Thirty-five years ago this week, Nature published a paper titled 'Man-made carbon dioxide and the “greenhouse” effect' by the eminent atmospheric scientist J. S. Sawyer (Nature 239, 23–26; 1972). In four pages Sawyer summarized what was known about the role of carbon dioxide in enhancing the natural greenhouse effect, and made a remarkable prediction of the warming expected at the end of the twentieth century. He concluded that the 25% increase in atmospheric carbon dioxide predicted to occur by 2000 corresponded to an increase of 0.6 °C in world temperature.

In fact the global surface temperature rose about 0.5 °C between the early 1970s and 2000. Considering that global temperatures had, if anything, been falling in the decades leading up to the early 1970s, Sawyer's prediction of a reversal of this trend, and of the correct magnitude of the warming, is perhaps the most remarkable long-range forecast ever made.

Sawyer's review built on the work of many other scientists, including John Tyndall's in the nineteenth century (see, for example, J. Tyndall Philos. Mag. 22, 169–194 and 273–285; 1861) and Guy Callender's in the mid-twentieth (for example, G. S. Callendar Weather 4, 310–314; 1949). But the anniversary of his paper is a reminder that, far from being a modern preoccupation, the effects of carbon dioxide on the global climate have been recognized for many decades.

Today, improved data, models and analyses allow discussion of possible changes in numerous meteorological variables aside from those Sawyer described. Hosting such discussions, the four volumes of the Intergovernmental Panel on Climate Change 2007 assessment run to several thousand pages, with more than 400 authors and about 2,500 reviewers. Despite huge efforts, and advances in the science, the scientific consensus on the amount of global warming expected from increasing atmospheric carbon dioxide concentrations has changed little from that in Sawyer's time.