in Australia.

Horehound clearwing moth

There is only one member of the family Sesiidae present in New Zealand: the exotic currant clearwing moth, Synanthedon tipuliformis. This is a pest of blackcurrants and redcurrants in New Zealand. Newly hatched larvae burrow into the stems and complete their life cycle there. Like HCM, the larvae of currant clearwing occur deep in the plant and are largely unavailable to natural enemies. Several parasitoids have been recorded, but their occurrence is rare.

Faunal associations A survey is in progress to assess the flora and fauna associated with horehound in New Zealand. The purpose of the survey is to determine whether HPM or HCM are already present in New Zealand, and to assess whether there are native species present on horehound that might be adversely affected by the successful control of horehound. In the South Island, three sites were visited in summer and another three in autumn. Almost no foliage damage was observed on the c.150 plants examined. Only four lepidopterous larvae were collected (L. Smith, Manaaki Whenua – Landcare Research, pers. comm.). Several of these larvae later died without feeding on horehound and may have been vagrant visitors on the horehound. Several adult hymenopterous parasitoids were captured, which may or may not be associated with lepidoptera. The generalist fauna of aphids and bugs normally associated with such herbs? was also poor, as were the populations of generalist predators such as spiders and coccinellids.

At some North Island survey sites the exotic sage leafhopper E. melissae was present in large numbers and caused noticeable damage. This leafhopper has a wide range of hosts. Damage caused

` December 2013 EPA0322 18

Application Form Approval to release a new organism by other herbivorous species was minimal. Significant numbers of predatory spiders, lacewings, ladybirds and mites were present at some sites, possibly as a result of high sage leafhopper populations. Larvae from several lepidopteran families were collected, but damage attributable to lepidopteran larvae was minimal. No pterophorid or sesiid larvae were found. No parasitoids have yet been reared from the lepidopteran larvae collected (C. Winks, Manaaki Whenua – Landcare Research, pers. comm.) .

Surveys indicate that HPM or HCM are not already present in New Zealand. There is no evidence that horehound is a significant host for any native insects. Horehound does not support high populations of lepidopteran larvae. Specialist natural enemies of lepidoptera do not appear to be abundant, and trophic relationships based around horehound appear to be simple. 3.1.4 Potential for safe biological control

Why biological control?

Horehound causes moderate adverse economic effects in the South Island of New Zealand. Biological control is an appropriate tactic to apply against horehound. Once established, introduced natural enemies colonise and damage the plant wherever it occurs, are widespread, and persist from year to year. Any benefits of biological control will accrue even in areas where it is not feasible to deploy other treatments.

The establishment of Wheeleria spilodactylus and Chamaesphecia mysiniformis would be an irreversible change. It would not be feasible to eradicate unwanted populations by the time they are detected. The agent must therefore be safe to introduce, and in particular must not:

• significantly affect populations of valued plants, whether native or non-native

• significantly displace native species in their natural habitat

• cause significant deterioration of natural ecosystems (see section 5.1).

Wheeleria spilodactylus and Chamaesphecia mysiniformis were approved for release and introduced to Australia from 1993 and 1997, respectively. Both have established widely. The effect of HPM and HCM on horehound varies from place to place. The biocontrol programme had not delivered measurable economic benefits by 2006 (Page & Lacey, 2006), but by 2012 it appeared that the existing agents may provide sufficient control and no further agents would be required (Weiss & Sagliocco 2012). This is still the prevailing opinion. There have been no formal assessments of control success, but farmers in some affected areas of Australia report that horehound has declined to insignificant levels as a result of biological control (G. Loxton, pers. comm.). There have been no reports of adverse economic or environmental effects of the two control agents in Australia.

Selection of plants for host-range testing

Marrubium vulgare L. is in the family Labiatae. Five genera in this family have representatives in New Zealand’s indigenous flora (Table 1): Scutellaria, Mentha, Plectranthus, Vitex, and Teucridium.

` December 2013 EPA0322 19

Application Form Approval to release a new organism

Although these genera belong to the same family as horehound, none belong to the same subfamily and none are closely related (Figure 6; Li et al., 2016). Marrubium is in the subfamily Lamioideae, Scutellaria is in Scutellarioideae, Teucridium is in Ajugoideae, Mentha and Plectranthus are in Nepetoideae, and Vitex is in Viticoideae. Of the native Labiatae genera, Scutellaria is therefore the most closely related to Marrubium; however, the subfamilies are significantly differentiated, and individuals in different subfamilies are not closely related. Each of the five genera with New Zealand indigenous species are represented by a single species (Table 1).

This considerable phylogenetic distance between M. vulgare and New Zealand’s indigenous flora suggests that potential biocontrol agents are unlikely to pose a threat to native species.

In Australia, plants to be tested against the two agents were selected according to the universally accepted centrifugal phylogenetic method of Wapshere (1974). This protocol is based on the premise that insects with a specialised diet are more likely to attack plants that are closely related to the observed host than plants that are less closely related (Briese & Walker, 2002; Sheppard et al., 2005). Plant selection centred on those species closely related to horehound, but also included (at the insistence of Australian regulatory authorities) many unrelated plants of commercial and domestic importance. The list of plants to be tested against each control agent varied slightly because of the taxonomic framework in place at the time, but this had no influence on the conclusions drawn from tests.

Figure 6. Phylogenetic relationships between subfamilies in the family Labiatae. The subfamily that contains Marrubium vulgare is highlighted in red, while the subfamilies that contain New Zealand native genera are highlighted in green. The subfamilies are significantly differentiated, and individuals in different subfamilies are only distantly related. Adapted from Li et al. 2016; subfamily placements for each genus were taken from http://www.mobot.org/MOBOT/Research/APweb/genera/lamiaceaegen.html

` December 2013 EPA0322 20

Application Form Approval to release a new organism

There are five native New Zealand plant species in the family Labiatae. One of these species is also present in Australia, and was tested. The other four have not been tested, but related species were included in the tests conducted in Australia before the agents were released there. These test plants can be regarded as surrogates for the New Zealand native species (Table 1). Manaaki Whenua Landcare Research considered that no additional testing was required to assess the risk to New Zealand native species.

Table 1. Surrogates for New Zealand native plant species included in host-range tests conducted in Australia

Subfamily Surrogate species tested in Australia against NZ native species Wheeleria spilodactylus Chamaesphecia mysiniformis

Mentha cunninghamii Nepetoideae Mentha diemenica Mentha diemenica Mentha spicata Mentha spicata Mentha australis Teucridium parvifolium Ajugoideae Teucrium racemosum Teucrium racemosum Teucrium corymbosum Ajuga australis Ajuga australis Scutellaria novae-zelandiae Ajugoideae Scutellaria humilis Scutellaria humilis Vitex lucens Viticoideae Vitex trifolia Vitex trifolia

Plectranthus parviflorus Ocimieae Tested Tested

The plants tested against Wheeleria spilodactylus are listed in Table 2, while those tested against Chamaesphecia mysiniformis are listed in Table 3.

Predicting the host range of Wheeleria spilodactylus (HPM) in New Zealand

Methods

Larval starvation tests were conducted in quarantine at the Keith Turnbull Research Institute at Frankston, Australia, in 1992/93. Ten first instar larvae were placed on the growing shoots of test plants. This was replicated eight times. Plants were examined after 10, 20 and 30 days and the number of larvae remaining was recorded. Development on horehound controls was completed by 30 days.

Results

The results are summarised in Table 2. Larval feeding was observed only on horehound controls and two of the test plants, both of which were closely related to Marrubium vulgare within the tribe Marrubieae. One other member of the tribe was not attacked. An average of 4.9 of the 10 larvae placed on horehound controls completed development at day 30 (range 2–10; n = 18). Development took longer on the closely related M. supinum, and only an average 2.1 of the 10 larvae applied completed development (range 0–7; n = 8). Mortality on black horehound, Ballota nigra, was high, with only 10% of larvae alive after 30 days. None had developed beyond third instar at the end of 30 days. Plants belonging to other subfamilies of the Labiatae were not attacked, nor were 30 test species

` December 2013 EPA0322 21

Application Form Approval to release a new organism belonging to 22 other plant families. There was no survival of larvae on plants in the sub-family Nepetoideae, which contains important culinary herbs such as thyme, sage, mint, and oregano. There was no attack on the native Plectranthus parviflorus, or on any of the plants designated as surrogates for other New Zealand native species.

Table 2. Survival and development of first instar Wheeleria spilodactyla larvae after 30 days, and damage to test plants

Mean % Larval survival Complete feeding to 30 development FAMILY Tribe Test species Common name observed days at 30 days? Subfamily

LABIATAE Lamioideae Marrubieae Marrubium vulgare Horehound Yes 49 100% (control) M. supinum Yes 21 76% Ballota nigra Yes 10 0% Leucadeae Leonotis nepetifolia No 0 No Nepetoideae Salveae Salvia officinalis Sage No 0 No S. uliginosa No 0 No Ornamental S. 'purple velvet' sage No 0 No Mentheae Mentha diemenica* Wild mint No 0 No M. spicata* spearmint No 0 No Origanum vulgare Oregano No 0 No Prunella vulgaris No 0 No Thymus vulgaris Thyme No 0 No Ocimieae Ociminum basilicum Basil No 0 No Plectranthus parviflorus* No 0 No Rosmarineae Rosmarinus officinalis Rosemary No 0 No Lavanduleae Lavandula spicalis Lavender No 0 No Ajugoideae Ajuga australis* No 0 No Teucrium racemosum* No 0 No T. corymbosum* No 0 No Scutellaria humilis* No 0 No Prostantheroideae Prostanthera ovalifolia No 0 No P. rotundifolia No 0 No No 0 No Hemiandra pungens No 0 No Westringia fruticosa No 0 No Viticoideae Vitex trifolia* No 0 No No 0 No OTHER FAMILIES No 0 No 22 families 30 species No 0 No

* New Zealand native or surrogate for New Zealand native species.

` December 2013 EPA0322 22

Application Form Approval to release a new organism

Conclusions

Wheeleria spilodactylus has only been reported on horehound and on Ballota nigra in its native Europe. The test results confirmed this narrow host range. Larvae did not feed on plants outside the subfamily Lamioideae. Within the subfamily, feeding was restricted to the tribe Marrubieae. One plant in the neighbouring tribe Leucadeae (Leonotis) was not susceptible to HPM larvae. Further, larvae developed faster on horehound controls than on the two other species in the tribe that supported larval feeding. Results therefore indicate that the moth is restricted to particular species within the tribe Marrubieae, and horehound is the preferred host. The risk of colonisation of plants other than Marrubium species following release is negligible.

Predicting the host range of Chamaesphecia mysiniformis (HCM) in New Zealand

Methods

Five newly emerged larvae were transferred onto each test plant using a fine brush. There were usually at least five replicates (25 larvae) for each plant species, though availability of plants limited some tests to four replicates. Test plants and controls were maintained in a glasshouse at 22–28°C. Root crowns were dissected after two to three weeks when the number of initial attacks and the survival of larvae were recorded (Table 3).

Results

Larvae of C. mysiniformis were found to attack and survive on four of the seven species of Marrubium tested. Of the larvae placed on M. vulgare, 84% were still alive after 14–21 days. Marrubium supinum is sufficiently closely related to M. vulgare that the two species can hybridise. Survival on this species was 36%. Lower survival was recorded on the other Marrubium species (Table 3). Three of the 25 larvae placed on black horehound, Ballota nigra, survived this period. All of these plants belong to the tribe Marrubieae in the subfamily Lamioideae. A single larva survived on Stachys arvensis (4%), which belongs to the related tribe Stachydeae. The larvae live inside root crowns and so test plants were destroyed during analysis. At the time of examination larvae on all attacked plant species had reached only first or second instar when the test was terminated. It is not known whether larvae found in plants other than M. vulgare would have completed development.

Small initial attacks were recorded on other plants as larvae tried to establish, but no larvae survived on plants of any other subfamilies of the Labiatae, or in plants from other families (Table 3; Sagliocco & Coupland, 1995). There was no survival on any species in the subfamily Nepetoideae, which includes most of the important culinary herbs such as thyme, sage, mint, oregano, basil, rosemary and lavender (Table 3). There was no survival of larvae on the plants designated as surrogates for New Zealand native species.

` December 2013 EPA0322 23

Application Form Approval to release a new organism

Table 3. Survival and development of first instar Chamaesphecia mysiniformis larvae after 14–21 days on a range of test plants (from Sagliocco & Coupland, 1995)

Labiatae Tribe Test species Common name Initial larvae % larval survival Subfamily to 14–21 days

Lamioideae Marrubieae Marrubium vulgare Horehound 25 84

(control) M. supinum 25 36 M. alysson 25 28 M. incanum 25 8 M. leonuroides 25 8 M. friwaldskyanum 20 0 M. velutinum 25 0 M. anisodon 25 0 Ballota nigra Black horehound 25 12 Stachydeae Stachys arvensis 25 4 S. alopecurus 25 0 S. densiflora 25 0 Leucadeae Leonotis oxymifolia 25 0 L. leonurus 25 0 Lamieae Lamium amplexicaule 25 0 Nepetoideae Salvia officinalis Sage 25 0 S. haematodes 25 0 Mentha australis* 25 0 M. diemenica* Wild mint 25 0 M. spicata* spearmint 25 0 Lycopus australis 25 0 Origanum vulgare Oregano 25 0 Prunella vulgaris 25 0 Thymus vulgaris Thyme 25 0 Dracocephalum ruyschiana 20 0 Ociminum basilicum Basil 25 0 Plectranthus parviflorus* 25 0 Rosmarinus officinalis Rosemary 25 0 Lavandula latifolia Lavender 25 0 Ajugoideae Ajuga australis* 50 0 Teucrium racemosum* 30 0

Scutellaria humilis* 25 0

Prostantheroideae Prostanthera ovalifolia 25 0 Hemiandra pungens 25 0 Westringia fruticosa 25 0 Viticoideae Vitex trifolia* 25 0

19 OTHER 25 species 10–25 0 FAMILIES * New Zealand native or surrogate for New Zealand native species.

` December 2013 EPA0322 24

Application Form Approval to release a new organism

Conclusions

Sagliocco and Coupland (1995) reported Chamaesphecia mysiniformis only from horehound in its native Europe. The test results confirmed this narrow host range. In tests, larvae only survived on plant species in the tribe Marrubieae of the subfamily Lamioideae. There are no native species or other species of value growing in New Zealand that belong to that tribe. With the exception of Stachys arvensis, on which 1/25 larvae survived after 14–21 days, species belonging to other tribes of the sub- family were not susceptible to C. mysiniformis, nor were species belonging to other subfamilies or families. Survival was much higher on Marrubium vulgare than on the other horehound relatives tested. Tests therefore indicate that the host range of C. mysiniformis is restricted to the horehound tribe and, functionally, may well be restricted to M. vulgare (as observed in its native range).

3.2. Regulatory status of the organism

Is the organism that is the subject of this application also the subject of:

an innovative medicine application as defined in section 23A of the Medicines Act 1981?

☐ Yes ☒ No

an innovative agricultural compound application as defined in Part 6 of the Agricultural Compounds and Veterinary Medicines Act 1997?

☐ Yes ☒ No

4. Māori engagement

Discuss any engagement or consultation with Māori undertaken and summarise the outcomes. Please refer to the EPA policy ‘Engaging with Māori for applications to the EPA’ on our website (www.epa.govt.nz) or contact the EPA for advice.

Māori views on this proposal have been sought by:

1. consultation with members of Te Herenga

2. consideration of issues raised in previous similar applications

3. consideration of generic issues by an EPA reference group

` December 2013 EPA0322 25

Application Form Approval to release a new organism

4. consultation with iwi and Treaty settlement authorities in the South Island, and a brief check of concordance with some iwi/hapū management plans

5. consultation with the Ngāpuhi HSNO komiti and the Ngāi Tahu HSNO komiti.

Consultation with Te Herenga

Te Herenga, the EPA’s national network, comprises approximately 80 iwi, hapū or Māori organisation representatives with national geographical and subject-matter coverage. Information about proposed biological control of horehound was distributed to members of Te Herenga in May 2017. The announcement directed readers to the Manaaki Whenua – Landcare Research website for further detail, and invited dialogue and feedback (http://www.landcareresearch.co.nz/science/plants-animals- fungi/plants/weeds/biocontrol/approvals/current-applications/horehound). It described how the applicant intended to assess the risks, costs and benefits associated with the proposed introductions and invited members to identify any issues they would like to have addressed in the applications.

No feedback has yet been received via this route. Any issues brought to the attention of the applicant before formal consideration of this application will be made available to the EPA. Members of Te Herenga will also be specifically informed by the EPA when each application is open for public submission, and will be able to comment on how the applicant has addressed issues raised during consultation.

Issues raised in previous consultations

This application is similar to other applications submitted over the last 10 years to introduce new biological control agents for weeds. Communications with Māori over previous applications are relevant here and have been summarised on the Manaaki Whenua – Landcare Research website (Manaaki Whenua – Landcare Research 2018d). The key areas identified in consultations over this period are:

• possible direct effects on native plant species (see section 5.1.2) • possible indirect effects on native flora and fauna, and other valued species (see sections 5.1.2, 5.3.2) • the need to monitor future effects (see section 7) • predictability of effects (see section 5.1.2) • specific benefits to Māori (see below) • effects on cultural and spiritual values (see below) • integration of control methods, and indigenous solutions (see below) • herbicides and biological control (see section 5.1.1) • aversion to the introduction of new organisms • lack of capacity precludes comment • is the weed present in our rohe?

` December 2013 EPA0322 26

Application Form Approval to release a new organism

Benefits accruing to New Zealand from the introduction of the two agents are explained in section 5. No benefits or costs of this proposal have been identified that are exclusive to Māori. Horehound is not listed in Manaaki Whenua – Landcare Research’s Māori plant use database (http://Māoriplantuse.landcareresearch.co.nz/WebForms/PeoplePlantSearch.aspx). There are five native members of the plant family Labiatae. The database notes that Mentha cunninghamii was used as a component of scents and to induce sweating. A range of uses for Vitex lucens are listed, from use as a timber to the treatment of ulcers. The introduction of the two moths would have no direct impact on the use of these plants as natural resources because the results of host range tests indicate that no native plants will be at risk from HPM or HCM (see section 3.1.4).

Any indirect impact of HPM and HCM through changes of relationships with other flora and fauna will be insignificant because no such interactions are expected, and because any interactions would be restricted to the immediate vicinity of horehound (see section 5.1).

Māori reference group

The EPA convened a Māori reference group (MRG) in 2015 to discuss the potential issues of significance to Māori surrounding the proposed applications. The MRG comprised four members with expertise and/or experience relevant to biocontrol proposals. After undertaking a review of the information available on the proposals, the MRG identified a number of initial draft principles or themes that apply to biological control proposals generally (Manaaki Whenua – Landcare Research 2018c):

• kaitiakitanga – responsibility of Māori to manage natural resources within and beyond hapū and iwi boundaries

• manaakitanga – ability of Māori to protect cultural rights and ownership within hapū and iwi boundaries

• whakapapa as the foundation for kaitiakitanga, and the need to consider the potential impacts of biocontrol agents across the breadth of trophic and ecosystem levels (see section 5.1.2).

• the requirement for applicants to provide comment and/or data to evaluate potential impacts (see section 7)

• the need to define the regional scope of effects, and effectively consider effects on iwi and hapū at a local level (see below)

• the desirability of making vegetation restoration an integral component of biocontrol

• the need to specifically address benefits to Māori.

With reference to the initial draft principles, the MRG noted that the proposed introduction of these control agents might have significant direct beneficial effects on culturally valued species, and indirect benefit to the wider native ecosystem. The MRG specifically commented that the presence of weeds

` December 2013 EPA0322 27

Application Form Approval to release a new organism of significant stature within the margins of the ngahere (forest) adversely affects our appreciation of the forest environment.

The focus of this application is to establish biological control of horehound to reduce large-scale environmental damage and control costs. Successful biological control would reduce the amount of herbicide currently applied to horehound (see section 5.1).

There were no benefits or costs identified that were exclusive to Māori. Benefits and costs would accrue generally to the market economy and to the environment (see sections 5.1, 5.3). Horehound is not yet encroaching significantly on Māori values, and successful biological control would mitigate the onset of those effects.

The addition of HPM and HCM would change the fauna within hapū and iwi boundaries. However, the moths are specific to horehound, with a low ecological footprint (see section 5.1.2). Current surveys of the insects already associated with horehound in New Zealand show that few species use this plant or are associated with it (see section 3.1.3). There is relatively little damage to horehound from indigenous species, and most damage is caused by generalist insects that have other hosts.

Surveys suggest there are no native species that would be displaced from horehound by the introduction of HPM and HCM, and the agents are not expected to materially change how ecosystems function.

There appear to be no indigenous natural enemies of horehound that could be developed or spread to achieve improved control (see section 3.1.3).

Regional consultation Horehound typically grows in dry environments. Although present throughout New Zealand, it is only common on the eastern side of New Zealand, south of Taupō (Figure 3). The most serious impacts are felt in the hill- and high-country farms of the South Island, including whenua managed by iwi and hapū. The Horehound Biocontrol Group is based in the Tekapo region, and the control agents are likely to be released in this region first. Regional and local consultation were therefore concentrated in the South Island.

Information about the proposed biological control programme was provided to the HSNO komiti of Ngāi Tahu for comment before this application was written. It was also provided to the eight iwi and hapū of Te Tau Ihu, with a request to get in touch if further dialogue was desirable. Information was also provided to the Ngāpuhi HSNO komiti.

No responses have been received via this route as yet. Any further information received before the consideration of this application will be passed to the EPA.

` December 2013 EPA0322 28

Application Form Approval to release a new organism

5. Risks, costs and benefits

Provide information of the risks, costs and benefits of the new organism(s).

These are the positive and adverse effects referred to in the HSNO Act. It is easier to regard risks and costs as being adverse (or negative) and benefits as being positive. In considering risks, cost and benefits, it is important to look at both the likelihood of occurrence (probability) and the potential magnitude of the consequences, and to look at distribution effects (who bears the costs, benefits and risks).

Consider the adverse or positive effects in the context of this application on the environment (e.g. could the organism cause any significant displacement of any native species within its natural habitat, cause any significant deterioration of natural habitats or cause significant adverse effect to New Zealand’s inherent genetic diversity, or is the organism likely to cause disease, be parasitic, or become a vector for animal or plant disease?), human health and safety, the relationship of Māori to the environment, the principles of the Treaty of Waitangi, society and the community, the market economy and New Zealand’s international obligations.

You must fully complete this section referencing supporting material. You will need to provide a description of where the information in the application has been sourced from, e.g. from in-house research, independent research, technical literature, community or other consultation, and provide that information with this application.

Horehound is a small shrub that grows to occupy space and displace grasses valued for grazing. It is an un-preferred fodder plant, and stock forced to graze horehound can suffer physiological damage (see section 2.3.2). Horehound produces barbed burs that attach to the wool of sheep, downgrading fleeces and leading to distress in lambs (see section 2.3.2). These adverse effects can be managed either by herbicide application or by removing grazing animals during seeding. Both solutions have adverse environmental and economic consequences. The aim of introducing the two control agents is to mitigate and/or ameliorate these effects.

The potential risks, costs and benefits of the proposed introduction to New Zealand of Wheeleria spilodactylus and Chamaesphecia mysiniformis have been identified by literature review, by review of issues raised in the last 18 applications to ERMA/EPA to introduce biocontrol agents for weeds, and by consultation with stakeholders.

All effects identified during this process are listed on the Manaaki Whenua – Landcare Research website (Manaaki Whenua – Landcare Research 2018b). Those effects considered to be potentially significant are highlighted on that list, and only those effects are addressed in detail in this section. Potential effects are associated with:

• permanent establishment in New Zealand of the two control agents

` December 2013 EPA0322 29

Application Form Approval to release a new organism

• reduction in the abundance and vigour of existing horehound stands

• reduction in the future effects of horehound.

5.1 Potential effects on the environment

5.1.1 Potential beneficial effects on the environment

Successful biological control of horehound would have beneficial effects on the environment if it resulted in ecosystem improvement on land currently invaded by horehound. Such improvement might come from less displacement of native species, reduced impacts of weed control, or improved air, water or soil quality from reduced herbicide usage.

Could successful biological control significantly reduce displacement of native vegetation by horehound?

Marrubium vulgare does not appear to be a problem in undisturbed native habitats in Australia (Weiss & Wills, 2000), but it can become a problem in disturbed or open native grasslands (Weiss & Sagliocco, 2012).

De Lange et al. (2010) recorded 27 New Zealand vascular plant taxa growing in short tussock grasslands alone that are either ‘at risk’ or threatened. Horehound is not currently recognised as a threat to the biodiversity values of natural areas in New Zealand, so successful biological control would not significantly reduce displacement of native species. However, horehound invasion of vulnerable montane ecosystems such as tussock grasslands is increasing (Groenteman et al., 2017) and its adverse effects on biodiversity in protected natural areas is likely to increase over time.

One horehound programme in Geraldine … Hard to kill and spreading along roadsides ….hardly any DOC records but widespread across country – under recorded. Sounds like a problem in high fertility sites… a lot of the dry land threatened plants are at risk from weeds according to de Lange et al so would expect some conservation benefit from a biocontrol agent. (D. Havell, DOC, pers. comm.)

As mentioned, horehound is most common on land ‘disturbed’ by grazing. Much of the extensive pasture grazed in the high country retains elements of indigenous flora, which presumably supports indigenous fauna. While grazed land may not have the highest conservation value, this land supports remnant native ecosystems. Those elements could be eliminated locally by horehound invasion. It is likely that successful control of horehound would help protect remnant ecosystems, but the overall benefit to New Zealand would be minor. Direct benefits to the conservation estate would be real at a local level but not significant nationally. Horehound is invasive in disturbed areas in Australia, and so in New Zealand it may pose a weed risk for amenity areas such as DOC camping grounds and carparks in dry parts of Otago and Canterbury.

` December 2013 EPA0322 30

Application Form Approval to release a new organism

Could successful biological control significantly reduce adverse effects of herbicide use in native vegetation?

Land managers are applying herbicides to retain the productive value of horehound-infested land. However, the only herbicide that is cost-effective in controlling horehound is deadly to associated dicotyledonous vegetation and is persistent, leading to loss of remaining native plants and erosion of soil from bared ground (Groenteman, 2018). Successful biological control would reduce or remove horehound plants slowly, without damaging other vegetation, reducing the risk of erosion. Reduced herbicide use would also reduce non-target deaths of native plants. However, Groenteman (2018) estimated that only about 3.8% of hill- and high-country land area is infested with horehound, so the benefit to the environment of replacing herbicide use over this small land area would be present but minor. Similarly, the amount of herbicide applied in this habitat is small, and any benefits of reduced contamination of air, soil and water from herbicide use would be minimal.

5.1.2 Potential adverse effects on the environment

The establishment of the horehound control agents in New Zealand would have adverse effects on the environment if feeding by larvae:

• reduced the populations of any native plant

• interfered significantly with ecosystem relationships such as trophic webs.

Biological control of horehound would have adverse effects if it:

• facilitated the establishment of worse weeds.

Do the control agents pose a risk to populations of any native plant?

Host range tests conducted before the two control agents were approved for release in Australia indicated that both were sufficiently host specific to be safely introduced to Australia (see section 3.1.4). The results of these tests were also considered adequate to assess the risk to native plant species in New Zealand.

In Australian tests, Wheeleria spilodactylus larvae did not feed on plants outside the subfamily Lamioideae (Table 2). One plant in the tribe Leucadeae that was tested (Leonotis) was not damaged. This suggests that the moth is restricted to the tribe Marrubieae. Further, the relative performances of larvae on controls and two other species in the tribe indicate that the risk of colonisation of plants other than Marrubium species is negligible.

Sagliocco and Coupland (1995) reported that Chamaesphecia mysiniformis larvae only survived on plant species in the tribe Marrubieae of the subfamily Lamioideae. With the exception of Stachys arvensis, on which 1/25 larvae were still alive after 14–21 days, species belonging to other tribes of the subfamily were not susceptible to newly hatched C. mysiniformis larvae, nor were species

` December 2013 EPA0322 31

Application Form Approval to release a new organism belonging to other subfamilies or families. Survival was much higher on M. vulgare than on the other horehound relatives tested. Tests therefore indicate that the host range of C. mysiniformis is restricted to the horehound tribe and, functionally, may well be restricted to the species M. vulgare (as observed in its native range). There are no native species or other species of value growing in New Zealand that belong to that tribe.

There are only five New Zealand native species in the same plant family as horehound (Labiatae), but none belong to the same subfamily (Lamioideae) (Table 1). One of the native species was tested in Australia and found not to be susceptible to either control agent. Testing of related species in Australia (Table 1) was sufficient to conclude that none of these species would be at risk from HPM or HCM.

It is highly unlikely that either HPM or HCM will damage any native plant following release, and the risk to populations of native plant species is considered to be minimal.

Could the agents directly or indirectly cause significant changes in the relationships between native species?

The introduction of the HPM or HCM could adversely affect existing ecosystem relationships if predation or parasitism of this new host resulted in significant population changes in other native species.

‘Apparent competition’ between two species can occur when both are preyed upon by the same natural enemy. For example, if species A and species B are both prey for a predator or parasitoid, a population increase of A could lead to an increase in predator or parasitoid numbers, which in turn could exert unnatural pressure on populations of species B. The introduction of a control agent would cause such adverse effects on trophic webs if populations developed that generated apparent competition in sensitive habitats, leading to significant displacement of valued native species.

A trophic web is the notional representation of all the biotic interactions affecting the population of a single species, in this case interactions between biocontrol agents and potential parasitoids, predators and diseases. Few trophic webs have been adequately described anywhere (but see http://www.landcareresearch.co.nz/publications/newsletters/biological-control-of-weeds/issue-69/food- web-inside-broom-galls ). It is therefore not possible to precisely define the effect of introducing a new organism to such a web. Kaser and Ode (2016) point out that trophic webs are complex, and that signs and strengths of interactions between elements of trophic webs can be difficult to measure or predict.

Frago (2016) and Kaser and Ode (2016) also point out that habitat complexity and fragmentation can modify interactions between members of a trophic web, and so ecosystem effects may be site-specific rather than general. Despite these reservations, the relative importance of HPM and HCM in web dynamics can be inferred from the biology and ecology of the agents and the target weed.

` December 2013 EPA0322 32

Application Form Approval to release a new organism

• The agents are host specific (see section 5.1.2) and will only reach high populations in the immediate vicinity of horehound plants.

• Interactions with native species will be localised around horehound plants.

• Surveys indicate that there are no native species currently using horehound as a host, and natural enemy associations appear to be simple ( see section 3.1.3)

• Although locally important, the total area of high- and hill-country farmland in the South Island heavily infested with horehound is estimated to be only 112,100 ha. It is not thought to be a key pest for natural habitats at present. The opportunity for control agents to significantly influence trophic webs is therefore limited.

• Horehound itself is an ecosystem modifier, and is likely to be the dominant influence on trophic webs within infested sites, outweighing any effects the control agents might have. Biological control aims to reduce the physical dominance of horehound and its influence on trophic webs.

• If the agents fail to establish, or do not achieve high population levels following release, it is unlikely that their presence could become a key factor in any trophic web.

• Overall effects of HPM and HCM on the quality of trophic webs are as likely to be beneficial as adverse (e.g. Kaser and Ode, 2016).

Little can be said about other potential trophic interactions, such as common diseases, but any apparent competition will be restricted to where the control agents are abundant – the vicinity of horehound infestations. Given the available information, Manaaki Whenua – Landcare Research concludes that Wheeleria spilodactylus and Chamaesphecia mysiniformis are unlikely to significantly influence the quality of trophic webs outside horehound infestations. Successful biological control would progressively reduce any influence of the agents on trophic webs and would also partially reverse any effects of horehound itself.

Could biological control of horehound lead to invasion of natural habitats by worse weeds?

Successful biological control is not expected to lead to replacement with a worse weed. Horehound is not considered to be a key factor in the integrity of strictly native habitats at present. In modified natural habitats of the Mackenzie/Otago region, alternative forage species are most likely to replace horehound. However, the weed most likely to replace horehound is Hieracium species, and these are already ubiquitous in the region.

No other potential adverse effects on the environment were considered significant (Manaaki Whenua – Landcare Research 2018b). There appear to be few native species using horehound as a host (see section 3.1.3) so reduction in the abundance or biomass of horehound would have no significant

` December 2013 EPA0322 33

Application Form Approval to release a new organism adverse effect on any native invertebrate species. Any effects on horehound abundance will develop gradually, and no catastrophic or rapid changes to invaded habitats are expected.

5.2 Potential effects on human health

5.2.1 Potential beneficial effects on human health

The adverse effects of horehound on farm productivity and animal welfare impose additional psychological stresses on those who farm affected areas (Appendix 2, in Groenteman, 2018). Successful biological control would ameliorate those stresses, but as horehound has limited distribution this effect would not be significant at a national scale. Successful biological control may reduce the future use of herbicides for horehound control. Herbicides applied to horehound according to label are unlikely to have significant adverse effects on operators, so the benefits of biological control are unlikely to be significant nationally. Any marginal benefit for human health is minimal because the herbicide applied specifically for horehound control is only a small proportion of overall exposure to herbicides in New Zealand (see section 5.3.1).

5.2.2 Potential adverse effects on human health

Neither the establishment of the two moth species nor the successful control of horehound would have significant adverse effects on human health in New Zealand. Successful biological control would not eliminate horehound or remove options for horehound cultivation should any new use for the plant be identified in the future. (The effect of biological control on the availability of horehound for extraction of phytomedicines is covered in section 5.3.2.) These moths may become abundant, but only around horehound stands.

The wings of all moths are covered with scales that are easily shed. Scales in the air can exacerbate respiratory issues in confined spaces, but these moths will rarely be common near human habitations. Moths are a common element of the New Zealand fauna, and the marginal increase in allergy risk from adding these two moths to the resident fauna in New Zealand is negligible.

5.3 Potential effects on the market economy

Groenteman et al. (2017) and Groenteman (2018) review the effects of horehound in New Zealand, and this analysis is summarised in section 2.3.2. Groenteman (2018) also records the experiences of 95 farmers who responded to a survey about the effects of horehound. Some of their comments are quoted here.

5.3.1 Potential beneficial effects on the market economy

Successful biological control would result in benefits to the market economy if it:

• reduced invasion or reinvasion of productive land

` December 2013 EPA0322 34

Application Form Approval to release a new organism

• reduced production losses in the primary sector

• reduced control costs for territorial authorities, regional councils, DOC, occupiers and other entities that manage land.

Could biological control reduce the invasion risk posed by horehound?

Of the respondents to the farmer survey (Groenteman, 2018), 64.5% indicated that horehound infestations were increasing, 27.5% thought it was stable, and 6% noted that the weed was declining. Respondents estimated that, left untreated, horehound infestations had doubled every 1 to 10 years. The median and the mean were consistent, estimating a three-year doubling period. Horehound spreads quickly by seeds within hooked burs that attach to passing animals or machinery (Figure 2), and the rate at which horehound reinvades cleared areas indicates a significant seedbank. Anecdotal evidence suggests that seed can last 7–10 years in the soil (Weiss & Sagliocco, 2000). The rate of spread is directly related to the number of seeds produced per unit area, and hence is determined by the vigour and abundance of parent plants.

Successful biological control would reduce the vigour and/or survival of horehound plants, reducing seed production. The number of burs would be reduced, and this would lead, over time, to a reduction in the seedbank in areas already infested. It would significantly reduce the ability of horehound to invade new areas, and possibly the ability to replace itself in some habitats. Partial control would also reduce seed production and the production of burs (also see section 3.2.3. Given the level of biological control achieved in Australia since 1998, this outcome is likely.

Could biological control reduce production losses caused by horehound?

Horehound is unpalatable to sheep and displaces better forage species in hill- and high-country pastures (Figure 7). It produces hooked burs that matt the wool of sheep, and this can cause distress to lambs and reduces the value of wool production in affected areas (see section 2.3.2). Lucerne forage crops are a key component of dryland farming, and horehound invasion seriously reduces productivity.

These production losses can be restored by the application of herbicides (see below). However, the use of herbicides to restore productivity of lucerne is problematic because the most effective and cost- effective herbicide is detrimental to the survival and persistence of lucerne, and herbicide persistence in the soil precludes replanting for several years. During this period it is usually horehound that re- establishes from the seedbank. There is evidence that invasion of lucerne and other forage by horehound can make dryland farming uneconomic (Groenteman, 2018).

` December 2013 EPA0322 35

Application Form Approval to release a new organism

Figure 7. Horehound displacing pasture in Otago

Another way to mitigate the effects of horehound on livestock is to destock land for the period that burs are present on plants. This reduces the effective carrying capacity of the land. The current costs of adapting farm management systems to manage the effects of horehound are unclear.

Could biological control reduce the costs of horehound management?

Successful biological control of horehound would benefit the market economy if the costs to occupiers, regional councils and DOC of controlling horehound were significantly reduced. Although it is present in the North Island, the major effects of horehound are found in the South Island

Horehound affects some farmers by taking up pasture space, thereby affecting economics. This is not huge, but some farmers do spray it. This is very hard to quantify but possibly $10,000 is spent by farmers controlling it chemically. Lost pasture might be in the vicinity of $20,000. These are both guesses. (D.Underhill, Hawke’s Bay Regional Council)

It seems to be wide spread in the southern parts of the Wairarapa although only scattered. We have never had any complaints or enquiries about the plant. (C. Carswell, Greater Wellington Regional Council)

Groenteman (2018) conservatively estimated that the annual cost of controlling horehound in high- and hill-country farming was at least $3.35 million (chemical control: $2.5 million; non-chemical control: $0.85 million) (see section 2.3.2).

Total biological control would remove the ability of horehound to compete with lucerne or other forage plants in farming systems, saving $3.35 million spent on control by occupiers. Given the level of biological control achieved in Australia since 1998, this outcome is likely. The threshold density or

` December 2013 EPA0322 36

Application Form Approval to release a new organism vigour of horehound plants below which control measures would not be considered necessary is uncertain. However, even partial biological control of horehound is likely to recover a proportion of the estimated control costs, and the projected benefit to dryland farmers would be moderate.

Responses from DOC and regional councils did not identify any significant expenditure on control measures for horehound (Manaaki Whenua – Landcare Research, 2018e).

5.3.2 Potential adverse effects on the market economy

Successful biological control of horehound would have adverse effects on the market economy if it significantly reduced the value of the plant as a medicinal herb, or as a nectar source for bees. The introduction of the two agents would have adverse effects on the market economy if they caused significant damage to non-target plants, including nursery species and culinary herb crops.

Could biological control of horehound significantly affect the wild harvest of horehound for medicinal use?

Marrubium vulgare is harvested from wild populations in New Zealand by medical herbalists. The parts of the plant most used are the leaves and flowering tops, either fresh or dried. The New Zealand Association of Medical Herbalists and the Herb Federation of New Zealand were consulted in May 2017, and they expressed opposition to the biocontrol programme. Their full responses can be found in Appendix 2 of Groenteman et al. 2017. The main concerns raised by medical herbalists include:

• the plant will be eradicated • accessibility to wild-crafted M. vulgare will be reduced due to the irreversible nature of biocontrol • the biocontrol agents would have an impact on other medicinal/culinary herbs in the family Labiatae once the populations of the target host had declined • the presence of medicinally valued plants on a farming property should be viewed as an additional source of income to the farmer • removal of M. vulgare from the environment undermines indigenous rights as well as national obligations under the WHO Traditional Medicine Strategy and the Convention on Biological Diversity • following control, the plant would be replaced by another opportunistic plant • targeting other weeds would make better use of resources.

If the two agents do not establish, then there will be no effect on the horehound harvest.

Maximum accumulation of marrubiin occurs in the glandular trichomes on fully expanded leaves just before the plant flowers in November, and that is the time herbalists tend to harvest the plant (Groenteman et al., 2017). If HPM became established, then harvest in November could coincide with peak populations of larvae, which could spoil the availability of clean foliage, even if control was not achieved. If the introduction of HPM and HCM resulted in successful biological control of horehound,

` December 2013 EPA0322 37

Application Form Approval to release a new organism then it is likely the abundance and size of horehound plants in the field would be significantly reduced and the quality of foliage would be compromised. Horehound would not be eradicated by biological control, but the accessibility of quality foliage would be significantly reduced.

It is clear that successful biological control of horehound would result in a transfer of benefits from current horehound harvesters to those benefiting from improved forage production. The value of that transfer is uncertain, because although one respondent estimated that one plant could currently provide enough foliage to make two bottles of extract (in Groenteman et al., 2017), the horehound foraging community has not provided information about the market size for horehound products in New Zealand.

We feel very strongly these insects NOT be imported and released, particularly, as there is significant potential for the commercialisation of this herb. There are manufacturers of herbal medicines in New Zealand, also there is distillery equipment in many places. This may produce another plant source to extend the use of their equipment. (K. Hilterman, Lavender Hill Herbals)

At present two of the main suppliers of herbal tinctures to the New Zealand market wholesale price of 1:2 200 ml tincture of Marrubium vulgare is $23.90 and $25.70 respectively. … 1 kg of herb is used in 2 litres. … So to make a 200 ml bottle of herbal extract/ tincture 100 grams of herb is used. Each healthy horehound plant … is likely to provide at least 200 grams of herb. … Two horehound plants can easily come to harvest in an area of one square foot. Minimal harvest volume of herb in this square foot would be 400 grams and worth say half the tincture price i.e. say $12.00 x 4 = $48.00 for 400grams. … so price per harvest could be $5,166,672 per hectare. So why would Mr …. persist in growing lucerne and running sheep and cattle on all of his ….hectares? Why not use some of this land to grow the medicinal herb? (see Groenteman et al. (2017) for full comment) We in the herbal medicine field use this invaluable respiratory herb all the time and buy it in from overseas. If this is indeed the medicinal plant that is growing perhaps there is an option to look at utilising it. (A. Palmer, herbalist, pers. comm.)

Losses to the horehound foraging community could be mitigated by:

• increasing foraging effort to achieve the same yield • protection of harvestable wild horehound plants from attack by control agents using organic pesticides applied in early spring, or by netting plants in autumn to prevent moth eggs being laid (see section 3.1.2) • instituting novel methods such as grazing to protect and possibly enhance horehound density in selected areas set aside for harvest • developing financial arrangements with landowners to actively promote and protect horehound in limited agreed areas.

Apart from its value as a medicinal herb, horehound has been used as a flavour in beer production overseas. Beer brewers might value M. vulgare for craft beer, but no such beers are produced

` December 2013 EPA0322 38

Application Form Approval to release a new organism commercially in New Zealand at present. Biological control would be unlikely to reduce the population of horehound to the extent that it could not be harvested for this niche purpose.

Could biological control significantly affect the commercial growing of culinary herbs?

The tests conducted to define the host range of HPM and HCM included a wide range of culinary herbs. The risk of damage to non-target plants, including culinary herbs that are grown commercially, is negligible (see section 3.1.4).

Could beekeeping be adversely affected by biological control of horehound?

In some parts of the world M. vulgare is valued by apiarists as an early season resource for honeybees. An initial conversation with Apiculture New Zealand indicated that M. vulgare is not a highly valued resource for apiarists in this country. Apiculture New Zealand has not responded to a more recent request for comment. It is unlikely that horehound has a significant role in New Zealand apiculture.

Could damage by horehound control agents reduce the viability of plant production in nurseries?

New Zealand Plant Producers Incorporated was consulted on this issue. They concluded that the host range of the agents appeared to be narrowly confined to target species, and saw no issues they would need to influence or contest (J. Liddle, NZPPI, pers. comm.).

5.4 Potential effects on society and communities

5.4.1 Potential beneficial effects on society and communities

Successful biological control of horehound would have benefits to societies and communities if replacement of horehound with other vegetation following successful biological control led to improved animal welfare and improved public conservation values.

Matting of fleeces by horehound burs not only reduces the value of the fleece but also causes discomfort and ill thrift, especially in lambs. Successful biological control would remove the risk of fleece matting. Any level of biological control would reduce the adverse animal welfare risks.

…sheep covered in horehound seed experience poor growth rates. Also there [are] adverse effects from forced consumption of horehound with tainted meat and liver problems. More commonly sheep are removed from high density areas and the land is removed from grazing. (G. Loxton, Sawdon Station)

While improvements to conservation values may be likely, current infestations of horehound are not considered to significantly downgrade conservation values, and so the potential benefits available are minimal.

` December 2013 EPA0322 39

Application Form Approval to release a new organism

5.4.2 Potential adverse effects on society and communities

The introduction of Wheeleria spilodactylus or Chamaesphecia mysiniformis would have a significant adverse effect on society and communities if attack by larvae reduced the value of plants currently growing in the community. Section 3.1.4 summarises tests designed to assess the susceptibility of a range of culinary herbs related to horehound to attack by HPM and HCM.

Horehound belongs to the subfamily Lamioideae. It is related to a wide range of common culinary herbs including thyme, oregano, mint, sage, basil, rosemary, and lavender. These belong to a different subfamily (Nepetoideae). A number of plant species belonging to this subfamily were tested before the control agents were released in Australia (Tables 1 & 2). None were susceptible to the control agents. Neither Chamaesphecia mysiniformis nor Wheeleria spilodactylus has been recorded to attack species in the subfamily Nepetoideae in its native range. Chamaesphecia mysiniformis has been established in Australia since 2001, and Wheeleria spilodactylus since 1995. There have been no reports of either species attacking culinary herbs in Australia since these agents established there. Experimental and observational evidence suggest that it is improbable that HPM or HCM would cause significant damage to culinary herbs related to horehound, and the risk is therefore minimal.

6. Pathway determination and rapid assessment

Under sections 38I and 35 of the HSNO Act your application may be eligible for a rapid assessment. The pathway for your application will be determined after its formal receipt, based on the data provided in this application form. If you would like your application to be considered for rapid assessment (as per the criteria below), we require you to complete one of the below sections. Fill in the section that is relevant to your application only.

6A. New organism that is or is contained within a veterinary or human medicine (section 38I)

6.1. Controls for organism Describe the controls you propose to mitigate potential risks (if any). Discuss what controls may be imposed under the ACVM Act (for veterinary medicines) or the Medicines Act (for human medicines)

Not applicable

6.2. Discuss if it is highly improbable (after taking into account controls if any):

• The doses and routes of administration of the medicine would have significant adverse effects on the health of the public or any valued species; and

` December 2013 EPA0322 40

Application Form Approval to release a new organism

• The organism could form an undesirable self-sustaining population and have significant adverse effects on the health and safety of the public, any valued species, natural habitats or the environment Do not include effects of the medicine or new organism on the person or animal being treated with the medicine

Not applicable

6B. New organism (excluding genetically modified organisms) (section 35)

6.3. Discuss if your organism is an unwanted organism as defined in the Biosecurity Act 1993

Neither Wheeleria spilodactylus nor Chamaesphecia mysiniformis is listed in the Ministry for Primary Industries Unwanted Organisms Register (https://www1.maf.govt.nz/uor/searchframe.htm)

6.4. Discuss if it is highly improbable, after taking into account the proposed controls, that the organism after release:

• Could form self-sustaining populations anywhere in New Zealand (taking into account the ease of eradication) • Could displace or reduce a valued species • Could cause deterioration of natural habitats • Will be disease-causing or be a parasite, or be a vector or reservoir for human, animal, or plant disease • Will have adverse effects on human health and safety or the environment

Self-sustaining populations The purpose of the application is to establish self-sustaining populations of Wheeleria spilodactylus and Chamaesphecia mysiniformis wherever horehound infestations exist in New Zealand. Both agents can disperse as adults. It is unlikely that either could be successfully eradicated once establishment was confirmed, so introduction should be considered irreversible (section 3.1). Neither agent is expected to have significant adverse economic or environmental effects (see section 5).

Displacement or reduction of a valued species Significant displacement of valued species is considered improbable for the following reasons. • The evidence presented in section 5 indicates that native plant species are not at significant risk of attack by Wheeleria spilodactylus or Chamaesphecia mysiniformis. • Other than horehound, no valued exotic plant species are at significant risk from Wheeleria spilodactylus or Chamaesphecia mysiniformis (sections 5.1.2, 5.3.2).

` December 2013 EPA0322 41

Application Form Approval to release a new organism

• It is improbable that any native plant or invertebrate species would be significantly displaced (section 5.1.2). There appear to be no native invertebrate species commonly associated with horehound that could be significantly displaced by Wheeleria spilodactylus or Chamaesphecia mysiniformis (Groenteman et al., 2017). • Any change in horehound abundance resulting from biological control is likely to be gradual over several years. It is highly improbable that this control agent will cause catastrophic decline in any horehound infestation that might lead to widespread rapid change in any native habitat. Successful biocontrol will tend to restore affected habitats to a pre-invasion state. • Permanent reductions in horehound biomass could potentially result in replacement by other invasive species in existing sites (see section 5.). This effect is considered insignificant because it is not likely that any other weeds would be more damaging than horehound in affected habitats (see section 5.3.1), and any effect is likely to be variable from place to place.

Deterioration of natural habitats Founding populations of HPM and HCM will be relatively small, and it will take several years before larval infestations reach damaging levels. Horehound is a perennial with significant reserves in the woody root. It is unlikely that the control agents will kill a plant within a single season. Change in horehound biomass and abundance will therefore be gradual, allowing surrounding vegetation time to regain the space occupied by horehound. Deterioration of natural habitats is therefore highly improbable. There are no published reports of such deterioration in Australia following the release of HPM and HCM.

Vector or reservoir of disease Neither Wheeleria spilodactylus nor Chamaesphecia mysiniformis is parasitic or capable of transmitting human or animal diseases. It is possible that the moths contaminated with spores could assist the spread of a plant disease, but they are no more likely to do so than any other moth. These species will make up a small proportion of the New Zealand moth fauna, and any additional risk following introduction would be insignificant. Larvae create lesions on the leaves or roots of host plants that might facilitate adventitious disease infection. However, damage to non-target hosts in tests was not significant (see sections 3 and 5.1.2), and any increase in disease incidence in non-target plants would be insignificant. Lepidopteran larvae do not form diseases in their own right and are not parasitic or predaceous. Lepidoptera are not known as significant vectors or reservoirs of plant diseases and do not pose a significant vector risk.

Human health Moth scales and hairs can adversely affect respiratory health when inhaled. Neither moth is expected to add to the risk of such effects, for two reasons. • Both moths are specific to horehound and will only be abundant around heavy horehound stands. Such stands are not typically associated with human habitations.

` December 2013 EPA0322 42

Application Form Approval to release a new organism

• Moths are abundant and ubiquitous in the New Zealand environment, and commonly fly into homes when attracted by light. The addition of these two species will not significantly increase the overall biomass of moths creating this effect.

7. Other information

Add here any further information you wish to include in this application including if there are any ethical considerations that you are aware of in relation to your application.

This application raises no known ethical considerations.

Post-release monitoring and measurement of impact

Manaaki Whenua – Landcare Research will provide populations of the control agent to farmers, and will oversee agent release and immediate follow-up. Monitoring will be done on a hierarchical basis. Initially, signs of establishment will be determined by visually searching for damage on horehound plants growing near the release sites. The presence of control agents on non-target plants will also be recorded. If damage to horehound plants becomes extensive and impacts are suspected, Manaaki Whenua – Landcare Research staff will seek support to undertake measurements of the agents’ dispersal and effects over time.

Wider monitoring of biological control in New Zealand

Manaaki Whenua – Landcare Research is focused on constant improvement in the practice of biological control of weeds in New Zealand, including world-leading research into minimising the interactions of introduced agents with existing trophic webs (Paynter et al., 2010; Fowler et al., 2012), better prediction of success (Paynter et al., 2012), how agents disperse (Paynter & Bellgard, 2011), the accuracy of host range testing in predicting eventual host range following release (Paynter et al., 2014), and monitoring the safety of biological control in New Zealand (Paynter et al., 2004). Manaaki- Whenua – Landcare Research undertake the above research in accordance with the priorities of stakeholders and as funds permit.

` December 2013 EPA0322 43

Application Form Approval to release a new organism

8. Checklist This checklist is to be completed by the applicant

Application Comments/justifications

All sections of the application form completed ☒ Yes ☐ No or you have requested an information waiver (If No, please discuss with an under section 59 of the HSNO Act Advisor to enable your application to be further processed)

Confidential data as part of a separate, ☐ Yes ☒ No N/A identified appendix

Supplementary optional information attached:

• Copies of additional references ☒ Yes ☐ No Provided separately

• Relevant correspondence ☒ Yes ☐ No See website

Administration Are you an approved EPA customer? ☒ Yes ☐ No If Yes are you an: Applicant: ☐ Agent: ☐

If you are not an approved customer, Paid in advance payment of fee will be by: • Direct credit made to the EPA bank ☒ Yes ☐ No account (preferred method of payment) ☐ Payment to follow Date of direct credit:

• Cheque for application fee enclosed ☐ Yes ☐ No ☐ Payment to follow

Electronic, signed copy of application e- ☒ Yes mailed to the EPA

` December 2013 EPA0322 44

Application Form Approval to release a new organism

Signature of applicant or person authorised to sign on behalf of applicant

☒ I am making this application, or am authorised to sign on behalf of the applicant or applicant organisation.

I have completed this application to the best of my ability and, as far as I am aware, the ☒ information I have provided in this application form is correct.

Signature Date3 May 2018

Request for information waiver under section 59 of the HSNO Act

I request for the Authority to waive any legislative information requirements (i.e. concerning ☐ the information that has been supplied in my application) that my application does not meet (tick if applicable).

Please list below which section(s) of this form are relevant to the information waiver request:

` December 2013 EPA0322 45

Application Form Approval to release a new organism

Appendices and referenced material (if any) and glossary (if required)

References

Briese, D.T., Walker, A. (2002) A new perspective on the selection of test plants for evaluating the host-specificity of weed biological control agents: the case of Deuterocampta quadrijuga, a potential insect control agent of Heliotropium amplexicaule. Biological Control 25: 273-287.

Clarke, C., Baker, J., Keller, M., Roush, R. (2000) Biological control of horehound: lessons from South Australia. Plant Protection Quarterly 15: 29-32. de Lange, P.J., Norton, D.A., Courtney, S.P., Heenan, P.B., Barkla, J.W., Cameron, E.K., Hitchmough , R., Townsend, A.J. (2010) Threatened and uncommon plants of New Zealand (2008 revision). New Zealand Journal of Botany 47:1, 61-96. doi: 10.1080/00288250909509794

Dugdale, J.S. (1988) Lepidoptera – annotated catalogue, and keys to family-group taxa. Fauna of New Zealand 14. (https://www.landcareresearch.co.nz/__data/assets/pdf_file/0017/26324/FNZ14Dugdale1988.pdf )

Fowler, S.V., Paynter, Q., Dodd, S., Groenteman, R. (2012) How can ecologists help practitioners minimize non- target effects in weed biocontrol? Journal of Applied Biology 49: 307-310.

Frago, E. (2016) Interactions between parasitoids and higher order natural enemies: intraguild predation and hyperparasitoids. Current Opinion in Insect Science 14: 81-86.

Gerard, P.J., Barringer, J.R.F., Charles, J.G., Fowler, S.V., Kean, J.M., Phillips, C.B., Tait, A.B., Walker, G.P. (2013) Potential effects of climate change on biological control systems: case studies from New Zealand. BioControl 58: 149-162.

Groenteman, G. (2018) Economics of horehound to New Zealand dryland farming. Unpublished report prepared for the Horehound Biocontrol Group.

Groenteman, G., Probst, C., Bellgard, S., Prebble, J. (2017) Feasibility for biological control of horehound, Marrubium vulgare L. Landcare Research Report LC3040 for the Horehound Biocontrol Group.

Kaser, J.M., Ode, P.J. (2016) Hidden risks and benefits of natural enemy-mediated indirect effects. Current Opinion in Insect Science 14: 105-111.

Li, B., Cantino, P.D., Olmstead, R.G., Bramley, G.L.C., Xiang, C.-L., Ma, Z.-H., Tan, Y.-H., Zhang, D.-X. (2016) A large scale chloroplast phylogeny of the Lamiacae sheds light on its sub-familial classification. Scientific Reports 6, article number 34343. doi: 10.1038/srep34343.

Manaaki Whenua Landcare Research (2018a) Host range testing of Wheeleria spilodactylus and Chamaesphecia mysiniformis. http://www.landcareresearch.co.nz/__data/assets/pdf_file/0008/156932/Summary-of- tests.pdf

` December 2013 EPA0322 46

Application Form Approval to release a new organism

Manaaki Whenua Landcare Research (2018b) Possible risks, costs and benefits of horehound control. http://www.landcareresearch.co.nz/__data/assets/pdf_file/0007/156931/potential-rcb-horehound.pdf

Manaaki Whenua Landcare Research (2018c) Report of the EPA Reference Group concerning biological control of weeds. http://www.landcareresearch.co.nz/__data/assets/pdf_file/0003/88338/Māori_Reference_Group_Report. pdf

Manaaki Whenua Landcare Research (2018cd) Summary of responses to consultation with Māori. http://www.landcareresearch.co.nz/__data/assets/pdf_file/0007/162277/horehound-biocontrol-pre- applcation-consultation-maori.pdf

Page, A.R., Lacey, K.L. (2006) Economic impact assessment of Australian weed biological control. Technical Series, CRC for Australian Weed Management 10, pp. 91-94.

Paynter, Q., Bellgard, S. (2011) Understanding dispersal rates of invading weed biocontrol agents. Journal of Applied Ecology 48: 407-414.

Paynter, Q., Fowler, S.V., Gourlay, A.H., Groenteman, R., Peterson, P.G., Smith, L., Winks, C.J. (2010) Predicting parasitoid accumulation on biological control agents of weeds. Journal of Applied Ecology 47: 575-582.

Paynter, Q., Fowler, S.V., Gourlay, A.H., Peterson, P.G., Smith, L.A., Winks, C.J. (2014) Relative performance on test and target plants in laboratory tests predicts the risk of non-target attack in the field for arthropod weed biocontrol agents. Biological Control 80: 133-142.

Paynter, Q., Overton, J.M., Hill, R.L., Bellgard, S.E., Dawson, M.I. (2012) Plant traits predict the success of weed biocontrol. Journal of Applied Ecology 49: 1140-1148.

Paynter, Q.E., Fowler, S.V., Gourlay, A.H., Haines, M.L., Harman, H.M., Hona, S.R., Peterson, P.G., Smith, L.A., Wilson-Davey, J.A., Winks, C.J. & Withers, T.M. (2004) Safety in New Zealand weed biocontrol: a nationwide survey for impacts on non-target plants. New Zealand Plant Protection 57: 102-107.

Sagliocco, J. (2000) The insect fauna associated with horehound (Marrubium vulgare L.) in western Mediterranean Europe and Morocco: potential for biological control in Australia. Plant Protection Quarterly 15: 21-25.

Sagliocco, J.-L., Coupland, J.B. (1995) Biology and host specificity of Chamaesphecia mysiniformis (Lepidoptera: Sesiidae), a potential biological control agent of Marrubium vulgare (Lamiaceae) in Australia. Biocontrol Science and Technology 5: 509-516.

Sagliocco, J.-L., Weiss, J. (2004) Importation, establishment and preliminary impact assessment of Chamaesphecia mysiniformis (Lepidoptera: Sesiidae) for the biological control of horehound in Australia. In: Proceedings of the 14th Australian Weeds Conference, 6–9 September 2004, Wagga Wagga, NSW. (Eds B.M. Sindel & S.B. Johnson), pp. 388-390. Weed Society of New South Wales, Sydney.

` December 2013 EPA0322 47

Application Form Approval to release a new organism

Sheppard, A.W., Heard, T.A., van Klinken, R.D. (2005) Scientific advances in the analysis of direct risks of biological control agents to non-target plants. Biological Control 35: 215-226.

Wapshere, A.J. (1974) A strategy for evaluating the safety of organisms for biological weed control. Annals of Applied Biology 77: 201-211.

Webb, C.J., Sykes, W.R., Garnock-Jones, P.J. 1988: Flora of New Zealand. Vol. IV. Naturalised Pteridophytes, Gymnosperms, Dicotyledons. Botany Division, DSIR, Christchurch.

Weiss, J., Sagliocco, J. (2000) Horehound (Marrubium vulgare): a comparison between European and Australian populations. Plant Protection Quarterly 15:18-20.

Weiss, J., Sagliocco, J.-L. (2012) Marrubium vulgare L. – horehound. In: Julien M, McFadyen R, Cullen J eds. Biological Control of Weeds in Australia. Pp. 360-367.

Weiss, J., Wills, E. (2000) Outcomes from the horehound workshop. Plant Protection Quarterly 15: 43-43.

` December 2013 EPA0322