Five new P1-(steroid-21-yl)-P2-(1-beta-D-arabinofuranosylcytosin-5'-yl)pyro phosphates (ara-CDP-steroids), five 1-beta-D-arabinofuranosylcytosine 5'-O-(alkyl)phosphates (ara-CMP-alkyl esters), and two P1-(alkyl)-P2-(1-beta-D-arabinofuranosylcytosin-5'-yl)pyrophosphat e (ara-CDP-alkyl esters) have been prepared and evaluated against L1210 lymphoid leukemia in culture and in mice (C3D2F1/J). These include ara-CDP-11-deoxycorticosterone (6a), ara-CDP-cortisone (6c), ara-CDP-corticosterone (6d), ara-CDP-cortexolone (6e), and ara-CDP-prednisone (6g), ara-CMP hexadecyl ester (7a), ara-CMP 1-cyclohexylmethyl ester (7b), ara-CMP 1-adamantylmethyl ester (7c), ara-CMP 2-(1-adamanthyl)ethyl ester (7d), ara-CMP 2-chloroethyl ester (7e), ara-CDP hexadecyl ester (9a), and ara-CDP 1-cyclohexylmethyl ester (9b). The in vitro antitumor results indicated that ara-CDP-steroids were as active as the previously reported ara-CMP-steroids and that ara-CMP and ara-CDP-alkyl esters were less growth inhibiting than ara-CDP-steroids and ara-C. However, the in vivo antitumor results indicated that ara-CDP-steroids were generally less effective than the previous monophosphate derivatives. Among them ara-CDP-corticosterone (6d) and the known ara-CDP-cortisol (6b) showed greater efficacy than ara-C with ILS value of 152% and 209%, respectively, at the optimal dose of 40 and 80 (mg/kg)/day for 9 days, while that of ara-C was 138% at the optimum dose of 9.2 (mg/kg)/day. Generally, ara-CMP alkyl esters (7a-e), given ip to the L1210 leukemic mice, were found to be toxic and ineffective. However, ara-CDP hexadecyl ester (9a) showed marginal activity (ILS, 38%). These preliminary results support the thesis that the ara-C conjugates of this type may require a lipophilic and naturally occurring moiety for improved efficacy.