Lab & Field Science · Kenya · June–July 2025

Reading predators
through their prey

Working in Lewa's research lab under the head of predator research, I contributed to scat analysis work identifying what lions and hyenas are eating — and what that means for the conservancy's endangered species.

Microscopy Hair Identification Predator-Prey Dynamics Decision Tree Development

Why it matters

Endangered species, predator pressure, and the data in between

Understanding what predators eat is foundational to conservation management. For Lewa and the neighbouring Borana Conservancy, this question has a specific urgency: both conservancies together host 317 of Kenya's remaining Grevy's zebra — out of a national population of just over 3,000. Knowing whether lions and hyenas are preying on Grevy's zebra, and at what rates, directly informs how the conservancy manages and protects the herd.

Scat analysis reveals what predators are consuming — which helps identify key prey species, track ecosystem interactions, and guide intervention where it matters most.

This work was led by Felix Kasyoki, head of predator research at Lewa, whose quarterly reports on predation dynamics across the landscape depend on consistent, accurate scat identification. During my placement, I worked alongside Felix to assist with this process — and contributed to improving its accuracy and documentation.

317
Grevy's zebra at Lewa/Borana
~3,000
Kenya's total Grevy's population
20
Hair strands per scat sample

The Lab Process

From field collection to microscope slide

🌿

Collection

Scat samples from lions and hyenas collected across the Lewa and Borana landscape during field days.

🌡️

Drying

Samples are dried to remove moisture and begin separating organic material for analysis.

💧

Rehydration & washing

Dried samples are rehydrated and washed to isolate hair strands and other identifiable material.

🔬

Microscope mounting

20 hair strands collected per sample, mounted on slides for microscopic examination of medulla and cortex structure.

Under the Microscope

Hair structure as a species identifier

Each animal species has a uniquely structured hair strand. Differentiation requires careful analysis of the medulla (the innermost, darker section) and the cortex (the outer layer) — looking at size, colour, breakage patterns, and overall morphology.

Hair strand under microscope — single strand showing medulla and cortex

Single hair strand under microscope — the dark inner section is the medulla, surrounded by the lighter cortex layer.

Multiple hair strands under microscope — showing structural comparison

Multiple strands mounted together — comparing structure, thickness, and tip morphology across specimens from the same sample.

Hair Identification

What you're actually looking at

Cortex Medulla Root

Cross-section diagram — hair strand anatomy

The medulla and cortex tell the story

Every mammal species produces hair with a distinct internal structure. The medulla — the dark inner core — varies in width, continuity, and patterning between species. The cortex surrounds it and differs in colour, texture, and thickness.

Identifying a hair strand to species level means examining: whether the medulla is continuous or fragmented, the relative width of medulla to cortex, the shape of the hair tip, the colour and texture of the cortex, and any distinctive features like segmentation.

The challenge: Many prey species share very similar hair structure — particularly Grevy's vs plains zebra, and different antelope species. Accurate identification requires cross-referencing with a reference library and applying consistent decision criteria.

My Contribution

Building a better identification framework

Drawing on guidance from Felix and existing reference material, I worked to refine and improve the identification process. This included improving reference sketches of known species, taking photographs of actual hair samples under the microscope for a visual reference library, and — the most significant contribution — developing a decision tree to systematise the identification logic.

My university training in earth science and mineral identification — which involves similar logic: classifying samples by physical properties under magnification — gave me a useful framework to apply here.

Hair identification decision tree for wild animal species at Lewa

The hair identification decision tree developed during the placement — classifying wild animal species by medulla structure, cortex properties, and tip morphology.

How the tree works: The first branch asks whether the hair's medulla is disjointed or continuous. This single property splits most species into two groups — those with fragmented medulla patterns (like Grevy's zebra variants, buffalo, impala) and those with continuous patterns (like plain zebra variants, giraffe, waterbuck). Subsequent branches narrow down by cortex colour, tip shape, and strand thickness until a species identification is reached.

What I contributed

Beyond the protocol

🌿

Decision tree development

Built a systematic identification framework to guide analysts through species classification by hair structure properties step-by-step.

📸

Visual reference library

Photographed actual hair samples under the microscope to create a photographic reference library to support future identification work.

✏️

Improved reference sketches

Refined the hand-drawn reference sketches used by analysts to compare microscope observations against known species profiles.

🎓

Cross-disciplinary application

Applied earth science and mineral identification methodology from university training to a biological classification challenge — an unexpected but effective transfer of skills.

← Also from this placement Wildlife AI Detection Model ← All Projects