From Hobbyist Fieldwork to Conservation Science: Detection Dogs Locating Eucalyptus Browse for Phasmid and Invertebrate Conservation

In applied conservation, some of the most effective innovations arise not from formal institutional design, but from the intersection of lived experience, field ecology, and practical problem-solving. The use of scent-detection dogs to locate specific eucalyptus trees for captive phasmid husbandry is a clear example of this principle. The approach originated with a dog owner undertaking postgraduate study in conservation and animal behaviour, who—faced with the ongoing, time-intensive task of sourcing suitable eucalyptus browse for captive phasmids—began exploring whether detection dogs could be trained to streamline and improve the reliability of field collection. 

What began as a pragmatic response to the limitations of manual plant identification and repeated hours in the field gradually developed into a structured methodology, and ultimately into a model with broader applications in conservation science.

Origins: a dog owner, postgraduate study, and an experimental idea

The first structured use of dogs for locating eucalyptus browse for phasmids began in the Wollongong region of New South Wales, where a dog owner completing postgraduate studies in conservation and animal behaviour began experimenting with the idea that detection dogs could be trained not only on animals or fungi, but on plant chemical profiles relevant to invertebrate conservation.

Working initially in a highly practical, field-based context, rescued dogs were incorporated into routine bushland collection trips several times a week. The original aim was simple: to support an amateur phasmid-keeping community by locating suitable eucalyptus species for feeding captive stick and leaf insects.

However, the system quickly developed beyond its initial scope. Over time, the dogs demonstrated not only the ability to locate target eucalyptus species, but also an emerging capacity to distinguish plant health status—leading to refinement in training toward locating non-diseased, high-quality browse sources suitable for sensitive invertebrate populations.

What began as an informal, postgraduate-driven applied research project gradually became a prototype for conservation detection methodology.

Why eucalyptus identification is complex: Corymbia and Angophora

A central challenge in this work is the taxonomic and ecological complexity of Australia’s eucalypt-dominated forests and parks. While the term “eucalyptus” is commonly used broadly, three closely related genera dominate these ecosystems:

• Eucalyptus (true eucalypts)
• Corymbia (bloodwoods and ghost gums)
• Angophora (commonly referred to as apples)

These genera are frequently mistaken for one another due to overlapping physical characteristics such as smooth bark, similar canopy structure, and lance-shaped juvenile leaves. In mixed woodland environments, especially under field conditions, reliable visual differentiation can be difficult even for trained observers.

However, for phasmids—many of which are highly host-specific—the differences between these genera are biologically significant. Subtle variations in leaf chemistry, including essential oil composition, tannin levels, and secondary metabolites, directly influence feeding behaviour, growth rates, and survival outcomes in captive populations.

• Corymbia species often contain distinct oil and resin profiles that can deter specialist insect feeders.
• Angophora species tend to have broader leaves and differing tannin structures, which may be unsuitable or rejected by certain phasmid species.
• Even within Eucalyptus, species-level variation can dramatically affect insect performance.

Because these genera frequently co-occur and visually converge in form, misidentification during manual collection poses a significant risk to captive breeding stability. Detection dogs address this by bypassing visual taxonomy entirely and responding instead to volatile organic compound (VOC) signatures associated with target plants and their physiological condition.

Expansion of role: from hobbyist system to conservation tool

As training protocols were refined, the dogs’ role shifted from simple eucalyptus location to a more sophisticated ecological function: identifying suitable, healthy host plants for invertebrate conservation use.

This included detecting:

• target eucalyptus species used in captive feeding programs
• trees free from disease, fungal infection, or stress-related chemical alteration
• consistent and high-quality foliage sources suitable for sensitive insect colonies

This refinement was particularly important for captive phasmid populations, where plant quality directly influences moulting success, reproduction, and long-term viability.

Endangered phasmids

The implications of this work extend to some of Australia’s most threatened invertebrates, including the critically endangered Lord Howe stick insect (Dryococelus australis), which was once believed extinct until a small wild population was rediscovered on Ball’s Pyramid.

Captive breeding and recovery programs for this species depend heavily on carefully controlled feeding regimes, where host plant suitability and consistency are essential. While not all feeding plants are eucalyptus-dependent, the broader methodology of ensuring precise, healthy, and chemically appropriate browse is directly relevant.

In this context, detection dogs provide a potential tool for identifying and sourcing ecologically appropriate vegetation for both current captive populations and future habitat restoration efforts associated with reintroduction planning.

Expansion into institutional conservation and captive breeding systems

What began as a postgraduate-led, field-based experiment supported by rescued dogs has since been adapted into broader conservation thinking, particularly in relation to captive invertebrate breeding programs.

Applications now being explored or implemented include:

• supporting captive phasmid colonies with verified eucalyptus browse
• improving sourcing accuracy for specialist herbivorous invertebrates
• assisting in ecological mapping of suitable host plant distributions
• enhancing biosecurity by identifying plant health and stress conditions before collection

These developments are particularly relevant for species with narrow dietary requirements and high sensitivity to plant chemistry variation.

Potential extension to folivorous mammals and koala conservation

There is also growing interest in extending this approach to folivorous mammal management, particularly in relation to captive and semi-managed koala populations.

Koalas require daily access to specific eucalyptus species, with preferences that vary not only by species but also by individual tree chemistry, seasonal variation, and regional adaptation. Maintaining consistent, appropriate browse is therefore a significant logistical challenge for many captive facilities.

While major institutions such as Taronga Zoo maintain established eucalyptus supply systems and plantation resources, many smaller facilities do not have access to dedicated eucalyptus farms and must rely on external sourcing networks.

In this context, detection dogs could potentially support:

• identification of high-quality koala feed trees in wild or managed landscapes
• mapping of sustainable harvesting zones for eucalyptus browse
• exclusion of stressed or chemically unsuitable trees
• long-term planning for captive feeding infrastructure and plantation development

Although still emerging as a concept, the underlying principle remains consistent with phasmid applications: ensuring precise, chemically and ecologically appropriate plant selection at scale.

Why this system works: scent as ecological data

The effectiveness of detection dogs in this context lies in their ability to interpret volatile chemical signals emitted by plants. Unlike visual identification, which is subject to seasonal variation, morphological similarity, and observer expertise, scent provides a stable and highly discriminative dataset.

In practice, this creates a three-part conservation system:

• human expertise defines ecological and taxonomic targets
• dogs translate plant chemistry into field-level detection
• captive animal outcomes validate plant suitability through biological response

This feedback loop allows refinement of both ecological sourcing and captive husbandry practices over time.

Conclusion

The use of detection dogs for locating eucalyptus species suitable for phasmid conservation began as an experimental idea within a postgraduate study context, supported by rescued dogs and an amateur insect-keeping community in the Wollongong region. What started as a practical solution to a sourcing problem has evolved into a broader conservation methodology.

Its relevance now extends from captive phasmid breeding systems to critically endangered species such as the Lord Howe stick insect, and potentially to folivorous mammal management in captive and semi-managed environments.

At its core, the approach reflects a shift in conservation practice: away from purely visual classification systems and toward multisensory ecological tools that integrate animal cognition, plant chemistry, and applied fieldwork.

In that sense, the dogs are not just locating trees—they are helping translate the chemical language of ecosystems into something conservation systems can use, one scent trail at a time.