For decades, cannabis growers have stripped the leaves off their plants and thrown them away. The flowers hold the cannabinoids that drive the legal market, and everything else is treated as agricultural waste. But a team of chemists at Stellenbosch University in South Africa just found 25 compounds in those discarded leaves that have never been documented in cannabis before, including 16 that belong to one of the rarest chemical classes in the entire plant kingdom.
Their results, published in the Journal of Chromatography A in early 2026, identified 79 phenolic compounds across the leaf and flower tissue of three commercial South African cannabis strains. Twenty-five of those compounds had never been reported in cannabis. And 16 of the 25 were classified as flavoalkaloids according to a Stellenbosch University institutional summary of the research; whether the peer-reviewed paper itself uses that exact count has not been independently confirmed from the full text. Flavoalkaloids are hybrid molecules that sit at the chemical crossroads of two well-known families: flavonoids (the pigments and antioxidants found in berries, tea, and citrus) and alkaloids (the potent nitrogen-containing compounds behind caffeine, morphine, and nicotine).
Flavoalkaloids are so scarce across all of botany that most plant scientists have never encountered one. Finding 16 of them concentrated in the leaves of a single cannabis strain is, analytically speaking, a significant haul.
How the compounds were found
The discovery hinged on a technique called online comprehensive two-dimensional liquid chromatography. In simplified terms, the method runs a plant extract through two different separation systems back to back, sorting molecules first by how well they dissolve in water and then by how strongly they cling to an oily surface. Pairing that two-stage separation with high-resolution mass spectrometry allowed the team to pull apart and tentatively identify compounds that would blur together on a conventional single-column instrument.
That word “tentatively” matters. The identifications are based on chromatographic behavior and mass spectral fingerprints, not on isolating each molecule and confirming its structure through nuclear magnetic resonance (NMR). The researchers provide retention-time maps and mass-accuracy values for every compound, which is strong analytical evidence, but some assignments could shift once pure reference standards become available. For the flavoalkaloids in particular, reference libraries are thin because so few of these molecules have been cataloged in any species.
Why the leaves matter
The commercial cannabis industry is built almost entirely around the flower. Extraction operations target cannabinoids like THC and CBD, and processing facilities routinely discard leaf material. Earlier profiling work that compared cannabis inflorescences, leaves, stem bark, and roots had already shown that leaves carry meaningful concentrations of non-cannabinoid metabolites, including flavonoids and terpenoids. But the sheer number of previously unknown phenolics now documented in leaf tissue raises a pointed question: how much potentially useful chemistry has the industry been composting?
A separate open-access study, published in Scientific Reports (note: this DOI references a 2026 publication whose availability should be verified independently), reinforces the pattern. Researchers using different analytical platforms on different hemp varieties documented rich chemical diversity in leaf material, arriving at conclusions that parallel the South African findings. The convergence of results from independent labs, different strains, and different continents strengthens the case that cannabis leaves hold far more chemical complexity than the industry has assumed.
On the cannabinoid side, a similar expansion is underway. GC-MS metabolite profiling of high-potency cannabis, described in a Universitat Politecnica de Catalunya press release (no peer-reviewed DOI or journal citation has been provided for this specific dataset), detected 43 cannabinoids in total, with 16 described as previously unreported homologous cannabinoids featuring altered side-chain lengths. The throughline is consistent: every time scientists point higher-resolution instruments at cannabis tissue, the catalog of known molecules grows substantially.
What the study does not show
No bioactivity data accompany the discovery. Whether these newly detected phenolics or flavoalkaloids have anti-inflammatory, antioxidant, antimicrobial, or any other pharmacological effects is entirely unknown at this stage. A Stellenbosch University summary of the research frames leaves as holding “medical potential,” but that language reflects a hypothesis, not a tested outcome. No cell-based assays, animal models, or clinical trials have been published for these specific molecules.
The geographic scope also limits how broadly the results apply. Only three commercial South African strains were analyzed. Cannabis chemistry varies widely by genotype, growing conditions, nutrient regimes, and harvest timing. Whether the same flavoalkaloids appear in North American, European, or Asian cultivars has not been tested. The hemp leaf study used different varieties and a different analytical platform; while it confirmed chemical richness in leaves, it did not report the same flavoalkaloid class. That gap could reflect genuine strain differences, divergent biosynthetic pathways, or simply differences in detection sensitivity between laboratories.
Historical context adds another layer of caution. A comprehensive review of phenolics and alkaloids previously reported in cannabis shows that the known catalog of non-cannabinoid constituents has grown steadily with each generation of instruments. “New to science” sometimes means “new to this level of detection” rather than truly absent from earlier plant material. Trace compounds that once fell below detection limits can now be resolved and annotated, and the boundary between known and unknown shifts with every methodological advance.
What comes next for cannabis leaf chemistry
The most defensible takeaway is narrow but consequential. Cannabis leaves, long treated as low-value biomass, contain a broad and largely uncharacterized array of phenolic compounds, including rare flavoalkaloids in at least some commercial strains. Modern multidimensional chromatography paired with high-resolution mass spectrometry can resolve dozens of these metabolites simultaneously, revealing chemical territory that older one-dimensional methods missed.
Turning those analytical findings into anything practical will require several steps that have not yet been taken: isolating individual compounds in sufficient quantity, confirming their structures through NMR, running targeted bioassays to test for pharmacological activity, and repeating the analysis across a wider range of cultivars and growing regions. Until that work is done, the South African study is best understood as a detailed chemical survey that opens new lines of inquiry rather than a direct guide to medical use or product development.
But the broader implication is hard to ignore. A part of the cannabis plant that the industry has systematically discarded turns out to harbor molecules that chemists have barely begun to catalog. Whether those molecules eventually prove useful in medicine, nutrition, or agriculture, the evidence published so far makes a strong case that the leaves deserve far more scientific attention than they have received. The next round of answers will depend on whether funding and regulatory frameworks catch up to what the instruments have already revealed.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
