Dissolving 100 mg Coomassie Blue G-250 in 50 mL of 95% ethanol, add 100 mL of 85% phosphoric acid, and dilute to one liter. The reagent needs to be filtered at least once and perhaps more, since it seems to precipitate dye over time. "Bradford reagents" are available commercially that use more stable formulations. I heard from someone that Sigma's formula uses 40 mL of methanol (final 4%) in place of ethanol and about 120 mL of phosphoric acid (final 10%); I tried this and I couldn't say it worked any better than the original. This reagent is said to be unstable, but I think I've used the same batch over a year or two without any problems.
To quantify protein, mix 0.25 mL of sample with 2.5 mL of Bradford reagent. After 5 minutes, measure the absorbance at 595 nm. One disadvantage to the reagent is that it gives a high blank which may affect subsequent readings because some reagent adheres to the cuvette. Another is that it is very sensitive to the presence of detergent, either from poorly-rinsed glassware or, heaven forbid, in the event you are studying detergent-solubilized membrane proteins.