8.1. An evolutionary cycle?
The Ledlow-Owen relation (Section 1.3) showed that a galaxy of a given optical luminosity can host either an FRI or FRII radio source. This resulted in renewed speculation in the 1990s that there may be an evolution between FRII and FRI activity controlled by external influences. Such speculation was supported by evidence of FRIIs associated with galaxy mergers (distorted isophotes and higher amounts of high-excitation ionized gas) and FRIs associated with galaxies in more relaxed dynamical states . Evolutionary ideas have also arisen from the so called `fundamental plane' that places AGN on an extension of the relationship between inner jet radio power, X-ray luminosity and black-hole mass found for X-ray binaries (XRBs) [146, 73]. It has been suggested that the changes in X-ray spectrum and jet luminosity that accompany changes in accretion characteristics in an XRB could apply to AGN, such that an individual object may go though transitions between an FRI and FRII, and indeed to becoming radio quiet (e.g., 123).
Observationally, kpc-scale jets accompany AGN with accretion flows that in the extreme are either geometrically-thick and radiatively inefficient or geometrically-thin and radiatively efficient, with the latter accompanied by high-excitation optical emission lines. It is possible that an AGN changes over the lifetime of a radio source, such that the observed kpc-scale radio structures are the result of ejection from an AGN evolving through different states. Some sort of intermittency of the central engine over timescales of ~ 104 - 106 years (shorter than the lifetime of radio sources, Section 1.4) gains support from observational and theoretical considerations (e.g., 164, 176, 111, 191). Multiple changes to the central structure over the lifetime of the radio source would be required to reconcile the claim that a geometrically-thick flow is needed to sustain a significant jet (with the most powerful requiring a spinning black hole)  with the observation that many AGN with powerful jets currently show geometrically-thin disks and high-excitation emission lines (see below).
Closer examination is needed of the extent to which the observed powers and structures of jets relate either to the accretion processes or to large-scale environmental effects. Both appear to play a rôle.
8.2. The rôle played by accretion processes
Broadly, powerful jets of FRII structure are associated with AGN showing high-excitation optical emission lines, while lower-power jets, normally but not always of FRI structure, are associated with AGN showing low-excitation lines. This suggests that the central engine has at least some influence on the power and large-scale structure of the jets (e.g., 9).
A correlation between the core radio emission and low-energy (~ 1 to 2 keV) nuclear X-ray output of radio galaxies has been known since the Einstein and ROSAT missions, and has been used to argue that the soft X-rays arise from pc-scale jets [70, 214, 37, 93]. An optical core is often seen with HST, and is interpreted as synchrotron emission from a similar small-scale emitting region [47, 94, 40, 49, 207]. Such pc-scales jets protrude from any gas and dust torus invoked by AGN unified models, and so this component should not be greatly affected by absorption, although relativistic effects will cause jet orientation to affect the level of X-ray flux observed.
Since jet emission dominates at low X-ray energies, it has been important to obtain sensitive spectral measurements that extend to the higher X-ray energies accessible to Chandra and XMM-Newton in order to probe the region closer to the SMBH and representative of the bolometric power of the central engine. At these energies any strong emission from the AGN should dominate jet emission even if it is largely absorbed at lower energies by a gas torus. Results find a number of radio galaxies showing clear evidence of a hard continuum, sometimes accompanied by Fe-line emission, and presumed to be emission associated with an accretion-disk corona (e.g., 204, 224, 90). Both the jet and central-engine X-ray components can sometimes be distinguished in the same spectrum (e.g., 54, 67, 225).
The hard component is more often detected in FRIIs than in FRIs. Of course, greater absorption from a torus could potentially combine with lower X-ray luminosity in causing the non-detection of the second component in most FRIs, and so particular reliability can be placed on the results of a study of nearby (z < 0.1) radio galaxies that has allowed for absorption in placing upper limits on the luminosity of undetected nuclear components . The radiative efficiency of the central engine was then found by correcting the X-ray luminosity to a bolometric luminosity and combining it with the inferred SMBH mass. In powerful FRIIs, radiatively-efficient accretion associated with a thin disk surrounded by an obscuring torus is normally inferred. FRII radio galaxies at z ~ 0.5 also show an absorbed X-ray component . In contrast, in z < 0.1 FRIs, all the nuclear X-ray emission can normally be interpreted as jet related, and usually only upper limits are found for accretion-related emission . Any X-ray luminosity associated with a non-jet central-engine component in low-power sources is normally sufficiently low to support earlier speculations based on the Ledlow-Owen relation that the physical difference between the two types of radio source arises from the different nature of their accretion disks and efficiency of accretion . Further support for these ideas comes from Spitzer results for the z < 0.1 sample  that show an additional component of hot dust only in FRIIs.
While results at first-look appear quite convincing of a connection between large-scale radio power and the structure of the central engine, there are sources which defy the trend. Both Cen A and NGC 4261 have large-scale FRI structures, and yet contain absorbed, hard, luminous X-ray components characteristic of the coronae of thin accretion disks seen through an obscuring torus [67, 225]. This might suggest that something relatively recent (perhaps the galaxy merger in the case of Cen A ) has provided additional material for accretion and affected the central engine in a way that has yet to be reflected in the power and structure of the large-scale radio emission. The difficulty is that merger and source-development time scales are expected to be comparable. A further complication is the tendency for any X-ray accretion-related components in FRII low-excitation radio galaxies to be less luminous than those seen in a typical FRII high-excitation radio galaxy , as was known for the optical continuum [49, 206]. This means that not all FRIIs have equivalent central engines. However, it is is hard to treat as a coincidence the tendency for the most powerful FRIIs with the least evidence for external disruption to arise from AGN showing high-excitation optical emission lines and evidence for thin accretion disks.
In the normally inferred absence of thin radiatively-efficient accretion disks in FRIs, it has been argued in several cases that sufficient X-ray-emitting hot gas is present in their galaxies and clusters to produce the required jet power through a geometrically-thick Bondi accretion flow (e.g., 62, 4). Here the jet power is inferred from the energy required to excavate the cavities observed in the X-ray-emitting gas, i.e., a more direct method than scaling from radio power (e.g., 211) as is normal in the absence of other information. Recent work confirms that the most luminous FRIIs also tend to lie in luminous X-ray clusters , and it is reasonable to assume that they experience similar or greater supplies of galaxy and cluster hot gas. However jet powers are also higher (how much so rests on uncertainties in speed and composition), consistent with requiring an extra energy source in the form of stars and gas clouds fuelling a thin accretion disk. A major outstanding problem is a full understanding of the mechanisms which convert gas infall into two different accretion structures. Jets are expected to be more strongly coupled to the structure of the host stellar system, and hence to play a more major rôle in feedback, if the accreting gas originates predominantly from the reservoir contained in the potential well of the system as a whole, whether it be hot (e.g., 4) or cold (e.g., 161) in origin.
8.3. The rôle of the environment
Assuming that jets are genuinely symmetric at production, the environment appears to be, at a minimum, a strong secondary factor (with jet power being the likely primary influence) in shaping large-scale jet structure. For example, some radio sources show what appears to be FRI morphology on one side and FRII on the other, and this has been used to argue for different environmental effects on the two sides .
VLBI proper-motion studies find few, if any, differences in the speed or morphology of FRI and FRII radio jets in their initial stages of development from the central engine [154, 88]. However, the radiative powers are higher in FRIIs, but not in linear proportion to their total radio powers (e.g., 87, 89), suggesting that on the small scale a radio source has knowledge of how it will evolve. Particularly compelling evidence that the environment does have some influence is the recent discovery that quasars, traditionally the hosts only of FRII structures, can host FRI radio structures, with evidence that denser, more clumpy, environments at higher redshift are allowing this to occur . The rôle of the X-ray-emitting environment in decelerating FRI jets was discussed in Section 4.
8.4. Information from beamed sources
The beamed counterparts of radio galaxies (quasars and BL Lac objects) do not allow the accretion structures to be probed in the X-ray, since the beamed jet emission swamps all other nuclear components; indeed it is sometimes dominant up to the TeV band. Multi-wavelength spectral energy distributions and variability time scales are used to probe the beaming parameters and the physical properties of the emitting regions (e.g., 84, 128, 193). Correlated flares are sometimes measured across wavebands, giving support to the presence of a dominant spatial region of emission (e.g., 205, 194), but otherwise uncertainties of size scales, geometries, and parameters for the competing processes of energy loss and acceleration often force the adoption of oversimplified or poorly-constrained models for individual jets. Much is published on the topic, and a review is beyond the scope of this work. Substantial progress in understanding is anticipated from multiwavelength programmes associated with the Fermi Gamma-ray Space Telescope.
VLBI radio-polarization studies have found systematic differences between powerful quasars (beamed FRIIs) and BL Lac objects (beamed FRIs) in core polarizations, the orientations of the magnetic fields in the inner jets, and in jet length, although it is difficult to separate intrinsic differences from the possible influence of the parsec-scale environment, such as the density and magnetic field contained in line-emitting gas .