The dilution refrigerator has become the workhorse for experiments in the 5 mK to 1 K range, and provide the starting temperature for nuclear demagnetization and the study of solids and liquids into the 30 nK range. The idea of refrigeration by the dilution of 3He by superfluid 4He was first suggested by Heinz London in 1951, following the proposal by L. Landau and I. Pomeranchuk (1948) that very diluted 3He in superfluid 4He would behave like a gas of quasi-particles. H. London, G. R. Clarke and E. Mendoza in 1962 in the discussion of measurements they presented of the osmotic pressure of the mixtures included sketches of two possible refrigerator designs that could work. Following their design, the first barely operating refrigerator (0.22K) was built in Leiden in 1965. In 1965-66 continuously operating 0.05K refrigerators were built in Moscow and in Manchester. In 1966 at the U. of Illinois a refrigerator built with improved heat exchangers reached 0.02K, and a single cycle refrigerator precooled by the continuous refrigerator reached 0.0044 K. Why did it take until 1965 to build the first refrigerators? Along the way 3He had to become reasonably available, properties of 3He-4He mixtures had to be studied starting with their vapor pressure, mixing enthalpy, phase separation, and the existence of dilute solutions down to 0.003 K (and theoretically 0 K). In this semi-historical “memoir” I’ll review the steps that led to the Illinois refrigerators and the development of the current commercial ones, including one recently installed in our Department.