Phil asked me to access Mann's MBH98 predictors from the ftp server in Virginia. Apparently there are 159. However, with no guidance I'm unable to identify a set of 159 data series from the thousands of files in the 'MBH98' directory structure. Furthermore, attempts to download the entire directory hierarchy for examination (it's only about 20Mb in total) have been fraught with failures - 3 hours on Wednesday afternoon and all this morning, and still incomplete. If Steve McIntyre's having the same difficulties, no wonder he's pissed off!