It is known that, up to about 21,000 years ago, the primary control on types of regional vegetation has been variation in the global climate. Does the same apply over much longer timescales? Or, for instance, are local ecological factors the main determinant? A well-dated vegetation record from the northwest of the United States, extending back 125,000 years, shows that, over this period too, global climate change determines vegetation types. As this record extends back to the last interglacial, about 125,000-115,000 years ago, it may give clues as to what happens when an interglacial episode ends and a glacial begins.