With the aid of the All Sky Automated Survey ( ASAS ) database on the Galactic field , we compare the iron abundances of fundamental mode RR Lyrae stars derived from the Fourier parameters with those obtained from low-dispersion spectroscopy .
We show from a set of 79 stars , distinct from the original calibrating sample of the Fourier method and selected without quality control , that almost all discrepant estimates are the results of some defects or peculiarities either in the photometry or in the spectroscopy .
Omitting objects deviating by more than 0.4 dex , the remaining subsample of 64 stars yields Fourier abundances that fit the spectroscopic ones with \sigma = 0.20 dex .
Other , more stringent selection criteria and different Fourier decompositions lead to smaller subsamples and concomitant better agreement , down to \sigma = 0.16 dex .
Except perhaps for two variables among the 163 stars , comprised of the ASAS variables and those of the original calibrating set of the Fourier method , all discrepant values can be accounted for by observational noise and insufficient data coverage .
We suggest that the agreement can be further improved when new , more accurate spectroscopic data become available for a test with the best photometric data .
As a by-product of this analysis , we also compute revised periods and select Blazhko variables .