We have tested the hypothesis of Berthet (1991A&A...251..171B) which foresees that Am stars become giant metallic A and F stars (defined by an enhanced value of the blanketing parameter {delta}m2 of the Geneva photometry) when they evolved. If this hypothesis is right, Am and metallic A-FIII stars need to have the same rate of binaries and a similar distribution of vsini. From our new spectroscopic data and from vsini and radial velocities in the literature, we show that it is not the case. The metallic giant stars are often fast rotators with vsini larger than 100 km/s, while the maximum rotational velocity for Am stars is about 100 km/s. The rate of tight binaries with periods less than 1000 days is less than 30 % among metallic giants, which is incompatible with the value of 75 % for Am stars (Abt & Levy, 1985ApJS...59..229A). Therefore, the simplest way to explain the existence of giant metallic F stars is to suggest that all normal A and early F stars might go through a short "metallic" phase when they are finishing their life on the main sequence. Besides, it is shown that only giant stars with spectral type comprised between F0 and F6 may have a really enhanced {delta}m2 value, while all A-type giants seem to be normal.