We investigate the diffuse absolute calibration of the InfraRed Array Camera on the Spitzer Space Telescope at 8.0 \mu m using a sample of 43 H ii regions with a wide range of morphologies near l =312 ^ { \circ } .
For each region we carefully measure sky-subtracted , point-source-subtracted , areally-integrated IRAC 8.0- \mu m fluxes and compare these with Midcourse Space eXperiment ( MSX ) 8.3- \mu m images at two different spatial resolutions , and with radio continuum maps .
We determine an accurate median ratio of IRAC 8.0- \mu m/MSX 8.3- \mu m fluxes , of 1.55 \pm 0.15 .
From robust spectral energy distributions of these regions we conclude that the present 8.0- \mu m diffuse calibration of the SST is 36 percent too high compared with the MSX validated calibration , perhaps due to scattered light inside the camera .
This is an independent confirmation of the result derived for the diffuse calibration of IRAC by the Spitzer Science Center ( SSC ) .

From regression analyses we find that 843-MHz radio fluxes of H ii regions and mid-infrared ( MIR ) fluxes are linearly related for MSX at 8.3 \mu m and Spitzer at 8.0 \mu m , confirming the earlier MSX result by Cohen & Green .
The median ratio of MIR/843-MHz diffuse continuum fluxes is 600 \times smaller in nonthermal than thermal regions , making it a sharp discriminant .
The ratios are largely independent of morphology up to a size of \sim 24 ^ { \prime } .
We provide homogeneous radio and MIR morphologies for all sources .
MIR morphology is not uniquely related to radio structure .
Compact regions may have MIR filaments and/or diffuse haloes , perhaps infrared counterparts to weakly ionized radio haloes found around compact H ii regions .
We offer two IRAC colour-colour plots as quantitative diagnostics of diffuse H ii regions .