NIII 57.330 or 57.317*
*controversial - see below
Good references for the CO wavelengths are Nolt etal. 1987, Journal of Molecular Spectroscopy, 125, 274 and Chackerian & Tipping, Journal of Molecular Spectroscopy, 99, 431, 1983. These wavelengths are very precisely known (see references). These lines are also good candidates for the wavelength calibration because they probably have the best chance of being narrow (in selected sources).
The NII, OI, CII, and SiI wavelengths have all been measured with Laser Magnetic Resonance techniques and are well known:
NII (121.8976, 205.1782) References are in Colgan etal. 1993, ApJ, 413, 237 They are Cooksy etal. 1986, J. Chem. Phys. 84, 6101 and Brown and Emerson 1993, in preparation - I'm tracking this down.
The [NII] lines have recently been measured in the lab via the LMR technique by John Brown et al. He told Rubin his preliminary frequencies are 1461.1314 and 2459.3815 GHz (estimated error 0.001) for the J = 1--> 0 and J = 2--> 1. This gives lambda of 205.1783 and 121.8975 microns. Andrew Cooksey told Rubin that there was a slight error in his earlier J = 2--> 1 measurement and that he concurs with the new value.
Update on [NII] line lab frequencies: 1461131.8 +- 0.7 MHz for J = 1--> 0. The reference is John Brown and K.M. Evenson, in preparation. This gives lambda of 205.1783 microns (Mike gets 205.1782). The J = 2--> 1 frequency is still preliminary, at 2459379.5 +- 0.6 MHz, yielding lambda of 121.8976 microns.
The OI line - 63.1700 um - Saykally etal. 1979, JChemPhys, 71, 1564 Watson etal. 1984, ApJ, 279, L1 says its 63.18372+0.00003 and refers to Evenson, 1982, private communication. This must supercede the 63.1700 value, but am checking it out.
The OI line - 145.5255 um - Davies etal. 1978, JChemPhys, 68, 1135.
CII 157.7409 um - Boreiko etal. ApJ, 353, 181 (1990) - astrophysical observations - Cooksy etal. ApJ, 305, L89 (1986) - laboratory LMR measurements
SiI 129.6817 - Inguscio et al. ApJL, 278, L127, 1984 (2311.7556 GHz). 68.472 - difference of upper level from C. Moore tables and above value by Rubin (not verified by Mike)
All these are presumably known to 0.0001 um or so. See references for details.
Surprisingly, the wavelengths of the doubly ionized lines are rather poorly known. The values we are using come from Moorwood etal. 1980, ApJ, 238, 565 and are very old and not all that well determined. We can only do slightly better with the KAO's CGS, having uncertainty of 1/2 channel or more, being, respectively, 0.004, 0.008, and 0.006 microns at 52, 88, and 57 microns. Many people use 57.317 for the wavelength of the NIII line, including the FIFI group, and this may fit the CGS data better, but we are not yet sure. We are also not sure where this number originated. Moorwood's paper is worth looking at just to see how FIR spectroscopy has come.