We present a re-evaluation of the hypothesis of a coupling between galactic cosmic rays, clouds and climate. We have used two independent estimates of low cloud cover from the International Satellite Cloud Climatology Project, covering 16.5 years of data. The cloud cover data are used in conjunction with estimates of galactic cosmic ray flux and measurements of solar irradiance. It is found that solar irradiance correlates better and more consistently with low cloud cover than cosmic ray flux does. The correlations are considerably lower when multichannel retrievals during daytime are used than retrievals using IR-channels only. Due to large autocorrelations, the statistical significance of the results is marginal. A mechanism is suggested whereby solar irradiance variations are amplified by interacting with sea surface temperature (SST), and subsequently low cloud cover. The feasibility of such a mechanism is supported by negative correlations between SSTs and low cloud cover in subtropical regions.