The solar Lyman α radiation is the dominant source of energy for the Earth's mesosphere; thus accurate knowledge of its variability is important for many studies of the middle and upper atmosphere. The measurements of the solar Lyman α irradiance taken during solar cycles 21 and 22 by the Atmospheric Explorer-E (AE-E), the Solar Mesospheric Explorer (SME), and the Upper Atmosphere Research Satellite (UARS) are compared and then combined to construct a long-term data set. The UARS measurements, from both the Solar Ultraviolet Spectral Irradiance Monitor (SUSIM) and Solar Stellar Irradiance Comparison Experiment (SOLSTICE), have the best photometric accuracy yet achieved with 2-&sgr; uncertainties of 10%; thus the SME and AE-E measurements, which have 2-&sgr; uncertainties of 40% and 60% respectively, are adjusted to agree with the UARS values. Because none of these satellite measurements overlapped in time, solar indices representative of the upper chromosphere are used in determining these adjustments. The solar indices considered for this analysis are the National Oceanic and Atmospheric Administration (NOAA) Mg II core-to-wing ratio, the National Solar Observatory (NSO) He 1083 nm equivalent width, the Penticton (Ottawa) 10.7 cm flux, and the Pioneer Venus Langmuir probe solar signal. For this long-term data set, the SME Lyman α irradiances are adjusted up by 25%, and the AE-E Lyman α irradiances are scaled down by a time dependent factor ranging from 10% to 40%. The relationships to the solar indices are also presented so that the estimates for the solar Lyman α irradiance can be extended back as far as 1947 using the solar 10.7 cm flux with an estimated uncertainty of 20%.¿ 1997 American Geophysical Union