Fe(III)-content ZSM-5 catalysts play important role in selective redox reactions, like biomimetic oxidations, or selective catalytic reactions (SCR) of pollutants, dangerous for the environment. There is a dispute in the literature whether the catalytic activity of these materials is due to isomorphously substituting framework (FW) iron, or extra-framework (EFW) oxide/hydroxide material of high dispersity, not incorporated into the lattice during the synthesis, or became ejected from the framework following the post-synthesis treatments (calcination, heat-treatment). In either case it is a must to know the FW/EFW ratio in order to correlate the catalytic activity with the number of active sites, or active surface area of the catalyst ingredient. In the case of EFW iron only (introduced, e.g. by ion-exchange) the chemical identity, particle size, location and thereby the probability of attainment from the fluid phase(s) are uncertain. Isomorphic substitution is more complicated, because in addition to these even the FW/EFW ratio is questionable. From both practical and academic interest it would be advantageous (e.g. from the point of view of above correlation) to have uniform FW, or EFW Fe(III) siting, or when it is mixed, to know the FW/EFW ratio. The authors approach this problem from several directions: they elaborate various synthesis methods to minimise EFW iron; extract unwanted EFW iron by reducing/complexing agents (SO2, hydroxylamine), or, if necessary, reintroduce EFW iron by (liquid and solid) ion-exchange and monitor the changes thus produced in the FW and EFW iron contents by X-band EPR and Mossbauer spectroscopy. Using mathematical methods and sophisticated analysis of the sub-spectra (intensity dependence on the iron content and temperature), the EPR spectrum-deconvolution made possible to reinterpret the X-band EPR Fe(III) spectra in weak ligand-field media, like zeolites. If the total iron contents of samples are known (e.g. from XRF spectra) the FW/EFW ratio can be computed on the basis of the deconvoluted EPR spectra alone. These values are in very good agreement with similar estimates obtained from the Mossbauer spectra, proving that the Mossbauer "loudness" for the various iron components is nearly equal.