Modelling future climates is both a very difficult task and one of key importance to our future well-being. Why then would we want to make it more difficult by shortening our perspective to the most recent period, ignoring historical data that would provide the long perspective on key processes? Palaeoenvironmental research has resulted in important advances in our understanding of global changes in oceanic, atmospheric and biotic systems during glaciation and the onset of the current glacial epoch. However, few palaeoenvironmental studies extend farther back than half a million years, and, with the exception of the recently completed Cape Roberts drilling project, vanishingly few consider events prior to the Eocene (55–38 million years ago).