X-Shaped Radio Galaxies (XRGs) Overview

• XRGs are defined by two misaligned radio lobe pairs—a bright primary with hotspots and fainter, diffuse wings representing older synchrotron plasma.
• Studies employ multi-frequency spectral mapping and simulations to uncover jet reorientation, backflow dynamics, and local re-acceleration processes in these sources.
• Large surveys reveal that most XRGs are FR II systems with characteristic wing orientations, reflecting a diverse origin including mergers, jet–shell interactions, and dual SMBH activity.

X-shaped radio galaxies (XRGs), also called winged radio galaxies, are a rare class of radio-loud active galactic nuclei whose large-scale emission comprises two misaligned pairs of lobes: an active, usually higher-surface-brightness primary pair and a fainter secondary pair of diffuse “wings,” producing an overall inversion-symmetric X-like morphology. In the standard phenomenology, the primary lobes are the currently powered structures, often with FR II-like edge brightening and terminal hotspots, whereas the wings are lower-surface-brightness, generally hotspot-free, and frequently interpreted as older synchrotron plasma. XRGs have long been treated as a diagnostic population for jet–environment coupling, jet-axis reorientation, merger-driven SMBH dynamics, and radio-plasma ageing, but recent observational and numerical work increasingly supports a heterogeneous rather than universal origin (Patra et al., 2023, Giri et al., 2024).

1. Morphological definition and classification

The defining observational property of an XRG is the coexistence of a bright primary lobe pair and a misaligned secondary pair of wings. In most descriptions, the primaries are edge-brightened and may terminate in hotspots, while the wings are diffuse, edge-darkened, and lack terminal hotspots. This contrast is central to the distinction between active and passive components of the source structure (Giri et al., 2023, Giri et al., 2024).

Operationally, recent wide-area surveys have separated XRGs from Z-shaped radio galaxies (ZRGs) by the location at which the wings appear to emerge. In FIRST- and LoTSS-based classifications, XRG wings emanate from the central or near-central region, commonly formalized as within approximately 25% of the primary jet length from the core, whereas ZRG wings emerge from the outer parts of the primary lobes. LoTSS DR2 adopted an equivalent inner-quarter geometric rule, while explicitly avoiding rigid cuts on wing-to-lobe length ratio or misalignment angle beyond the X/Z split itself (Bera et al., 2020, Bera et al., 24 Apr 2026).

The distribution of wing–lobe angles is strongly weighted toward large misalignments. FIRST-based work found that most XRGs have angles in the 70–90 degree range, peaking near 75 degrees, and TGSS ADR1 reported that 60% of XRGs lie in the 80–90 degree bin. Review work synthesizing earlier samples similarly places the typical misalignment around 75 degrees or greater. Wing-to-primary length ratios are usually below unity, with a median around 0.9 in the review literature, but clear exceptions exist, including sources with wings longer than primaries and ratios up to about 2.8 (Bera et al., 2020, Bhukta et al., 2020, Giri et al., 2024).

Although XRGs are often described as lying near the FR I/FR II transition, the population is morphologically dominated by FR II systems. This asymmetry matters because backflow-based models require powerful, hotspot-forming jets, whereas FR I XRGs, though well documented, are comparatively uncommon and often demand additional explanations such as reorientation or jet–shell interactions (Bera et al., 24 Apr 2026, Bruno et al., 2019).

2. Demography and survey-defined populations

Large-area radio surveys have transformed XRGs from a small set of archetypes into a statistically tractable population. FIRST-based visual and semi-automated searches produced a sequence of increasingly inclusive catalogs, culminating in 296 winged systems in one full-footprint FIRST study, of which 161 were classified as XRGs and 135 as ZRGs. An earlier extended FIRST catalog listed 290 XRG candidates, split into 106 “strong” and 184 “probable” systems, nearly doubling the then-known population (Bera et al., 2020, Yang et al., 2019).

Low-frequency surveys have expanded this census further by recovering diffuse, steep-spectrum wings that are faint or partially resolved out at 1.4 GHz. TGSS ADR1 identified 58 new winged sources, including 40 XRGs, and LoTSS DR1 found 33 winged systems, of which 21 were XRGs. The largest step came with LoTSS DR2, which reported 621 confirmed winged radio galaxies and 403 additional candidates over about 5,700 square degrees; among the confirmed systems, 382 were XRGs and 239 ZRGs, giving an XRG-to-ZRG ratio of about 1.6 (Bhukta et al., 2020, Pal et al., 2021, Bera et al., 24 Apr 2026).

The LoTSS DR2 sample also quantified the large-scale nature of the population. For confirmed winged sources, the median linear size is 498 kpc, about 16% exceed 0.7 Mpc, and 25 sources exceed 1 Mpc. Redshifts are available for 72% of confirmed LoTSS DR2 winged sources, with median $z \approx 0.519 \pm 0.296$, substantially above typical FIRST-based medians. These numbers indicate that low-frequency selection is recovering both more distant and more extended systems (Bera et al., 24 Apr 2026).

Population-level radio properties are likewise consistent with lobe-dominated radio galaxies. In LoTSS DR2, nearly all winged sources are steep spectrum, with mean $\alpha_{144}^{1400} = -0.84 \pm 0.16$ overall and $-0.86 \pm 0.15$ for XRGs; about 88% of confirmed winged sources are morphologically FR II, and about 95% of classified XRGs are FR II. FIRST- and TGSS-based samples show the same broad trend, though with a higher relative fraction of FR I and complex morphologies in older, higher-frequency catalogs (Bera et al., 24 Apr 2026, Bera et al., 2020, Bhukta et al., 2020).

3. Spectral structure, ageing, and particle re-acceleration

Spatially resolved spectral-index mapping has become one of the primary tools for discriminating XRG formation scenarios. The key logic is straightforward: in most backflow or jet-reorientation models, wing plasma should on average be older than the associated primary-lobe plasma and therefore spectrally steeper. By contrast, a contemporaneously fueled dual-AGN interpretation predicts no systematic wing–lobe spectral difference (Patra et al., 2023).

The standard two-frequency spectral index may be written, depending on sign convention, either through $S_\nu \propto \nu^\alpha$ or $S_\nu \propto \nu^{-\alpha}$. For the 144 MHz–1.4 GHz pair, one common form is

$$\alpha_{144}^{1400} = -\frac{\ln\!\left(S_{1400}/S_{144}\right)}{\ln\!\left(1400/144\right)}.$$

A widely used synchrotron-age estimator is

$$t_{\rm syn} \approx 1590\,\frac{B^{1/2}}{B^2 + B_{\rm IC}^2}\,\nu_b^{-1/2}\ {\rm Myr},$$

with $B_{\rm IC} \approx 3.25(1+z)^2\,\mu{\rm G}$, although several XRG studies explicitly caution that two-frequency data alone do not constrain full JP/KP/CI ageing fits, break frequencies, or reliable absolute ages (Patra et al., 2023, Giri et al., 2023, Giri et al., 2024).

A matched-resolution LoTSS-DR2 plus FIRST/uGMRT study of 25 XRGs derived spectral-index maps and ridge-line profiles for 40 lobe–wing systems. Its dominant result was a progressive steepening from hotspot to wing in the majority of measured ridge-lines; flatter-spectrum wings were found in at most one clear case, J1015+5944, corresponding to 1 out of 25 sources or 0–1 out of 40 ridge-lines. This substantially weakens the claim that spectrally similar wings and lobes are a generic property of XRGs (Patra et al., 2023).

At the same time, later work showed that spectral steepening is not strictly monotonic everywhere. In a 40-XRG sample with 68 sufficiently resolved wings, 11 wings showed cessation of steepening or local flattening of the lobe-to-wing spectral profile, usually beginning near the structural bend where the lobe plasma stream turns into the wing. That study interpreted the effect as evidence for in-situ acceleration or re-acceleration associated with flow-deflection shocks, shear, reconnection, or turbulence. The result does not negate the older-on-average character of wings; rather, it localizes departures from simple ageing to dynamically special regions (Patra et al., 11 Nov 2025).

Case studies reinforce the same point. In Abell 3670, multifrequency JVLA and uGMRT work found progressive steepening from lobes to wings and age differences of about 20–30 Myr or more, with maximum radiative ages reaching $\sim 80$ Myr in the outer wings; in 4C32.25 and 4C61.23, 400 MHz plus archival broadband analysis again showed that wings are the oldest components, though the favored dynamical interpretation differed source by source (Bruno et al., 2019, Bruno et al., 2024, Retana-Montenegro, 3 Oct 2025).

A major caveat is that the wing–lobe spectral-index difference, $\Delta \alpha$, is not by itself a robust model discriminator. RMHD simulations of backflowing, shock-filled cocoons show that projection, spectral curvature, and localized re-acceleration can produce both positive and negative $\alpha_{144}^{1400} = -0.84 \pm 0.16$0, making single-band or two-band comparisons intrinsically degenerate (Giri et al., 2022).

4. Formation channels and their observational signatures

Hydrodynamic backflow in an anisotropic ambient medium remains the classical XRG model. In this picture, FR II jets terminate in hotspots, shocked plasma flows back into an over-pressured cocoon, and the cocoon plasma is diverted along the steepest pressure gradient, typically near the host minor axis. This framework naturally explains the predominance of FR II primaries, the diffuse character of wings, and the strong statistical tendency for wings to avoid the host major axis (Giri et al., 2023, Bera et al., 27 Apr 2026).

Jet reorientation models instead attribute the wings to relic plasma left behind after a substantial change in jet axis, whether through spin-flip, stochastic accretion, precession, or SMBH binary evolution. These models are particularly attractive for sources with very long or well-collimated wings, wings aligned along the host major axis, or clear evidence that the wing system predates the active lobes by tens of Myr. In MRC 2011-298, broadband radio spectral analysis found that the observed properties are consistent with an abrupt reorientation by about 90 degrees, while the S-shaped jets are plausibly driven by precession with $\alpha_{144}^{1400} = -0.84 \pm 0.16$1 Myr (Bruno et al., 2024).

Jet–shell interaction models occupy an intermediate position by linking XRG morphology to post-merger stellar and gaseous shells. In this scenario, jets propagating near the host major axis are temporarily deflected by rotating shells, producing wings without requiring a spin-axis change. A generalized review argued that the bulk of the evidence favored either jet–shell interaction or spin-flip hypotheses for most XRGs, while the FR I XRG in Abell 3670 was interpreted specifically in terms of jet–shell interaction because its wings are older than the lobes, its lobes align with the optical major axis, its wings with the minor axis, and the source lacks the FR II hotspots expected for strong backflow (Gopal-Krishna et al., 2010, Bruno et al., 2019).

Dual-AGN models have occupied a special place in the debate because they predict two contemporaneously powered jet pairs. Population-level spectral mapping now disfavors this as a general explanation: the rarity of flatter-spectrum wings is inconsistent with spectrally coeval lobe pairs in most systems (Patra et al., 2023). Yet dual or binary SMBHs are not irrelevant to XRGs. A spectroscopic study of 187 XRGs found a roughly 30% incidence of double-peaked [O III] narrow emission lines and concluded that about 95% of DPNEL XRGs with usable line diagnostics are dual-AGN candidates, while J0725+5835 was directly associated with an AGN pair separated by about 100 kpc and showed multiple signatures of jet re-orientation. The consistent inference is therefore not that dual SMBHs are unimportant, but that they likely characterize a subset rather than the dominant formation channel (Ghosh et al., 15 Sep 2025, Yang et al., 2022).

A further possibility is that some XRG-like structures emerge without mergers or strong ambient asymmetry at all. A 3D GRMHD simulation bridging Bondi to event-horizon scales found that intermittent, wobbling jets in the run-up to a magnetically arrested disk state can inflate cavity pairs in different directions and generate transient X-shaped morphologies. This suggests that at least some XRGs may arise from chaotic accretion-flow dynamics internal to the central engine (Lalakos et al., 2022).

5. Host galaxies, environments, and multiwavelength constraints

XRG hosts are predominantly massive ellipticals, and the optical-radio alignment statistics are among the most persistent empirical constraints on formation models. In a study of 106 strong XRG candidates, 41 sources had sufficiently reliable optical major-axis measurements for direct comparison with radio structure. About 76% had wing offsets greater than 30 degrees from the optical major axis, with median $\alpha_{144}^{1400} = -0.84 \pm 0.16$2, while the primary radio axis tended to lie closer to the optical major axis. Six counter-examples, however, had wings closer to the optical major axis, arguing against a single universal hydrodynamic picture (Joshi et al., 2019).

LoTSS DR2 extended this alignment analysis to a much larger sample. For 270 XRGs with reliable optical and wing position angles, 75% showed wing–optical major-axis offsets greater than 30 degrees, the median offset was about 54 degrees, and 17% were nearly perpendicular. At the same time, 25% had offsets below 30 degrees and about 8% were nearly aligned, again implying that a substantial minority of XRGs require something beyond simple minor-axis backflow diversion (Bera et al., 27 Apr 2026).

Host-galaxy properties reinforce the merger and gas-rich context of at least part of the population. In the 2019 optical study, XRGs had slightly lower mean black-hole masses than a control FR II sample and a much higher fraction of red mid-IR colors, with about 80% satisfying $\alpha_{144}^{1400} = -0.84 \pm 0.16$3. In the LoTSS DR2 multiwavelength follow-up, 94% of XRGs had $\alpha_{144}^{1400} = -0.84 \pm 0.16$4, XRGs were more radio-loud than ZRGs, and only 12% of XRGs were classified as LERGs by the adopted WISE criteria, compared with 36% of ZRGs. These results suggest that XRGs preferentially occupy massive elliptical hosts with substantial cold dust or cold-gas signatures and relatively powerful AGN activity (Joshi et al., 2019, Bera et al., 27 Apr 2026).

Large-scale environments are typically modest rather than rich. The optical-environment analysis of 2019 found that XRGs and FR II radio galaxies inhabit similarly poor environments within about 1 Mpc, with median richnesses 8.94 and 11.87, respectively. The LoTSS DR2 study measured local galaxy surface densities of 5.7 and 4.8 galaxies Mpc$\alpha_{144}^{1400} = -0.84 \pm 0.16$5 as the mean and median for XRGs, again indicating moderate-density environments rather than massive clusters as the default setting (Joshi et al., 2019, Bera et al., 27 Apr 2026).

X-ray diagnostics offer an independent view of the ambient medium. NGC 326 shows a likely shock front, hot rims, and a wing cavity in Chandra data, while numerical simulations predict that reorientation should generate four elongated X-ray cavities and multiple quasi-orthogonal shocks, whereas backflow should yield one dominant axial cavity pair plus weaker off-axis depressions. These X-ray observables are among the clearest proposed discriminants between formation scenarios, but deep data exist for only a small number of archetypes (Hodges-Kluck et al., 2011, Giri et al., 2023).

6. Numerical modeling, synthesis, and open problems

Recent simulations have shifted the field from qualitative cartoons to explicit multiwavelength predictions. A 3D RMHD comparison of backflow and jet-reorientation models, run in the same ellipsoidal atmosphere with PLUTO, showed that both mechanisms can generate bona fide XRGs. The predicted differences are quantitative: backflow produced $\alpha_{144}^{1400} = -0.84 \pm 0.16$6 in the reference run and broader, less collimated wings, while reorientation commonly produced $\alpha_{144}^{1400} = -0.84 \pm 0.16$7, more collimated structures, more isotropic ICM heating, bow shocks with $\alpha_{144}^{1400} = -0.84 \pm 0.16$8–3 in orthogonal directions, and four elongated X-ray cavities aligned with both wings and lobes (Giri et al., 2023).

Backflow-specific RMHD plus particle-transport simulations add a complementary emission-side perspective. In thermally dominated cocoons propagating through triaxial atmospheres, diverted backflows can form wings aligned with the ambient minor axis, while internal shocks re-energize electrons and generate localized flatter-spectrum patches. A central conclusion of that work is that projection and re-acceleration make $\alpha_{144}^{1400} = -0.84 \pm 0.16$9 highly degenerate, so spatially resolved, multi-frequency spectral maps are required for any serious test of wing ageing (Giri et al., 2022).

More speculative but important is the GRMHD result that intermittent jet wobble before the onset of a magnetically arrested disk can itself create transient X-shaped morphologies, with the rare XRG phase arising because the wobbling state occupies only a small fraction of the duty cycle. This mechanism does not eliminate merger- or environment-driven models, but it broadens the causal space and reinforces the conclusion that XRGs are not a single-physics population (Lalakos et al., 2022).

The resulting synthesis is increasingly clear. The majority of XRGs appear consistent with hydrodynamic backflow or related diversion processes in anisotropic halos, especially given the FR II dominance, the steepening from hotspot to wing in most resolved spectral maps, and the minor-axis preference of wings. A significant minority, however, requires jet-axis change, jet–shell interaction, or both; very long or major-axis wings, strong age differences, orthogonal cavity systems, and some direct dual-AGN detections are difficult to subsume under pure backflow. This suggests a plural taxonomy in which morphology records the interplay of jet power, jet stability, accretion history, ambient anisotropy, and, in some systems, SMBH multiplicity (Giri et al., 2024, Bera et al., 27 Apr 2026).

The most pressing observational agenda follows directly from these conclusions: broader frequency coverage to constrain spectral curvature and fit JP/KP/CI ageing models; high-resolution polarization and Faraday-rotation mapping to trace shocks, shear, and magnetic topology; deep X-ray imaging to count cavities and measure Mach numbers; optical and IFU work to quantify host triaxiality, shells, and merger debris; and VLBI follow-up of candidate dual or binary nuclei. As current LoTSS-, FIRST-, and SKA-pathfinder samples continue to expand, XRGs are increasingly serving not as exemplars of one mechanism, but as a controlled laboratory in which several distinct routes to large-scale radio-axis misalignment can be disentangled on a source-by-source basis (Patra et al., 2023, Giri et al., 2023, Ghosh et al., 15 Sep 2025).