New Zealand botanicals in skincare, ingredient by ingredient

Published 6 August 2026Zane Hitchcox, editor

The New Zealand flora was isolated for a very long time, which is a good reason to expect defensive chemistry that exists nowhere else and a poor reason to assume it does anything useful on a human face.

Three questions, asked of every ingredient

Is there a plausible mechanism? Something in the extract has to act on a cell, a microbe or a lipid membrane, and somebody has to have identified what. Is there human evidence on skin? A compound that kills Candida in a petri dish, or protects a mouse, or works when swallowed, has not been shown to work when rubbed on. And could the amount in a typical product plausibly deliver it? A botanical listed below the preservative is present at one percent or less, often much less.

A standing point follows from all three. An extract potent enough to be worth putting on a label for a reason other than romance is a pharmacologically active substance, and pharmacologically active substances irritate and sensitise skin. That is the trade nobody advertises. When 140 patients at a contact dermatitis clinic were patch-tested against a tray of 47 botanical extracts, the commonest relevant positive reaction was to tea tree oil — a botanical chosen by consumers precisely because it is active.¹ A 2020 paper in Science Immunology supplied part of the mechanism: small, very hydrophobic molecules of the kind that fill essential oils and oleoresins can bind inside CD1a, an abundant protein in human skin, and trigger T cell responses without behaving like conventional peptide antigens.² "Natural" and "gentle" are unrelated properties, and for the more interesting plants here they are close to opposites.

A New Zealand label need not say how much of the hero plant is in the jar, nor even use the international ingredient nomenclature.

That rule is most of the distance between a front-of-pack story and a back-of-pack fact, and reading the list properly is the only way to close it.

Mānuka — Leptospermum scoparium: honey and oil are different products

Two entirely different materials come off this shrub, and they are routinely conflated. The honey is made by bees from mānuka nectar. The oil is steam-distilled from branches and leaves and appears on labels as Leptospermum Scoparium Branch/Leaf Oil. They share a species name and almost nothing else.

The honey's chemistry is settled. In 2008 Mavric and colleagues at TU Dresden measured methylglyoxal in six New Zealand mānuka honeys at 38 to 761 mg/kg — up to a hundred times the level in conventional honeys — and showed the honeys inhibited E. coli and S. aureus at dilutions corresponding to 1.1 to 1.8 mM methylglyoxal, matching the minimum inhibitory concentration of methylglyoxal itself.⁴ That is the MGO number on the jar, and it is a real measurement of a real antibacterial.

What has been tested on people is wounds. The 2015 Cochrane review pooled 26 trials and 3,011 participants and found high-quality evidence, from two trials of 992 people, that honey dressings heal partial-thickness burns faster than conventional dressings; beyond burns and infected post-operative wounds it judged the evidence low or very low quality.⁵ The nearest thing to a facial-skin study is a 14-patient within-person comparison in atopic dermatitis whose authors said it needs confirming by randomised controlled trials.⁶ Certainty for mānuka honey on healthy or cosmetically ageing facial skin: not established, and the full reading of those trials is set out separately.

Mānuka oil is a separate case, and a variable one. Porter and Wilkins characterised commercial New Zealand mānuka oil in Phytochemistry in 1999: sesquiterpene hydrocarbons predominate at 60% or more, oxygenated sesquiterpenes and triketones together account for up to 30%, and the antimicrobial activity of the commercial oil tracked a single fraction containing six triketones.⁷ Non-commercial oils from different sites differed so widely that the authors sorted them into four groups, so "mānuka oil" on a label is not a specification. A 2005 Planta Medica paper found a triketone-rich mānuka oil virucidal against herpes simplex virus types 1 and 2 in cell culture.⁸ The nearest thing to a human trial of either oil is not a skin study at all: a 2009 feasibility trial in European Journal of Oncology Nursing gave 19 patients a gargle of mānuka and kānuka oils for radiotherapy-induced oral mucositis, and called its own sample size a significant limitation.⁹ Mucosa is not skin. Certainty for mānuka oil on skin: not established, with the added point that a triketone-rich essential oil is exactly what the patch-test literature above is about.

Kānuka — Kunzea: the one with a positive trial

Kānuka is a different genus from mānuka, and the name on the label is usually out of date: de Lange's 185-page revision in PhytoKeys in 2014 split the New Zealand Kunzea ericoides complex into ten species, all endemic, seven of them new.¹⁰ A jar labelled Kunzea ericoides may hold Kunzea robusta or several others. Kānuka essential oil is also chemically unlike mānuka oil: the same paper characterises it by α-pinene at over 50%, with viridiflorol and viridiflorene below 10%.⁷ No triketone appears among the components those authors identified in it, so the fraction that carried mānuka oil's antimicrobial activity is not part of kānuka oil's described composition. Nothing reads across from one oil to the other.

The evidence that matters here is about kānuka honey, not oil. The Medical Research Institute of New Zealand ran a randomised controlled trial of a 90% medical-grade kānuka honey and 10% glycerine preparation in rosacea across five primary-care sites, published in BMJ Open in June 2015.¹¹ It is the best-evidenced claim any New Zealand botanical has on facial skin.

The same group ran the same preparation against acne the following year, and that trial was negative.¹² Reporting only the first would be dishonest.

Certainty for kānuka honey in rosacea: low. For acne: low, and pointing at no useful effect on top of antibacterial soap — the trial found no benefit, but with five responders in total it is too imprecise to prove there is none. For anything else — ageing, dryness, general "calming" — not established. The rosacea trial reads differently once the withdrawal figures and the comparator are taken into account, which is why it gets a page of its own.

Kawakawa — Piper excelsum: heavily sold, barely studied on skin

Kawakawa balm is a New Zealand staple, and it rests on no published skin evidence whatever. Balms are made by infusing the heart-shaped leaves — by convention the ones most holed by the kawakawa looper moth — into a carrier oil or beeswax. On an ingredient list it appears as Piper Excelsum Leaf Extract, or under the older synonym Macropiper excelsum.

The chemistry has been done properly, by Plant & Food Research. A 2019 Journal of Ethnopharmacology paper identified the key metabolites of kawakawa leaf and brewed tea as diayangambin, elemicin, myristicin, unidentified lignans and amides, and ran 14-day and 28-day rat safety studies at doses equivalent to up to four cups of tea a day, finding no adverse effects under those conditions.¹³ Human work exists too, and all of it involves swallowing the plant: a 2024 Molecular Nutrition & Food Research study from the Liggins Institute gave kawakawa tea to 36 volunteers across two studies and identified 26 urinary metabolites associated with drinking it.¹⁴

The gap is the important part. There is no published randomised trial of kawakawa balm on human skin for any condition — not eczema, not dryness, not wound healing, not itch. What is published concerns tea. Certainty for kawakawa on skin: not established, and the longer account of what a balm actually is does not improve that answer.

Horopito — Pseudowintera colorata and polygodial

This is the one with real pharmacology. Horopito is the red-blotched shrub sometimes called the New Zealand pepper tree, and its active constituent, polygodial, is a sesquiterpene dialdehyde with antifeedant properties. The red is not decoration: a 2012 New Phytologist study found that wild leaves with wider red margins carried more polygodial and had taken less herbivore damage, and that caterpillars ate the green-margined leaves preferentially under white light but not when colour discrimination was defeated.¹⁵ The colour advertises the chemistry. Polygodial itself was characterised as an anti-Candida agent by McCallion and colleagues in Planta Medica in March 1982, the second paper in a series on antibiotic substances from New Zealand plants.¹⁶ That work was laboratory microbiology: activity against yeast in culture is not a finding about people.

Two human trials exist, and are almost always misdescribed. Both tested an oral tablet, coded K-712, whose 100 mg contains 10 mg of Pseudowintera colorata oleoresin standardised to 30% polygodial, in recurrent vulvovaginal candidiasis, against itraconazole.¹⁷ Neither was topical, neither concerned skin, and both came from one research group in one journal outside dermatology.

A 2013 follow-up by the same group in 122 women reported fewer relapses on the horopito preparation than on weekly itraconazole (22 versus 39) over two years, with comparable overall mycological cure.¹⁸ Together these are a reasonable basis for further work on an oral antifungal and no basis at all for a claim about skin. Certainty for horopito applied topically, for any indication: not established. And the only published report of horopito applied to a human body site is not a benefit study: a 2007 case report in Acta Dermato-Venereologica of contact vulvitis attributed to Pseudowintera colorata in a topical herbal medicament.¹⁹ One case is not a rate, but it points the opposite way from the marketing.

Harakeke — Phormium tenax seed oil

Harakeke is New Zealand flax, unrelated to linseed flax (Linum). The cosmetic ingredient is a cold-pressed oil from the black seeds, listed as Phormium Tenax Seed Oil. It is generally described as rich in linoleic acid, and the compositional chemistry behind that description is old — Morice's 1970 survey of the seed fats of New Zealand and Australian monocotyledons in Phytochemistry.²⁰ That is a lipid survey, not a dermatological study, and this page does not repeat the percentage figures that circulate in supplier literature: certificates of analysis are not published data.

The mechanism claimed for a linoleic-rich oil is real in outline. Linoleic acid is the fatty acid esterified into acylceramide and acylglucosylceramide in the epidermis, and feeding studies in essential-fatty-acid-deficient rats established decades ago that supplementing linoleate — but not oleate, and not α-linolenate — restores the epidermal water permeability barrier, independently of linoleate's role as the precursor to arachidonate.²¹ That does not establish that spreading a linoleic-rich seed oil on intact adult skin repairs anything, or that harakeke oil does it better than sunflower oil, which is also linoleic-rich, far cheaper, and equally untested against it.

Published New Zealand research on Phormium mostly concerns other parts of the plant — a 1987 Journal of Natural Products paper isolated antifungal compounds from the roots²² — and there is no published human trial of harakeke seed oil on skin at all. Certainty: not established. As an emollient it is a perfectly reasonable ingredient; as a differentiated active it is a story.

Tōtara — Podocarpus totara and totarol

Tōtara heartwood resists rot, and the compound most often credited with it is totarol, a diterpenoid phenol named for the tree. It has a substantial antibacterial literature, one strand of it pointing straight at skin. Kubo and colleagues, testing six diterpenoids against twelve organisms in the Journal of Natural Products in 1992, found totarol bactericidal only against Gram-positive bacteria, and the most sensitive of those was Propionibacterium acnes — the acne organism, since reclassified as Cutibacterium acnes.²³ A 1996 Planta Medica paper located the mechanism in the bacterial respiratory chain, near coenzyme Q.²⁴ And in 1999 a Massey University group in FEMS Microbiology Letters showed that totarol and eight other diterpenes potentiate methicillin against methicillin-resistant Staphylococcus aureus, one of the nine cutting the minimum inhibitory concentration 256-fold.²⁵

Every one of those is a microbiology result; none involved human skin, and killing C. acnes in a well plate is a long way from reaching it inside a follicle at whatever concentration a cosmetic delivers. The species point matters too: the 1992 and 1996 papers both used totarol isolated from Podocarpus nagi, an Asian conifer, not from tōtara, so "tōtara extract" on a label and "totarol" in the literature are not automatically the same material. Nor is totarol content something a label discloses: an extract sitting below the preservative is 1% or less, and how much of that is totarol goes unstated. Certainty for tōtara or totarol on human skin, for any cosmetic indication: not established.

Native seaweeds and algal extracts

Seaweed is the vaguest category on any New Zealand label, because "native seaweed complex" is a marketing phrase and not a specification. The large brown algae here include Macrocystis pyrifera (giant kelp), Durvillaea antarctica (rimurapa, bull kelp) and Ecklonia radiata. The one New Zealand has written a commercial harvesting policy for, though, is Undaria pinnatifida — not native at all. It is a kelp native to Japan, classified as an unwanted organism under the Biosecurity Act 1993. The Ministry for Primary Industries prohibits harvesting it from natural surfaces except as pest control, permits harvest from artificial surfaces and beach-cast material, and allows farming in three areas only — Wellington, Marlborough and Banks Peninsula.²⁶ A "native seaweed" ingredient can therefore be an invasive species scraped off a marine-farm structure under a biosecurity permit.

The chemistry of interest is fucoidan, a sulfated polysaccharide, characterised from New Zealand-grown Undaria specifically: a 2014 Frontiers in Nutrition paper from Auckland University of Technology isolated it from sporophylls grown in the Marlborough Sounds, measured 15.02% sulfate and a molecular weight above 150 kDa, and tested the fractions against cancer cell lines.²⁷ The endpoint has nothing to do with skin. Skin-relevant work is preclinical. HaCaT keratinocytes and HFF-1 fibroblasts pretreated with Undaria fucoidan showed less senescence and less reactive oxygen species after UV exposure in a 2021 study.²⁸ And extracts of Macrocystis pyrifera and Durvillaea antarctica — both of which grow in New Zealand, though the work was done in Chile — inhibited HSV-1 and HSV-2 in culture in 2020 and, formulated topically in a mouse model of HSV-1 skin infection, reduced the severity and duration of lesions by more than aciclovir.²⁹ That mouse result is the strongest topical finding on this page for any seaweed — and it is an animal study, against a virus, for an indication no cosmetic may claim. A keratinocyte UV result is not a sunscreen claim. Certainty for topical seaweed or fucoidan on human skin: not established.

Kiwifruit and other food-crop extracts

New Zealand's horticultural exports supply a second tier: kiwifruit (Actinidia deliciosa and Actinidia chinensis), blackcurrant, grape and apple extracts, usually sold as antioxidant or gently exfoliating. The exfoliation story has a plausible basis — kiwifruit contains actinidin, a cysteine protease that digests protein, which is why it tenderises meat and will not set gelatine.

The same protein is most of the problem. Pastorello and colleagues identified the 30 kDa major allergen of kiwifruit as actinidin itself, in the Journal of Allergy and Clinical Immunology in 1998.³⁰ Contact urticaria from kiwifruit on eczematous skin is documented too, though thinly: a 2015 report worked three patients up with in-vivo component-resolved diagnosis, described them as the first such cases, and concluded that the allergen responsible was thermolabile, gastrosensitive and none of the thirteen kiwifruit allergens then known.³¹ The skin-contact culprit, in other words, is not established to be actinidin. What is established is the shape of the trade described at the top of this page: the thing that makes kiwifruit interesting in a cosmetic — a protein active enough to digest other proteins — is inseparable from the thing that makes it an allergen, which is that it is a protein. Compromised skin is where both effects are largest. There is no published trial of a topical kiwifruit extract for any cosmetic endpoint. Certainty: not established, with a documented allergy risk better established than any benefit.

Merino wool keratin: a real biotechnology, thin dermatology

This one is not a botanical, and the engineering behind it is the most substantial thing on the page. Wool is keratin, and from the early 2000s New Zealand companies worked out how to solubilise it into peptide fractions with controlled molecular weight and cysteine chemistry — the ingredients that appear on labels as hydrolysed keratin or keratin amino acids.

The published skin evidence is thinner than the engineering, and it comes down to two papers by Barba and colleagues, both reporting in vivo volunteer studies. The 2008 paper in Skin Research and Technology applied a wool-derived hydrolysed keratin peptide of molecular weight under 1000 Da, in an aqueous solution and in an internal-wool-lipid liposome suspension, and reported increases in hydration and elasticity against untreated control sites, plus improved water-holding capacity.³² The 2007 paper in the Journal of Cosmetic Science used a different material — a high-cystine keratin fraction in the S-sulphonated form — on hands, and reported improved hydration and elasticity in volunteers with dry skin and some protection against surfactant-induced damage.³³

Three limits matter more than the results. The outcomes are biophysical instrument readings, not clinical endpoints a person would notice, and the published abstracts state neither participant numbers nor effect sizes, so this page quotes none. Both papers carry the same six authors, two of whom give the addresses of the New Zealand keratin companies Canesis Network Ltd and Keratec Ltd. That is normal in ingredient science and still a conflict worth knowing about. And a hydrated, more elastic stratum corneum is what any competent humectant-and-occlusive moisturiser produces; nothing here shows keratin peptides do it better than glycerin. Certainty for wool-derived keratin peptides improving skin hydration and elasticity: low. For anything beyond that, including the "rebuilds skin" framing the ingredient usually carries: not established.

Where a claim crosses a legal line

None of this stops a product carrying a botanical on the front of the pack. What it bears on is what the pack may say the botanical does. Cosmetics and medicines are separate categories in New Zealand, and what moves a product between them is the claim, not the formula.

It is also why the vaguer phrasing — "soothing", "restoring", "supports the skin barrier" — is so widespread. Those words survive regulatory scrutiny precisely because they mean nothing testable.