The discovery that the Galactic SGR 1935$+$2154 emitted FRB 200428 simultaneous with a gamma-ray flare demonstrated the common source and association of these phenomena. If FRB radio emission is the result of coherent curvature radiation, the net charge of the radiating "bunches" or waves may be estimated. A statistical argument indicates that the radiating bunches must have a Lorentz factor $\gtrapprox 10$. The observed radiation frequencies indicate that their phase velocity (pattern speed) corresponds to Lorentz factors $\gtrapprox 100$. Coulomb repulsion implies that the electrons making up these bunches may have yet larger Lorentz factors, limited by their incoherent curvature radiation. These electrons also Compton scatter in the soft gamma-ray field of the SGR. In FRB 200428 the power radiated coherently at radio frequencies exceeded that of Compton scattering, but in more luminous SGR outbursts Compton scattering dominates, precluding the acceleration of energetic electrons. This explains the absence of a FRB associated with the giant 27 December 2004 outburst of SGR 1806$-$20. SGR with luminosity $\gtrsim 10^{42}$ ergs/s do not emit FRB, while those of lesser luminosity can do so.

Flyby images of (486958) Arrokoth (Ultima Thule, 2014 MU$_{69}$) show a comparatively bright "necklace" in the neck, or cleft between its two lobes, in contrast to its generally low albedo. The necklace is found in the most (orbitally averaged) shaded part of its surface. It may consist of clean, high albedo, ice condensed from vapor sublimed by dirty, low albedo, ice elsewhere; clean ice accumulates where the maximum temperatures are the lowest. Ammonia and propane have the necessary vapor pressure. Surrounding gas in the proto-Solar System would facilitate redeposition of molecules sublimed by warmer parts of the surface into the cleft, as well as smoothing the surface and explaining, by hydrodynamic drag, Arrokoth's slow (compared to its breakup rate) rotation.

Does the Galaxy contain sources of micro-FRB, lower energy events resembling the known FRB but detectable only at Galactic distances? The answer to this question is essential to determining the nature of FRB sources. At typical ($10$ kpc) Galactic distances a burst would be about 117 dB brighter than at a "cosmological" ($z = 1$) distance. The radiation of Galactic micro-FRB, if they exist, could be detected after Lunar reflection, or an upper bound on their rate set, by a modest (20 m at 1.4 GHz) radio telescope staring at the Moon. It would have all-sky sensitivity. The delay between detection of direct (by STARE2 or dipoles) and Lunar-scattered radiation would restrict a burst's position to a narrow arc.

Identifying the sources of the highest energy cosmic rays requires understanding how they are deflected by the stochastic, spatially intermittent intergalactic magnetic field. Here we report measurements of energetic charged-particle propagation through a laser-produced magnetized plasma with these properties. We characterize the diffusive transport of the particles experimentally. The results show that the transport is diffusive and that, for the regime of interest for the highest-energy cosmic rays, the diffusion coefficient is unaffected by the spatial intermittency of the magnetic field.

Popular models of repeating Fast Radio Bursts (and perhaps of all Fast Radio Bursts) involve neutron stars because of their high rotational or magnetostatic energy densities. These models take one of two forms: giant but rare pulsar-like pulses like those of Rotating RAdio Transients and outbursts like those of Soft Gamma Repeaters. Here I collate the evidence, recently strengthened, against these models, including the absence of Galactic micro-FRB, and attribute the 16 day periodicity of FRB 180916.J0158+65 to the precession of a jet produced by a massive black hole's accretion disc.

Repeating and apparently non-repeating fast radio bursts (FRB) differ by orders of magnitude in duty factors, energy and rotation measure. Extensive monitoring of apparently non-repeating FRB has failed to find any repetitions. This suggests the two types differ qualitatively, rather than in repetition rate, and are produced by distinct kinds of sources. The absence of periodicity in repeating FRB argues that they are not produced by neutron stars. They may originate in dilute relativistic jets produced by low luminosity black hole accretion. Non-repeating FRB may be produced by catastrophic events such as the collapse of an accreting magnetic neutron star to a black hole or of an accreting magnetic white dwarf to a neutron star, during which a disappearing magnetic moment radiates dipole radiation that accelerates electrons in nearby matter. If they are emitted by collimated beams of relativistic particles or charge "bunches" with Lorentz factor $\gamma$, their radiation is collimated into a solid angle $\sim \gamma^{-2}$ sterad, reducing the energy requirement. If powered by magnetic reconnection, a pulse of length $\Delta t$ may draw on the magnetic energy in a volume $\sim \gamma^4 (c \Delta t)^3$, although only a fraction $\sim 1/\gamma^2$ of this may be dissipated without decollimation.

Micheli, et al. (2018) reported that a seven-parameter fit to the orbit of 1I/2017 U1 `Oumuamua indicated a non-gravitational acceleration in the anti-Solar direction, and attributed it to recoil from comet-like outgassing. The implied gas to dust ratio is at least 100 times greater than that of known Solar System comets. The reported collapse of the scatter of nearly contemporaneous coordinate residuals upon inclusion of the non-gravitational term in the orbital fits is difficult to understand. There are grounds for skepticism.

Bodies in the outer Solar System may consist of, or be covered by, solid ammonia. Their temperature depends on its emissivity. We calculate that emissivity as a function of frequency, angle and polarization and convolve it with a Planck function to obtain its integrated emissivity.

Curvature radiation is a natural candidate for the emission mechanism of FRB. However, FRB spectra have structure with $\Delta \nu/\nu \sim \text{0.03--0.2}$, inconsistent with the very smooth spectrum of curvature radiation. Although this spectral structure might be attributed to chromatic scintillation or lensing, in four FRB high spectral resolution data indicate scintillation decorrelation bandwidths much narrower than the observed $\sim \text{30--300}$ MHz spectral structure. Some of the observed structure may be intrinsic to the radiation mechanism. I suggest that the observed spectral structure reflects the spatial structure of a clumpy radiating charge distribution, and that the characteristic curvature radiation frequency may be much higher than the observed frequencies. In this coupled plasma-curvature radiation process the radiated spectra are the product of the spectra of the plasma wave and that of incoherent curvature radiation. The argument applies to all coherent radiation processes, including those that produce pulsar nanoshots. The implied FRB "clump" charges are large, and produce electrostatic potentials that suggest electron Lorentz factors $\gtrsim 10^2$. The result applies generally to coherently radiating sources.

More than a decade after their discovery, astronomical Fast Radio Bursts remain enigmatic. They are known to occur at "cosmological" distances, implying large energy and radiated power, extraordinarily high brightness and coherent emission. Yet their source objects, the means by which energy is released and their radiation processes remain unknown. This review is organized around these unanswered questions.

The recently discovered first interstellar object 1I/2017 U1 (`Oumuamua) has brightness that varies by a factor of 10, a range greater than that of any Solar System asteroid, a spectrum characteristic of Type D asteroids, and no evidence of evaporating volatiles, contrary to expectation for exo-Oort clouds. This object was the first example of the proposed "Jurads", objects depleted in volatiles and ejected from planetary systems during the post-main sequence evolution of their parent star. I suggest that heating by the star's giant stage fluidized a precursor cometary object as volatiles escaped, causing it to assume the Jacobi ellipsoidal shape of a self-gravitating incompressible liquid. The collision that produced the inferred tumbling motion must have occurred thousands of years after the formation of 1I/2017 U1 `Oumuamua. Jacobi ellipsoids have a unique relation among rotation rate, density and axial ratio. The inferred axial ratio $\gtrapprox 5$ implies a lower bound on the density of 1.6 g/cm$^3$, excluding an icy interior unless it is almost entirely frozen CO$_2$. This object is the first Jurad to be discovered and may be related to accreting objects that pollute white dwarf atmospheres and may make Soft Gamma Repeaters.

The repeating FRB 121102 emitted a pair of discrete bursts separated by 37 ms and another pair, 131 days later, separated by 34 ms, during observations that detected FRB at a mean rate of $2 \times 10^{-4}$/s. Here I assume that these events are separate bursts rather than multiple peaks from longer single bursts and consider their implications. They then are inconsistent with Poissonian statistics. The measured intervals constrain any possible periodic modulation underlying the highly episodic emission. If more such short intervals are measured a period may be determined or periodicity may be excluded. Sources of narrow wandering beams predict pairs (or higher multiples) of bursts closely spaced in time compared to the mean event rate.

The distribution of FRB fluxes and fluences is characterized by a few very bright events and a deficiency of fainter events compared to expectations for a homogeneous space-filling distribution. I define a metric to quantify this, and apply it to the 17 presently known Parkes FRB, products of a comparatively homogeneous search. With 98\% confidence we reject the hypothesis of a homogeneous distribution in Euclidean space. Possible explanations include a reduction of fainter events by cosmological redshifts or evolution or a cosmologically local concentration of events. The former is opposed by the small value of the one known FRB redshift. The latter contradicts the Cosmological Principle, but may be explained if the brighter FRB originate in the Local Supercluster.

Could the dips of "Boyajian's Star" (KIC 8462852) have been caused by matter in our Solar System? The interval between periods of deep dips is nearly twice the orbital period of the Kepler satellite. I consider heliocentric obscuring rings in the outer Solar System that graze the line of sight to the star once per orbit of Kepler. The hypothesis predicts that future dips may be observed from Earth during windows separated by a year, although their detailed structure depends on the distribution of particles along the ring. Dips observed at terrestrial sites separated by > 1000 km in a direction perpendicular to the Earth's orbital motion may be uncorrelated.

The repeating FRB 121102, the only FRB with an accurately determined position, is associated with a variable persistent radio source consistent with a low luminosity active galactic nucleus. I suggest that FRB originate in the accretion disc funnels of intermediate mass black holes. Narrowly collimated radiation is emitted along the wandering instantaneous angular momentum axis of accreted matter. We observe this emission as a fast radio burst when it sweeps across the direction to the observer. This model constrains the mass of the black hole to values below those of galactic nuclei. It predicts, in contrast to neutron star (pulsar or SGR) models, that repeating FRB will not be periodic and will be co-located with persistent but variable radio sources resulting from the off-axis emission. The model is analogous, on smaller spatial, lower mass and accretion rate and shorter temporal scales, to AGN making double radio sources, with FRB corresponding to blazars in which the jets point toward us.

Are some neutron stars produced without a supernova, without ejecting mass in a remnant? Theoretical calculations of core collapse in massive stars often predict this. The observation of the repeating FRB 121102, whose dispersion measure has not changed over several years, suggests that dark core collapses are not just failures of computer codes, but may be real. The existence of one repeating FRB with unchanging dispersion measure is not conclusive, but within a decade hundreds or thousands of FRB are expected to be discovered, likely including scores of repeaters, permitting useful statistical inferences. A na\"{\i}ve supernova remnant model predicts observable decline in dispersion measure for 100 years after its formation. If an upper limit on the decline of 2 pc/cm$^3$-y is set for five repeating FRB, then the na\"{\i}ve model with nominal parameters is rejected at the 95\% level of confidence. This may indicate dark neutron star formation without a supernova or supernova remnant. This hypothesis may also be tested with LSST data that would show, if present, a supernova at an interferometric FRB position if it occurred within the LSST epoch.

There are striking phenomenological similarities between Fast Radio Bursts and lightning in the Earth's and planetary atmospheres. Both have very low duty factors, $\lesssim 10^{-8}$--$10^{-5}$ for FRB and (very roughly) $\sim 10^{-4}$ for the main return strokes in an active thundercloud. Lightning occurs in an electrified insulating atmosphere when a conducting path is created by and permits current flow. FRB may occur in neutron star magnetospheres whose plasma is believed to be divided by vacuum gaps. Vacuum is a perfect insulator unless electric fields are sufficient for electron-positron pair production by curvature radiation, a high-energy analogue of electrostatic breakdown in an insulating gas. FRB may be "electrars" powered by the release of stored electrostatic energy, counterparts to Soft Gamma Repeaters powered by the release of stored magnetostatic energy (magnetars). This frees pulsar FRB models from the constraint that their power not exceed the instantaneous spin-down power. Energetic constraints imply that the sources of more energetic FRB have shorter spin-down lifetimes, perhaps shorter than the three years over which FRB 121102 has been observed to repeat.

Magnetic fields are ubiquitous in the Universe. Extragalactic disks, halos and clusters have consistently been shown, via diffuse radio-synchrotron emission and Faraday rotation measurements, to exhibit magnetic field strengths ranging from a few nG to tens of $\mu$G. The energy density of these fields is typically comparable to the energy density of the fluid motions of the plasma in which they are embedded, making magnetic fields essential players in the dynamics of the luminous matter. The standard theoretical model for the origin of these strong magnetic fields is through the amplification of tiny seed fields via turbulent dynamo to the level consistent with current observations. Here we demonstrate, using laser-produced colliding plasma flows, that turbulence is indeed capable of rapidly amplifying seed fields to near equipartition with the turbulent fluid motions. These results support the notion that turbulent dynamo is a viable mechanism responsible for the observed present-day magnetization of the Universe.

The universe is permeated by magnetic fields, with strengths ranging from a femtogauss in the voids between the filaments of galaxy clusters to several teragauss in black holes and neutron stars. The standard model behind cosmological magnetic fields is the nonlinear amplification of seed fields via turbulent dynamo to the values observed. We have conceived experiments that aim to demonstrate and study the turbulent dynamo mechanism in the laboratory. Here we describe the design of these experiments through simulation campaigns using FLASH, a highly capable radiation magnetohydrodynamics code that we have developed, and large-scale three-dimensional simulations on the Mira supercomputer at Argonne National Laboratory. The simulation results indicate that the experimental platform may be capable of reaching a turbulent plasma state and study dynamo amplification. We validate and compare our numerical results with a small subset of experimental data using synthetic diagnostics.

The emission of the white dwarf-M dwarf binary AR Sco is driven by the rapid synchronization of its white dwarf, rather than by accretion. This requires a comparatively large magnetic field $\sim 100$ gauss at the M dwarf and $\sim 10^8$ gauss on the white dwarf, larger than the fields of most intermediate polars but within the range of fields of known magnetic white dwarfs. The spindown power is dissipated in the atmosphere of the M dwarf by magnetic reconnection, accelerating particles that produce the observed synchrotron radiation. The displacement of the optical maximum from conjunction may be explained either by dissipation in a bow wave as the white dwarf's magnetic field sweeps past the M dwarf or by a misaligned white dwarf's rotation axis and oblique magnetic moment. In the latter case the rotation axis precesses with a period of decades, predicting a drift in the orbital phase of maximum. Binaries whose emission is powered by synchronization may be termed synchronars, in analogy to magnetars.

It is generally assumed that the sources of Fast Radio Bursts radiate roughly isotropically, so that the observed low duty cycle of any individual source indicates a similar low duty cycle of its radio-frequency emission. An alternative hypothesis is that the radiative duty cycle is ${\cal O}(1)$, but that the radiation is emitted in a beam with a solid angle comparable to the observed duty cycle, whose direction wanders or sweeps across the sky. This hypothesis relaxes the extreme power demands of isotropically radiating models of FRB at the price of multiplying the number of sources. The constraints on pulsar models are relaxed; rather than being unprecedentedly fast-spinning and highly-magnetized with short spin-down times, their parameters may be closer to those of typical radio pulsars. The nanoshots of Galactic pulsars pose analogous energetic problems that may be resolved if their emission also is beamed, suggesting a relation between these phenomena, despite their vastly different energy scales.

There are several phenomenological similarities between Soft Gamma Repeaters and Fast Radio Bursts, including duty factors, time scales and probable repetition. The sudden release of magnetic energy in a neutron star magnetosphere, as in popular models of SGR, can meet the energy requirements of FRB but requires both the presence of magnetospheric plasma in order that dissipation occur in a transparent region and a mechanism for releasing much of that energy quickly. FRB sources and SGR are distinguished by long-lived (up to thousands of years) current-carrying coronal arches remaining from formation of the young neutron star and their decay ends the phase of SGR/AXP/FRB activity even though "magnetar" fields may persist. Runaway increase in resistance when the current density exceeds a threshold releases magnetostatic energy in a sudden burst and produces high brightness GHz emission of FRB by a coherent process; SGR are produced when released energy thermalizes as an equilibrium pair plasma. Failures of some alternative FRB models and the non-detection of SGR~1806-20 at radio frequencies are discussed in appendices.

Fast Radio Bursts are millisecond bursts of radio radiation at frequencies of about 1 GHz, recently discovered in pulsar surveys. They have not yet been definitively identified with any other astronomical object or phenomenon. The bursts are strongly dispersed, indicating passage through a high column density of low density plasma. The most economical interpretation is that this is the interglactic medium, indicating that FRB are at "cosmological" distances with redshifts in the range 0.3--1.3. Their inferred brightness temperatures are as high as $10^{37\,\circ}$K, implying coherent emission by "bunched" charges, as in radio pulsars. I review the astronomical sites, objects and emission processes that have been proposed as the origin of FRB, with particular attention to Soft Gamma Repeaters and giant pulsar pulses.

The widths, dispersion measures, dispersion indices and fluences of Fast Radio Bursts (FRB) impose coupled constraints that all models must satisfy. Observation of dispersion indices close to their low density limit of $-2$ sets a model-independent upper bound on the electron density and a lower bound on the size of any dispersive plasma cloud. The non-monotonic dependence of burst widths (after deconvolution of instrumental effects) on dispersion measure excludes the intergalactic medium as the location of scattering that broadens the FRB in time. Temporal broadening far greater than that of pulsars at similar high Galactic latitudes implies that scattering occurs close to the sources, where high densities and strong turbulence are plausible. FRB energetics are consistent with supergiant pulses from young, fast, high-field pulsars at cosmological distances. The distribution of FRB dispersion measures is inconsistent with expanding clouds (such as SNR). It excludes space-limited distributions (such as the local supercluster), but agrees with a homogeneous cosmological distribution with intergalactic dispersion. The FRB $\log{N}$--$\log{S}$ relation also indicates a cosmological distribution, aside from the anomalously bright Lorimer burst.

The pulse widths, dispersion measures and dispersion indices of Fast Radio Bursts (FRB) impose coupled constraints that all models must satisfy. We show that if the dispersion measures resulted from propagation through the intergalactic medium at cosmological distances and the pulse widths were a consequence of scattering by single thin screens, then the screens' electron densities were $\gtrsim 20$/cm$^3$, $10^8$ times the mean intergalactic density. This problem is resolved if the radiation scattered close to its source, where high densities are possible. Observation of dispersion indices close to their low density limit of $-2$ sets a model-independent upper bound on the electron density and a lower bound on the size of the dispersive plasma cloud, excluding terrestrial or Solar System origin. The scattering and much of the dispersion measures may be attributed to scattering regions about 1 AU from the sources, with electron densities $\sim 3 \times 10^8$/cm$^3$. The inferred parameters are only marginally consistent; re-examination of the assumed relation between dispersion measure and distance is warranted. Origin in an ionized starburst or protogalaxy is suggested, but statistical arguments exclude compact young SNR in the Galactic neighborhood. An appendix applies these arguments to PSR J1745-2900 at the Galactic Center. We suggest that its pulse width and angular broadening may be reconciled if we are near a caustic or focal point produced by refraction, rather than by the classic thin sheet scattering model.

The fast (ms) radio bursts reported by Thornton, {\it et al.} have extremely high brightness temperatures if at the inferred cosmological distances. This implies coherent emission by "bunches" of charges. We model the emission region as a screen of dipole radiators resonant at the frequency of observation and calculate the necessary charge bunching. From this we infer the minimum electron energy required to overcome electrostatic repulsion. The FRB, like the giant pulses of the Crab pulsar, display banded spectra that may be harmonics of plasma frequency emission by plasma turbulence. If FRB are the counterparts of comparatively frequent Galactic events, these may be detected by radio telescopes in their far sidelobes or by small arrays of dipoles.

Chirped radio pulses known as perytons are consistent with plasma propagation, but an astronomical origin has been excluded. TE and TM waveguide modes have dispersion relations formally similar to those of plasmas, and emission at the plasma frequency of a plasma recombining by two-particle processes also produces a chirp $\delta t \propto \omega^{-2}$, but these hypotheses fail quantitatively. The chirp might be approximately fitted to three-body recombination, cyclotron radiation in exponentially decaying fields, or to the resonance of a distributed LC circuit whose component conductors move, but no specific source has been identified.

Large contributions to the near closure of the Universe and to the acceleration of its expansion are due to the gravitation of components of the stress-energy tensor other than its mass density. To familiarise astronomers with the gravitation of these components we conduct thought experiments on gravity, analogous to the real experiments that our forebears conducted on electricity. By analogy to the forces due to electric currents we investigate the gravitational forces due to the flows of momentum, angular momentum, and energy along a cylinder. Under tension the gravity of the cylinder decreases but the 'closure' of the 3-space around it increases. When the cylinder carries a torque the flow of angular momentum along it leads to a novel helical interpretation of Levi-Civita's external metric and a novel relativistic effect. Energy currents give gravomagnetic effects in which parallel currents repel and antiparallel currents attract, though such effects must be added to those of static gravity. The gravity of beams of light give striking illustrations of these effects and a re-derivation of light bending via the gravity of the light itself. Faraday's experiments lead us to discuss lines of force of both gravomagnetic and gravity fields. A serious conundrum arrises if Landau and Lifshitz's definition of Gravitational force is adopted.

The recently observed antiglitch of AXP 1E 2259+586 is inconsistent with magnetar models, but may be explained as the consequence of sudden accretion of retrograde matter or "propeller" interaction with surrounding matter. AXP/SGR are explained as single neutron stars accompanied by fallback matter from their natal supernovae. Their phenomenology may be partly accretional and partly the result of dissipation of magnetic energy.

We study gravitational waves to first and second order in amplitude in vacuum asymptotically flat spacetimes. The Einstein equations are solved to first order and these solutions are superposed to form a time-symmetric ingoing and then outgoing pulse regular everywhere. The waves are assumed to have odd-parity and a non-vanishing angular momentum which keeps them away from the axis at all times. The averaged energy of the waves is evaluated. The relevant Einstein equation is then solved to second order in the amplitude. The influence of the angular momentum of the waves on the rotation of local inertial frames with respect to the frames at great distances is analyzed. The rotation of the frames occurs even in the region around the origin where spacetime is almost flat. The rotation is without time delay as it follows from the constraint equation. The results are illustrated graphically for various values of the "harmonic index m" corresponding to azimuthal rotation and the "harmonic index l" describing the latitudinal rotation of the waves. The apparent motions of the fixed stars on the celestial sphere as seen through rotating waves from the local inertial frame are calculated and displayed.

A new formula is given for the fast linear gravitational dragging of the inertial frame within a rapidly accelerated spherical shell of deep potential. The shell is charged and is electrically accelerated by an electric field whose sources are included in the solution.

I consider how physical processes scale over eight orders of magnitude in black hole mass, from stellar masses in gamma-ray bursts (GRB) and black-hole X-ray binaries (BHXRB) to supermassive active galactic nuclei (AGN). Accretion rates onto stellar mass black holes range over more than sixteen orders of magnitude, from the lower luminosity BHXRB to GRB. These enormous parameter ranges correspond to qualitative as well as quantitative differences in behavior. The fundamental questions involve the balance between nonequilibrium and thermalized plasmas. When energy fluxes exceed a critical value $\sim 10^{29}$ erg/cm$^2$s, as in GRB, a black-body equilibrium pair plasma forms. At the lower fluxes found in AGN, BHXRB and microquasars, accretion power electrodynamically accelerates a small number of very energetic particles, explaining their non-thermal spectra and the high energy gamma-ray emission of blazars. Ultra-high energy cosmic rays may be accelerated by massive black holes, otherwise undetectable, with very low thermal luminosities. New-born fast high-field pulsars may be in the black-body equilibrium regime, resembling SGR in permanent outburst. I also consider the question, significant for the acceleration of nonthermal particles in GRB outflows, of whether collisionless plasmas interpenetrate rather than forming hydrodynamic shocks, and propose this as an alternative to internal shock models of GRB.

Gamma-ray bursts (GRBs) have been an enigma since their discovery forty years ago. However, considerable progress unraveling their mysteries has been made in recent years. Developments in observations, theory, and instrumentation have prepared the way so that the next decade can be the one in which we finally answer the question, "What are gamma-ray bursts?" This question encompasses not only what the progenitors are that produce the GRBs, but also how the enormous luminosity of the GRBs, concentrated in gamma rays, is achieved. Observations across the electromagnetic spectrum, from both the ground and space, will be required to fully tackle this important question. This white paper, mostly distilled from a recent study commissioned by the Division of Astrophysics of the American Physical Society, focuses on what very high energy (~100 GeV and above) gamma-ray observations can contribute. Very high energy gamma rays probe the most extreme high energy particle populations in the burst environment, testing models of lepton and proton acceleration in GRBs and constraining the bulk Lorentz factor and opacity of the outflow. Sensitivity improvements of more than an order of magnitude in the very high energy gamma-ray band can be achieved early in the next decade, in order to contribute to this science.

This is a report on the findings of the extragalactic science working group for the white paper on the status and future of TeV gamma-ray astronomy. The white paper was commissioned by the American Physical Society, and the full white paper can be found on astro-ph (arXiv:0810.0444). This detailed section discusses extragalactic science topics including active galactic nuclei, cosmic ray acceleration in galaxies, galaxy clusters and large scale structure formation shocks, and the study of the extragalactic infrared and optical background radiation. The scientific potential of ground based gamma-ray observations of Gamma-Ray Bursts and dark matter annihilation radiation is covered in other sections of the white paper.

This is a report on the findings of the gamma ray burst working group for the white paper on the status and future of TeV gamma-ray astronomy. The white paper is an APS commissioned document, and the overall version has also been released and can be found on astro-ph. This detailed section of the white paper discusses the status of past and current attempts to observe gamma ray bursts at GeV-TeV energies. We concentrate on the potential of future ground-based gamma-ray experiments to observe the highest energy emission ever recorded for GRBs, particularly for those that are nearby and have high Lorentz factors in the GRB jet. It is clear that major advances are possible and that the detection of very high energy emission would have strong implications for GRB models, as well as cosmic ray origin.

This is a short report on the preliminary findings of the gamma ray burst (GRB) working group for the white paper on the status and future of very high energy (VHE; >50 GeV) gamma-ray astronomy. The white paper discusses the status of past and current attempts to observe GRBs at GeV-TeV energies, including a handful of low-significance, possible detections. The white paper concentrates on the potential of future ground-based gamma-ray experiments to observe the highest energy emission ever recorded for GRBs, particularly for those that are nearby and have high Lorentz factors in the GRB jet. It is clear that the detection of VHE emission would have strong implications for GRB models, as well as cosmic ray origin. In particular, the extended emission phase (including both afterglow emission and possible flaring) of nearby long GRBs could provide the best possibility for detection. The difficult-to-obtain observations during the prompt phase of nearby long GRBs and short GRBs could also provide particularly strong constraints on the opacity and bulk Lorentz factors surrounding the acceleration site. The synergy with upcoming and existing observatories will, of course, be critical for both identification of GRBs and for multiwavelength/multimessenger studies.

Linear perturbations of Friedmann-Robertson-Walker universes with any curvature and cosmological constant are studied in a general gauge without decomposition into harmonics. Desirable gauges are selected as those which embody best Mach's principle: in these gauges local inertial frames can be determined instantaneously via the perturbed Einstein field equations from the distributions of energy and momentum in the universe. The inertial frames are identified by their `accelerations and rotations' with respect to the cosmological frames associated with the `Machian gauges'. In closed spherical universes, integral gauge conditions are imposed to eliminate motions generated by the conformal Killing vectors. The meaning of Traschen's integral constraint vectors is thus elucidated. For all three types of FRW universes the Machian gauges admit much less residual freedom than the synchronous or generalized harmonic gauge. Mach's principle is best exhibited in the Machian gauges in closed spherical universes. Independent of any Machian motivation the general perturbation equations and discussion of gauges are useful for cosmological perturbation theory.

Explicit expressions are found for the axisymmetric metric perturbations of the closed, flat and open FRW universes caused by toroidal motions of the cosmic fluid. The perturbations are decomposed in vector spherical harmonics on 2-spheres, but the radial dependence is left general. Solutions for general odd-parity $l$-pole perturbations are given for either angular velocities or angular momenta prescribed. In particular, in case of closed universes the solutions require a special treatment of the Legendre equation.

If perturbations beyond the horizon have the velocities prescribed everywhere then the dragging of inertial frames near the origin is suppressed by an exponential factor. However if perturbations are prescribed in terms of their angular momenta there is no such suppression. We resolve this paradox and in doing so give new explicit results on the dragging of inertial frames in closed, flat and open universe with and without a cosmological constant.

This work assembles some basic theoretical elements on thermal equilibrium, stability conditions, and fluctuation theory in self-gravitating systems illustrated with a few examples. Thermodynamics deals with states that have settled down after sufficient time has gone by. Time dependent phenomena are beyond the scope of this paper. While thermodynamics is firmly rooted in statistical physics, equilibrium configurations, stability criteria and the destabilizing effect of fluctuations are all expressed in terms of thermodynamic functions. The work is not a review paper but a pedagogical introduction which may interest theoreticians in astronomy and astrophysicists. It contains sufficient mathematical details for the reader to redo all calculations. References are only to seminal works or readable reviews. (Except in the ultimate sub-section.) Delicate mathematical problems are mentioned but are not discussed in detail.

We consider the four dimensional discontinuity generated by two identical pieces of a five dimensional space pasted along their edge (that is a "brane" in a "$Z_2$-symmetric" "bulk"). Using a four plus one decomposition of the Riemann tensor we write the equations for gravity on the brane and recover in a simple manner a number of known "brane world" scenarios. We study under which conditions these equations reduce, exactly or approximately, to the four dimensional Einstein equations. We conclude that if the bulk is imposed to be only an Einstein space near the brane, Einstein's equations can be recovered approximately on the brane, but if it is imposed to be strictly anti-de Sitter space then the Einstein equations cannot hold, even approximately, on a quasi-Minkowskian brane, unless matter obeys a very contrived equation of state.

We consider cosmological models where the universe, governed by Einstein's equations, is a piece of a five dimensional double-sided anti-de Sitter spacetime (that is, a "$Z_2$-symmetric bulk") with matter confined to its four dimensional Robertson-Walker boundary or "brane". We study the perturbations of such models. We use conformally minkowskian coordinates to disentangle the contributions of the bulk gravitons and of the motion of the brane. We find the restrictions put on the bulk gravitons when matter on the brane is taken to be a scalar field and we solve in that case the brane perturbation equations.

Isolated isothermal spheres of N gravitationally interacting points with equal mass are believed to be stable when density contrasts do not exceed 709. That stability limit does, however, not take into consideration fluctuations of temperature near the onset of instability. These are important when N is finite. Here we correlate {\it global mean quadratic temperature fluctuations} with onset of instability. We show that such fluctuations trigger instability when the density contrast reaches a value near $709\cdot\exp(-3.3N^{-1/3})$. These lower values of limiting density contrasts are significantly smaller than 709 when N is not very big and this suggests (i) that numerical calculations with small N may not reflect correctly the onset of core collapse in clusters with big N and (ii) that a greater number of globular clusters than is normally believed may already be in an advanced stage of core collapse because most of observed globular clusters whose parameters fit quasi-isothermal configurations are close to marginal stability.

Gamma-ray bursts, if they are generated in the process of interaction between relativistic strongly magnetized winds and an ambient medium, may be accompanied by very short pulses of low-frequency radio emission. The bulk of this emission is expected to be at the frequencies of $\sim 0.1-1$ MHz and cannot be observed. However, the high-frequency tail of the low-frequency radio emission may reach a few ten MHz and be detected, especially if the strength of the magnetic field of the wind is extremely high.

When gamma-rays emerge from a central source they may undergo Compton scattering in surrounding matter. The resulting Compton-scattered electrons radiate. Coherent radiation by such Compton electrons follows nuclear explosions above the Earth's atmosphere. Particle acceleration in instabilities produced by Compton electron currents may explain the radio emission by SN1998bw. Bounds on coherent radiation are suggested for supernovae and gamma-ray bursts; these bounds are very high, but it is unknown if coherent radiation occurs in these objects.

I consider radio emission from the remarkable SN1998bw. Ni-56 and Co-56 decays produce a gamma-ray flux whose Compton-scattered electrons naturally explain the observed mildly relativistic exapnsion of the radio source and its double- peaked history. Such models require a surrounding plasma, perhaps produced by the supernova progenitor, whose interaction with the nonrelativistic debris may account for the observed X-ray source. The radio spectrum appears to be self- absorbed. This interpretation determines the brightness temperature, and hence the energy of the radiating electrons, implying a surprisingly large magnetic field. Attempts to avoid this conclusion by interpreting the spectrum as the result of inverse bremsstrahlung absorption do not lead to significantly lower fields. The large inferred field may have several explanations: radiation from a central pulsar, a turbulent hydrodynamic dynamo or an aspherical Compton current, but a frozen-in field from the supernova progenitor is not adequate. The electron-ion and particle-field equipartition problems are discussed. Compton electrons also explain the inferred expansion speed of SN1987A's spots.

When does a GRB stop and its afterglow begin? A GRB may be defined as emission by internal shocks and its afterglow as emission by an external shock, but it is necessary to distinguish them observationally. With these definitions irregularly varying emission (at any frequency) must be the GRB, but smoothly varying intensity is usually afterglow. The GRB itself and its afterglow may overlap in time and in frequency, and distinguishing them will, in general, require detailed modeling.

When a charged insulating spherical shell is uniformly accelerated, an oppositely directed electric field is produced inside. Outside the field is the Born field of a uniformly accelerated charge, modified by a dipole. Radiation is produced. When the acceleration is annulled by the nearly uniform gravity field of an external shell with a 1 + beta cos theta surface distribution of mass, the differently viewed Born field is static and joins a static field outside the external shell; no radiation is produced. We discuss gravitational analogues of these phenomena. When a massive spherical shell is accelerated, an untouched test mass inside experiences a uniform gravity field and accelerates parallelly to the surrounding shell. In the strong gravity regime we illustrate these effects using exact conformastatic solutions of the Einstein-Maxwell equations with charged dust. We consider a massive charged shell on which the forces due to nearly uniform electrical and gravitational fields balance. Both fields are reduced inside by the ratio of the g_00 inside the shell to that away from it. The acceleration of a free test particle, relative to a static observer, is reduced correspondingly. We give physical explanations of these effects.

The reported anomalous acceleration may be explained as the recoil of radiated waste RTG heat scattered by the back of the high gain antenna.

Is it possible to create a small volume of equilibrium pair plasma in the laboratory with an intense laser pulse? No. The required power would exceed 10^{18} W. The number of seed particles required to absorb the laser energy would approach the number of pairs in the desired plasma, and their radiation in a visible laser field would occur at extremely high frequencies.

Slowly rotating collapsing spherical shells have flat spaces inside and the inertial frames there rotate at omega_s(t) relative to infinity. As first shown by Lindblom & Brill the inertial axes within the shell rotate rigidly without time delays from one point to another. Although the rotation rate of the inertial axes is changing the axes are inertial so, relative to them, there is neither an omega dot_s times r (Euler) fictitious force nor any other. However, Euler and other fictitious forces arise in the frame which is at rest with respect to infinity. An observer at the center who looks in one direction (phi = 0, theta = pi/2 say) fixed to infinity will see that the sky appears to rotate and can compare its apparent rotation with those of the local inertial frame and of the shell itself. By contrast in the electromagnetic analogue there is a time delay in the propagation of the magnetic field inside a rotating collapsing charged shell in flat space. We demonstrate this time delay by devising a null experiment in which the Larmor precession of a charged oscillator would be exactly cancelled by the rotation of the inertial frame but for the delay. In the ``combined'' problem of a collapsing charged shell we show that, due to the coupling of electromagnetic and gravitational perturbations, the instantaneous rotation of inertial frames inside the shell can be caused by pure electric currents in a non-rotating shell.

The discovery by BeppoSAX and coordinated ground-based observations of persistent X-ray, visible and radio counterparts to GRB has successfully concluded a search begin in 1973. The problem of explaining the mechanisms of GRB and their persistent counterparts remains. There are two classes of models: 1) GRB continue weakly for days at all frequencies; 2) GRB emission shifts to lower frequencies as relativistic debris sweeps up surrounding gas (in an ``external shock'') and slows. In 1) the visible afterglow is accompanied by continuing gamma-ray emission, as hinted by the high energy emission of GRB940217 and the ``Gang of Four'' bursts of October 27--29, 1996; the persistent emission may fluctuate. Behavior of this sort may be found in ``internal shock'' models. Models of class 2) have been the subject of several theoretical studies which disagree in assumptions and details but which predict that at each frequency the flux should rise and then decline, with the maximum coming later at lower frequencies. Some of this behavior has been observed, but data from GRB970508 show that its afterglow cannot be simply extrapolated from its gamma-ray emission. It is likely that both classes of processes occur in most GRB. Comparisons between GRB show that they are not all scaled versions of the same event. These results suggest that most gamma-ray emission is the result of ``internal shocks'' while most afterglow is the result of ``external shocks'', and hint at the presence of collimated outflows. Self-absorption in the radio spectrum of GRB970508 permitted the size of the radiating surface to be estimated, and in future GRB it may be possible to follow the expansion of the shell in detail and to construct an energy budget.

The gamma-ray burst source GRB970228 has been observed after a delay of 8--12 hours in X-rays and after one day in visible and near infrared light. This marks the first detection of emission at lower frequencies following the gamma-ray observation of a GRB and the first detection of any visible counterpart to a GRB. We consider possible delayed visible and X-ray emission mechanisms, and conclude that the intrinsic gamma-ray activity continued at a much reduced intensity for at least a day. There are hints of such continued activity in other GRB, and future observations can decide if this is true of GRB in general. The observed multi-band spectrum of GRB970228 agrees with the predictions of relativistic shock theory when the flux is integrated over a time longer than that required for a radiating electron to lose its energy.

Relativistic shock models of gamma-ray bursts may be tested by comparison of their predicted low energy asymptotic spectral indices s to observations. Synchrotron radiation theory predicts that the instantaneous spectrum has s = - 1/3 and the spectrum integrated over the radiative decay of the electrons' energies has s = 1/2, with other cases lying between these limits. We examine the spectra of 11 bursts obtained by the Large Area Detectors on BATSE. One agrees with the predicted instantaneous spectrum, as does the initial portion of a second, and three are close to the predicted integrated spectrum. All the observed asymptotic spectral slopes lie in the predicted range. This evidence for relativistic shocks is independent of detailed models of bursts and of assumptions about their distances. Radiation observed with the predicted instantaneous spectrum has a comparatively smooth time dependence, consistent with the necessarily long radiation time, while that with the predicted integrated spectrum has a spiky time dependence, consistent with the necessarily short radiation time.

Sari and Piran have demonstrated that the time structure of gamma-ray bursts must reflect the time structure of their energy release. A model which satisfies this condition uses the electrodynamic emission of energy by the magnetized rotating ring of dense matter left by neutron star coalescence; GRB are essentially fast, high field, differentially rotating pulsars. The energy densities are large enough that the power appears as an outflowing equilibrium pair plasma, which produces the burst by baryon entrainment and subsequent internal shocks. I estimate the magnetic field and characteristic time scale for its rearrangement, which determines the observed time structure of the burst. There may be quasi-periodic oscillations at the rotational frequencies, which are predicted to range up to 5770 Hz (in a local frame). This model is one of a general class of electrodynamic accretion models which includes the Blandford and Lovelace model of AGN, and which can also be applied to black hole X-ray sources of stellar mass. The apparent efficiency of nonthermal particle acceleration is predicted to be 10--50%, but higher values are possible if the underlying accretion flow is super-Eddington. Applications to high energy gamma-ray observations of AGN are briefly discussed.

The recent discovery of a planet in an orbit with eccentricity $e = 0.63 \pm 0.08$ around the Solar-type star 16 Cyg B, together with earlier discoveries of other planets in orbits of significant eccentricity, raises the question of the origin of these orbits, so unlike the nearly circular orbits of our Solar system. In this paper I consider close encounters between two planets, each initially in a nearly circular orbit (but with sufficient eccentricity to permit the encounter). Such encounters are described by a two-body approximation, in which the effect of the attracting star is neglected, and by the approximation that their separation vector follows a nearly parabolic path. A single encounter cannot produce the present state of these systems, in which one planet is in an eccentric orbit and the other has apparently been lost. Even if the requirement that the second planet be lost is dropped, nearly circular orbits cannot scatter into eccentric ones.

The outstanding task of gamma-ray burst astronomy is to test the hypothesis that they are at cosmological distances. This can be done by determining the coordinates of at least a few bursts to ~ 15'' or better. If they are in distant galaxies, the next task is to determine from which component of the galaxies the bursts originate; sub-arcsecond positions would answer this question. I outline ground-based systems which can accomplish these tasks, given BATSE/BACODINE coordinates and conservative assumptions about visible counterparts; these systems would be a ground-based upgrade to GRO which would accomplish the chief objectives of HETE.

Sillanp\"a\"a, {\it et al.} (1996) have demonstrated that the AGN OJ287 has intensity peaks which recur with a period of about 12 years. I suggest that this is the result of the sweeping of a precessing relativistic beam across our line of sight. In analogy to Her~X-1 and SS433, precession is attributed to the torque exerted by a companion mass on an accretion disc. Secondary maxima observed 1.2 years after two of these peaks may be evidence of nodding motion.

We report evidence from the 3B Catalogue that long ($T_{90} > 10$ s) and short ($T_{90} < 10$ s) gamma-ray bursts represent distinct source populations. Their spatial distributions are significantly different, with long bursts having $\langle V/V_{max} \rangle = 0.282 \pm 0.014$ but short bursts having $\langle V/V_{max} \rangle = 0.385 \pm 0.019$, differing by $0.103 \pm 0.024$, significant at the $4.3 \sigma$ level. Long and short bursts also differ qualitatively in their spectral behavior, with short bursts harder in the BATSE (50--300 KeV) band, but long bursts more likely to be detected at photon energies > 1 MeV. This implies different spatial origin and physical processes for long and short bursts. Long bursts may be explained by accretion-induced collapse. Short bursts require another mechanism, for which we suggest neutron star collisions. These are capable of producing neutrino bursts as short as a few ms, consistent with the shortest observed time scales in GRB. We briefly investigate the parameters of clusters in which neutron star collisons may occur, and discuss the nuclear evolution of expelled and accelerated matter.

Data from the 3B Catalogue suggest that short and long GRB are the results of different classes of events, rather than different parameter values within a single class: Short bursts have harder spectra in the BATSE bands, but chiefly long bursts are detected at photon energies over 1 MeV, implying that their hard photons are radiated by a process not found in short bursts. The values of \langle V/V_{max} \rangle for short and long bursts differ by 4.3 \sigma, implying different spatial distributions. Only the soft gamma-ray radiation mechanisms are the same in both classes.

We report evidence from the 3B Catalogue that short (T_90 < 10 s) and long (T_90 > 10 s) GRB represent different populations and processes: Their spectral behavior is qualitatively different, with short bursts harder in the BATSE range, but chiefly long bursts detected at higher photon energies; \langle V/V_max \rangle = 0.385 \pm 0.019 for short GRB but \langle V/V_max \rangle = 0.282 \pm 0.014 for long GRB, differing by 0.103 \pm 0.024. Long GRB may be the consequence of accretion-induced collapse, but this mechanism fails for short GRB, for which we suggest colliding neutron stars.

A relative mechanics with no absolute space is shown to be equivalent to Newtonian mechanics applied in a universe of zero net angular momentum. Closed spaces in General Relativity have no angular momentum and shrivel to one point as the mass-energy contained tends to zero, so obeying Mach's principle on the origin of inertia.

The following principle of minimum energy may be a powerful substitute to the dynamical perturbation method, when the latter is hard to apply. Fluid elements of self-gravitating barotropic flows, whose vortex lines extend to the boundary of the fluid, are labelled in such a way that any change of trial configurations automatically preserves mass and circulation. The velocity field is given by a mass conserving Clebsch representation. With three independent Lagrangian functions, the total energy is stationary for all small variations about a flow with fixed linear and angular momenta provided Euler's equations for steady motion are satisfied. Thus, steady flows are stable if their energy is minimum. Since energy is here minimized subject to having local and global contants of the motion fixed, stability limits obtained that way are expected to be close to limits given by dynamical perturbation methods. Moreover, the stability limits are with respect to arbitrary, not necessary small, perturbations. A weaker form of the energy principle is also given which may be easier to apply. The Lagrangian functional, with the same three Lagrange variables is stationary for the fully time dependent Euler equations. It follows that the principle of minimum energy gives stability conditions that are both necessary and sufficient if terms linear in time derivatives (gyroscopic terms) are absent from the Lagrangian. The gyroscopic term for small deviations around steady flows is given explicitly. Key words: Energy variational principle; Self-gravitating systems; Stability of fluids.

Observed intensities and inferred distances of soft gamma repeaters imply luminosities in excess of the nominal (electron-scattering opacity) Eddington limit by four to six orders of magnitude. I review the physical basis of this limit. Accretional luminosities may exceed it if energy is hydrodynamically coupled from accreting matter to closed field lines where it forms a pair gas. This magnetically confined pair gas radiates roughly a black body spectrum with $k_B T_e \approx 23$ KeV, consistent with observations, at a luminosity up to $\sim 3 \times 10^{44}$ erg s$^{-1}$ for a surface field of $10^{13}$ gauss. Magnetic transparency is not required. I discuss the minutes-long continuing emission of March 5, 1979, steady counterparts to SGR, their spin periods and the recoil problem.

The delayed hard (up to 25 GeV) photons observed more than an hour following a gamma-ray burst on February 17, 1994 may result from the collisions of relativistic nucleons with a dense cloud, producing $\pi^0$. The required cloud density is $\sim 2 \times 10^{11}$ cm$^{-3}$. This cloud may be the remains of the disrupted envelope of a neutron star, and may survive as an excretion disc of $\sim 10^{14}$--$10^{15}$ cm radius around the coalescing binary.

We develop a model of SGR in which a supernova leaves planets orbiting a neutron star in intersecting eccentric orbits. These planets will collide in $\sim 10^4$ years if their orbits are coplanar. Some fragments of debris lose their angular momentum in the collision and fall onto the neutron star, producing a SGR. The initial accretion of matter left by the collision with essentially no angular momentum may produce a superburst like that of March 5, 1979, while debris fragments which later lose their angular momentum produce an irregular pattern of smaller bursts.

Particles with energies below the mean energy $E_0$ in relativistic shocked plasmas should assume an equilibrium energy distribution. This leads to a synchrotron spectrum $F_\nu \propto \nu^{1/3}$ up to approximately the critical frequency $\nu_0$ of an electron with the energy $E_0$. Application to GRBs implies that a burst with $10^{-5}$ erg/cm$^2$s of soft gamma-rays and $h \nu_0 = 300$ KeV should be about 18th magnitude in visible light and a few $\mu$Jy at 1 GHz (less if self-absorbed).

Relativistic blast wave models predict the spectrum of the emitted synchrotron radiation. The electrons in the shocked region are heated to a Wien distribution whose ``temperature'' is $1/3$ of the mean electron energy. This energy scale determines a characteristic (break) frequency of synchrotron radiation. At much lower frequencies a spectrum $F_nu \propto \nu^{1/3}$ is predicted independently of details of the emitting region. This is consistent with the observed soft X-ray spectra of GRB. It implies low visible and radio intensities, unless there are collective emission processes.

Gamma-ray burst statistics are best explained by a source population at cosmological distances, while spectroscopy and intensity histories of some individual bursts imply an origin on Galactic neutron stars. To resolve this inconsistency I suggest the presence of two populations, one at cosmological distances and the other Galactic. I build on ideas of Shemi and Piran (1990) and of M\'esz\'aros and Rees (1993) involving the interaction of fireball debris with surrounding clouds to explain the observed intensity histories in bursts at cosmological distances. The distances to the Galactic population are undetermined because they are too few to affect the statistics of intensity and direction; I explain them as resulting from magnetic reconnection in neutron star magnetospheres.

Self-consistent models of gamma-ray burst source regions at 100 Kpc distance are possible if the radiating plasma is confined to very thin sheets, and I estimate parameters. Energy sources might be elastic (by starquakes) or magnetic (by reconnection), but mechanisms remain obscure. I discuss a very speculative model involving collisions between comets in a hypothetical inner Oort cloud.

I present a model for acceleration of protons by the second-order Fermi process acting on randomly scrambled magnetic flux arches above an accretion disc. The accelerated protons collide with thermal protons in the disc, producing degraded energetic protons, charged and neutral pions, and neutrons. The pions produce gamma-rays by spontaneous decay of $\pi^0$ and by bremsstrahlung and Compton processes following the decay of $\pi^\pm$ to $e^\pm$.

Dust is observed to form in nova ejecta. The grain temperature is determined by the diluted nova radiation field rather than the gas kinetic temperature, making classical nucleation theory inapplicable. We used kinetic equations to calculate the growth of carbon nuclei in these ejecta. For expected values of the parameters too many clusters grew, despite the small sticking probability of atoms to small clusters, and the clusters only reached radii of about 100\AA\ when the carbon vapor was depleted. We then included the effects of cluster photodissociation by ultraviolet radiation from the nova. This suppresses nucleation, but too well, and no grains form at all. Finally we suggest that a few growing carbon nuclei may be protected from photodissociation by a sacrificial surface layer of hydrogen.

We have calculated gamma-ray radiative transport in regions of high energy density, such as gamma-ray burst source regions, using a discrete ordinate, discrete energy group method. The calculations include two-photon pair production and annihilation, as well as three-photon pair annihilation. The radiation field itself acts as an absorbing medium, and the optical depth depends on its intensity, so the problem is intrinsically nonlinear. Spherical divergence produces effective collimation of the flux. At high optical depth the high energy ($E > 1$ MeV) portion of the emergent spectrum assumes a nearly universal form. An approximate limit is derived for the high energy flux from a gamma-ray burst source region of given size, and the implications of this limit for the distance to the March 5, 1979 event are briefly discussed. We discuss more generally the problem of very luminous bursts, and implications of Galactic halo distances for flare models.

The BATSE experiment on GRO has demonstrated the isotropic arrival directions and flat $\log N$ {\it vs.} $\log S$ of cosmic gamma-ray bursts. These data are best explained if the burst sources are distributed throughout an extended spherical Galactic halo, as previously suggested by Jennings. The halo's radius is at least 40 Kpc, and probably is more than 100 Kpc. I consider possible origins of this halo, including primordial formation and neutron stars recoiling from their birthplaces in the Galactic disc. A simple geometrical model leads to a predicted relation between the dipole and quadrupole anisotropy. I suggest that neutron stars born with low recoil become millisecond pulsars, while those born with high recoil become the sources of gamma-ray bursts; these populations are nearly disjoint. Quiescent counterparts of gamma-ray bursts are predicted to be undetectably faint.