Given a black, clear bottom 96-well plate, round wells, with read from the bottom, does the well depth and liquid volume affect the fluorescence read in the plate reader?
It may, but it depends on the plate reader and what you are measuring (i.e. adherent cells or bulk fluid). It is worth experimenting with different volumes and reader settings (if the reader can adjust focal depth) to find the conditions that give the greatest signal-to-background ratio for your measurement. You could also compare round-bottom and flat-bottom plates.
My preference for measuring bulk fluorescence would be top reading in a solid black plate.
Ideally, for precise fluorescence measurements, flat-bottom plates are recommended. The ideal situation would be solid black plates with a top reader as Dr. Shapiro clearly mentioned. If you gonna use round-bottom plates, you can optimize the conditions with various depths and volumes for the highest signal-to-noise ratio.
Based on my experience, it is better to use flat bottom multiwell plates for fluorescence measurement. I am not sure if that stands correct for any type of optical activity readouts and perhaps depends on what type of plate reader you are using. For our resazurin based cell viability assay we used 384 well plates with 36 uL of liquid in it for fluorescence measurement from the bottom (capacity for each well was in the range of ~80 uL). Regarding liquid volume, I would say as long as you are covering the whole surface of each well, you should be fine. However I would suggest to have 1/4 to 1/2 of liquid of each wells capacity. You can check this by trying different volume, but should be consistent throughout your plate and with proper controls in the plate, being in that volume range should work well. Any time we had any confusion regarding the plate reader or the performance of the liquid handling robots, we always tested it with diluted dye and reading it in the plate reader. Best of luck!