Most of the sources detected in the extreme ultraviolet (EUV; 100 Ang to 600 Ang) by the Rosat WFC and EUVE all-sky surveys have been identified with active late-type stars and hot white dwarfs that are near enough to escape absorption by interstellar gas. However, about 15% of EUV sources are as of yet unidentified with any optical counterparts. We examine whether the unidentified EUV sources may consist of the same population of late-type stars and white dwarfs. We present B and R photometry of stars in the fields of seven of the unidentified EUV sources. We detect in the optical the entire main-sequence and white-dwarf population out to the greatest distances where they could still avoid absorption. We use colour-magnitude diagrams to demonstrate that, in most of the fields, none of the observed stars have the colours and magnitudes of late-type dwarfs at distances less than 100 pc. Similarly, none are white dwarfs within 500 pc that are hot enough to be EUV-emitters. The unidentified EUV sources we study are not detected in X-rays, while cataclysmic variables, X-ray binaries, and active galactic nuclei generally are. We conclude that some of the EUV sources may be a new class of nearby objects, that are either very faint at optical bands or which mimic the colours and magnitudes of distant late-type stars or cool white dwarfs. One candidate for optically faint objects is isolated old neutron stars, slowly accreting interstellar matter. Such neutron stars are expected to be abundant in the Galaxy, and have not been unambiguously detected.

We report the optical identification of an $R=18.3$ mag, $z=2.432$ quasar at the position of a 6 cm radio source and a faint \rosat PSPC X-ray source. The quasar lies within the error circles of unidentified extreme-UV (EUV) detections by the \euve and \rosat WFC all-sky surveys at $\sim 400$ \AA\ and $\sim 150$ \AA, respectively. A 21 cm H I emission measurement in the direction of the quasar with a $21'$-diameter beam yields a total H I column density of $N_{H}=3.3\times 10^{20}$ cm$^{-2}$, two orders of magnitude higher than the maximum allowed for transparency through the Galaxy in the EUV. The source of the EUV flux is therefore probably nearby ($\ltorder 100$ pc), and unrelated to the quasar.

Unidentified extreme UV sources, detected in the {\it EUVE} and {\it ROSAT} WFC all-sky surveys, could be isolated old neutron stars, accreting material from the ISM. The closest neutron stars, that are located in the local ISM bubble of unusually low density, are faint and cool ($L \sim 10^{27} {\rm \ erg \ s^{-1}}$, $T \sles 6 $eV). The extreme UV spectrum of these sources is very sensitive to the H I column density, since large fraction of the emission is emitted just below the hydrogen Lyman edge. The {\it EUVE} sources with large count rate in the long-wave bandpass (600\AA) seem to be the most promising candidates. These sources should have low H I column density ($N_{HI} \sles 10^{18} {\rm \ cm^{-2}}$), constraining their distance to a few tens of parsec. Otherwise, their spectra would be significantly modified by ISM absorption, inevitably become stronger in the short wavelength (100\AA, 200\AA) {\it EUVE} bandpasses. If these unidentified objects are familiar EUV sources rather than neutron stars, either white dwarfs, late type stars or cataclysmic variables, they are expected to be identifiable, generally brighter than $\sim 14$.

Interstellar matter that is accreted onto isolated magnetic neutron stars in the Galaxy ($\sim 10^9$ by number) is accelerated and reflected back by MHD shocks, which envelope the stars. The integrated power in the Galaxy $L_{cr,ns}$ is $ \sgreat 10^{40} {\rm erg \ s^{-1} }$, the energy distribution is a power law of spectral index $> 2$, and the particle energy can be raised to $10^6$ GeV, consistent with the power and spectrum of primary cosmic rays in the Galaxy. The major contribution for $L_{cr,ns}$ comes from a minority of $\sim 10^7$ isolated neutron stars which are located within dense clouds. Sources in these clouds, that are generally spread within the Galactic disk, can explain the concentration of high-energy cosmic rays in the Galactic plane, as deduced from pion decay spectra in gamma-ray observations. The soft X-ray luminosity from these neutron stars is consistent with the Galactic X-ray background. The accretion may be associated with ion-neutral bias, that is further enhanced by ion confinement in frozen-in magnetic fields, which can raise the relative abundance of first ionization potential (FIP) elements in the cosmic rays.

Relativistic flows resulting from sudden explosive events upscatter ambient interstellar photons of local radiation fields. For Lorentz factor $ > 100$ and dense optical - UV radiation fields the emergent signal is a typical gamma ray burst. Presumably the explosions occur in dense globular clusters or in galactic nuclei, at cosmological distances.

We examine analytically and numerically the evolution of a relativistic fireball. We show that, after an early rearrangement phase, most of the matter and energy in the fireball is concentrated within a narrow shell. The shell propagates at nearly the speed of light, with a frozen radial profile and according to a simple set of scaling laws. The spectrum of the escaping radiation is harder at early times and softer later on. The results are applicable to models of $\gamma$-ray bursts. Submitted to Monthly Notices of the RAS Figures available on request from the authors

If gamma ray burst sources are in the galactic halo they inevitably involve a creation of an opaque pair plasma fireball, just like in cosmological sources. We find that the typical physical conditions in a galactic halo fireball are: optical depth $\ap 10^8$, thermal energy $\ap 100 KeV$, maximal relativstic expansion $\G \ap 300 $ and a maximal baryonic load of $\ap 10^{-15} M_\odot$. This does not rule out galctic halo models but it poses an addtional severe constraint on all such sources. A comparison of these conditions with the physical conditions at cosmological fireballs reveal that galactic halo fireballs are less favorable than cosmological ones as sources of gamma ray bursts.