We present a search for FRI radio galaxies between 1 < z < 2 in the COSMOS field. In absence of spectroscopic redshift measurements, the selection method is based on multiple steps which make use of both radio and optical constraints. The basic assum
ptions are that 1) the break in radio power between low-power FRIs and the more powerful FRIIs does not change with redshift, and 2) that the photometric properties of the host galaxies of low power radio galaxies in the distant universe are similar to those of FRIIs in the same redshift bin, as is the case for nearby radio galaxies. We describe the results of our search, which yields 37 low-power radio galaxy candidates that are possibly FRIs. We show that a large fraction of these low-luminosity radio galaxies display a compact radio morphology, that does not correspond to the FRI morphological classification. Furthermore, our objects are apparently associated with galaxies that show clear signs of interactions, at odds with the typical behavior observed in low-z FRI hosts. The compact radio morphology might imply that we are observing intrinsically small and possibly young objects, that will eventually evolve into the giant FRIs we observe in the local universe. One of the objects appears as point-like in HST images. This might belong to a population of FRI-QSOs, which however would represent a tiny minority of the overall population of high-z FRIs. As for the local FRIs, a large fraction of our objects are likely to be associated with groups or clusters, making them beacons for high redshift clusters of galaxies. Our search for candidate high-z FRIs we present in this paper constitutes a pilot study for objects to be observed with future high-resolution and high-sensitivity instruments (shortened)
