1 Introduction

At present the only completely identified sample of radio sources is the 3CRR survey (Laing et al. 1983). The parent 3C radio sample has been in existence for nearly 35 years and remains the most important single sample of radio sources. This is because it is (i) complete - it has a flux limit and has virtually complete redshift information - and (ii) unbiased - the sample is selected at low radio frequency, and is therefore not influenced by flat-spectrum components, for which Doppler beaming distorts the statistics by enhancing the flux of radio sources whose axes just happen to be pointing in our direction.

However, it is likely that the 3C sample, although it makes finding galaxies at high redshift easy, is not a representative sample of high-redshift galaxies because the presence of powerful radio AGN affects the properties of the optical galaxy. Specifically, the discovery of UV-optical extended continuum (McCarthy et al. 1987) aligned with the radio axis is evidence for very strong interaction between the radio jet and the intragalactic medium. It has been claimed (Dunlop & Peacock 1993) that this alignment effect tends to disappear at radio powers only a factor 5-10 lower than the 3C survey. In this case we might expect to pick up passively evolving giant ellipticals by selecting lower-luminosity radio galaxies. There is, however, much debate about this point (e.g. Eales & Rawlings 1996).

These observations are part of an ongoing project (Rhee et al. 1996; hereafter Paper I) to identify and study radio galaxies and quasars from the 4C survey (Pilkington & Scott 1965; Gower et al. 1967) between declinations of $-4^{\circ}$ and $+4^{\circ}$. The 4C survey is approximately a factor of 5 deeper than the 3C survey, and has a limiting flux density of 2 Jy at 178 MHz. In order to get a complete sample, large enough to study the properties of distant galaxies, to determine their space density and to be able to compare the properties of the steep spectrum radio galaxies with those of normal radio galaxies and quasars, we selected the 4C sources whose angular sizes are smaller than 30 arcseconds using the 365-MHz Texas survey (Douglas et al. 1980). Our selection is independent of the radio spectral index and the complete radio sample will be published elsewhere. The subset of the 4C sample whose declinations are between $-4^{\circ}$ and $+4^{\circ}$ has a lot of radio observations, providing good radio spectral indices and has the advantage of being observable from both hemispheres. In this article, we have concentrated on steep radio spectrum $\alpha\gt.9\footnote{The spectral index is defined by Flux Density $\propto$\space (Frequency)$^{-\alpha}.$}$ objects which are not identified on the Palomar Sky Survey in order to improve positional information for followup optical spectroscopy. The sources have also been selected, as in Paper I, to have angular sizes of less than 30$\hbox{$^{\prime\prime}$}$.Here we present radio maps of a further 144 radio sources (19 of which are common with Paper I). The sources were chosen as far as possible to complete the mapping of RA ranges according to the LST ranges at which the observations were scheduled. Hence most sources have ${\rm RA}<5{\rm h}$, $8{\rm h}<{\rm RA}<11{\rm h}$, $15{\rm h}<{\rm RA}<18{\rm h}$ and ${\rm RA}\gt 20{\rm h}$.