I will try and illustrate a comment I made on Steve Trigg's post below from some of my Notocelia photos. I usually concentrate on the costal stigulae pattern to separate these two. I just hope that Martin Harvey's recent comment about current research into this pair doesn't result in unsafe separation by wing markings!
To me, trimaculana has three pairs of white costal stigulae, fairly broad and quite perpendicular to the wing edge, whereas rosaecolana has an additional pair of fine white costal stigulae towards the centre of the wing edge, all of which appear a bit more oblique in direction.
N. trimaculana at the top (from this morning) and N. rosaecolana from my archive below:
I hope that helps.