HONEY · GRAYANOTOXIN
Mad honey — Xenophon, Pompey, and a 2,400-year-old battlefield neurotoxin
Grayanotoxin in honey is documented in two of the most famous incidents in ancient military history — Xenophon's army in 401 BCE and Pompey's troops in 65 BCE. The chemistry is the same today. The FDA does not test US honey for it.
The Anabasis, 401 BCE
The earliest detailed account of grayanotoxin poisoning is Xenophon's Anabasis, recording the retreat of the Ten Thousand Greek mercenaries through what is today northeastern Turkey, on the Black Sea coast. The soldiers, finding hives of local honey, ate it in quantity. Xenophon describes the result with clinical precision: those who ate a little appeared drunk; those who ate more became maddened, lay on the ground unable to stand, and seemed near death. By the next day the symptoms began to lift, and on the third day most could rise.
The region — the southeastern coast of the Black Sea, home of Rhododendron ponticum and Rhododendron luteum — is the source of deli bal, the "mad honey" still produced and traded in Turkey today.
Pompey's defeat, 65 BCE
Three and a half centuries later, the same plants were used against Roman troops in the same region. During Pompey's campaign pursuing King Mithridates VI of Pontus (66–63 BCE), the local Heptacomitae tribe placed bowls of mad honey along the Roman line of march. Pompey's soldiers, eager and undisciplined, drank the honey. The Greek geographer Strabo (c. 64 BCE – 24 CE) recorded the result:
They mixed bowls of the crazing honey, placed them in the roads, and when the soldiers drank the mixture and lost their senses, they attacked them and easily disposed of them.
Strabo (Geographica 12.3.18) reports three maniples of Roman troops attacked while incapacitated. The episode is one of the earliest documented deliberate uses of a biological neurotoxin as a battlefield weapon.
Mithridates VI himself — the man Pompey was pursuing — maintained an entire toxicological library, studied poisons systematically as both weapons and medicines, and produced the Mithridatium, an antidote formula that survived as a real pharmaceutical preparation in European apothecaries for over 1,500 years.
The mechanism
Grayanotoxin is a class of diterpenoid neurotoxins produced by plants in the Ericaceae family — rhododendrons, azaleas, mountain laurel. Bees that forage these plants concentrate the toxins in the honey they produce. In the human body, grayanotoxin binds to voltage-gated sodium channels in nerve cell membranes and prevents normal depolarisation.
Acute high-dose presentation: dizziness, weakness, vomiting, hypotension, bradycardia, sometimes complete cardiac block, sometimes fatal. Chronic low-dose presentation: peripheral neuropathy, tingling and numbness in the hands and feet, dizziness, weakness, persistent low-grade impairment — a symptom profile that overlaps a range of modern neurological and anxiety presentations, almost never investigated as food-borne neurotoxin exposure.
What the FDA tests honey for
Grayanotoxin: not tested.
Pyrrolizidine alkaloids (PA) — a separate class of liver-toxic and DNA-damaging alkaloids that the USDA itself calls "the most widespread poisonous plant problem affecting humans, animals, and insects worldwide" — not tested.
Quinolizidine alkaloids from lupines (the class documented in the Water vector): not tested.
Tremetol (in the case of bees foraging white snakeroot or rayless goldenrod near dairy regions): not tested.
Geographic origin tracing in ultra-filtered honey (which has had pollen removed): no longer reliably possible.
Estimates of fraudulent adulteration (added sugar syrups, mis-labelled origin) of imported US honey vary widely by methodology. Independent NMR studies have reported incidence in the 30–70 percent range across surveyed samples. The FDA's own surveillance, based on HPLC methods with different sensitivity and sample frames, typically reports lower single-digit figures. The numerical gap reflects methodology sensitivity, sample selection, and product category — and is separate from the natural-toxin contamination question (grayanotoxin, PAs, quinolizidine alkaloids), which falls into a distinct testing gap because no routine US honey panel covers those analytes at all.
The geographic load
In the United States today, large azalea and rhododendron concentrations are documented along the Texas/Louisiana/Mississippi/Alabama/Florida Gulf Coast belt; through the Carolinas and Virginia; up the Mid-Atlantic; and across the Pacific Northwest. State flowers of Washington, West Virginia, and (on both sides of a contested border) Kashmir are rhododendron varieties — including, in Kashmir's case, Rhododendron ponticum, the same species responsible for Xenophon's incident.
The plants are not hidden. The toxin is documented. The historical record stretches back 2,400 years. The current regulatory testing regime is what it is.