On the phone from Geneva, Mr. Kirkby says that Mr. Svensmark’s hypothesis “started me thinking: There’s good evidence that pre-industrial climate has frequently varied on 100-year timescales, and what’s been found is that often these variations correlate with changes in solar activity, solar wind. You see correlations in the atmosphere between cosmic rays and clouds—that’s what Svensmark reported. But these correlations don’t prove cause and effect, and it’s very difficult to isolate what’s due to cosmic rays and what’s due to other things.”

In 1997 he decided that “the best way to settle it would be to use the CERN particle beam as an artificial source of cosmic rays and reconstruct an artificial atmosphere in the lab.” He predicted to reporters at the time that, based on Mr. Svensmark’s paper, the theory would “probably be able to account for somewhere between a half and the whole” of 20th-century warming. He gathered a team of scientists, including Mr. Svensmark, and proposed the groundbreaking experiment to his bosses at CERN.

Then he waited. It took six years for CERN to greenlight and fund the experiment. Mr. Kirkby cites financial pressures for the delay and says that “it wasn’t political.”

Mr. Svensmark declines entirely to guess why CERN took so long, noting only that “more generally in the climate community that is so sensitive, sometimes science goes into the background.”