CYCLIC AMP has been implicated in the regulation of cell growth. However, the molecular mechanism of cyclic AMP action, especially that involving the control of tumour growth in vivo is not clear. Our previous studies on two cell populations of Walker 256 mammary carcinoma (W256), one regressing (responsive), the other growing (unresponsive) under dibutyryl (db) cyclic AMP treatment, showed a correlation between altered cyclic AMP-binding of the tumour cytosol and db cyclic AMP-unresponsiveness1. Deficient cyclic AMP-binding in other malignant growth has also been reported2–4. It is suggested that the defective cyclic AMP-binding proteins may be the possible cause for the loss of growth control by cyclic AMP. To determine the relationship between the qualitative difference of cyclic AMP-binding proteins and db cyclic AMP-responsiveness in vivo, we have used various experimental tumour models, such as, transplantable mammary carcinoma (MTW9) in Wistar-Furth inbred female rats, a hormone-dependent tumour; primary, N-nitrosomethylurea-induced mammary carcinoma (NMU) and transplantable mammary adenocarcinoma (R3230AC) in Fischer female rats, hormone-responsive tumours; and hepatoma 5123 in Buffalo/N female rats.