The binding of iron to transferrin has been studied by electron paramagnetic resonance and optical absorption spectroscopy and by measurements of proton magnetic relaxation rates under conditions in which carbon dioxide (and therefore bicarbonate) is excluded. Even in the absence of bicarbonate, specific binding of Fe3+ to transferrin may occur. Nitrilotriacetate, ethylenediaminetetraacetate, and oxalate were found capable of replacing bicarbonate to form colored complexes with iron and transferrin. Since the nitrilotriacetate and oxalate complexes, in particular, have optical absorption spectra similar to those of the usual iron-transferrin-bicarbonate complex, their recognition depends on correlating electron paramagnetic resonance and optical studies. The electron paramagnetic resonance spectrum of the iron-transferrin-nitrilotriacetate complex consists of a line, 30 gauss in width, centered near the apparent g value, 4.25, whereas the oxalate complex shows a broad absorption, from 700 to 1400 gauss. In contrast, the bicarbonate complex has a spectrum centered near the apparent g value, 4.22, with a total width of 120 gauss. This spectrum is not influenced by changes in pH in the range, 6.1 to 10.5. Of a number of metal-complexing agents tested, only those with 2 or more carboxyl groups were found capable of replacing bicarbonate to form ternary complexes with transferrin and iron. Azide and thiocyanate were found ineffective, while a 100-fold excess of cyanide did not perturb the spectrascopic properties of iron-transferrin-bicarbonate. These findings suggest that bicarbonate is not simply coordinated to iron in iron-transferrin-bicarbonate. This view is corroborated by proton relaxation rate measurements, which show little change between iron-transferrin and iron-transferrin-bicarbonate. Cu2+ was also found to bind specifically to apotransferrin in the absence of bicarbonate, as evidenced by the electron paramagnetic resonance spectrum of the Cu2+-transferrin complex, which is quite different from the spectrum of the Cu2+-transferrin-bicarbonate complex. Ternary complex formation among Fe3+, transferrin, and chelate was observed almost immediately after addition of the ferric chelates to apotransferrin in air. On standing in the presence of an excess of bicarbonate these complexes were gradually replaced, the time depending on the chelate, by the bicarbonate complex. It appears likely, then, that the initial and rapid step in the formation of iron-transferrin-bicarbonate is the formation of the ternary complex, iron-transferrin-chelate, while the slow reaction entails displacement of the chelate by bicarbonate.