COHMAP (Cooperative Holocene Mapping Project) had its start in the 1970s with the intent to compare atmospheric general circulation models with quantitative palaeoclimatic reconstructions based on stratigraphic pollen analysis. As the sophistication of the models increased and as additional palaeoclimatic indicators (lake levels, marine microfossils) were incorporated in the reconstructions, the programme expanded to global coverage through the participation of numerous collaborators. The carefully nurtured dual approach of modelling and data analysis was strengthened by continued testing of one against the other, with emphasis on interdisciplinary understanding fostered by workshops. With the development of new computer models that have higher spatial resolution and improved boundary conditions, and with diversification and enlargement of the geographic and temporal coverage of field and analytical data, substantial refinements can be anticipated in the understanding of global climatic changes since the last glacial maximum.