There are a wide variety of mounting media to chose from for fluorescence microscopy. I have tried several recipes and essentially use two types almost exclusively. The first is a mixture of 90% glycerol in 0.1 M sodium borate, pH 9 supplemented with 3% N-propyl gallate. This concentration of glycerol has a refractive index near to that of microscope immersion oil to prevent spherical aberration of the images resulting from R.I. mismatch. It is buffered at an alkaline pH because fluorescein (a fluorescent dye I commonly use) exhibits a pH dependent maximal fluorescence under slightly basic conditions. Rhodamines, in contrast, do not have this property. N-propyl gallate is included as an antifading agent to prevent photobleaching (see below). I usually mount my samples in a drop of this medium and seal the coverslips to the slide using clear nail polish. Colored nail polish should never be used as the dyes employed in their manufacture are themselves fluorescent and may leech into the medium over time, producing a fluorescent background. Also, when mounting samples expressing green fluorescent protein, never mount samples using nail polish as the organic solvents somehow kill its fluorescence.
The second is a commercially available preparation from Molecular Probes called ProLong. This stuff comes packaged as two components - a liquid mountant and a dry anti-fading agent. The two are mixed together and the samples mounted and the preparation hardens in a few hours, making a semi-permanent preparation. The advantage of ProLong are that it retards photobleaching better than any other anti-fade agent I've ever seen. This property makes it extremely useful for fluorescence microscopy - especially for optical sectioning on the confocal microscope. ProLong is very expensive, however - it is packaged as individual vials (which must be used immediately after reconstitution) and costs approximately $5 per experiment. It is the best anti-fade I have tried - so you get what you pay for. The other disadvantage of ProLong is that it is difficult to obtain good phase-contrast or Nomarski images of samples prepared with it, probably due to its refractive index. It is often desirable to correlate fluorescence images with bright field - this can be problematic with ProLong.
As a final note, after mounting, fluorescent specimens should be stored in the refrigerator and protected from light to maximize the lifetime of the fluorescent probes.