Jets are presumed to obtain much of their energy from the infall of matter into a supermassive black hole. It is then natural to suppose that electromagnetic radiation would carry much of the energy from the system on the smallest scales, since a plausible mechanism for the extraction of energy is the twisting of magnetic field linked to the accretion disk (e.g., 138). Fast interactions with the plasma environment and efficient particle acceleration should load the field with matter. In the resulting magnetohydrodynamic flow much of the momentum would be carried by particles, although Poynting flux may carry a significant fraction of the total energy [163, 6].
Polarized radiation is the signpost to significant energy in relativistic particles and magnetic fields. Jet plasma must be neutral, on average, to remain collimated, but this can be achieved by various combinations of relativistic and cold electrons, positrons, and protons. Alternatively, it has been suggested that some of the energy is transported in a decaying neutral beam of ultra-high-energy neutrons and -rays .
Several quantities are available to help sort out the jet composition.
Radiation drag and observational constraints on Comptonized radiation by cold electrons and positrons seriously hamper electron-positron jets formed close to the central black hole [185, 186]. In the cores of some quasars the radiated power is too large to be met by that contained in a jet of magnetic field and relativistic leptons close to minimum energy, and observational constraints on Comptonized radiation limit the density of cold leptons, so that a significant proton component is required if the energy carrier is indeed particles . Thus, an electron-proton plasma is usually favoured when jets are discussed.
The presence of relativistic protons is supported for some FRI radio galaxies: the lobes, if assumed to be lepton-dominated and radiating at minimum energy (Section 2.2) would collapse under the pressure of the X-ray-emitting medium unless there is an additional pressure source and, although there are several ways of boosting the internal pressure in such a situation, magnetic dominance would make the sources unusual, electron dominance is unlikely from constraints on inverse-Compton scattering of the CMB, and non-relativistic protons are disfavoured on grounds of Faraday rotation, leaving a relativistic proton component most likely (e.g., 53). However, decreased filling factors cannot be ruled out (e.g., 65), except perhaps where the radio structure has excavated a clear cavity in the X-ray-emitting atmosphere (e.g., 24). If indeed the extra pressure is from relativistic protons, it is uncertain as to how much arises from entrained material accelerated in the shear layer of the decelerating jet as compared to particles transported from the core (see Section 4).
FRII jets transport more energy to larger distances, and thus have more need than FRI jets for a non-radiating energy carrier with high momentum transport. Relativistic hydrodynamic simulations find that the key parameter in preventing jets from strongly decelerating in an external boundary layer is density contrast with the external medium, in the sense that denser jets can propagate further . A more dominant relativistic proton content could provide this. Protons are also required if the low-frequency spectral turn-over in hotspots is the result of cyclotron resonant absorption (e.g., 91). On the other hand, pressure balance has been used to argue against relativistic protons in some FRII lobes. It is argued that the presence of relativistic protons is improbable since (a) the lobe magnetic field based on synchrotron and inverse Compton emission agrees with that from minimum energy calculated using relativistic leptons alone, and (b) that, even in the absence of such protons, the source is in pressure balance with the external medium (e.g., 10, 54). However, these calculations ignore possible dynamical effects in FRII lobes, and there are considerable additional sources of uncertainty (see Section 2.2 and item 4 above).
In the context of the beamed iC-CMB model for quasar jets (Section 3.2), it is possible to extend an argument limiting the density of cold electron-positron pairs  to kpc-scale regions . In the case of PKS 0637-752, upper limits on Comptonized CMB radiation from Spitzer are sufficiently low to place stringent limits on the mass flux carried by cold lepton pairs, with the implication that this jet is indeed made electrically neutral through a strong presence of protons . However, this argument relies on the beamed iC-CMB model being correct, with a large kinematic power being sustained throughout the jet (see Section 3.3).
Jet composition remains uncertain, and various degeneracies between physical quantities and observable parameters render it difficult to make watertight arguments. However, X-ray measurements continue to provide important clues to the puzzle.