The vital ultraviolet- (UV-) protective and photosynthetic pigments of cyanobacteria and lichens (microbial symbioses) that dominate primary production in Antarctic desert ecosystems auto-fluoresce at short wavelengths. We therefore use a long-wavelength (1064 nm) infrared laser for non-intrusive in situ Raman spectrometry of their ecologically significant compounds (especially pigments). To confirm that the power loss at this longer wavelength is justified to avoid swamping by background fluorescence, we compared Raman spectra obtained with excitation at 1064, 852, 830, 785, 633 and 515 nm. These are typical of lasers used for Raman spectroscopy. We analysed communities of the cyanobacterium Nostoc commune and the highly pigmented lichens Acarospora chlorophana and Caloplaca saxicola. These require screening compounds (e.g. pigments such as scytonemin in cyanobacteria and rhizocarpic acid in the fungal symbiont of lichens). They are augmented by quenching pigments (e.g. carotenoids) to dissipate the energy of free radicals generated by penetrating UV. We also analysed organisms having avoidance strategies (e.g. endolithic communities within translucent rocks, including the common cyanobacterium Chroococcidiopsis). These require accessory pigments for photosynthesis at very low light intensities. Although some organisms gave useable Raman spectra with short-wavelength lasers, 1064 nm was the only excitation that was consistently excellent for all organisms. We conclude that a 1064 nm Raman spectrometer, miniaturized using an InGaAs detector, is the optimal instrument for in situ studies of pigmented microbial communities at the limits of life on Earth. This has practical potential for the quest for biomolecules residual from any former surface life on Mars.