I have just read the column concerning late elution on HPLC (LC/GC March 99, v17(3) pp. 230-234). I have another solution for handling this problem. Time the runs so that the late peak doesn't interfere with the peaks of interest in the run. I analyze Lysine by HPLC, and there is a late peak at 45 minutes. I time the run for 22 min. With the time that the autoinjector takes to inject the sample, this makes the late peak come out after my lysine peak on the third run and for every sample thereafter.