David, my 3458A has a T.C. for DCV, 10V of about +0.45ppm/K, i.e. about +0.25ppm/K for the A/D and +0.2ppm/K for the LTZ1000A reference. These are typical, specified values for this instrument. Other 6.5 .. 8.5 digit DMMs have very similar T.C. specifications. Therefore a temperature stability of 0.2 °C is required to achieve 0.1ppm stability, which is a reasonable criterion for 7.5 and 8.5digit DMMs. For ratio and indirect comparisons of references that temperature constancy is necessary on a minutes timescale. For time and temperature stability measurements on DUTs, that needs to be maintained over up to 24h, during which I monitor the 3458A internal temperature. An example is the characterization of my DIY LTZ1000 references, which are all trimmed to < 0.03ppm/K by means of the 3458A. These measurements also run over several hours at +/- 0.1°C stable room temperature, but varying the DUTs temperature over +/- 10°C. Only our basement allows such stable conditions at 21 +/- 1 °C seasonal changes. Temperature variations greater than +/- 0.5°C make precise DCV measurements nearly impossible. Frank