1. Experiments are reported which aimed at determining whether transferrin loses sialyl residues from the carbohydrate side-chains during the biological lifetime of the molecule. To explore this possibility, transferrin fractions of relatively high sialic acid content (referred to as sialotransferrin) were prepared from purified rabbit and bovine transferrin by preparative polyacrylamide-gel electrophoresis. After labelling with 125I, the preparations were injected into a group of three rabbits each. From the plasma samples obtained between 1 h and 6–8 days after injection, transferrin was partially purified, mixed with 131I-labelled asialotransferrin of the corresponding species and run in preparative polyacrylamide-gel electrophoresis. In each specimen examined, the 125I radioactivity migrated ahead of the marker asialotransferrin, and no portion of the dose was detected with the electrophoretic mobility of asialotransferrin.
2. Evidence is presented that bovine transferrin desialylated in vitro remains detectable in the plasma of rabbits for intervals which are comparable with those found in previous studies with rabbit asialotransferrin.
3. A mathematical model is described for the computation of asialo- to sialotransferrin radioactivity ratios in the plasma, continuous desialylation of pulse-injected sialotransferrin being assumed. Calculations were made at various hypothetical rates of desialylation.
4. On the basis of the experimental data and the model it is concluded that transferrin (both rabbit and bovine) is not subjected to systematic desialylation in rabbits. Random desialylation of some transferrin could take place at rates less than 5% of the fractional catabolic rate of transferrin, which would be without any biological significance.