Several people from SAS helped me with this, and I appreciate all their work. It looks like, as I always suspect, that you can do most anything in SAS even if not in a totally direct way. Here is the answer that worked for me, given by Sanjay Matange: data faculty;   set faculty;   a=.;   if _n_ = 1 then a=0;   run; proc sgplot data=faculty noautolegend;       where uasrank in ('1','2','3');       series x=uasyear y=uassug50/group=uasrank                                                 markers                                                 markerattrs=(symbol=circlefilled)                                                 lineattrs=(pattern=solid color=black)                                                 curvelabel    curvelabelpos=start;                series x=uasyear y=uasuaavg/group=uasrank                                                 markers                                                 markerattrs=(symbol=squarefilled)                                                 lineattrs=(pattern=solid color=CX990000)   curvelabel   curvelabelpos=end;    series x=uasyear y=a / lineattrs=graphdata1(color=black pattern=solid)              name='a' legendlabel="SUG 50th Percentile";    series x=uasyear y=a / lineattrs=graphdata2(color=CX990000 pattern=solid)              name='b' legendlabel="UA Average";       keylegend "a" "b";       xaxis label=' ';       yaxis label=' ' min=50000;;       format  uasyear $year. uasrank $fac_rank. uassug50 dollar12. uasuaavg dollar12.; run; ... View more