I have a hard time reading this graph: http://lh3.googleusercontent.com/Dqp5VJzn4SRzlKiWcVdavOWhYIuLGpzrnWTtVTq9FXZlxghcAizk-J8qSa1bls9jFtsKoXO8FiFtMqZplcYh9VT

I have a hard time reading this graph: http://lh3.googleusercontent.com/Dqp5VJzn4SRzlKiWcVdavOWhYIuLGpzrnWTtVTq9FXZlxghcAizk-J8qSa1bls9jFtsKoXO8FiFtMqZplcYh9VT_ZL22=s0 http://lh3.googleusercontent.com/Dqp5VJzn4SRzlKiWcVdavOWhYIu... I can see that on either side of the graduated tiles, the diagonally adjacent plots are mirrors of one another. But the x/y dimensions on the quasi-random plots seem so distorted, that information is destroyed. The fact that none of the points are color-coded in any of those tiles, means I can't intuitively draw conclusions by noting corresponding dots across second and third tier composites. each time we pick a point, we should pick it as far away as possible from every other point across all dimensions I don't understand how you can assign a plot like that, and still constrain the mass (M0) and initial helium (Y0) points to the graduated markings on their respective tiles, and come up with a diamond-shaped quasi-random pattern, which still communicates the accurate mass/helium relationship for each dot? Why is the upper right corner in those two M0/Y0 composites sparse, but there are still diamond grid clusters across the rest of those particular tiles?