1.Tissues weighing from 0·006 to 3·0 g were dissected and put directly into 15 ml screw cap polypropylene tubes with a conical base (maximum of 3 g per tube). It proved unnecessary to mince tissue, even though smaller pieces may aid quicker digestion. Five millilitres of 2 M KOH in 99 % (IMS) ethanol with 0·5 % Tween-80 was then added. Tissue digestion was usually completed in 2-4 h using a dry heating block held at 60°C, with intermittent shaking. Samples were routinely processed using fresh tissues, although storage of frozen tissue (in the dark at -20°C) introduced no detectable error in BF estimation and tended to aid tissue maceration.2.After digestion was completed the tubes were centrifuged at 3000 r.p.m. (1500 g) for 15 min.3.The supernatant was carefully aspirated until < 500 µl was left, thereby reducing the possibility of accidental loss of microspheres.4.After 1 ml dH2O was added the tubes were quickly vortexed to prevent microsphere flocculation and aid resuspension of remaining pellets, while the subsequent addition of ethanoic Tween (100 % ethanol + 0·5 % Tween-80) allowed complete sedimentation by centrifugation.5.Nine millilitres of ethanoic Tween-80 was added, and the tubes were vortexed and spun at 1500 g for 15 min.6.The supernatant was aspirated as above (Step 3).7.Five millilitres of 100 mM phosphate buffer (pH 7·00) was added to neutralize the pellet as alkaline solutions quench fluorescence. Using aqueous solutions increased the possibility of microsphere loss by adhesion to the surface of tubes or aggregation, which could readily be seen by eye when tubes were examined under bright light, and hence the buffer was followed by addition of 4 ml absolute ethanol and further vortexing, before spinning (1500 g) for 20 min.8.The supernatant was aspirated, leaving up to 300 µl depending on the amount of tissue residue, and the remaining microspheres and pellet were quickly vortexed to ensure complete resuspension.9.The tubes were left to evaporate in an oven at 60°C and then briefly vortexed during this period to disperse more of the flocculent, until around 100 µl fluid remained. This improves solvent extraction of microspheres, which may be less efficient in a completely dry pellet.10.Two or three millilitres of solvent (di(ethylene glycol) ethyl ether acetate, 98 %; Aldrich Chemical Co., Poole, Dorset, UK) was added according to the expected fluorescence intensity, and vortexed several times over 3-5 min. After 30 min the tubes were sonicated in a water bath for 5 min to complete dye extraction by the solvent. The tubes were kept as far as possible in the dark after the solvent was added to avoid photo-bleaching of the fluorescent dyes.11.After sonication, the tubes were occasionally spun (1500 g) once more to sediment any undigested/undissolved material, though this was rarely necessary in our experiments. Readings should be completed within 1 h to avoid loss of signal intensity.