Svensmark and Friis-Christensen  proposed a ``cosmic ray-cloud cover'' hypothesis that cosmic ray flux, modulated by solar activity, may modify global cloud cover and thus global surface temperature by increasing the number of ions in the atmosphere, leading to enhanced condensation of water vapor and cloud droplet formation. We evaluate this idea by extending their period of study and examining long-term surface-based cloud data (from national weather services and the Global Telecommunication System) as well as newer satellite data (International Satellite Cloud Climatology Project (ISCCP) D2, 1983-1993). No meaningful relationship is found between cosmic ray intensity and cloud cover over tropical and extratropical land areas back to the 1950s. The high cosmic ray-cloud cover correlation in the period 1983-1991 over the Atlantic Ocean, the only large ocean area over which the correlation is statistically significant, is greatly weakened when the extended satellite data set (1983-1993) is used. Cloud cover data from ship observations over the North Atlantic, where measurements are denser, did not show any relationship with solar activity over the period 1953-1995, though a large discrepancy exists between ISCCP D2 data and surface marine observations. Our analysis also suggests that there is not a solid relationship between cosmic ray flux and low cloudiness as proposed by Marsh and Svensmark .