Picture the scene as the demonologists described it: a moonlit clearing, a fire at the centre, and thirteen figures in a ring. One leads — sometimes described as a man in black, sometimes as the Devil himself — and twelve follow. The image is so fixed in Western imagination that it feels ancient, inevitable, almost mathematical. It is also, in large part, the invention of a single early twentieth-century anthropologist whose central thesis has been almost entirely dismantled. The thirteen-witch coven is a story about how a number acquires meaning, and how difficult it is to take that meaning back once it has settled in.
The Scholar Who Built the Myth
The specific claim that witches organised themselves into groups of thirteen was popularised — some would say fabricated — by Margaret Murray, an Egyptologist at University College London who turned her attention to European witch trials in the years following the First World War. Her 1921 book The Witch-Cult in Western Europe argued that the accused witches of the early modern period were not devil-worshippers or deluded peasants but practitioners of a continuous pre-Christian fertility religion, organised into cells of exactly thirteen members.
Murray called these cells "covens" — a word that existed before her but which she codified — and she argued that the number thirteen was deliberate and sacred: twelve initiates gathered around a single leader, a conscious inversion of Christ and his apostles, a Satanic mirror of the Church's most resonant image. The theory was elegant, internally consistent, and enormously influential. It was also, as historians spent the subsequent decades demonstrating, built on selective reading, invented connections, and a fundamental misreading of the trial records she claimed as evidence.
Murray's coven of thirteen has no more basis in the historical record than the broomstick. Both are images that tell us what observers expected to see, not what the accused actually did. — Norman Cohn, Europe's Inner Demons, 1975
The actual trial records Murray cited showed no such consistency. Accused groups ranged from three individuals to several dozen. When interrogators extracted confessions of organised gatherings — almost always under torture or the threat of it — the numbers varied wildly and corresponded to local circumstances: who had been named, who had been arrested, who had agreed to implicate a neighbour. The thirteen was not there in the evidence. Murray placed it there.
What the Trials Actually Recorded
The North Berwick witch trials of 1590, among the most documented of the Scottish cases, described gatherings of accused witches that ran to over a hundred individuals at their largest claimed sabbath. The Pendle trials of 1612 involved twelve accused — tantalisingly close to a coven, but not quite. The Salem proceedings of 1692 named over two hundred people before they were done. Real persecutions did not follow Murray's tidy arithmetic.
What the trials did consistently show was the power of suggestion during interrogation. Accused persons, asked to name their accomplices, often named the same number as the interrogator seemed to expect — a phenomenon that modern psychologists of coercive questioning would recognise immediately. When demonological manuals like the Malleus Maleficarum (1487) described sabbaths, they were describing what inquisitors believed they would find, not what witnesses independently reported. The thirteen entered interrogation rooms before the accused did.
The Lunar Argument
There is, separate from Murray's thesis, a more poetic explanation for the number's association with witchcraft — one grounded in the calendar rather than in ecclesiastical inversion. The lunar year contains thirteen full moons. Pre-Christian agricultural societies tracked seasons by the moon, and the thirteenth month — the irregular one that the solar calendar couldn't quite absorb — was associated with liminality, with times outside normal time, with the kind of uncanny threshold that witch-belief clustered around.
If covens did gather by the lunar calendar, as some scholars of folk religion have argued, then thirteen would have been a natural organisational unit: one meeting per moon, one leader per cycle. The number would have been practical before it was symbolic. This theory has the advantage of not requiring the accused to have read demonological literature, but it has the disadvantage of being almost entirely unprovable from existing records. It is a plausible speculation in a field full of plausible speculations.
Gerald Gardner and the Modern Inheritance
Whatever its origins, the thirteen-member coven received its most consequential endorsement in the mid-twentieth century, when Gerald Gardner — a retired colonial civil servant with an interest in occultism — founded the modern religion of Wicca and codified its practices in a series of texts published in the 1950s. Gardner drew heavily on Murray's scholarship, which had by then been absorbed into popular culture and repeated often enough to feel traditional. He specified thirteen as the ideal coven size: large enough to raise and direct what practitioners called the Cone of Power, small enough to maintain secrecy and trust.
Wicca grew, spread, and diversified, and the thirteen-member structure went with it. Contemporary witchcraft communities debate the number actively — many modern covens are smaller, some reject the structure entirely — but the image persists. Thirteen witches in a circle appears on tarot cards, in films, in novels, in the vocabulary of anyone who has ever half-jokingly called their friend group a coven. Murray's discredited scholarship became Gardner's spiritual practice became everyone's cultural shorthand.
The number thirteen did not need the witches to become unlucky. It was already freighted before the first demonologist put quill to parchment. But witchcraft gave it a new geometry: the circle, the fire, the count that stops just past the edge of the ordinary. Whether or not thirteen witches ever actually gathered anywhere, the image of them doing so is now older than anyone alive — which is, in the arithmetic of superstition, more than old enough to count as real.