THE LAST OF the American witchcraft trials is now nearly 300 years past, but the witches continue to cast their spell. Young and old alike succumb to their fascination, as John Demos discovered when he was visiting a New England home to look at a family manuscript about Goodwife Eunice Cole, accused of witchcraft in 17th-century Hampton, Massachusetts. As he was reading over the document he heard two children playing, one pretending to be a town official with the job of whipping a "mean old witch," the other pretending to be the very same Goody Cole whose life Demos was researching. "It is a long way from her time to ours," Demos realized, "but Goody Cole has made the whole journey." And so, we might add, have the other witches.

The witches still intrigue us because they seeem so exotic and out of date in 17th-century New England, an odd throwback to earlier centuries, a strange transplant to one part of America at a time when witchcraft was dying out elsewhere in the world. They haunt us, too, because the mere mention of witchcraft immediately brings to mind one of the most dramatic events of our early history--the Salem witch trials of 1692. But it is exactly these assumptions that Demos challenges. Witchcraft was not, as he argues, a rare and anachronistic feature of New England life, "no meandering sideshow, isolated from the history of New England. On the contrary; it belonged to--and in-- that history from beginning to end." It was not confined to Salem but "belonged to the regular business of life in pre- modern times."

Accordingly, Demos mentions the Salem trials only briefly and concentrates instead on witchcraft in other parts of New England. His book is a rich combination of biographical stories of individual witches and their accusers on the one hand and analytical chapters approaching witchcraft from the perspective of different disciplines on the other. He gives us astonishingly full studies of people like Goody Rachel Clinton of Ipswich, accused of sending a child into fits, appearing in the form of a turtle, and mysteriously emptying a barrel of beer; Goody Elizabeth Morse of Newbury, charged with making pots and pans fly and with making "a great stone . . . remove from place to place"; and Goody Elizabeth Garlick of Easthampton, Long Island, charged with using witchcraft to kill five people.

Interspersed with these life stories are the more general chapters. "Biography" gives a collective profile of the accused witches, their age, sex, social and marital status, and so on. "Psychology" is a study of the accusers--the emotions, drives, and inner tensions that led one neighbor to charge another with being the source of his own misfortunes. "Sociology" seeks to identify the features that made New England more fertile for witch hunts than other areas of the world and the characteristics that made some communities within New England particularly susceptible to witch hunting. "History," the final section, asks why communities produced witches at some times and not at others.

Demos' profile indicates that the "typical" witch was a middle-aged, lower-class, quarrelsome woman associated in some way with medicine and previously accused of another crime. She did not get along with people who lived near her; her neighbors, in turn, were disturbed because their dislike of being with her conflicted with their desire for a tightly knit community. In one Demos scenario a fortyish housewife might ask a favor of her neighbor, to have some left-over food, perhaps, or some babysitting time. The neighbor would refuse but feel guilty about being unneighborly. Soon after this encounter a catastrophe would befall the selfish neighbor, possibly a death or illness or "fits" in his family, but something that could be interpreted as retribution sent by the rejected housewife. Who could be capable of such mysterious retribution? Only a witch. This was frightening, but it was also a great relief: there was nothing unneighborly in refusing to help a witch, so the accuser's conscience was now clear.

Demos found such scenarios more common in New England than elsewhere because New Englanders saw their own history as an epic battle between God and Satan; witches were a vital force on Satan's side. He found them more common in some towns than in others because witchcraft was likely to flourish in towns after they had just gone through a period of war or internal conflict and during a time when they were suffering from natural disasters. He found witches more likely to be taken to court if they had offended community leaders than if they had bothered lesser folk and he concluded that New Englanders stopped prosecuting witches at the end of the 17th century not because they ceased to believe in witchcraft but because the community leaders had begun to doubt that sufficient legal evidence could be found to prove the charges against particular individuals.

Much of this interpretation is not surprising; we suspected it all along. But Demos has backed it up with methods borrowed from psychology and sociology that historians have not applied to this subject before, and has filled it in with a wealth of detail from new sources. This wealth is at once a strong point and a weak point of the book.

The book is worth reading as a magnificent lesson in the use of legal sources to recreate the everyday lives of ordinary people in the past, individuals who would otherwise remain quite obscure. But Demos has occasionally been almost too successful in his fact-finding: he ends up telling us more about some of the witches than we really need to know in order to understand their identification with witchcraft. Take for example the biographical sketch of Goody Rachel Clinton. We learn of her unsettled childhood, her mother's insanity and her own unhappy marriage, all of which clearly help us to understand how she might have acted in ways that made her be taken for a witch. But what is the relevance of the name of the English village in which she was born, the name of the ship in which she crossed the ocean, or the activities of her husband long after he had left her? All of these might be relevant, in certain contexts. But when Demos does not give us these contexts, the witches, rather than coming to life, are buried under detail.

It is curious to find other potentially useful perspectives on witchcraft left out of Entertaining Satan. In a book whose main characters are mostly women it is strange to find only a passing mention of feminism and the status of women in 17th-century New England. Witchcraft must have been affected by the roles women were supposed to play and the inability of some women to fit those roles. In a book that is about New Englanders' perceptions of witches it is surprising to find no explanation of what New Englanders thought witchcraft was and how it actually worked. In a book that discusses physical symptoms--headaches, fits, stiff legs and the like--it is odd to have only psychoanalytical explanations of the symptoms with virtually no reference to possible disease or dietary problems.

Two minor points are puzzling. The book concludes that "Witchcraft and conflict exclude one another at any given point in time . . . caught up in the roils of controversy a community would not (could not?) simultaneously sponsor witchcraft proceedings." It is startling, then, to find in one of the few references to Salem that the witch hunt there was particularly intense because "concern with witchcraft was there joined to internal factionalism." Finally, a question: can 93 witchcraft accusations over 60 years in a population that exeeded 100,000 by the end of the period be sufficient to show that witchcraft "belonged to the reguleathar business of life" throughout New England?

Nevertheless the book gives promise of achieving popular success far beyond the small community of historians who might ordinarily be expected to read a scholarly study. If it does, that will be a tribute both to Demos' extraordinary sleuthing and to the magic the witches continue to work.