You claim is correct, but I wasn't comparing disks of the same data density, I was comparing disks of the same data capacity, i.e. the smaller disk has a higher data density.
Having said that, I've realised that my point is only accurate if the track widths remain the same on both disks, which probably isn't the case.
If you picture the platter of a 3.5" disk, you could compact its data into the centre 2.5" by moving all the tracks closer together, or by putting more bits into each of the tracks that lie inside the centre 2.5".
If the former method is used, each track has the same number of bits in it as before. Each head reads the same number of bits per rotation, so the required RPM is the same for both disks.
If the latter method is used, each track has more bits in it. Each head reads more bits per rotation, so the required RPM is lower for the smaller, denser disk.
It is likely that disk capacities are increased by both narrowing the track widths and packing more data into each track, so the 72% I quoted is inaccurate. However, I'd submit that it might be easier to put more bits in each track than to increase the track count, since the latter requires making the heads narrower.
Whatever the balance between the two methods, 2.5" disks of the same capacity only need to spin somewhere between 72 and 100% the speed of 3.5" disks to maintain the same data rate. There's also the added bonus of the heads not needing to move as far when switching tracks.
So I'm sorry for overstating the difference in my first post, but I stand by my claim that larger disks of the same capacity need a higher RPM to maintain the same performance.