WHEN mouse or rabbit globin mRNA is injected into oocytes of Xenopus laevis both α and β mRNAs are translated but only about one fifth as much α globin is synthesised as β globin1–3. On the other hand, the same message sample results in about equal amounts of α and β globin synthesis in a reticulocyte cell-free system and is believed to contain roughly equal amounts of α and β globin mRNA. Giglioni et al.3 have shown that if haemin is injected into oocytes at the same time as globin mRNA, there is an equal ratio of α:β globin synthesis, and their experiments strongly suggest that this is achieved by increasing the efficiency of translation of α globin mRNA. These experiments thus identified a factor which has a different effect on two kinds of mRNA, and further demonstrated that the microinjection of living cells can reveal regulatory effects not readily seen in the cell-free systems used so far. The experiments reported here were also carried out by injecting mRNA into frog cells. They show that during development a component is formed that alters the ratio of α: β rabbit globin synthesis, a change not seen when mouse globin mRNA is injected. This is therefore a second example of a translational factor revealed by a living cell assay system. This component shows an even higher degree of selectivity in its action than haemin.