>Thanks to Jim for raising an issue that has long been a question I have
>never heard an adequate answer to.  What exactly are the beads that people
>use for these measurements. I presume they are not hollow spheres but a
>semi-solid or open matrix.  Are they optically transparent?  I doubt it.
>If not, then how can you expect to get clean data from the far side of teh
>sphere-ie the top of the object away from the light source.  It has always
>seemed to me that the only way you could accurately do this is to make
>spheres out of a refractive index matched material and then resuspend those
>beads in fluorescent dye and image the black holes. Haven't I seen "isn't
>science cool" demonstrations of materials that disappear in water due to
>index matching?  ANyone know what material has that property?  Dave
>
 
 
I'll post a full discussion of Jim's comments in a day or two when I've
run a couple of tests, but just to make one point clear here, the crucial
thing about these bead tests is that they must be WELL below the resolution
limit one is trying to measure.  Hence the question of transparency is
irrelevant, and so is the question of whether the dye is at the surface
or all the way through.  When one is focussed just below the bead, there is
a steeply converging (included angle ~150 degrees) cone of light coming
down to a patch about 250nm in diameter and more than twice that in height.
The effect of a tiny blob, sitting in that huge cone, is negligible.
 
The key question is whether a 100nm bead is small enough to count as
being well below the resolution limit.  Its cross-sectional area is
only one sixth that of the Airy Disk.  The theoreticians would say that
50nm is better - as I practical man I'd say that I'd not be able to get
a useful measurement if I made my s/n ratio worse by a factor of 4!  More
on this point later.
 
The other answer is that 100nm (or even 1 um) of either PVC or PVA (or
any other likely bead material) certainly is transparent, but equally
certainly isn't index-matched in water - though it's probably not far off
in glycerol.
 
Dave's proposed test would be interesting but certainly would not measure
Airy / Rayleigh type resolution.  Years back Nick White did some 'Negative
fluorescence) of this sort.  But most cell biologists are looking at
sub-resolution bright structures on a dark background.
 
                                                        Guy Cox