Stacking of an arene D on caffeine (1) near 7-Me and 8-H is known ( 1 H NMR shifts, CCl 4 , AUS concept) to yield the same association constant K for each signal of 1. Toluene (T) and 1 now yielded the same K from 7-Me and 8-H while K from 1-Me (and 3-Me) is smaller indicating a minor stacking centre near N-1 that forms both 1T and 1T 2 ; this 1T 2 disturbs computations of K. The main stacking centre forms no 1T 2 since its 1T is stabilized by dipole-dipole interaction. Diphenylmethane Dp, a substituted T that cannot form a dimer of type T 2 , does not form 1Dp 2 . Ethyl of 1-ethyltheobromine (2) is proven to stand perpendicular so that the above minor complex can only arise on one face of 2 halving K minor and making the disturbance of computations by 2T 2 insignificant; apart from this minor 2T there are cis-trans isomers of the main complex. Stacking of 1 and 2 with larger arenes is discussed in terms of relative molecular sizes under consideration of the perpendicularly placed ethyl group of 2. Stacking of D near N-3 and N-9 of 1,3,7,9-tetramethyluric acid (4) is known. 4 and T now provided K = 0.218 l mol −1 from 3-Me and 9-Me but K = 0.157 l mol −1 from 1-Me and 7-Me. This and the complex induced shifts IK indicate formation of stacking 4T + 4T 2 and of two edge-on 4T standing on 3-Me and 9-Me under dipole-dipole interaction with a neighbouring oxygen. The oxygens of 4 (and probably of 1, too) form edge-on complexes with hexafluorobenzene; this underlines the importance of polar effects.