The experimental situation of the density spectrum is a confused one. Cloud chambers offer the most direct access to the spectrum; their measurements, including new data, and their limitations are examined here. Experimental evidence shows that a mass above a chamber as thick as 0.6 radiation lengths can seriously distort a spectrum from that measured by a 'thin'-walled chamber of 0.1 radiation lengths. Thin chambers at mountain altitude and sea level show a sudden steepening of the exponent at high densities that is statistically significant. Monte-Carlo simulations of density spectra from an assumed primary spectrum reveal that normal interactions of proton primaries cannot reproduce the results. Other density spectra are examined and triggering selection bias emerges as a strong reason why some spectra are discordant.