Effect of indirect sample preparation procedures on the apparent concentration of asbestos in settled dusts.