The illusion that ethnography is a matter of sorting strange and irregular facts into familiar and orderly categories--this is magic, that is technology--has long since been exploded. What it is instead, however, is less clear. That it might be a kind of writing, putting things to paper, has now and then occurred to those engaged in producing it, consuming it, or both. But the examination of it as such has been impeded by several considerations, none of them very reasonable. One of these, especially weighty among the producers, has been simply that it is an unanthropological sort of thing to do. What a proper ethnographer ought properly to be doing is going out to places, coming back with information about how people live there, and making that information available to the professional community in practical form, not lounging about in libraries reflecting on literary questions. Excessive concern, which in practice usually means any concern at all, with how ethnographic texts are constructed seems like an unhealthy self-absorption--time wasting at best, hypochondriacal at worst. The advantage of shifting at least part of our attention from the fascinations of field work, which have held us so long in thrall, to those of writing is not only that this difficulty will become more clearly understood, but also that we shall learn to read with a more percipient eye. A hundred and fifteen years (if we date our profession, as conventionally, from Tylor) of asseverational prose and literary innocence is long enough.