In this paper we compare lava temperatures measured using Cr-Al thermo
couples or infrared spectrometry with estimated quenching temperatures
based on the glass geothermometry calibration of Helz and Thornber (1
987). Comparative data are available for the April 1982 and September
1982 summit eruptions, the Pu'u 'O'o east rift eruption (1983-1986), a
ll three eruptions at Kilauea, and the 1984 Mauna Loa eruption. The re
sults show that quenching temperatures, based on the MgO contents of K
ilauean glasses (T-MgO), lie within +/- 10 degrees C of field measurem
ents using the infrared spectrometer for 85% of the samples. Where a C
r-Al thermocouple was used, 90% of the field measurements lie within 1 degrees to -11 degrees C of T-MgO for samples with T-field > 1130 de
grees C. Samples where T-field < 1130 degrees C show larger divergence
. The uncertainty in T-MgO by itself is +/- 10 degrees C, so the level
of agreement between held measurements and T-MgO is very good for Kil
auean lavas. Systematic comparison of field measurements df temperatur
e with glass geothermometry for the 1984 Mauna Loa eruption suggests t
hat, although the field and glass temperatures lie within +/- 10 degre
es C of each other, the Kilauean T-MgO calibration is nevertheless not
appropriate for Mauna Loa glasses and that actual quenching temperatu
res for Mauna Loa samples will lie 10 degrees-20 degrees C higher than
would be predicted from the Kilauea calibration curve. Consideration
of possible effects of variable volatile content suggest that in most
cases these are small. Samples erupted early in an eruption may reflec
t preeruptive water contents different enough to affect T-MgO signific
antly, but later spatter samples and all flow samples appear to have e
quilibrated at low enough water contents for the calibration to be app
licable. We conclude that the MgO-based geothermometer can be applied
to glassy Kilauean samples to give temperatures that generally will li
e within +/- 10 degrees C of a field measurement. Plots of glass MgO c
ontent versus time, if a suitable sample base is available, should giv
e a thorough, quantitative record of the thermal history of any Kilaue
an eruption.