A sample of 6C radio sources designed to find objects at redshift z>4– III. Imaging and the radio galaxy K–z relation

In this paper, the third and final of a series, we present complete K -band imaging and some complementary I -band imaging of the filtered 6C* sample. We find no systematic differences between the K – z relation of 6C* radio galaxies and those from complete samples, so the near-infrared properties of luminous radio galaxies are not obviously biased by the additional 6C* radio selection criteria (steep spectral index and small angular size). The 6C* K – z data significantly improve delineation of the K – z relation for radio galaxies at high redshift ( z >2) . Accounting for non-stellar contamination, and for correlations between radio luminosity and stellar mass, we find little support for previous claims that the underlying scatter in the stellar luminosity of radio galaxies increases significantly at z >2 . In a particular spatially flat universe with a cosmological constant (Ω_M=0.3 and Ω_Λ=0.7) , the most luminous radio sources appear to be associated with galaxies with a luminosity distribution with a high mean (≈5  L *), and a low dispersion ( σ ∼0.5 mag) which formed their stars at epochs corresponding to z ≳2.5 . This result is in line with recent submillimetre studies of high-redshift radio galaxies and the inferred ages of extremely red objects from faint radio samples.     

The redshift distribution of FIRST radio sources at 1 mJy

We present spectra for a sample of radio sources from the FIRST survey, and use them to define the form of the redshift distribution of radio sources at mJy levels. We targeted 365 sources and obtained 46 redshifts (13 per cent of the sample). We find that our sample is complete in redshift measurement to R ∼18.6, corresponding to z ∼0.2. Galaxies were assigned spectral types based on emission-line strengths. Early-type galaxies represent the largest subset (45 per cent) of the sample and have redshifts 0.15≲ z ≲0.5; late-type galaxies make up 15 per cent of the sample and have redshifts 0.05≲ z ≲0.2; starbursting galaxies are a small fraction (∼6 per cent), and are very nearby ( z ≲0.05). Some 9 per cent of the population have Seyfert 1/quasar-type spectra, all at z ≳0.8, and 4 per cent are Seyfert 2 type galaxies at intermediate redshifts ( z ∼0.2).
Using our measurements and data from the Phoenix survey (Hopkins et al.), we obtain an estimate for N ( z ) at S _1.4 GHz≥1 mJy and compare this with model predictions. At variance with previous conclusions, we find that the population of starbursting objects makes up ≲5 per cent of the radio population at S ∼1 mJy.     

The radio galaxy 3C 356 and clues to the trigger mechanisms for powerful radio sources

We present deep near-infrared images, taken with the Subaru Telescope, of the region around the   z =1.08  radio source 3C 356 which show it to be associated with a poor cluster of galaxies. We discuss evidence that this cluster comprises two subclusters traced by the two galaxies previously proposed as identifications for 3C 356, which both seem to harbour active galactic nuclei, and which have the disturbed morphologies expected if they underwent an interpenetrating collision at the time the radio jets were triggered. We explain the high luminosity and temperature of the diffuse X-ray emission from this system as the result of shock heating of intracluster gas by the merger of two galaxy groups. Taken together with the results on other well-studied powerful radio sources, we suggest that the key ingredient for triggering a powerful radio source, at least at epochs corresponding to   z ∼1  , is a galaxy–galaxy interaction which can be orchestrated by the merger of their parent subclusters. This provides an explanation for the rapid decline in the number density of powerful radio sources since   z ∼1  . We argue that attempts to use distant radio-selected clusters to trace the formation and evolution of the general cluster population must address ways in which X-ray properties can be influenced by the radio source, both directly, by mechanisms such as inverse Compton scattering, and indirectly, by the fact that the radio source may be preferentially triggered at a specific time during the formation of the cluster.     

Optical spectroscopy of two overlapping, flux-density-limited samples of radio sources in the North Ecliptic Cap, selected at 38 and 151 MHz

We present the results of optical spectroscopy of two flux-density-limited samples of radio sources selected at frequencies of 38 and 151 MHz in the same region around the North Ecliptic Cap, the 8C-NEC and 7C- iii samples respectively. Both samples are selected at flux density levels ≈20 times fainter than samples based on the 3C catalogue. They are amongst the first low-frequency selected samples with no spectral or angular size selection for which almost complete redshift information has been obtained, and contain many of the lowest-luminosity z >2 radio galaxies so far discovered. They will therefore provide a valuable resource for understanding the cosmic evolution of radio sources and their hosts and environments. The 151-MHz 7C- iii sample is selected to have S ₁₅₁≥0.5 Jy and is the more spectroscopically complete; out of 54 radio sources fairly reliable redshifts have been obtained for 44 objects. The 8C sample has a flux limit of S ₃₈≥1.3 Jy and contains 58 sources of which 46 have fairly reliable redshifts. We discuss possible biases in the observed redshift distribution, and some interesting individual objects, including a number of cases of probable gravitational lensing. Using the 8C-NEC and 7C- iii samples in conjunction, we form the first sample selected on low-frequency flux in the rest-frame of the source, rather than the usual selection on flux density in the observed frame. This allows us to remove the bias associated with an increasing rest-frame selection frequency with redshift. We investigate the difference this selection makes to correlations of radio source properties with redshift and luminosity by comparing the results from traditional flux-density selection with our new method. We show in particular that flux-density-based selection leads to an overestimate of the steepness of the correlation of radio source size with redshift.     

The 6C** sample of steep-spectrum radio sources – II. Redshift distribution and the space density of high-redshift radio galaxies

We use the 6C** sample to investigate the comoving space density of powerful, steep-spectrum radio sources. This sample, consisting of 68 objects, has virtually complete K -band photometry and spectroscopic redshifts for 32 per cent of the sources. In order to find its complete redshift distribution, we develop a method of redshift estimation based on the K – z diagram of the 3CRR, 6CE, 6C* and 7CRS radio galaxies. Based on this method, we derive redshift probability density functions for all the optically identified sources in the 6C** sample. Using a combination of spectroscopic and estimated redshifts, we select the most radio luminous sources in the sample. Their redshift distribution is then compared with the predictions of the radio luminosity function of Jarvis et al. We find that, within the uncertainties associated with the estimation method, the data are consistent with a constant comoving space density of steep-spectrum radio sources beyond z ≳ 2.5, and rule out a steep decline.     

Morphology and surface brightness evolution of z∼1.1 radio galaxies

We use K '-band (2.1-μm) imaging to investigate the angular size and morphology of 10 6C radio galaxies, at redshifts 1≤ z ≤1.4. Two radio galaxies appear to be undergoing mergers, another contains, within a single envelope, two intensity peaks aligned with the radio jets, while the other seven appear consistent with being normal ellipticals in the K band.
Intrinsic half-light radii are estimated from the areas of each radio galaxy image above a series of thresholds. The 6C galaxy radii are found to be significantly smaller than those of the more radio-luminous 3CR galaxies at similar redshifts. This would indicate that the higher mean K -band luminosity of the 3CR galaxies reflects a difference in the size of the host galaxies, and not solely a difference in the power of the active nuclei.
The size–luminosity relation of the z ∼1.1 6C galaxies indicates a 1.0–1.6 mag enhancement of their rest frame R -band surface brightness relative to either local ellipticals of the same size or FRII radio galaxies at z <0.2. The 3CR galaxies at z ∼1.1 show a comparable enhancement in surface brightness. The mean radius of the 6C galaxies suggests that they evolve into ellipticals of L ∼ L * luminosity, and is consistent with their low-redshift counterparts being relatively small FRII galaxies ∼25 times lower in radio luminosity, or small FRI galaxies ∼1000 times lower in radio luminosity. Hence the 6C radio galaxies appear to undergo as much optical and radio evolution as the 3CR galaxies.     

The relationship between star formation rate and radio synchrotron luminosity at 0 < z < 2

We probe the relationship between star formation rate (SFR) and radio synchrotron luminosity in galaxies at  0 < z < 2  within the northern Spitzer Wide-area Infrared Extragalactic survey (SWIRE) fields, in order to investigate some of the assumptions that go into calculating the star formation history of the Universe from deep radio observations. We present new 610-MHz Giant Metrewave Radio Telescope (GMRT) observations of the European Large-Area ISO Survey-North 2 (ELAIS-N2) field, and using this data, along with previous GMRT surveys carried out in the ELAIS-N1 (North 1) and Lockman Hole regions, we construct a sample of galaxies which have redshift and SFR information available from the SWIRE survey. We test whether the local relationship between SFR and radio luminosity is applicable to   z = 2  galaxies, and look for evolution in this relationship with both redshift and SFR in order to examine whether the physical processes which lead to synchrotron radiation have remained the same since the peak of star formation in the Universe. We find that the local calibration between radio luminosity and star formation can be successfully applied to radio-selected high-redshift, high-SFR galaxies, although we identify a small number of sources where this may not be the case; these sources show evidence for inaccurate estimations of their SFR, but there may also be some contribution from physical effects such as the recent onset of starburst activity, or suppression of the radio luminosity within these galaxies.     

Radio–optical alignments in a low radio luminosity sample

We present an optically based study of the alignment between the radio axes and the optical major axes of eight z ∼0.7 radio galaxies in a 7C sample. The radio galaxies in this sample are ≈20 times less radio‐luminous than 3C galaxies at the same redshift, and are significantly less radio-luminous than any other well-defined samples studied to date. Using Nordic Optical Telescope images taken in good seeing conditions at rest frame wavelengths just longward of the 4000-Å break, we find a statistically significant alignment effect in the 7C sample. Furthermore, in two cases where the aligned components are well separated from the host we have been able to confirm spectroscopically that they are indeed at the same redshift as the radio galaxy. However, a quantitative analysis of the alignment in this sample and in a corresponding 3C sample from HST archival data indicates that the percentage of aligned flux may be lower and of smaller spatial scale in the 7C sample. Our study suggests that alignments on the 50-kpc scale are probably closely related to the radio luminosity, whereas those on the 15‐kpc scale are not. We discuss these results in the context of popular models for the alignment effect.     

FR II radio galaxies with z < 0.3 — I. Properties of jets, cores and hotspots

In previous papers we have discussed high-resolution observations of a large sample of powerful radio galaxies with z  < 0.3. Jets are detected in up to 80 per cent of the sample, and radio cores in nearly all the objects; in addition, we are able to resolve the hotspots in most sources. In this paper we present measurements of the radio properties of these components.   The prominences of the jets detected do not appear to be a function of radio luminosity, providing the clearest evidence yet that the reported low detection rate of jets in radio galaxies has been an artefact of low-sensitivity observations. We find a positive correlation between the total source length and core prominence in the narrow-line radio galaxies. We have found evidence for a relationship between hotspot size and total source size, but few other significant relationships between hotspot properties and those of the jets or lobes. We compare our measurements with those of Bridle et al., based on observations of a sample of quasars, and argue that the results are consistent with a modification of the unified model in which the broad-line radio galaxies are the low-luminosity counterparts of quasars, although the situation is complicated by contamination with low-excitation radio galaxies which appear to have radio properties different from the high-excitation objects. We discuss the classes of empirical model that can be fitted to the data set.     

CENSORS: A Combined EIS–NVSS Survey Of Radio Sources – I. Sample definition, radio data and optical identifications

A new sample of radio sources, with the designated name CENSORS (A Combined EIS–NVSS Survey Of Radio Sources), has been defined by combining the National Radio Astronomy Observatory Very Large Array Sky Survey (NVSS) at 1.4 GHz with the ESO Imaging Survey (EIS) Patch D, a 3° by 2° region of sky centred at RA     , Dec. −21°00'00' (J2000). New radio observations of 199 NVSS radio sources with NVSS flux densities   S _{1.4 GHz} > 7.8 mJy  are presented, and are compared with the EIS I -band imaging observations which reach a depth of   I ∼ 23  ; optical identifications are obtained for over two-thirds of the ∼150 confirmed radio sources within the EIS field. The radio sources have a median linear size of 6 arcsec, consistent with the trend for lower flux density radio sources to be less extended. Other radio source properties, such as the lobe flux density ratios, are consistent with those of brighter radio source samples. From the optical information, 30–40 per cent of the sources are expected to lie at redshifts   z ≳ 1.5  .
One of the key goals of this survey is to accurately determine the high-redshift evolution of the radio luminosity function. These radio sources are at the ideal flux density level to achieve this goal; at redshifts   z ∼ 2  they have luminosities which are around the break of the luminosity function and so provide a much more accurate census of the radio source population at those redshifts than the existing studies of extreme, high radio power sources. Other survey goals include investigating the dual-population unification schemes for radio sources, studying the radio luminosity dependence of the evolution of radio source environments, and understanding the radio power dependence of the K – z relation for radio galaxies.     

Jet–gas interactions in z∼ 2.5 radio galaxies: evolution of the ultraviolet line and continuum emission with radio morphology

We present an investigation into the nature of the jet–gas interactions in a sample of 10 radio galaxies at  2.3 < z < 2.9  using deep spectroscopy of the ultraviolet (UV) line and continuum emission obtained at Keck II and the Very Large Telescope. Kinematically perturbed gas, which we have shown to be within the radio structure in previous publications, is always blueshifted with respect to the kinematically quiescent gas, is usually spatially extended, and is usually detected on both sides of the nucleus. In the three objects from this sample for which we are able to measure line ratios for both the perturbed and quiescent gases, we suggest that the former has a lower ionization state than the latter.
We propose that the perturbed gas is part of a jet-induced outflow, with dust obscuring the outflowing gas that lies on the far side of the object. The spatial extent of the blueshifted perturbed gas, typically ∼35 kpc, implies that the dust is spatially extended at least on similar spatial scales.
We also find interesting interrelationships between UV line, UV continuum and radio continuum properties of this sample.     

A submillimetre survey of the star formation history of radio galaxies

We present the results of the first major systematic submillimetre survey of radio galaxies spanning the redshift range 1< z <5. The primary aim of this work is to elucidate the star formation history of this sub class of elliptical galaxies by tracing the cosmological evolution of dust mass. Using SCUBA on the JCMT, we have obtained 850-μm photometry of 47 radio galaxies to a consistent rms depth of 1 mJy, and have detected dust emission in 14 cases. The radio galaxy targets have been selected from a series of low-frequency radio surveys of increasing depth (3CRR, 6CE, etc.), in order to allow us to separate the effects of increasing redshift and increasing radio power on submillimetre luminosity. Although the dynamic range of our study is inevitably small, we find clear evidence that the typical submillimetre luminosity (and hence dust mass) of a powerful radio galaxy is a strongly increasing function of redshift; the detection rate rises from ≃15 per cent at z <2.5 to ≳75 per cent at z >2.5, and the average submillimetre luminosity rises at a rate ∝(1+ z )³ out to z ≃4. Moreover, our extensive sample allows us to argue that this behaviour is not driven by underlying correlations with other radio galaxy properties such as radio power, radio spectral index, or radio source size/age. Although radio selection may introduce other more subtle biases, the redshift distribution of our detected objects is in fact consistent with the most recent estimates of the redshift distribution of comparably bright submillimetre sources discovered in blank field surveys. The evolution of submillimetre luminosity found here for radio galaxies may thus be representative of massive ellipticals in general.     

HST and UKIRT imaging observations of z∼ 1 6C radio galaxies – II. Galaxy morphologies and the alignment effect

Powerful radio galaxies often display enhanced optical/ultraviolet emission regions, elongated and aligned with the radio jet axis. The aim of this series of papers is to investigate separately the effects of radio power and redshift on the alignment effect, together with other radio galaxy properties. In this second paper, we present a deeper analysis of the morphological properties of these systems, including both the host galaxies and their surrounding aligned emission.
The host galaxies of our 6C subsample are well described as de Vaucouleurs ellipticals, with typical scale sizes of  ∼10 kpc  . This is comparable to the host galaxies of low- z radio sources of similar powers, and also the more powerful 3CR sources at the same redshift. The contribution of nuclear point source emission is also comparable, regardless of radio power.
The 6C alignment effect is remarkably similar to that seen around more powerful 3CR sources at the same redshift in terms of extent and degree of alignment with the radio source axis, although it is generally less luminous. The bright, knotty features observed in the case of the z ∼ 1 3CR sources are far less frequent in our 6C subsample; neither do we observe such strong evidence for evolution in the strength of the alignment effect with radio source size/age. However, we do find a very strong link between the most extreme alignment effects and emission-line region properties indicative of shocks, regardless of source size/age or power. In general, the 6C alignment effect is still considerably stronger than that seen around lower redshift galaxies of similar radio powers. Cosmic epoch is clearly just as important a factor as radio power: although aligned emission is observed on smaller scales at lower redshifts, the processes which produce the most extreme features simply no longer occur, suggesting considerable evolution in the properties of the extended haloes surrounding the radio source.     

The active nuclei of z < 1.0 3CRR radio sources

We combine Chandra and XMM–Newton X-ray data from our previous papers with new X-ray observations and with Spitzer mid-infrared (mid-IR) data in order to study the nature of the nuclei of radio galaxies and radio-loud quasars with   z < 1.0  from the 3CRR sample. The significant increase in sample size over our previous work, the reduction of bias in the sample as a result of new observations and the availability of more mid-IR data allow us to show conclusively that almost all objects classed as low-excitation radio galaxies in optical spectroscopic studies lack a radiatively efficient active nucleus. We show that the distribution of absorbing columns in the narrow-line radio galaxies differs from the population of X-ray-selected radio-quiet type 2 quasars and from that in local Seyfert 2s. We comment on the current evidence for the nature of the soft X-ray component in radio-galaxy nuclear spectra, concluding that a jet origin for this component is very hard to evade. Finally, we discuss the recently discovered 'fundamental plane' of black hole activity, showing that care must be taken when placing radio-loud active galactic nucleus (AGN) on such diagnostic diagrams.     

Radio imaging of the Subaru/XMM–Newton Deep Field – II. The 37 brightest radio sources

We study the 37 brightest radio sources in the Subaru/ XMM–Newton Deep Field. We have spectroscopic redshifts for 24 of 37 objects and photometric redshifts for the remainder, yielding a median redshift z _med for the whole sample of   z _med≃ 1.1  and a median radio luminosity close to the 'Fanaroff–Riley type I/type II (FR I/FR II)' luminosity divide. Using mid-infrared (mid-IR) ( Spitzer MIPS 24 μm) data we expect to trace nuclear accretion activity, even if it is obscured at optical wavelengths, unless the obscuring column is extreme. Our results suggest that above the FR I/FR II radio luminosity break most of the radio sources are associated with objects that have excess mid-IR emission, only some of which are broad-line objects, although there is one clear low-accretion-rate object with an FR I radio structure. For extended steep-spectrum radio sources, the fraction of objects with mid-IR excess drops dramatically below the FR I/FR II luminosity break, although there exists at least one high-accretion-rate 'radio-quiet' QSO. We have therefore shown that the strong link between radio luminosity (or radio structure) and accretion properties, well known at z ∼ 0.1, persists to z ∼ 1. Investigation of mid-IR and blue excesses shows that they are correlated as predicted by a model in which, when significant accretion exists, a torus of dust absorbs ∼30 per cent of the light, and the dust above and below the torus scatters ≳1 per cent of the light.     

Quasars from the 7C survey – I. Sample selection and radio maps

We describe the selection of candidate radio-loud quasars obtained by cross-matching radio source positions from the low-frequency (151-MHz) 7C survey with optical positions from five pairs of EO POSS-I plates scanned with the Cambridge Automatic Plate-measuring Machine (APM). The sky region studied is centred at RA 10^h 28^m, Dec.+41° and covers ≈0.057 sr. We present VLA observations of the quasar candidates, and tabulate various properties derived from the radio maps. We discuss the selection criteria of the resulting '7CQ' sample of radio-loud quasars. The 70 confirmed quasars, and some fraction of the 36 unconfirmed candidates, constitute a filtered sample with the following selection criteria: 151-MHz flux density S ₁₅₁>100 mJy; POSS-I E -plate magnitude E ≈ R <20; POSS-I colour ( O E )<1.8; the effective area of the survey drops significantly below S ₁₅₁≈200 mJy. We argue that the colour criterion excludes few if any quasars, but note, on the basis of recent work by Willott et al., that the E magnitude limit probably excludes more than 50 per cent of the radio-loud quasars.     

MAMBO observations at 240 GHz of optically obscured Spitzer sources: source clumps and radio activity at high redshift

Optically very faint  ( R > 25.5)  sources detected by the Spitzer Space Telescope at 24 μm represent a very interesting population at redshift   z ∼ (1.5–3)  . They exhibit strong clustering properties, implying that they are hosted by very massive haloes, and their mid-infrared emission could be powered by either dust-enshrouded star formation and/or by an obscured active galactic nucleus (AGN). We report observations carried out with the Max Planck Millimetre Bolometer (MAMBO) array at the IRAM 30-m antenna on Pico Veleta of a candidate protocluster with five optically obscured sources selected from the 24-μm Spitzer sample of the First-Look Survey. Interestingly, these sources appear to lie on a high-density filament aligned with the two radio jets of an AGN. Four out of five of the observed sources were detected. We combine these measurements with optical, infrared and radio observations to probe the nature of the candidate protocluster members. Our preliminary conclusions can be summarized as follows: the spectral energy distributions (SEDs) of all sources include both AGN and starburst contributions; the AGN contribution to the bolometric luminosities ranges between 14 and 26 per cent of the total. Such a contribution is enough for the AGN to dominate the emission at 5.8, 8 and 24 μm, while the stellar component, inferred from SED fitting, prevails at 1.25 mm and at  λ < 4.5 μ  m. The present analysis suggests a coherent interplay at high z between extended radio activity and the development of filamentary large-scale structures.     

The IPHAS-POSS-I proper motion survey of the Galactic plane

We present millimetre observations of a sample of 12 high-redshift ultraluminous infrared galaxies (ULIRGs) in the extended growth strip (EGS). These objects were initially selected on the basis of their observed mid-IR colours (  0.0 < [3.6]−[4.5] < 0.4  and  −0.7 < [3.6]−[8.0] < 0.5  ) to lie at high redshift  1.5 ≲ z ≲ 3  , and subsequent 20–38 μm mid-IR spectroscopy confirms that they lie in a narrow redshift window centred on   z ≈ 2  . We detect 9/12 of the objects in our sample at high significance  (>3σ)  with a mean 1200 μm flux of  〈 F _1200 μm〉= 1.6 ± 0.1  mJy. Our millimetre photometry, combined with existing far-IR photometry from the Far-IR Deep Extragalactic Legacy Survey (FIDELS) and accurate spectroscopic redshifts, places constraints both sides of the thermal dust peak. This allows us to estimate the dust properties, including the far-IR luminosity, dust temperature and dust mass. We find that our sample is similar to other high- z and intermediate- z ULIRGs, and local systems, but has a different dust selection function than submillimeter-selected galaxies. Finally, we use existing 20-cm radio continuum imaging to test the far-IR/radio correlation at high redshift. We find that our sample is consistent with the local relation, implying little evolution. Furthermore, this suggests that our sample selection method is efficient at identifying ultraluminous, starburst-dominated systems within a very narrow redshift range centred at   z ∼ 2  .     

The cluster environments of the z ∼ 1 3CR radio galaxies

An analysis of the environments around a sample of 28 3CR radio galaxies with redshifts 0.6< z <1.8 is presented, based primarily upon K -band images down to K ∼20 taken using the UK Infrared Telescope (UKIRT). A net overdensity of K -band galaxies is found in the fields of the radio galaxies, with the mean excess counts being comparable to that expected for clusters of Abell Class 0 richness. A sharp peak is found in the angular cross-correlation amplitude centred on the radio galaxies that, for reasonable assumptions about the luminosity function of the galaxies, corresponds to a spatial cross-correlation amplitude between those determined for low-redshift Abell Class 0 and 1 clusters.
These data are complemented by J -band images also from UKIRT, and by optical images from the Hubble Space Telescope . The fields of the lower redshift ( z ≲0.9) radio galaxies in the sample generally show well-defined near-infrared colour–magnitude relations with little scatter, indicating a significant number of galaxies at the redshift of the radio galaxy; the relations involving colours at shorter wavelengths than the 4000 Å break show considerably greater scatter, suggesting that many of the cluster galaxies have low levels of recent or on-going star formation. At higher redshifts the colour–magnitude sequences are less prominent owing to the increased field galaxy contribution at faint magnitudes, but there is a statistical excess of galaxies with the very red infrared colours ( J − K ≳1.75) expected of old cluster galaxies at these redshifts.
Although these results are appropriate for the mean of all of the radio galaxy fields, there exist large field-to-field variations in the richness of the environments. Many, but certainly not all, powerful z ∼1 radio galaxies lie in (proto)cluster environments.     

Deviations from passive evolution – star formation and the ultraviolet excess in z∼ 1 radio galaxies

Galaxy colours are determined for two samples of 6C and 3CR radio sources at z ∼ 1, differing by a factor of ∼6 in radio power. Corrections are made for emission-line contamination and the presence of any nuclear point source, and the data analysed as a function of both redshift and the radio source properties. The galaxy colours are remarkably similar for the two populations, and the ultraviolet excess evolves with radio source size similarly in both samples, despite the fact that the alignment effect is more extensive for the more powerful 3CR radio galaxies. These results seem to suggest that the alignment effect at these redshifts does not scale strongly with radio power, and is instead more closely dependent on galaxy mass (which is statistically comparable for the two samples). However, it is likely that the presence of relatively young (≲several times 10⁸-yr-old) stellar populations has considerably contaminated the K -band flux of these systems, particularly in the case of the more powerful 3CR sources, which are ∼0.5 mag more luminous than the predictions of passive evolution models at z ∼ 1. The higher luminosity of the 3CR alignment effect is balanced by emission at longer wavelengths, thereby leading to comparable colours for the two samples.