Reply to Popandsonminers:
I wasn't actually paying much attention to the actual sizes of the fragments when I was getting ready to post the pictures, so after your question I went back and took a closer look. In each picture there is a double arrow-ended scale that shows the scale of the picture. In each case for these pictures the scale marker is 0.5 mm long. My apologies to those folks who like to work in inches or thousandths of inches - I tried going that way, but it makes me uncomfortable as I have been a metric guy for far too many years. However, I know a lot of people relate to Mesh Size, particularly in North America, so the 0.5 mm scale is the equivalent of 35 mesh.
Liberation size is one of those wishy-washy terms that can mean different things depending on the application. But for most general applications the "liberation size" is the size that ore needs to be crushed down to to separate the good stuff from the gangue. Liberation size is extremely dependent on the character of the ore. For example it depends on both mineral particle size and gangue particle size. Not only that, but it depends on the particle size distribution of each. It also depends on the location of the good stuff withing the gangue. For example, particles of good stuff that are occluded within gangue particles will have a significantly different liberation size than the same size of good stuff that is always on the outside of the gangue particles. To complicate the matter even more, the liberation size may depend on the type of process that is being used to recover the good stuff. For example, the liberation size for a leaching process only requires that all teh particles of good stuff be exposed to the leaching fluid - not necessarily completely separated from the gangue. On the other hand, flotation processes require that the good stuff and the gangue are completely separated to allow recovery of the good stuff - so a different liberation size.
OK, that got sort of long winded.
Anyway, for the pyrite pictures above, the frame showing the larger fragments clearly had lots of pyrite clinging to quartz. The average size of particles in that sample was about 2.00 mm or 10 mesh. So the liberation size defined as the size needed to separate pyrite from quartz was smaller than 10 mesh. In the photo showing the smaller particles the average size was about 0.5 mm or 35 mesh. In this case, there was almost complete separation of the pyrite from the quartz. so the liberation size was greater than 35 mesh.
So there we have a rough petrographic study telling us that for this ore sample the liberation size is somewhere between 10 and 35 mesh. If this was ore for a new mine it would not require a lot more study to pin this liberation size down even closer for purposes of designing crushing equipment for pyrite recovery.