An article published by Friis-Christensen and Lassen [1991> appeared to indicate an association between solar cycle lengths (SCLs) and climate. It attracted worldwide attention and has since been extensively referred to. We here present an updated analysis using a recent temperature reconstruction with the time period of comparison considerably expanded. The correlation is found to be weak. In the light of this new result we analyze the question how the article by Friis-Christensen and Lassen was able to create the impression of a 'strikingly good agreement,' as the authors described it. We show that the main reason is an unacceptable mixing of filtered and nonfiltered data in the graphical representation. Hereby, an artificial agreement of the solar data with the global warming since 1970 was established. The article by Friis-Christensen and Lassen has created and still creates confusion both in scientific and public discussions on climate change. We have therefore found it relevant to deliver the present analysis. ¿ 2000 American Geophysical Union