Under the auspices of the Inter-American Metrology System (SIM), the National Institute of Standards and Technology (NIST) initiated a regional comparison for type K thermocouples from (100 to 1,100) degrees C with 11 participating countries. The use of type K material above approximately 200 degrees C is considered destructive. Therefore, each participating laboratory was sent new, unused wire from a lot of material characterized by NIST. The uniformity of the lot was remarkable, especially at temperatures above 500 degrees C; the standard deviation of the thermocouple emf values of multiple cuts tested at NIST was 2.7 mu V or less over the full temperature range. The high uniformity eliminated any need to correct for variations of the transfer standard among the laboratories, greatly simplifying the analysis. The level of agreement among the laboratories' results was quite good. Even though test procedures and equipment varied significantly among the participants, the standard deviation of all emf values at each test temperature was less than the equivalent of 0.20 degrees C at 200 degrees C and below, and less than 0.60 degrees C from (400 to 1,100) degrees C. Of the 380 total bilateral combinations of the data at the eight test temperatures, only 13 (i.e., 3.4% of all combinations) are outside the k = 2 limits, and of these 13, only 3 are outside k = 3 limits. All the outliers occur at temperatures of 800 degrees C and below, which suggests that drift of the type K wire due to high-temperature oxidation did not cause changes in the thermocouple emf comparable to or larger than the claimed uncertainties.