Estimates of cloud cover over the South Pole are presented from five different data sources: routine visual observations ; C vis ), surface-based spectral infrared (IR) data (2001; C PAERI ), surface-based broadband IR data (1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003); C pyr ), the Extended Advanced Very High Resolution Radiometer (AVHRR) Polar Pathfinder (APP-x) dataset (1994-99; C APP-x ), and the International Satellite Cloud Climatology Project (ISCCP) dataset (1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003); C ISCCP ). The seasonal cycle of cloud cover is found to range from 45%-50% during the short summer to a relatively constant 55%-65% during the winter. Relationships between C pyr and 2-m temperature, 10-m wind speed and direction, and longwave radiation are investigated. It is shown that clouds warm the surface in all seasons, 0.5-1 K during summer and 3-4 K during winter. The annual longwave cloud radiative forcing is 18 W m Ϫ2 for downwelling radiation and 10 W m Ϫ2 for net radiation. The cloud cover datasets are intercompared during the time periods in which they overlap. The nighttime bias of C vis is worse than previously suspected, by approximately Ϫ20%; C ISCCP shows some skill during the polar day, while C APP-x shows some skill at night. The polar cloud masks for the satellite data reviewed here are not yet accurate enough to reliably derive surface or cloud properties over the East Antarctic Plateau. The best surface-based source of cloud cover in terms of the combination of accuracy and length of record is determined to be C pyr . The use of the C pyr dataset for further tests of satellite retrievals and for tests of polar models is recommended.