Dear All, A client has asked to assess temporal trends in a multivariate data set (chemical measurements) collected at k different locations. The number of measurements per location (38) is of course greater than the number of stations (15) and number of years (10). One approach is to look at each station separately but that requires 15 analyses and fails to account for the almost certain relationship between stations. Any thoughts on integrating all information would be appreciated. Will summarize and respond to group. Sincerely, Barry Zajdlik