The melting curve of iron is reliably determined to 105 GPa using the laser-heated diamond cell. In particular, these and other melting data obtained with the same technique are in close agreement with independent measurements using piston-cylinder and large-volume presses (5--20 GPa range), as well as shock wave experiments (50--250 GPa range). In order to obtain reliable melting data from the internally heated diamond cell, whether by laser heating or by Joule (resistance) heating, temperature gradients across the sample must be quantitatively measured. An analysis of results derived from Joule heating inside the diamond cell, and consideration of the resulting heat flow and temperature distribution, documents that such ''wire heating'' experiments can lead to significant underestimates of the melting temperature and its pressure dependence. The best estimate of the high-pressure melting curve of iron, as derived from the laser-heated diamond cell and Hugoniot temperature measurements, yields melting temperatures of 4800 (¿ 200) K and 6700 (¿400) K at 133 GPa and 243 GPa, respectively. These values are compatible with the Lindemann relation. We conclude that the triple point between the &ggr;, &egr;, and liquid phases of iron lies in the vicinity of 300 GPa and 7500 K; no other as yet unobserved high-temperature, high-pressure phases need be invoked based on the presently available data. ¿1991 American Geophysical Union