The first known systemic sign of the existence of a viable embryo prior to implantation is the presence of a lymphocyte modifying activity which is detectable in maternal serum within hours of fertilisation. The study of this so called 'early pregnancy factor (EPF)' activity has generated considerable interest in the field of reproductive biology due not only to the potential to develop new, extremely early methods of pregnancy detection but also to monitor embryo viability prior to implantation. However, the basis for this activity expression has defied molecular definition although early studies have assumed it to involve the production of a single pregnancy specific protein. Recent studies provide evidence that this is not so. An objective assessment of the basic observations can only be interpreted as 'EPF' activity expression being due to the action of a complex multifactorial system. As such a new paradigm has been proposed by Clarke (1992) to explain the capacity of pregnancy sera to induce increased RITs i.e. to express so called 'EPF' activity. Briefly, it has been proposed that all sera (whether pregnancy or non-pregnancy sera) contain inherently active low molecular weight moieties, and in addition possess the capacity to stimulate the production of even more when applied to the lymphocytes in the rosette inhibition assay. However, other components in sera render the lymphocytes refractory to the actions of these moieties preventing them from exerting their effects. It is proposed that pregnancy sera (but not non-pregnancy sera) contain a functional form of thioredoxin which reverses / prevents the refractory state allowing the active moieties to exert their effects resulting in the expression of increased R1Ts i.e. so called 'EPF' activity. The present studies extend this model by defining the role of thioredoxin in the 'EPF' phenomenon with site directed mutagenesis studies determining the molecular features of the thioredoxin molecule required for its action in the rosette inhibition assay. Thioredoxin alone was shown to be incapable of inducing an elevated BIT in the rosette inhibition assay but will cooperate with non-pregnancy sera or another appropriate cell stimuli (e.g. PAF) to allow for activity expression. Site directed mutagenesis studies presented within this thesis clearly show that thioredoxins function in the rosette inhibition assay is not dependent upon its redox activity. Instead these studies have pin-pointed cysteine 72 as being of particular importance and that this residue must be reduced and protected from oxidation for thioredoxin to function in its regulatory role. Indeed, mutants lacking an intact cysteine 72 were shown to have an inhibitory effect, counteracting the ability of wild type thioredoxin to allow for the induction of elevated BITs. Consequently these studies indicate possible differences which may occur between the active forms of thioredoxin that are present in pregnancy sera and the inactive forms present in non-pregnancy sera. Any new paradigm should also be capable of accounting for previously reported phenomena. Consequently, studies were undertaken to determine whether the new paradigm can adequately explain early ammonium sulfate fractionation and gel permeation results. Studies described within this thesis have reexamined the susceptibility of pregnancy sera to 40% ammonium sulfate fractionation and indicate that a system of components is required for pregnancy sera to induce elevated RITs. Examination of this phenomenon clearly demonstrated that active moieties of low molecular weight are released from association with macromolecular components and that these macromolecular components are functionally equivalent to thioredoxin in that they allow for the reversal of the refractory state. Likewise, the association of the ability to induce elevated RITs with multiple molecular species obtained on gel permeation fractionation of pregnancy sera is also explainable within the new paradigm. Studies presented in Chapter 6 using ammonium sulfate fractionation and antibody adsorption unequivocally demonstrated the presence of low molecular weight active moieties as well as functional thioredoxin / thioredoxin-like molecules in each of the multiple molecular weight fractions. These studies provide further appreciation of the molecules and mechanisms at work in the rosette inhibition assay and provide substantiation of a clearly defined paradigm to guide future research into the so called 'EPF' system of early pregnancy sera.