Quantitative Isolation of Sterols.

I n 1909, Windaus published the important discovery that cholesterol, sitosterol, and P-cholestanol form sparingly soluble addition compounds with digitonin, whereas a-cholestanol and cholesterol esters do not (1). Ever since, this reaction has been a major tool of steroid chemists; it has been used in many methods for cholesterol determination and in many separations of steroids with the beta configuration at carbon 3 from those with the alpha configuration. In 1937, I reported experiments in which this reaction had been employed in various ways in a study of the cholesterol of the blood serum (2). One of these (Procedure F1) was the method used by the late Rudolph Schoenheimer and which I applied under his guidance. It had been developed from the experience of Schoenheimer and Dam (3), and it involved precipitation in SO'% ethanol by digitonin in a t least 50% excess, filtration, washing with 80% ethanol and ether, drying, and weighing. In preparing to apply this method to the unsaponifiable lipids of brain, I tested it on pure cholesterol with only two changes: (a ) the amount per sample was decreased from an average of 13 mg to about 2 mg, and (b) centrifugation instead of filtration was used in the isolation and washing of the precipitate. Unexpectedly, recoveries were not quantitative; the digitonide weights accounted for only about 90-95y0 of the cholesterol taken for analysis. A series of experiments, in which several variations of the procedure were tried, led to the finding of conditions under which satisfactory results were consistently obtained. Good recoveries of sterols from the digi-