Virtually all of the N detected by 15 N cross polarization (CP) NMR spectra of four HF-treated soil clay fractions is amide N. However, the intensity of this 15 N CP NMR signal (per unit N) is 27-57% lower than detected for a wheat protein, gliadin. There are two possible explanations -either the amide N in the soil clay fractions produces proportionately less NMR signal than does the amide N in gliadin, or part of the N in the soil clay fractions produces little or no NMR signal. The cross polarization dynamics of the gliadin amide resonance and amide resonances detected for the soil clay fractions are very similar and thus should produce similar amounts of signal, ruling out the first possibility. Therefore up to half or even more of the organic N in these soil clay fractions must be in a form that is insensitive to NMR detection. For a model compound (caffeine), non-protonated heterocyclic N produced less than 20% of the signal of an equivalent amount of amide N in gliadin. Results from several 13 C NMR techniques provide further evidence that much of the undetected N in the soil clay fractions may be heterocyclic.