Abstract.The title compound is synthesized from the corresponding 2',3'-O-isopropylidene nucleoside with POCl in trimethylphosphate. The anomers are separated by preparative HPLC.3

The reaction of¯avanone with Cr(CO) 6 gives two diastereomeric complexes 1 (cis) and 2 (trans) with the general formula [Cr(CO) 3 (C 15 H 12 O 2 )]. Due to the existence of an asymmetric carbon atom in the -pyrone ring, the two isomers are chiral. Both complexes were characterized by elemental microanalysis as well as by IR, NMR, UV, and mass spectroscopy. The X-ray analysis of 1 has shown that the Cr(CO) 3 tripod is complexed to the benzene ring in an eclipsed conformation. The phenyl group adopts a cis con®guration with respect to the tricarbonylchromium moiety, and the -pyrone ring is oriented in a 1,2-diplanar conformation.

The hydrogenation products of humulinic acids A and B are described and mechanisms for the reactions are advanced. The isolated 2,4-diisopentylcyclopentane-l,3-dione (X), truns-l,3-diisopentyl cyclopentane (XV) and 2,5-diisopentyl-2-cyclopenten-1-one (XII) have probably not been isolated before.Humulinic acid (') (11) is obtained on alkaline hydrolysis of humulone (I) as a mixture of the two stereoisomers IIa (HAA) and IIb (HAB) ( 2 ) . Hydrogenation of these humulinic acids A and B yields a mixture of dihydrohumulinic acids A and B (DHHAA and DHHAB) with structures 111, dihydrodeoxohumulinic acids A and B (DHDHAA and DHDHAB) with structures IV, and higher hydrogenated reaction products (').The separation of HAA and HAB, necessary to study this hydrogenation reaction, was carried out by SSD (43 5 ) .HAA can consume up to 7.15 moles of hydrogen when Adams' catalyst is activated by addition of chloroplatinic acid to the reaction medium. The mixture so obtained contains the same substances as obtained when reducing DHHAA or DHDHAA in similar conditions. This is shown by gas chromatographic analysis and is an indication that the hydrogenation proceeds in a consecutive way.

Methyl allohumulinic acid D (I) was isolated after acid treatment of either humulinic A (HAA) or humulinic acid B (HAB). Some specific reductions on this and other humulinic acids are reported. Dihydromethylhumulinic acid D (IV) is described. A new route to dihydro-HAB was developed.When humulinic acid A (HAA) or humulinic acid B (HAB) is boiled in acidic methanol about 32% of humulinic acid C (HAC) can be isolated by counter current distribution -C.C.D. (2) Redistribution at pH 7.47 of the tail fraction containing methylhumulinic acid D (Me-HAD) yields a second product with K = 1.66 (carbon tetrachloride/ phosphate buffer) in a total yield of 3%. Physical and chemical properties indicate that it is compound I. This is called (methy1)alZohumulinic acid D, because it is derived from the D isomer (11) by a double bond shift. The "allo" nomenclature is in accordance with other compounds encountered in this field and which have migrated double bonds (394).

Steady State Distribution (SSD) is a continuous separation method based on liquid extraction and the technique and theory of the new method have been described extensively by us before ('9 2 v '). We now intend to report some new possibilities of SSD as well as the practical aspects of some separations that have been carried out in our laboratory.In SSD two immiscible solvents migrate in opposite directions and this in a train of cells, thus achieving discontinuous operation. This allows the separation to be stopped and restarted at will. Each shift of the solvent phases is called a transfer and a sma'll quantity of the mixture is added at each transfer in the middle of the train. The method is in principle only suited for binary mixtures, and the phase composition must be such that one of the components has a partition coefficient below 1, the other then higher than 1 ; in fact the ideal situation is obtained when JK,.K, = 1In this case the flow of both solvents may be equal or in other words the ratio of upper phase transfers (x) to lciwer phase transfers (y) is 1. In this case a so called "cycle" consists of 1 upper and 1 lower transfer. This ideal situation can normally only be obtained by so much effort that it is not warranted to try for it, especially since that even when for example JKA. KB > 1 the distribution can still be carried out by adopting an x/y transfer ratio < 1 and vice versa. The exact values should be so that x/y = JK,.K, and the cycle becomes then something like 3 upper against 2 lower phase transfers.